Burkholderia Ambifaria Sp. Nov., a Novel Member of the Burkholderia Cepacia Complex Including Biocontrol and Cystic ﬁbrosis-Related Isolates

International Journal of Systematic and Evolutionary Microbiology (2001), 51, 1481–1490 Printed in Great Britain

Burkholderia ambifaria sp. nov., a novel member of the Burkholderia cepacia complex including biocontrol and cystic ﬁbrosis-related isolates

1 Laboratorium voor Tom Coenye,1 Eshwar Mahenthiralingam,2 Deborah Henry,3 Microbiologie, 4 1 1 Universiteit Gent, K. L. John J. LiPuma, Severine Laevens, Monique Gillis, Ledeganckstraat 35, David P. Speert3 and Peter Vandamme1 B-9000 Ghent, Belgium 2 School of Biosciences, Cardiff University, Cardiff, Author for correspondence: Tom Coenye. Tel: j32 9 264 51 14. Fax: j32 9 264 50 92. UK e-mail: Tom.Coenye!rug.ac.be 3 Department of Pediatrics, Division of Infectious and A polyphasic taxonomic study, including amplified fragment length Immunological Diseases, University of British polymorphism (AFLP) fingerprinting, DNA–DNA hybridizations, DNA base-ratio Columbia, Vancouver, determinations, phylogenetic analysis, whole-cell fatty acid analyses and an British Columbia, Canada extensive biochemical characterization, was performed on 19 Burkholderia 4 Department of Pediatrics cepacia-like isolates from the environment and cystic fibrosis (CF) patients. and Communicable Several of the environmental isolates have attracted considerable interest due Diseases, University of Michigan Medical School, to their biocontrol properties. The polyphasic taxonomic data showed that the Ann Arbor, MI, USA strains represent a new member of the B. cepacia complex, for which the name Burkholderia ambifaria sp. nov. is proposed. The type strain is strain LMG 19182T. B. ambifaria can be differentiated from the other members of the B. cepacia complex by means of AFLP fingerprinting, whole-cell fatty acid analysis, biochemical tests (including ornithine and lysine decarboxylase activity, acidification of sucrose and β-haemolysis) and a newly developed recA gene-based PCR assay. 16S rDNA-based RFLP analysis and PCR tests allowed differentiation of B. ambifaria from Burkholderia multivorans, Burkholderia vietnamiensis and B. cepacia genomovar VI, but not from B. cepacia genomovars I and III and Burkholderia stabilis. The finding that this new taxon includes both strains isolated from CF patients and potentially useful biocontrol strains supports the general consensus that the large-scale use of biocontrol strains belonging to the B. cepacia complex would be ill-advised until more is known about their potential pathogenic mechanisms.

Keywords: Burkholderia ambifaria, Burkholderia cepacia complex, cystic ﬁbrosis, biocontrol, taxonomy

INTRODUCTION least six different genomic species, referred to col- lectively as the B. cepacia complex (Vandamme et al., The taxonomy of Burkholderia cepacia-like organisms 1997; Coenye et al., 2001). One of these genomic has been evolving rapidly and it has been shown that species was identified as the previously described presumed B. cepacia strains isolated from cystic Burkholderia vietnamiensis (Gillis et al., 1995), and the fibrosis (CF) patients and the environment belong to at names Burkholderia multivorans and Burkholderia stabilis have been proposed for strains respectively ...... classified previously as B. cepacia genomovars II and Abbreviations: AFLP, amplified fragment length polymorphism; CF, IV (Vandamme et al., 1997, 2000). B. cepacia genom- cystic fibrosis. ovars III and VI still await formal binomial species The GenBank accession numbers for the recA gene sequences of B. ambifaria strains LMG 19182T, R-8863 and R-5142 and B. cepacia genomovar name assignment, pending the availability of differ- VI strain LMG 18943 are AF323985, AF143975, AF143801 and AF323971, ential phenotypic tests (Vandamme et al., 2000; respectively. Coenye et al., 2001).

01692 # 2001 IUMS 1481 T. Coenye and others

Originally known as a plant pathogen, B. cepacia has DNA preparation. DNA used for AFLP fingerprinting, emerged as an important opportunistic pathogen and DNA–DNA hybridizations and determination of the DNA plays a significant role in patients with CF and chronic base composition was prepared as described by Pitcher et al. granulomatous disease (Gilligan, 1991; Speert et al., (1989). DNA used for PCR amplification of 16S rRNA and 1994; Govan & Deretic, 1996; Govan et al., 1996; recA genes was prepared by heating one or two colonies (picked from a plate grown overnight) at 95 C for 15 min in LiPuma, 1998a, b). B. cepacia has also attracted m 20 µl lysis buffer containing 0n25% (w\v) SDS and 0n05 M considerable commercial interest as a biological con- NaOH. Following lysis, 180 µl distilled water was added to trol, bioremediation and plant-growth-promoting the lysis buffer and the DNA solutions were stored at agent (Holmes et al., 1998; LiPuma & Mahen- k20 mC. thiralingam, 1999). Field tests have shown that B. AFLP fingerprinting. The preparation of template DNA cepacia-like organisms can colonize the rhizosphere of (using restriction enzymes ApaI and TaqI), amplification several economically important crops, including corn, [using primers A00 and T00 in the pre-selective PCR and maize, rice, pea, sunflower and radish, and can thereby primers B07 (labelled with the fluorescent dye 6-FAM) and increase the crop yield significantly (Parke et al., 1991; T11 in the selective PCR], separation of the fragments using McLoughlin et al., 1992; Bowers & Parke, 1993; an ABI Prism 377 automated DNA sequencer and numerical Hebbar et al., 1998). However, the risks associated analysis were performed as described previously (Coenye et with the use of B. cepacia are not yet clear, and there is al., 1999). a general consensus that the large-scale use of these 16S rRNA gene RFLP. Primers UNI1 and UNI2 were used for organisms would be ill-advised until more is known the amplification of a 1 kb fragment of the 16S rRNA gene about background environmental levels of the various (Mahenthiralingam et al., 2000). Restriction with the restriction enzyme DdeI (New England Biolabs) and agarose members of the B. cepacia complex, the fate of gel electrophoresis were performed as described before biocontrol strains after environmental release, patho- (Segonds et al., 1999). genic mechanisms, clinical outcomes and the inter- rRNA-based PCR assays. Assays for the identification of action of introduced biocontrol strains with environ- members of the B. cepacia complex based on the rRNA mental and clinical strains (Govan & Vandamme, operon were described previously (LiPuma et al., 1999). The 1998; Holmes et al., 1998; LiPuma & Mahen- following primers were used: RHG-F and RHG-R (specific thiralingam, 1999; Vidaver et al., 1999; Govan et al., for all of the Burkholderia and Ralstonia species), PC-SSF 2000). and PC-SSR (specific for B. cepacia genomovars I and III and B. stabilis), BC-GII and BC-R (specific for B. multi- One of the most studied biocontrol isolates is B. vorans) and BC-V and BC-R (specific for B. multivorans and cepacia AMMD (l LMG 19182). This strain was B. vietnamiensis). isolated from apparently healthy pea plants in Determination of the DNA base composition. DNA was Wisconsin (USA) in 1985. Several studies have shown enzymically degraded into nucleosides as described by that strain LMG 19182 is very effective in controlling Mesbah et al. (1989). The nucleoside mixture obtained was phytopathogenic Pythium species (responsible for pre- then separated by HPLC using a Waters SymmetryShield C8 and post-emergence damping-off in peas) (Parke, column thermostatted at 37 mC. The solvent was 0n02 M 1990; Parke et al., 1991; Xi et al., 1996) and Aphano- NH%H#PO% (pH 4n0) with 1n5% acetonitrile. Non-meth- myces euteiches (responsible for root-rot and the main ylated lambda phage DNA (Sigma) was used as the factor limiting pea production in the American Mid- calibration reference. west) (King & Parke, 1993). DNA–DNA hybridizations. DNA–DNA hybridizations were performed with photobiotin-labelled probes in microplate In an ongoing survey of B. cepacia-like isolates by wells as described by Ezaki et al. (1989), using an HTS7000 means of whole-cell protein analysis, amplified frag- Bio Assay Reader (Perkin-Elmer) for the fluorescence ment length polymorphism (AFLP) fingerprinting and measurements. The hybridization temperature was 50 mC. sequence analysis of the recA gene, 18 isolates from Phylogenetic analysis of 16S rRNA gene sequences. The 16S human clinical and environmental specimens exhibited rDNA sequence of B. ambifaria strain LMG 19182T was similarity towards B. cepacia LMG 19182. This report retrieved from the GenBank database (accession no. describes the polyphasic taxonomic study that was AF043302) and compared with the published 16S rDNA used for the further characterization of these isolates. sequences of other Burkholderia species. A phylogenetic tree The genotypic and phenotypic characteristics allowed was constructed with the  2.1 software package the classification of these strains as a new member of (Applied Maths), based on the neighbour-joining method the B. cepacia complex, for which we propose the name (Saitou & Nei, 1987). Approximately 1460 bases were used Burkholderia ambifaria sp. nov. and all unknown bases were excluded from the calculations. recA gene sequencing and phylogenetic analysis. The sequences of the recA genes of B. ambifaria strains LMG METHODS 19182T, R-5142 and R-8863 were determined as described Bacterial strains and growth conditions. B. ambifaria strains previously (Mahenthiralingam et al., 2000) and compared used in this study are listed in Table 1. Reference strains of with previously determined sequences. A phylogenetic tree other Burkholderia species and genomovars have been was constructed as described above. Approximately 1040 described previously (Vandamme et al., 1997, 2000; Coenye bases were used and all unknown bases were excluded from et al., 2001). All strains were grown aerobically on Trypticase the calculations. soy agar (BBL) and incubated at 28 mC unless indicated Fatty acid methyl ester analysis. After incubation for 24 h at otherwise. 28 mC, a loopful of well-grown cells was harvested and fatty

1482 International Journal of Systematic and Evolutionary Microbiology 51 Burkholderia ambifaria sp. nov...... teborg, $ , biocontrol strain Sesbania exaltata ...... J. L. Parke Pea rhizosphere (WI, USA, 1985), biocontrol strain LMG Bacteria Collection, Laboratorium voor Microbiologie Gent, Universiteit Gent, Ghent, \ T CCUG 44356 ...... teborg, Sweden; LMG, BCCM $ T strains studied AMMD B. ambifaria T List of Strain Designation as received Other strain designation(s) Depositor Source LMG 19182 R-10027R-10028R-10029LMG 19466 – – – – ES0020 ES0033 AU0212 ES0050 Own isolate Own isolate Own isolate Own isolate Snap bean rhizosphere (USA) Commercial soil (USA) CF patient Sedge (USA) rhizosphere (USA) LMG 17829LMG 17828LMG 11351LMG 6991R-697R-698 ATCC 53267R-5140 ATCC 53266R-5142 PHP7R-8863 CCUG 2856R-9914R-9917 CEP0102, C2965 PC526R-9927 FC662 PC406LMG 19467 J82LMG 19465 M54 ATCC 17760, Ral-3 FC655 – – ATCC – – – ATCC 51671 CCUG J2744, FC766 J2743, FC767 ATCC HI2345, FC623 HI2347, FC625 HI2347, FC627 FC661, R-9945 CEP0516 J. Corn L. roots T. Parke (WI, Heulin USA), CEP0617 J. biocontrol L. CEP0996, strain Parke R-9935 CEP0958 Stine Forest Microbial soil Products Stine (Trinidad, Microbial 1958) Products E. Pedersen ATCC Corn roots (WI, USA), Biocontrol biocontrol strain strain Biocontrol strain P. Taylor P. Taylor Corn rhizosphere (WI, J. Maize USA) Burns rhizosphere Corn rhizosphere (WI, P. USA) Taylor Corn rhizosphere, biocontrol strain Leaves of CF patient (Australia) CF patient, respiratory tract (Australia) CF patient, sputum CF (USA) patient, sputum (Australia) Belgium. Department of Clinical Bacteriology, Go Table 1...... Culture collections are abbreviated as: ATCC, American Type Culture Collection, Manassas, VA, USA; CCUG, Culture Collection of the University of Go

International Journal of Systematic and Evolutionary Microbiology 51 1483 T. Coenye and others acid methyl esters were prepared, separated and identified ovar VI but were indistinguishable from B. cepacia using the Microbial Identification System (Microbial ID) as genomovars I and III and from B. stabilis (data not described before (Vandamme et al., 1992). shown). Phenotypic characterization. Tests were performed as described previously (Henry et al., 1997). Briefly, pure cultures rRNA-based PCR assays were stored at k70 mC in Mueller–Hinton broth (Difco) with 8% DMSO. Frozen isolates were subcultured to DNA fragments of the expected size were amplified for Columbia agar containing 5% sheep blood (PML Micro- all of the B. ambifaria isolates investigated with the biologicals) before testing. The primary identification system primer pair RHG-F and RHG-R (specific for the used was the API Rapid NE system (bioMe! rieux Vitek), genera Burkholderia and Ralstonia) and with the supplemented with glucose, maltose, lactose, xylose, sucrose primer pair PC-SSF and PC-SSR (specific for B. and adonitol oxidation\fermentation (OF) sugars (Hugh & Leifson, 1953) and an adaptation of Moeller lysine and cepacia genomovars I and III and for B. stabilis). ornithine decarboxylases (Difco). In brief, the 2 ml test Using primer pairs BC-GII and BC-R (specific for B. volume was overlaid with 0n5 ml light mineral oil and a multivorans) and BC-GV and BC-R (specific for B. heavy, visible inoculum was used. A tube without amino multivorans and B. vietnamiensis), no amplification acid was included as a negative control. OF sugars were product was obtained (data not shown). incubated for up to 7 d and other tests for 2 d. Incubations were done at 35 mC except where mentioned otherwise. DNA–DNA hybridizations and determination of the Growth on Trypticase soy agar (Becton Dickinson) at 35 GjC content and 42 mC was observed for appearance and pigment. The hybridization results revealed that DNA–DNA T RESULTS binding values between B. ambifaria LMG 19182 and B. cepacia genomovar I LMG 1222T (38%), B. AFLP fingerprinting multivorans LMG 13010T (56%), B. cepacia genomovar III LMG 16656 (24%), B. stabilis LMG 14294T Reproducibility was checked by preparing PCR prod- T ucts in duplicate and was always higher than 92% (37%), B. vietnamiensis LMG 10929 (44%), B. cepacia genomovar VI LMG 18941 (50%) and B. (data not shown). Following numerical analysis, nine T clusters could be delineated, while Burkholderia pyrro- pyrrocinia LMG 14191 (58%) were intermediate or cinia T low. DNA–DNA hybridization experiments carried LMG 14191 occupied a separate position (Fig. T 1). The B. ambifaria isolates investigated appeared as a out between B. ambifaria isolates LMG 19182 ,LMG homogeneous group that formed cluster A. B. cepacia 11351, R-9917, R-8863 and R-5142 revealed binding genomovar VI constituted cluster B. Reference isolates levels between 68 and 95% (Table 2). All B. ambifaria of B. multivorans, B. vietnamiensis and B. stabilis isolates investigated had GjC contents between 65n5 formed clusters C, D and F, respectively. Clusters E and 66n8 mol% (Table 2). and G respectively comprised three and four B. cepacia genomovar III reference isolates. Representatives of Cellular fatty acid analysis Burkholderia gladioli and B. cepacia genomovar I formed clusters H and I, respectively. The following fatty acids were present in B. ambifaria: 14:0 (3n4p0n4%), 16:0 (15n0p3n0%), 17:0 cyclo (5 1 2 9%), 16:1 2-OH (1 5 0 7%), 16:0 2-OH Phylogenetic analysis n p n n p n (trace amounts), 16:0 3-OH (5n9p0n5%), 18:0 The 16S rRNA gene sequence of strain LMG 19182T (1n6p2n1%), 19:0 cyclo ω8c (1n7p1n2%), 18:1 2-OH (GenBank accession no. AF043302) was compared (2n4p1n0%), summed feature 2 (comprising 14:0 3- with the available 16S rRNA gene sequences of OH, 16:1 iso I, an unidentified fatty acid with representatives of the β-Proteobacteria. Strain LMG equivalent chain length of 10n928 or 12:0 ALDE or 19182T was closely related to members of the B. any combination of these fatty acids) (7n2p0n5%), cepacia complex, with similarity levels ranging from summed feature 3 (comprising 16:1ω7c or 15 iso 2-OH 98n1to98n8%. Similarity levels towards other Burk- or both) (22n5p4n4%) and summed feature 7 (com- holderia species were below 97n7%. Similarity levels prising 18:1ω7c,18:1ω9t or 18:1ω12t or any com- towards representatives of other taxa belonging to the bination of these fatty acids) (32n1p4n7%). Summed β-Proteobacteria were below 95n1% (Fig. 2). Based on feature 3 and summed feature 7 probably correspond the recA gene sequences, the three B. ambifaria isolates to 16:1ω7c and 18:1ω7c, respectively, as these fatty investigated formed a tight phylogenetic cluster (in- acids have been reported in Burkholderia species ternal similarities " 99n4%). Similarities towards (Stead, 1992). other representatives of the B. cepacia complex were between 94n2 and 95n6% (Fig. 3). Phenotypic characterization

16S rRNA gene RFLP The following characteristics were present in all the isolates investigated: oxidase, β-galactosidase and Using restriction enzyme DdeI, the B. ambifaria lysine decarboxylase activity, acidiﬁcation of glucose, isolates investigated could be distinguished from B. lactose, maltose, xylose and adonitol and assimilation multivorans, B. vietnamiensis and B. cepacia genom- of glucose, -arabinose, -mannose, -mannitol, N-

1484 International Journal of Systematic and Evolutionary Microbiology 51 Burkholderia ambifaria sp. nov.

...... Fig. 1. Dendrogram derived from the unweighted pair group average linkage of correlation coefficients between the AFLP patterns of the strains studied. The correlation coefficient is expressed as percentage similarity for convenience. acetylglucosamine, -gluconate, caprate, adipate, - β-haemolysis, pigment production, denitrification, malate, citrate and phenylacetate. The following char- aesculin hydrolysis and gelatin liquefaction. acteristics were absent in all strains investigated: ornithine decarboxylase, tryptophanase, arginine di- B. ambifaria-specific PCR hydrolase and urease activity and assimilation of maltose. The following characteristics were strain- Using the nucleotide sequence alignment of the pre- dependent: growth at 42 mC, acidification of sucrose, viously determined recA genes (Mahenthiralingam et

International Journal of Systematic and Evolutionary Microbiology 51 1485 T. Coenye and others

...... Fig. 2. Neighbour-joining phylogenetic tree showing the position of B. ambifaria within the β-Proteobacteria based on 16S rDNA sequence comparisons. Pelistega europaea was used as an outgroup. Bar, 5% sequence dissimilarity.

...... Fig. 3. Neighbour-joining phylogenetic tree showing the position of B. ambifaria within the B. cepacia complex based on recA gene sequence comparisons. Bordetella pertussis was used as an outgroup. Bar, 10% sequence dissimilarity.

1486 International Journal of Systematic and Evolutionary Microbiology 51 Burkholderia ambifaria sp. nov.

Table 2. DNA–DNA binding values and GjC content of all strains examined

Strain GjC DNA-binding values (%) with B. ambifaria strain: content (mol%) 1 2345

B. ambifaria 1. LMG 19182T 66n8 100 2. R-8863  68 100 3. R-5142  78 100 4. R-9917 65n8 95 100 5. LMG 11351 65n5 89 91 100 B. cepacia genomovar I LMG 1222T 66n738 B. multivorans LMG 13010T 67n7* 56 B. cepacia genomovar III LMG 16656 66n124 B. stabilis LMG 14294T 65n837 B. vietnamiensis LMG 10929T 66n2 44425340 B. cepacia genomovar VI LMG 18941 67n4* 50 B. pyrrocinia LMG 14191T 66n158

* Data from Coenye et al. (2001). , Not determined.

al., 2000) and the B. ambifaria recA sequences de- as a new member of the B. cepacia complex, B. termined in this study, specific primers were designed ambifaria sp. nov. to detect B. ambifaria strains. A primer pair was designed to match all sequences within the B. ambifaria recA phylogenetic group and to be mismatched at the Taxonomic position of B. ambifaria 3h base with all other B. cepacia complex recA sequences. The forward primer BCRGC1 (5h-GTC- It has been shown previously that AFLP, a genomic GGGTAAAACCACGCTG-3h) corresponds to pos- fingerprinting technique based on the selective ampli- itions 207–225 in the recA sequence of R-8863 fication and visualization of restriction fragments from (AF143795) and the reverse primer BCRGC2 (5h- a total digest of genomic DNA, allowed differentiation ACCGCAGCCGCACCTTCA-3h) corresponds to of Burkholderia species and genomovars and that its positions 999–1016 in the recA sequence of R-8863 results are generally consistent with DNA–DNA (base positions in primer sequences that were mis- hybridization levels (Coenye et al., 1999, 2001; Van- matched in other genomovars are underlined). damme et al., 2000). In this study, all B. ambifaria Approximately 20 ng DNA was incorporated into isolates investigated formed a single homogeneous 25 µl reactions that also contained 1 U Taq polymerase cluster and could be differentiated easily from other (Qiagen), 250 mM each dNTP (Pharmacia), 5 µlQ Burkholderia species and genomovars (Fig. 1). solution (Qiagen) and 1i PCR buffer (Qiagen). The Comparison of the 16S rDNA sequence of strain following thermal profile was applied for amplifi- T cation: initial denaturation at 94 mC for 2 min; 30 LMG 19182 with the available sequences of other cycles of 30 s at 94 mC, 45 s at 62 mC and 60 s at 72 mC; Burkholderia species indicated that its closest relatives and a final extension of 10 min at 72 mC. Using primers were members of the B. cepacia complex (similarity BCRGC1 and BCRGC2 and the PCR conditions levels ranging from 98n1to98n8%). The DNA GjC described above, an 810 bp product was amplified content (65n5–66n8 mol%) is also consistent with that from all B. ambifaria strains, while no product was of other members of the B. cepacia complex (66– amplified from strains belonging to other taxa (data 69 mol%) (Vandamme et al., 1997; Coenye et al., not shown). 2001). A main drawback of the comparison of 16S rRNA genes is that these genes evolve so slowly that DISCUSSION very recently diverged species may not be recognizable (Fox et al., 1992; Palys et al., 2000) and the sequencing We performed a polyphasic taxonomic study to of several other genes (including gyrB, rpoB, infB and characterize 19 B. cepacia-like isolates from the en- recA) has been proposed to complement the phylo- vironment and from CF patients. Several of the genetic information obtained from 16S rDNA se- environmental strains have attracted considerable quences (Eisen, 1995; Yamamoto & Harayama, 1995, interest due to their biocontrol properties. The results 1998; Venkateswaran et al., 1998; Hedegaard et al., of the present study allowed us to classify these strains 1999; Mare! chal et al., 2000). Comparison of the recA

International Journal of Systematic and Evolutionary Microbiology 51 1487 T. Coenye and others

Table 3. Phenotypic characteristics useful for the differentiation of B. ambifaria from other members of the B. cepacia complex and from B. gladioli ...... For characteristics other than fatty acid content, data were taken from Henry et al. (2001), except for B. cepacia genomovar VI (Coenye et al., 2001) and for B. ambifaria (this study), and represent the percentage of strains giving a positive reaction. The numbers of strains investigated in fatty acid analysis were 19 (B. ambifaria), 8 (B. cepacia genomovar I), 14 (B. multivorans), 54 (B. cepacia genomovar III), 11 (B. stabilis), 7 (B. vietnamiensis), 11 (B. cepacia genomovar VI) and 4 (B. gladioli). Data for fatty acid content are percentages of total fatty acids and were taken from this study (B. ambifaria) or from Vandamme et al. (1997, 2000) and Coenye et al. (2001) (other taxa).

Characteristic B. ambifaria B. cepacia B. multivorans B. cepacia B. stabilis B. vietnamiensis B. cepacia B. gladioli (19 strains) genomovar I (110 strains) genomovar III (27 strains) (36 strains) genomovar VI (27 strains) (23 strains) (142 strains) (14 strains)

Growth at 42 mC 26 43 100 83 0 100 100 4 β-Galactosidase activity 100 100 100 100 0 100 100 100 Decarboxylation of: Lysine 100 100 54 99 100 100 0 0 Ornithine 0 30 0 72 100 0 0 0 Acidiﬁcation of sucrose 95 91 0 90 0 94 0 0 β-Haemolysis 84 9 0 4 0 36 0 22 (weak) Fatty acid content: 16:0 15n0p3n026n8p2n828n9p3n526n1p3n525n6p4n319n5p2n526n1p1n429n0p1n1 17:0 cyclo 5n1p2n917n9p3n618n2p4n921n3p4n217n8p3n314n0p4n916n0p2n217n2p1n9 Summed feature 3* 22n5p4n44n6p2n56n5p1n53n0p1n02n7p0n89n8p2n54n2p2n05n1p2n3 Summed feature 7* 31n1p4n711n8p3n511n5p4n89n3p4n410n7p1n819n7p5n59n9p4n311n0p3n7

* Summed feature 3 comprises 16:1ω7c or 15 iso 2-OH or both. Summed feature 7 comprises 18:1ω7c,18:1ω9t or 18:1ω12t or any combination of these fatty acids.

sequence of B. ambifaria isolates LMG 19182T,R- of other members of the B. cepacia complex (Table 3). 5142 and R-8863 with sequences of other members of In particular, the small relative amounts of 16:0 and the B. cepacia complex confirmed the close relationship 17:0 cyclo and the large relative amounts of summed of B. ambifaria to the other members of the B. cepacia feature 3 and summed feature 7 present in all B. complex, but showed that B. ambifaria formed a ambifaria strains investigated are good characteristics different phylogenetic subgroup. In addition, DNA– for identification. All B. ambifaria strains investigated DNA hybridizations between B. ambifaria LMG could be identified unambiguously using the recA 19182T and representative strains of the other members gene-derived PCR primers BCRGC1 and BCRGC2. of the B. cepacia complex revealed only low or In addition, B. ambifaria strains can be identified by intermediate binding values (24–58%), while DNA– AFLP fingerprinting, DNA–DNA hybridizations and DNA binding values towards other B. ambifaria recA gene sequencing. Using the recently developed strains (R-8863, R-5142, R-9917 and LMG 11351) rRNA-based PCR assays (LiPuma et al., 1999) and were significantly higher (68–95%) (Table 2). 16S rDNA DdeI RFLP (Mahenthiralingam et al., 2000), B. ambifaria strains can be distinguished from The phylogenetic allocation and intermediate DNA– B. multivorans, B. vietnamiensis and B. cepacia genom- DNA binding values towards members of the B. ovar VI but not from the other members of the B. cepacia complex indicated unambiguously that B. cepacia complex. ambifaria represents a seventh genomovar within the B. cepacia complex. Description of Burkholderia ambifaria sp. nov. Identification of B. ambifaria Burkholderia ambifaria (am.bi.fahria. M.L. adj. ambi- farius that has two sides, of double meaning, am- Biochemically, B. ambifaria is very similar to the other biguous; referring to the fact that isolates of this members of the B. cepacia complex and to B. gladioli. species can be used for biological control purposes but However, several tests allowed the identification of B. also cause infections in humans). ambifaria strains. The most important differential tests are growth at 42 mC, β-galactosidase, lysine decarb- Cells are Gram-negative, non-sporulating, straight oxylase and ornithine decarboxylase activity, acidi- rods. All strains grow on MacConkey agar. Oxidase, fication of sucrose and β-haemolysis (Table 3). Pre- β-galactosidase and lysine decarboxylase activities are viously, it has been reported that fatty acid analysis is present. No ornithine decarboxylase, tryptophanase, generally not a good method for the differentiation of arginine dihydrolase or urease activity. Growth is members of the B. cepacia complex (Vandamme et al., observed at 30 and 37 mC. Positive for assimilation of 1997; Coenye et al., 2001). However, the fatty acid glucose, -arabinose, -mannose, -mannitol, N- profile of B. ambifaria is markedly different from those acetylglucosamine, -gluconate, caprate, adipate, -

1488 International Journal of Systematic and Evolutionary Microbiology 51 Burkholderia ambifaria sp. nov. malate, citrate and phenylacetate, but not maltose. broecke, K., Gillis, M., Speert, D. P. & Vandamme, P. (2001). Positive for acidification of glucose, maltose, lactose, Burkholderia cepacia genomovar VI, a new member of the xylose and adonitol. The GjC content is 65n5– Burkholderia cepacia complex isolated from cystic fibrosis 66n8 mol%. The small relative amounts of fatty acids patients. Int J Syst Evol Microbiol 51, 271–279. 16:0 and 17:0 cyclo and the large relative amounts of Eisen, J. A. (1995). The RecA protein as a model molecule for summed feature 3 and summed feature 7 present in all molecular systematic studies of bacteria: comparison of trees of strains investigated are good characteristics for iden- RecAs and 16S rRNAs from the same species. J Mol Evol 41, tification (see Table 3). Strains have been isolated from 1105–1123. the environment and from the respiratory tract of CF Ezaki, T., Hashimoto, Y. & Yabuuchi, E. (1989). Fluorometric patients. The finding that this new taxon includes both deoxyribonucleic acid-deoxyribonucleic acid hybridization in strains isolated from CF patients and potentially useful microdilution wells as an alternative to membrane filter hybridization in which radioisotopes are used to determine biocontrol strains supports the general consensus that genetic relatedness among bacterial strains. Int J Syst Bacteriol the large-scale use of biocontrol strains belonging to 39, 224–229. this taxon is ill-advised until more is known about their Fox, G. E., Wisotzkey, J. D. & Jurtshuk, P., Jr (1992). How close is potential pathogenic mechanisms. close: 16S rRNA sequence identity may not be sufficient to T T The type strain is LMG 19182 (l CCUG 44356 ). guarantee species identity. Int J Syst Bacteriol 42, 166–170. Phenotypic characteristics of the type strain are the Gilligan, P. H. (1991). Microbiology of airway disease in patients same as those described above for the species. In with cystic fibrosis. Clin Microbiol Rev 4, 35–51. addition, the type strain is characterized by the Gillis, M., Van Van, T., Bardin, R. & 7 other authors (1995). presence of nitrate reduction, gelatinase activity, Polyphasic taxonomy in the genus Burkholderia leading to an growth at 42 mC and acidification of sucrose and the emended description of the genus and proposition of Burk- absence of aesculin hydrolysis and pigment pro- holderia vietnamiensis sp. nov. for N#-fixing isolates from rice in duction. Its GjC content is 66n8 mol% and its 16S Vietnam. Int J Syst Bacteriol 45, 274–289. rDNA and recA gene sequences are available in Govan, J. R. W. & Deretic, V. (1996). Microbial pathogenesis in GenBank under accession nos AF043302 and cystic fibrosis: mucoid Pseudomonas aeruginosa and Burk- AF323985, respectively. Strain LMG 19182T was holderia cepacia. Microbiol Rev 60, 539–574. isolated from the rhizosphere of pea plants in Govan, J. R. W. & Vandamme, P. (1998). Agricultural and medical Wisconsin, USA, in 1985. Strains LMG 19182T,LMG microbiology: a time for bridging gaps. Microbiology 144, 6991, LMG 19465, LMG 19466, LMG 19467, LMG 2373–2375. 11351, LMG 17828 and LMG 17829 have been Govan, J. R. W., Hughes, J. E. & Vandamme, P. (1996). Burk- deposited in the BCCM\LMG Bacteria Collection, holderia cepacia: medical, taxonomic and ecological issues. J Universiteit Gent, Ghent, Belgium. Med Microbiol 45, 395–407. Govan, J. R. W., Balandreau, J. & Vandamme, P. (2000). Burk- holderia cepacia – friend AND foe. ASM News 66, 124–125. ACKNOWLEDGEMENTS Hebbar, K. P., Martel, M. H. & Heulin, T. (1998). Suppression of T.C. acknowledges the support received from the Vlaams pre- and postemergence damping-off in corn by Burkholderia Instituut voor Bevordering van Wetenschappelijk-techno- cepacia. Eur J Plant Pathol 104, 29–36. logisch onderzoek in de Industrie (Belgium) in the form of a Hedegaard, J., Steffensen, S. A. de A., Nørskov-Lauritsen, N., bursary for advanced study. P.V. and M.G. are indebted to Mortensen, K. K. & Sperling-Petersen, H. U. (1999). Identification the Fund for Scientific Research-Flanders (Belgium) for a of Enterobacteriaceae by partial sequencing of the gene en- position as a postdoctoral fellow and for research and coding translation initiation factor 2. Int J Syst Bacteriol 49, personnel grants, respectively. J.L. and D.P.S. acknowledge 1531–1538. the financial support received from the Cystic Fibrosis Henry, D. A., Campbell, M. E., LiPuma, J. J. & Speert, D. P. (1997). Foundation (USA) and the Canadian Cystic Fibrosis Foun- Identification of Burkholderia cepacia isolates from patients dation, respectively. We acknowledge the financial support with cystic fibrosis and use of a simple new selective medium. J received from the Cystic Fibrosis Trust (UK) (grant RS15 to Clin Microbiol 35, 614–619. P.V. and grant PJ472 to E.M.) and wish to thank Johan Goris for the determination of the GjC content and Henry, D. A., Mahenthiralingam, E., Vandamme, P., Coenye, T. & Stefanie Van Trappen and Julie Fadden for excellent Speert, D. P. (2001). Phenotypic methods for determining assistance. genomovar status of the Burkholderia cepacia complex. J Clin Microbiol 39, 1073–1078. Holmes, A., Govan, J. & Goldstein, R. (1998). Agricultural use of REFERENCES Burkholderia (Pseudomonas) cepacia: a threat to human health? Bowers, J. H. & Parke, J. L. (1993). Epidemiology of Pythium Emerg Infect Dis 4, 221–227. damping-off and Aphanomyces root rot of peas after seed Hugh, R. & Leifson, E. (1953). The taxonomic significance of treatment with bacterial agents for biological control. Phyto- fermentative versus oxidative metabolism of carbohydrates by pathology 83, 1466–1473. various gram-negative bacteria. J Bacteriol 66, 24–26. Coenye, T., Schouls, L. M., Govan, J. R. W., Kersters, K. & King, E. B. & Parke, J. L. (1993). Biocontrol of Aphanomyces root Vandamme, P. (1999). Identification of Burkholderia species and rot and Pythium damping-off by Pseudomonas cepacia AMMD genomovars from cystic fibrosis patients by AFLP finger- on four pea cultivars. Plant Dis 77, 1185–1188. printing. Int J Syst Bacteriol 49, 1657–1666. LiPuma, J. J. (1998a). Burkholderia cepacia. Management issues Coenye, T., LiPuma, J. J., Henry, D., Hoste, B., Vandemeule- and new insights. Clin Chest Med 19, 473–486.

International Journal of Systematic and Evolutionary Microbiology 51 1489 T. Coenye and others

LiPuma, J. J. (1998b). Burkholderia cepacia epidemiology and Segonds, C., Heulin, T., Marty, N. & Chabanon, G. (1999). pathogenesis: implications for infection control. Curr Opin Differentiation of Burkholderia species by PCR-restriction Pulm Med 4, 337–341. fragment length polymorphism analysis of the 16S rRNA gene LiPuma, J. J. & Mahenthiralingam, E. (1999). Commercial use of and application to cystic fibrosis isolates. J Clin Microbiol 37, Burkholderia cepacia. Emerg Infect Dis 5, 305–306. 2201–2208. LiPuma, J. J., Dulaney, B. J., McMenamin, J. D., Whitby, P. W., Speert, D. P., Bond, M., Woodman, R. C. & Curnutte, J. T. (1994). Stull, T. L., Coenye, T. & Vandamme, P. (1999). Development of Infection with Pseudomonas cepacia in chronic granulomatous rRNA-based PCR assays for identification of Burkholderia disease: role of nonoxidative killing by neutrophils in host cepacia complex isolates recovered from cystic fibrosis patients. defense. J Infect Dis 170, 1524–1531. J Clin Microbiol 37, 3167–3170. Stead, D. E. (1992). Grouping of plant-pathogenic and some McLoughlin, T. J., Quinn, J. P., Bettermann, A. & Bookland, R. other Pseudomonas spp. by using cellular fatty acid profiles. Int (1992). Pseudomonas cepacia suppression of sunflower wilt J Syst Bacteriol 42, 281–295. fungus and role of antifungal compounds in controlling the Vandamme, P., Vancanneyt, M., Pot, B. & 10 other authors disease. Appl Environ Microbiol 58, 1760–1763. (1992). Polyphasic taxonomic study of the emended genus Mahenthiralingam, E., Bischof, J., Byrne, S. K., Radomski, C., Arcobacter with Arcobacter butzleri comb. nov. and Arcobacter Davies, J. E., Av-Gay, Y. & Vandamme, P. (2000). DNA-based skirrowii sp. nov., an aerotolerant bacterium isolated from diagnostic approaches for identification of Burkholderia cepacia veterinary specimens. Int J Syst Bacteriol 42, 344–356. complex, Burkholderia vietnamiensis, Burkholderia multivorans, Vandamme, P., Holmes, B., Vancanneyt, M. & 8 other authors Burkholderia stabilis, and Burkholderia cepacia genomovars I (1997). Occurrence of multiple genomovars of Burkholderia and III. J Clin Microbiol 38, 3165–3173. ! cepacia in cystic fibrosis patients and proposal of Burkholderia Marechal, J., Clement, B., Nalin, R., Gandon, C., Orso, S., Cvejic, multivorans sp. nov. Int J Syst Bacteriol 47, 1188–1200. J. H., Bruneteau, M., Berry, A. & Normand, P. (2000). A recA gene phylogenetic analysis confirms the close proximity of Frankia to Vandamme, P., Mahenthiralingam, E., Holmes, B., Coenye, T., Acidothermus. Int J Syst Evol Microbiol 50, 781–785. Hoste, B., De Vos, P., Henry, D. & Speert, D. P. (2000). Iden- tification and population structure of Burkholderia stabilis sp. Mesbah, M., Premachandran, U. & Whitman, W. B. (1989). Precise nov (formerly Burkholderia cepacia genomovar IV). J Clin measurement of the GjC content of deoxyribonucleic acid by Microbiol 38, 1042–1047. high-performance liquid chromatography. Int J Syst Bacteriol 39, 159–167. Venkateswaran, K., Dollhopf, M. E., Aller, R., Stackebrandt, E. & Nealson, K. H. (1998). Shewanella amazonensis sp. nov., a novel Palys, T., Berger, E., Mitrica, I., Nakamura, L. K. & Cohan, F. M. metal-reducing facultative anaerobe from Amazonian shelf (2000). Protein-coding genes as molecular markers for ecologic- Int J Syst Bacteriol 48 ally distinct populations: the case of two Bacillus species. Int J muds. , 965–972. Syst Evol Microbiol 50, 1021–1028. Vidaver, A. K., Doyle, M. P., Gerone, P. J. & 7 other authors Parke, J. L. (1990). Population dynamics of Pseudomonas cepacia (1999). Burkholderia cepacia – friend or foe? ASM News 65, in the pea spermosphere in relation to biocontrol of Pythium. 587. Phytopathology 80, 1307–1311. Xi, K., Stephens, J. H. G. & Verma, R. (1996). Application of Parke, J. L., Rand, R. E., Joy, A. E. & King, E. B. (1991). Biological formulated rhizobacteria against root rot of field pea. Plant control of Pythium damping-off and Aphanomyces root rot of Pathol 45, 1150–1158. peas by application of Pseudomonas cepacia or P. fluorescens to Yamamoto, S. & Harayama, S. (1995). PCR amplification and seed. Plant Dis 75, 987–992. direct sequencing of gyrB genes with universal primers and their Pitcher, D. G., Saunders, N. A. & Owen, R. J. (1989). Rapid application to the detection and taxonomic analysis of Pseudo- extraction of bacterial genomic DNA with guanidium thio- monas putida strains. Appl Environ Microbiol 61, 1104–1109. cyanate. Lett Appl Microbiol 8, 109–114. Yamamoto, S. & Harayama, S. (1998). Phylogenetic relationships Saitou, N. & Nei, M. (1987). The neighbor-joining method: a new of Pseudomonas putida strains deduced from the nucleotide method for reconstructing phylogenetic trees. Mol Biol Evol 4, sequences of gyrB, rpoD and 16S rRNA genes. Int J Syst 406–425. Bacteriol 48, 813–819.

1490 International Journal of Systematic and Evolutionary Microbiology 51