Edited by W. Wade
A novel strategy for the isolation and identification of environmental Burkholderia cepacia complex bacteria
Version of Record online: 9 JAN 2006
FEMS Microbiology Letters
Volume 249, Issue 2, pages 303–307, August 2005
How to Cite
Vanlaere, E., Coenye, T., Samyn, E., Van den Plas, C., Govan, J., de Baets, F., de Boeck, K., Knoop, C. and Vandamme, P. (2005), A novel strategy for the isolation and identification of environmental Burkholderia cepacia complex bacteria. FEMS Microbiology Letters, 249: 303–307. doi: 10.1016/j.femsle.2005.06.026
- Issue online: 9 JAN 2006
- Version of Record online: 9 JAN 2006
- Received 22 March 2005, Revised 7 June 2005, Accepted 13 June 2005
- Burkholderia cepacia complex;
- Cystic fibrosis;
The purpose of this study was to develop a novel strategy for the isolation and identification of Burkholderia cepacia complex bacteria from the home environment of cystic fibrosis (CF) patients. Water and soil samples were enriched in a broth containing 0.1%l-arabinose, 0.1%l-threonine, and a mixture of selective agents including 1 μg ml−1 C-390, 600 U ml−1 polymyxin B sulfate, 10 μg ml−1 gentamycin, 2 μg ml−1 vancomycin and 10 μg ml−1 cycloheximide. On selective media (consisting of the same components as above plus 1.8% agar), several dilutions of the enrichment broth were inoculated and incubated for 5 days at 28°C. Isolates with different randomly amplified polymorphic DNA patterns were inoculated in Stewart's medium. Putative B. cepacia complex bacteria were confirmed by means of recA PCR and further identified by Hae III-recA restriction fragment length polymorphism analysis. Our results suggest that these organisms may be more widespread in the home environment than previously assumed and that plant associated soil and pond water may be reservoirs of B. cepacia complex infection in CF patients.
Cystic fibrosis (CF) is the most common, life threatening and recessively inherited disease in Caucasian populations, with a carrier rate of 1 in 25 and an incidence of 1 in 2500 live births . CF is the result of a mutation affecting the cystic fibrosis transmembrane conductance regulator (CFTR) gene, which encodes a chloride channel. The CFTR protein is essential for the regulation of salt and water movements across membranes. Improper or faulty regulation results in thickened secretions in organs such as the lung and pancreas . There is a wide range of clinical presentations, but most of morbidity and more than 90% of the mortality in CF is related to chronic pulmonary disease [2,3]. Typical organisms causing infections in CF patients are Haemophilus influenzae, Staphylococcus aureus, Pseudomonas aeruginosa, Stenotrophomonas maltophilia and Burkholderia cepacia complex [4,5]. Since the 1980s, it has been recognized that B. cepacia (first isolated as a plant pathogen in the 1940s) is an important pathogen for individuals with CF. Recent studies have shown that organisms presumptively identified as B. cepacia constitute different species, collectively known as the B. cepacia complex [6,7]. The B. cepacia complex comprises at least nine species: B. cepacia, Burkholderia multivorans, Burkholderia cenocepacia, Burkholderia stabilis, Burkholderia vietnamiensis, Burkholderia dolosa, Burkholderia ambifaria, Burkholderia anthina and Burkholderia pyrrocinia. Although B. cepacia complex organisms only infect a small proportion of CF patients, they have a significant impact on morbidity and mortality . Furthermore, B. cepacia complex infection can lead to the development of ‘cepacia syndrome’, a rapidly fatal pneumonia with associated bacteraemia [10,11]. Each of the nine B. cepacia complex species contains isolates recovered from respiratory secretions of CF patients, but B. multivorans and B. cenocepacia are predominant. Several studies confirmed the evidence for inter-patient spread and nosocomial infections [8,12] and as a result, guidelines were issued to reduce the risk of acquisition (including segregation of colonized patients) [11,13]. These patient management efforts have a significant psycho-social impact on the CF community [11,14–16]. Despite these infection control measures, novel infections with B. cepacia complex organisms continue to occur.
B. cepacia complex bacteria occur naturally in soil and in water [12,17–23]. Most probably the environment acts as a reservoir for novel B. cepacia complex infections [24,25], but their prevalence and distribution in the environment is not clear. Studies addressing this issue have been hampered by difficulties in isolating B. cepacia complex organisms from the environment in spite of the availability of several selective media [22,26–31]. The main problems of all these media are a too limited selectivity and sensitivity.
Vermis et al.  examined the assimilation of unusual carbon sources, the antimicrobial susceptibilities and the ability to grow on B. cepacia selective media for a large variety of clinical and environmental isolates belonging to the nine B. cepacia complex species. B. cepacia complex organisms were very heterogeneous in the utilization of carbon sources and in their susceptibilities for antimicrobial components. Overall, a combination of l-arabinose and l-threonine or l-arabinose and d-cellobiose proved the most suitable carbon sources to support growth of all B. cepacia complex strains. Based on these findings, Vermis et al.  developed a selective enrichment broth comprising two carbon sources. Selectivity was based on the unusual nature of the carbon sources and on C-390 and polymyxin B. Following enrichment, putative B. cepacia complex organisms were subcultured on a diagnostic O/F agar supplemented with gentamycin.
The purpose of the present study was use the data by Vermis et al. as a starting point for a new strategy for the isolation of B. cepacia complex bacteria from the environment and to examine potential sources of B. cepacia complex infections in CF patients.
2Materials and methods
2.1Isolation of B. cepacia complex strains
Water and soil samples were collected from the house and garden of six Belgian B. cepacia complex positive CF patients. From each patient's house, tap water was collected and, if available, water from an aquarium, garden pond or flower vases. Soil was collected from different sites in the garden such as soil of flowering plants and vegetables and also from plants grown inside the house. For the water samples, 100 ml was filtered (0.45 μm HA filter type, Millipore) (except for tap water of which 1 l was used) and the filter was placed in 20 ml of a selective enrichment broth. For the soil samples, 5 g of soil were placed directly in 90 ml of the enrichment broth. The selective enrichment broth contained 0.1%l-arabinose (Sigma), 0.1%l-threonine (Sigma), 0.5% NaCl (Merck), 0.02% MgSO4 (UCB), 0.1% NH4H2PO4 (Sigma), 0.1% K2HPO4 (Merck), 0.002% yeast extract (Oxoid) and a mixture of selective agents including 1 μg ml−1 C-390 (Biosynth AG), 600 U ml−1 polymyxin B sulfate (Sigma), 10 μg ml−1 gentamycin (Sigma), 2 μg ml−1 vancomycin (Sigma) and 10 μg ml−1 cycloheximide (Sigma). Following incubation at 28°C for 5 days, serial dilutions were made on selective agar plates (formula as above plus 1.8% agar) using a spiroplater (Led Techno). Subsequently, colonies differing in colour and morphology were subcultured on Tryptone Soy Agar (Oxoid).
2.2Identification and typing of B. cepacia complex strains
DNA was prepared from all isolates and randomly amplified polymorphic DNA (RAPD) typing was performed as described previously [33,34]. Only isolates with different patterns (assessed by visual comparison) were further investigated.
Isolates were subsequently inoculated in Stewart's medium [35,36] to investigate the utilization of glucose and arginine. All nine B. cepacia complex species oxidise glucose but do not hydrolyse arginine (reaction pattern in the tube is a yellow slope and a green butt) (manuscript submitted elsewhere). They were further identified using Hae III-recA restriction fragment length polymorphism (RFLP) analysis. RFLP analyses were performed as described previously . Electrophoretic separation of the restriction fragments was however performed by polyacrylamide gel electrophoresis. Briefly, 10 ml of a 40% (w/v), 29:1 acrylamide: bisacrylamide monomer solution (National Diagnostics) was mixed with 5 ml 10X TBE buffer and 34.5 ml sterile MilliQ Ultrapure water to obtain an 8% polyacrylamide gel. PhiX174 DNA/Hinf I (Promega) comprising 20 DNA fragments ranging from 24 to 726 base pairs was used as a size marker and as reference marker and internal standard for normalization using Bionumerics 4.0 software package (Applied Maths). For all recA PCR positive isolates, the obtained RAPD pattern was compared with a database stored in Bionumerics 4.0 containing RAPD patterns of all B. cepacia complex isolates recovered from Belgian CF patients. All non-B. cepacia complex isolates were further identified by means of cellular fatty acid methyl ester analysis (MIS system) and protein electrophoresis as described previously [38,39].
3.1Isolation of B. cepacia complex strains
Using the approach described above, we recovered 267 isolates from 44 environmental samples (26 water samples and 18 soil samples). After a visual comparison of the obtained RAPD patterns, 170 isolates showed unique patterns and were thus considered as distinct strains that were further investigated. Following incubation in Stewart's medium, 37 strains were identified as putative B. cepacia complex bacteria, as they oxidised glucose but did not hydrolyse arginine.
3.2Identification and typing of B. cepacia complex strains
Eleven strains were confirmed as B. cepacia complex. They represented four genomovars (B. multivorans, B. cenocepacia, B. pyrrocinia and B. anthina) as determined by using recA-RFLP and were recovered from the home environment of four patients (see Table 1). The strains were recovered from soil samples in- and outside the house, except one from pond water. The remaining 26 putative B. cepacia complex strains showed no amplification product after recA PCR (false positives after Stewart) and were further identified, together with the other 133 strains by means of cellular fatty acid analysis and protein profiling. Table 2 shows all genera identified by cellular fatty acid analyses (a total number of 134 strains) and the subsequent species level identifications obtained after protein profiling. Twenty-six strains were not further investigated with protein profiling, because there were no reference strains for these bacteria in the database. Sixteen of these isolates were identified as Sphingobacterium sp. with a very high identification score (>0.7); ten were identified as Zoogloea sp. with intermediate identification scores (>0.4). None of the environmental B. cepacia complex strains showed RAPD patterns that were similar to those recovered from patients (data not shown).
|Patient 1||B. cenocepacia||H||Soil of a grass plant on the garden table|
|B. cenocepacia||I||Soil of a grass plant on the garden table|
|B. cenocepacia||I||Soil of a grass plant on the garden table|
|Patient 2||B. anthina||T||Soil of a green plant on the window sill inside the house|
|Patient 3||B. anthina||AS||Soil of a succulent in the kitchen|
|B. multivorans||F||Soil of a succulent in the veranda|
|B. multivorans||F||Soil of a succulent in the veranda|
|B. anthina||T||Soil of a composed flowerpot|
|B. anthina||AS||Soil of a papyrus plant|
|Patient 4||B. pyrrocinia||Se6||Soil near the pond in the garden|
|B. pyrrocinia||Se6||Pond water|
|Genus identification after FAME||Number of strains (n= 34)||Species identification after protein profiling||Number of strains|
|Variovorax||26 (19.5%)||Comamonas acidovorans||22/26|
|Burkholderia||25 (18.5%)||Pandoraea sp. nov.b||14/25|
|Ralstonia||22 (16.5%)||Ralstonia basilensis||21/22|
|Chryseobacterium||18 (13.5%)||Chryseobacterium sp. nov.b||18/18|
|Sphingobacterium||16 (12%)||Not investigated||16/16|
|Zoogloea||10 (7.5%)||Not investigated||10/10|
|Pseudomonas||10 (7.5%)||Herbaspirillum lusitanum||4/10|
|Herbaspirillum sp. nov.b||1/10|
|Aquaspirillum||7 (5%)||H. lusitanum||7/7|
The isolation and identification strategy described above allowed the recovery of eleven B. cepacia complex strains from four different genomovars from the home environment of four Belgian CF patients. They were found in seven soil samples and one water sample. B. cepacia complex bacteria were not recovered from samples of the home environment of two patients. There is no apparent reason for this discrepancy. Following incubation in the selective enrichment broth strains belonging to a limited number of other genera (Pandoraea, Chryseobacterium, Comamonas, Ralstonia, Herbaspirillum and Pseudomonas) were also recovered from the selective agar plates.
The strategy, based on RAPD fingerprinting, growth in Stewart's medium, Hae III-recA-RFLP, cellular fatty acid analysis and protein profiling was effective, but still time-consuming and laborious. The use of RAPD fingerprinting and Stewart's medium reduced the number of putative B. cepacia complex strains to be investigated and thus the workload. Strains of only a few other genera showed the same reaction pattern in Stewart's medium as B. cepacia complex strains, i.e. Herbaspirillum sp. and Sphingobacterium sp. Non-B. cepacia complex strains were first analysed by cellular fatty acid analysis. This is an expensive and rapid screening method for which an extensive commercial database (MIS system) is available [38,39]. In our experience, the technique typically yields reliable identification results up to the genus level and provides only a tentative species level identification. Following cellular fatty acid analysis, the strains were further identified to the species level using whole cell protein profiling and an in-house developed database. Sodium dodecyl sulphate–polyacrylamide gel electrophoresis of whole cell proteins is a suitable and relatively simple technique for species level identification, but not very time effective . The protein electrophoretic identification results demonstrated that identification results at the genus level by using fatty acid data were not always reliable, in particular for a number of closely related genera belonging to the β-Proteobacteria.
The key question, whether colonised patients contaminate their home environment or whether the home environment functions as a reservoir for infection of the patient remains difficult to answer. However, the considerable diversity of B. cepacia complex bacteria in the limited number of samples tested, and the absence of the patient's isolate in the environmental samples favour the latter possibility. Therefore the present evidence suggests that plant associated soil and water from house and garden may be reservoirs of B. cepacia complex infection in CF patients.
We would like to thank the ‘Belgische Vereniging voor Strijd tegen Mucoviscidose’ for financial support. T.C. and P.V. are indebted to the Fund for Scientific Research – Flanders (Belgium) for a position as postdoctoral fellow and research grants, respectively.
- Ursing, J.B., Rossello-Mora, R.A., Garcia-Valdes, E., Lalucat, J.Taxonomic note: a pragmatic approach to the nomenclature of phenotypically similar genomics groups Int. J. Syst. Bacteriol. 45 1995 604
- UK Cystic Fibrosis Trust Infection Control Group. The Burkholderia cepacia complex. Suggestions for prevention and Infection Control. Cystic Fibrosis Trust, Bromley, September 2004.
- 2000) Burkholderia cepacia-friend and foe. Am. Soc. Microbiol. News 66, 124–125., , (
- Burbage, D.A., Sasser, M.A medium selective for Pseudomonas cepacia Phytopath 76 1982 706
- Govan, J.R.W.Pseudomonas, Stenotrophomonas, Burkholderia Collee, J.G., Fugse, A.G., Masmion, B.P., Simmons, A., Eds. Practical Medical Microbiology 14th ed. 1996 Churchill Livingstone Edinburgh 413 422
- 1993) Transfer of Erwinia ananas (synonym, Erwinia uredovora) and Erwinia stewartii to the genus Pantoea emend. as Pantoea ananas (Serrano 1928) comb. nov. and Pantoea stewartii (Smith 1898) comb. nov., respectively, and description of Pantoea stewartii subsp. indologenes subsp. nov. Int. J. Syst. Bacteriol. 43, 162–173., , (
- Pot, B., Vandamme, P., Kersters, K. (1994) Analysis of electrophoretic whole-organisms protein fingerprinting. In: Chemical Methods in Prokaryotic Systematics (Goodfellow, M., O'Donnell, A.G., Eds.), pp.493–521 Wiley, Chichester, UK.