- Top of page
- Materials and methods
Twenty-one strains of Burkholderia cepacia isolated from the environment, and 21 clinical strains isolated principally from sputum of cystic fibrosis (CF) patients, were characterized genotypically by macrorestriction analysis (genome fingerprinting) and PCR ribotyping, and phenotypically by susceptibility to antibiotics and the ability to macerate onion tissue. The plasmid content of the strains was also investigated. Environmental isolates showed a high degree of genetic variability, all strains differing from both one another and the CF isolates. The CF isolates were less variable, with common strains found in patients attending three geographically distinct CF centres. Phenotypic variation was found both within and between CF and environmental strains. Generally, CF isolates displayed higher levels of antibiotic resistance, while the ability to macerate onion tissue was more prevalent amongst environmental isolates. Plasmids were more frequently found in CF isolates, but were of similar size in both groups of strains. Such variability is not surprising in view of the existence of multiple genomovars within the B. cepacia complex.
The Gram-negative bacterium Burkholderia cepacia was originally described as a phytopathogen of onion ( Burkholder 1950) and subsequently as a saprophyte in soils and waters ( Morris & Roberts 1959). More recently, it has emerged as an opportunistic pathogen of man, particularly among cystic fibrosis (CF) and chronic granulomatous disease patients ( Govan et al. 1996 ), as a nosocomially acquired infection in a hospital intensive care unit ( Pegues et al. 1996 ), and amongst oncology patients ( Pegues et al. 1993 ). Burkholderia cepacia displays intrinsic resistance to a range of antibiotics ( Wilkinson & Pitt 1995), shows diverse degradative properties including the degradation of recalcitrant polychlorinated aromatic compounds, and produces a number of compounds antagonistic to phytopathogenic fungi. Such properties have led to interest in B. cepacia as a possible agent of bioremediation or biocontrol ( Govan et al. 1996 ).
Characterization of B. cepacia infecting CF patients in epidemiological studies in the UK ( Govan et al. 1996 ; Pitt et al. 1996 ) and the USA ( Johnson et al. 1994 ) has shown strong evidence of an epidemic strain of B. cepacia, and of person-to-person transmission amongst CF patients. The risk posed to CF patients from environmental reservoirs of B. cepacia strains and the relationship between clinical and environmental isolates is unclear, with contradictory evidence for phenotypic differences between isolates from CF patients and the environment ( Govan et al. 1996 ). Isolates from CF patients and nosocomially acquired infections have been shown to be incapable of causing the maceration of onion tissue in vitro whereas environmental isolates have been shown to macerate onion tissue and display high levels of pectinolytic activity ( Gonzalez 1979 ; Bevivino et al. 1994 ). However, a number of other CF isolates, including the epidemic strain, have been shown to cause maceration of onion by other workers ( Butler et al. 1995 ). The inability of environmental isolates to adhere to human uroepithelial cells has been suggested as evidence that environmental B. cepacia is incapable of causing infection in man ( Bevivino et al. 1994 ), though more recent evidence has suggested that the environment may be a potential source of infection to CF patients ( Cazzola et al. 1996 ). Such contradictory evidence has led to caution being advised in the use of B. cepacia as an agent of bioremediation or biocontrol until the relationship between clinical and environmental isolates of B. cepacia has been fully assessed ( Butler et al. 1995 ; Govan et al. 1996 ).
Phenotypic differences between B. cepacia strains such as those described above, along with genetic variability in both environmental and clinical isolates ( Pitt et al. 1996 ; Wise et al. 1996 ), have led to suggestions that the phylogeny of the species should be revised ( Govan et al. 1996 ; Vandamme et al. 1997 ). Yohalem & Lorbeer (1994 ) placed B. cepacia isolates into four groups or phenoms on the basis of a range of phenotypic properties. Genomic fingerprinting by pulsed field gel electrophoresis (PFGE) and arbitrarily primed PCR (AP-PCR) found four distinct genetic groups or genomovars amongst clinical and environmental isolates of B. cepacia ( Vandamme 1995; Revets et al. 1996 ). Genomovars have been defined as genetically different groups of bacteria that are indistinguishable by current phenotypically based typing methods ( Ursing et al. 1995 ). Recent evidence has suggested that B. cepacia is formed by a complex of five genomic species: B. cepacia genomovars I, II and IV, B. vietnamensis and the newly proposed species, B. multivorans ( Vandamme et al. 1997 ).
Although strains of B. cepacia infecting CF patients in the UK and North America have been well characterized phenotypically and by molecular typing methods ( Pitt et al. 1996 ; Henry et al. 1997 ), there has been little charaterization of environmental isolates found in the UK. Butler et al. (1995) investigated the phenotypic properties of 12 environmental isolates of B. cepacia and undertook genomic analysis by macrorestriction digests followed by PFGE. These isolates were found to differ both genetically and phenotypically from each other and from isolates infecting CF patients. However, most of these isolates were obtained from tropical houses in botanical gardens. These isolates could be described as coming from an unusual habitat in terms of the environmental population of B. cepacia that occurs in the UK, and CF patients are much less likely to come into contact with such isolates than with those occurring in the natural rural and urban environment of the UK. In this study, isolates obtained from the environment (soils, rhizosphere and waters) from both urban and rural areas were characterized phenotypically by their ability to macerate onion tissue in vitro ( Lelliott & Stead 1987) and their susceptibility to a range of antibiotics, genetically through molecular typing by PCR ribotyping, PFGE macrorestriction analysis, and through their plasmid content and size. These isolates were compared with B. cepacia strains isolated from CF patients living in the South Wales area, and with isolates from other European CF centres, to investigate relationships and variations between B. cepacia isolates from CF patients and those in the environment to which the patients may be exposed.
- Top of page
- Materials and methods
All environmental isolates tested in this study were found to differ genetically from each other, indicating that there is great genetic variability in the environmental population of B. cepacia in South Wales. High levels of variability have previously been described in environmental populations of B. cepacia ( Wise et al. 1996 ). Sixty-five unique electropherotypes (ETs) were found by multilocus enzyme electrophoresis in 217 B. cepacia isolates collected from a 5 km stretch of stream over a period of 32 d. It is also notable that genetic differences were found in isolates obtained from sites close together. Examples of this are PW1, PW2 and PW6, isolated over a 100 m stretch of stream, and PW5 and PW7, isolated from soil samples taken a few metres apart. The 12 environmental isolates from botanical gardens described by Butler et al. (1995) gave 11 different profiles by macrorestriction analysis, though it was notable that isolates from the same source differed in profile. This evidence strongly indicates high genetic variability amongst environmental populations of B. cepacia. More detailed genetic and phenotypic investigation, such as the use of fatty acid methyl ester analysis (FAME), would allow the genomovar status of the isolates obtained from the environment in this study to be determined.
No identity was found between the environmental B. cepacia isolates from South Wales and the B. cepacia isolates from patients attending the Cardiff CF centre, indicating that the environmental isolates differ genetically from those infecting CF patients. The environmental isolates were also found to differ from those obtained from other CF centres, from the species type strain NCPPB 2993 (ACTC 25416), and also from other environmental isolates obtained from the National Collection of Industrial and Marine Bacteria (NCIMB) and the National Collection of Plant Pathogenic Bacteria (NCPPB). The differences would suggest that environmental B. cepacia isolates may form a separate population to those infecting CF patients.
With the techniques used in this study, identity between all but one of the Cardiff CF isolates was shown. These isolates can be further subdivided by Taq I digestion of the PCR products ( Ryley et al. 1995 ). Nevertheless, the identity shown in this study between these strains, and colleagues from Scotland and Strasbourg, both by PCR ribotyping and macrorestriction analysis, would indicate that they are closely related if not identical. The presence of a common strain amongst these centres would indicate both a common initial source, and strong evidence of person-to-person transmission ( Govan et al. 1996 ).
Phenotypic differences in antibiotic resistance and the ability to cause onion maceration both between CF and environmental isolates, and within the populations, were also observed ( Tables 1 and 2). Levels of antibiotic resistance were generally higher amongst CF isolates, as found by Butler et al. (1995) . However, it is not clear whether such differences are fundamentally related to the biology of the strains themselves, or are due to the development of acquired resistance amongst the CF isolates which will have received considerable exposure to such drugs. Maceration of onion tissue was more common and generally more severe by environmental isolates, but was not exclusively caused by these isolates.
Plasmids were found in 55% of B. cepacia isolates, though they were more commonly found amongst the CF isolates (76% of isolates) than in environmental isolates (33% of isolates). Plasmids in excess of 100 kbp were harboured by both clinical and environmental isolates. Lennon & DeCicco (1991) suggested that large plasmids were associated with clinical isolates and high levels of antibiotic resistance but in this study, there was no evidence of large plasmids being more frequently harboured in CF isolates; indeed, many of the largest plasmids were found in environmental isolates. Strains which appeared identical by ribotyping and genomic fingerprinting were nevertheless found to differ in their plasmid content, e.g. the Cardiff CF isolates. The transfer, acquisition and assimilation of plasmid-borne genes is extensively documented, and the fact that closely related strains differ in their plasmid content could reflect differences in selective pressure, such as antibiotic treatment regimes, and differences in the transferable plasmid-bearing lung microflora of other species in CF patients. It is not known if the plasmids are sufficiently stable to be an aid in strain identification; the plasmids in our collection are as yet cryptic.
Three main points emerge from this study. First, the environmental population is genetically very diverse. There is also considerable phenotypic diversity amongst these strains. Secondly, the clinical population of B. cepacia is genetically much more homogeneous and is distinct from the environmental population. The final point that can be made is that the variation both between and within environmental and clinical groups may be the result of B. cepacia being a complex formed by five genomic species, with the clinical population being predominantly formed by B. cepacia genomovar III, B. vietnamensis and B. multivorans, while many environmental isolates, particularly phytopathogenic isolates, fall into genomovar I ( Govan et al. 1996 ; Revets et al. 1996 ; Vandamme et al. 1997 ). However, it would appear that such a grouping is not rigid, with genomovar I being described in CF patients and other clinical sources and some strains predominantly found in CF patients also occurring in soil ( Vandamme et al. 1997 ). Although this study suggests that environmentally occurring B. cepacia forms a separate population to that infecting CF patients, it is limited to only 21 environmental isolates and 21 clinical isolates. A more extensive study, particularly including a wider range of less related CF isolates, would be needed to confirm the validity of these findings. The apparent acquisition of B. cepacia from the environment by CF patients ( Cazzola et al. 1996 ), the close genetic relationship between the species type strain and the ET12 CF epidemic strain ( Johnson et al. 1994 ), the ability of some CF strains to cause onion maceration, and the overlap of genomic species between the environment and CF patients ( Vandamme et al. 1997 ), suggest that a greater understanding of relationships between clinical and environmental B. cepacia is needed. The occurrence of genomic species infecting CF patients in the environment and the ability of primarily environmental genomovars to infect man needs further investigation. The use of molecular typing techniques such as PCR and PFGE, together with animal models such as the cftrm1HGU mouse which develops lung disease when challenged with B. cepacia ( Davidson et al. 1995 ), may allow both a greater understanding of the relationship between clinical and environmental isolates of B. cepacia and their potential pathogenicity. Such information may answer the question of whether environmentally occurring B. cepacia pose a risk of infection to CF patients.