Diversity and significance of Burkholderia species occupying diverse ecological niches

Authors


*For correspondence. E-mail Tom.Coenye@UGent.be; Tel. (+32) 9 2645114; Fax (+32) 9 2645092.

Summary

Members of the genus Burkholderia are versatile organisms that occupy a surprisingly wide range of ecological niches. These bacteria are exploited for biocontrol, bioremediation and plant growth promotion purposes, but safety issues regarding human infections, especially in cystic fibrosis patients, have not been solved. This minireview gives an overview of the taxonomic and ecological diversity of the genus with particular emphasis on strains belonging to the Burkholderia cepacia complex and addresses the important question whether ‘good’ and ‘bad’ strains are actually the same.

The genus Burkholderia contains over 30 species, which occupy remarkably diverse ecological niches, ranging from contaminated soils to the respiratory tract of humans. The Burkholderia cepacia complex is ubiquitous in nature and can be found in soil, water (including sea water), the rhizosphere of plants, in humans and various animal species and in the hospital environment. Burkholderia cepacia complex isolates have been exploited for various purposes, including biological control of plant pathogens, bioremediation of recalcitrant xenobiotics and plant growth promotion. Unfortunately, some Burkholderia species have been involved in human infections, and safety issues regarding these human infections are hampering the wide-spread biotechnological applications. The purpose of this minireview is to give an overview of the remarkable diversity of the genus Burkholderia and to shed some light on the important question whether or not clinical and environmental B. cepacia complex isolates are the same. A phylogenetic tree based on 16S rRNA gene sequences, showing the positions of all of the Burkholderia species and representatives of related genera, is shown in Fig. 1.

Figure 1.

16S rDNA based phylogenetic tree showing the positions of all of the Burkholderia species and representatives of related genera. Scalebar indicates 5% sequence dissimilarity.

Diversity of the genus Burkholderia

The genus Burkholderia was created by Yabuuchi et al. (1992) to accommodate the former rRNA group II pseudomonads, excluding Pseudomonas pickettii and Pseudomonas solanacearum, which were transferred to the genus Ralstonia (Yabuuchi et al., 1995). Traditionally, Burkholderia species are known as plant pathogens and soil bacteria with two important exceptions, B. mallei and B. pseudomallei, which are primary pathogens for humans and animals. Our present knowledge on the natural diversity of members of this genus indicates that the range of interactions between these bacteria and their hosts is more complex, diverse, and, often, contradictory. The interactions of some species seem restricted to one type of host, whereas others have a much wider host range. The type of interaction may be that of a pathogen but can also be symbiotic, or both.

Burkholderia caryophylli, Burkholderia plantarii, Burkholderia glumae and Burkholderia andropogonis are the species that are known as plant pathogenic Burkholderia species. Burkholderia caryophylli is pathogenic for carnations (Dianthus caryophyllus) but also causes onion rot (Ballard et al., 1970; Palleroni, 1984). Burkholderia plantarii causes seedling blight of rice and forms the disease-causing substance tropolone, a non-benzoid aromatic compound with a seven-membered ring (Azegami et al., 1987; Urakami et al., 1994). Burkholderia glumae causes rot of rice grains and seedlings and has emerged during the past two decades as the most important bacterial pathogen of rice in Japan, Korea and Taiwan (Goto and Ohata, 1956; Kurita and Tabei, 1967; Uematsu et al., 1976; Urakami et al., 1994). Burkholderia andropogonis is the causative agent of stripe disease of sorghum (Andropogon sp.) and leaf spot of velvet bean (Stizolobium deeringianum). Two different pathovars (B. andropogonis pv. andropogonis and pv. stizolobii) have been described (Palleroni, 1984). Pseudomonas woodsii, an important pathogen of carnation (Smith, 1911), was recently shown to be a junior synonym of B. andropogonis (Coenye et al., 2001a).

The majority of Burkholderia species however, are known as soil bacteria, which exhibit different types of non-pathogenic interactions with plants. The ecological role of species like Burkholderia glathei, Burkholderia graminis, Burkholderia phenazinium, Burkholderia caribensis, Burkholderia caledonica, Burkholderia hospita, Burkholderia terricola and Burkholderia sacchari is largely unknown. Burkholderia glathei was proposed for strains isolated from fossil lateritic soils in Germany (Zolg and Ottow, 1975). Burkholderia phenazinium strains were isolated from soil and were characterized by the production of iodinin (Bell and Turner, 1973; Viallard et al., 1998). Strains belonging to B. graminis were first isolated by Viallard et al. (1998) from the rhizosphere of wheat, corn and pasture grasses during a field survey of French and Australian soils. Burkholderia caribensis was described by Achouak et al. (1999) for a group of exopolysaccharide-producing strains isolated from a vertisol on the island of Martinique. These bacteria constituted the dominant cultivable population in this soil. B. caledonica strains were isolated from the rhizosphere of different plants in Scotland (Coenye et al., 2001b). Some of these soil-borne Burkholderia species were detected in several soil types as transconjugants that acquired the plasmids pJP4 or pEMT1 (containing genes required for the degradation of 2,4-dichlorophenoxyacetic acid) following inoculation of a donor strain (Newby et al., 2000; Goris et al., 2002). A more detailed study of some of these transconjugants entailed the description of B. hospita and B. terricola (Goris et al., 2002). Both species were isolated as pJP4 or pEMT1 containing transconjugants of agricultural soil in Belgium; the latter has also been isolated from soil samples in Scotland, Italy and Spain (P. Vandamme, unpubl. obs.). The name B. sacchari was proposed for a single strain recovered from the soil of a sugar-cane plantation in Brazil (Brämer et al., 2001). This organism can produce polyhydroxyalkanoates from a wide range of carbon sources (Brämer et al., 2001; 2002).

Several Burkholderia species have developed intimate beneficial interactions with plants and colonize roots, stems and leaves. The ability to fix atmospheric nitrogen has been established in several burkholderias including named species like Burkholderia vietnamiensis (see below) and Burkholderia kururiensis[the name B. kururiensis was proposed for a single strain isolated from an acquifer polluted with trichloroethylene (Zhang et al., 2000)] (Gillis et al., 1995; Magalhaes Cruz et al., 2001; Estrada-De Los Santos et al., 2001), tentatively named species like ‘Burkholderia tropicalis’ and ‘Burkholderia brasilensis’, and unnamed strains probably representing several novel species (Estrada-De Los Santos et al., 2001). ‘Burkholderia tropicalis’ and ‘B. brasilensis’ have not been formally described but strains of these putative novel species have been recovered from various parts of banana and pineapple plants in South-America. Burkholderia tuberum and Burkholderia phymatum are novel species (Vandamme et al., 2002) reported to be capable of nitrogen fixation and nodulation in tropical legume plants (Moulin et al., 2001). The former species was isolated from root nodules of Aspalathus carnosa in South Africa; the latter from root nodules of Machaerium lunatum in French Guiana. The capacity to induce nodule formation was reported for two B. caribensis and one B. dolosa (B. cepacia genomovar VI, Vermis et al., 2003) isolate as well (Moulin et al., 2001). However, recent experiments with this B. dolosa isolate revealed that the nodA and nif genes could not longer be detected (C. Boivin-Masson, pers. comm.) suggesting that the isolate lost this capacity or that the original observation was based on experimental error. Although all of these species can be considered endosymbionts, they are all culturable on agar slants in normal laboratory conditions. Plant endosymbionts belonging to the genus Burkholderia, which remain uncultured despite numerous attempts have been reported in several studies. Endosymbionts of leaf galls of Psychotria kirkii were classified as Candidatus Burkholderia kirkii (Van Oevelen et al., 2002). Uncultured bacterial endosymbionts of the arbuscular mycorrhizal fungi belonging to the Gigasporaceae (Bianciotto et al., 1996; 2000) were recently formally classified as Candidatus Glomeribacter gigasporarum (Bianciotto et al., 2003). These bacteria are close relatives of the genus Burkholderia and were also reported to contain putative nif genes (Minerdi et al., 2001). Most remarkable, Van Borm et al. (2002) recently showed that the microflora inhabiting a pouch-shaped organ at the junction of the midgut and the intestine of Tetraponera ants partially consists of Burkholderia species, which are most likely involved in the oxidative recycling of nitrogen-rich metabolic waste. These organisms are phylogenetically closely related to B. fungorum and B. caledonica but have so far not been cultured and their exact taxonomic position remains uncertain.

Interactions between burkholderias and humans or animals are traditionally known for B. mallei and B. pseudomallei. The former causes glanders in horses, mules and donkeys and was one of the first biological weapons of the 20th century, being used by Germany during World War I (Wheelis, 1998). Glanders in humans is acquired from infected animals or by contact with organisms causing glanders via ingestion or inhalation (Palleroni, 1984). Burkholderia pseudomallei is the aetiological agent of melioidosis in humans and animals, and is endemic to South-east Asia, Northern Australia and temperate regions that border the equator (Dance, 1991). Humans can be infected by soil contamination of skin abrasions, ingestion or inhalation (Dance, 2000). Sporadically the organism is acquired via person-to-person or animal-to-person spread (Dance, 2000). Burkholderia pseudomallei is a saprophytic organism that routinely can be isolated from environmental niches like water, moist soil and rice paddies (Brook et al., 1997). These two species were shown to represent a single genomic species by DNA–DNA hybridization criteria (Rogul et al., 1970). However, the major differences in their biochemical activities and in the clinical symptoms and epidemiology justify this two-species concept. Recently, several environmental B. pseudomallei-like organisms were formally classified as Burkholderia thailandensis (Brett et al., 1998). In contrast to B. pseudomallei, B. thailandensis strains are not correlated with human disease and are avirulent in the Syrian golden hamster animal model (Brett et al., 1997). One of the B. thailandensis isolates which was obtained from a surface soil along a roadside in Thailand, was later reclassified as Burkholderia ubonensis (Yabuuchi et al., 2000). Phylogenetically, the latter species belongs to the B. cepacia complex (see below, Fig. 1).

Finally, several Burkholderia species occupy multiple niches, may have both pathogenic and symbiotic interactions with plants, and have become known as opportunistic pathogens in humans. Although they are not considered important pathogens for the normal human population, some are considered serious threats for specific patient groups such as cystic fibrosis patients. These species include all Burkholderia cepacia complex bacteria, Burkholderia gladioli and Burkholderia fungorum. Burkholderia gladioli strains have been isolated from decayed onions, Gladiolus sp., Iris sp. and rice, for which this species is believed to be pathogenic, and from various human clinical sources (Palleroni, 1984; Miyagawa and Kimura, 1989; Ura et al., 1996). Burkholderia gladioli consists of two pathovars (pv. gladioli and pv. allicola) as it appeared that it comprised organisms with different pathogenic capacities (Palleroni, 1984). Strains belonging to Burkholderia cocovenenans (also known as ‘Pseudomonas farinofermentans’) produce toxoflavin and bongkrekic acid and are responsible for cases of food poisoning caused by consumption of fermented corn flour (Zhao et al., 1995). Burkholderia cocovenenans was shown to be a junior synonym of B. gladioli (Coenye et al., 1999); the latter having nomenclatural priority. Similarly, also Pseudomonas antimicrobica[a name proposed for a single isolate that was antagonistic to a wide range of phytopathogenic bacteria and fungi (Attafuah and Bradbury, 1989; Walker et al., 1996)] was demonstrated to be a junior synonym of B. gladioli (Coenye et al., 2000). Burkholderia fungorum too has been recovered from a wide range of ecological niches, including root nodules of plants, fungi, and clinical specimens of humans and animals (Coenye et al., 2001b; P. Vandamme, unpubl. obs.).

The Burkholderia cepacia complex

Like many other burkholderias, B. cepacia was originally known as a plant pathogenic pseudomonad. Pseudomonas cepacia was described by Burkholder in 1950 as the causative agent of bacterial rot of onion bulbs, causing a disease called sour skin (Burkholder, 1950). In subsequent years however, the group of human opportunistic pathogens known as ‘eugonic oxidisers group 1′ and, later, Pseudomonas kingii, and the environmental organism known as Pseudomonas multivorans (Morris and Roberts, 1959; Stanier et al., 1966; Jonsson, 1970), were shown to represent the same species (Ballard et al., 1970; Snell et al., 1972; Samuels et al., 1973; Sinsabaugh and Howard, 1975). During the past decade, international collaborative polyphasic-taxonomic studies (Coenye et al., 2001c) have demonstrated that strains identified as B. cepacia actually represent a complex of closely related species or genomovars. This group, collectively referred to as the B. cepacia complex, comprises at least nine species (Table 1) sharing a high degree of 16S rDNA (98–100%) and recA (9495%) sequence similarity, and moderate levels of DNA–DNA hybridization (30–50%) (Vandamme et al., 1997; 2000; 2002; 2003; Coenye et al., 2001d; 2001e). Apart from B. cepacia, also genomovars V and IX were identified as established Burkholderia species, i.e. B. vietnamiensis (Gillis et al., 1995) and B. pyrrocinia (Imanaka et al. 1965) respectively. Isolates from B. multivorans (genomovar II), B. cenocepacia (genomovar III) and B. stabilis (genomovar IV) were present in historical P. multivorans or P. kingii strain collections (Stanier et al., 1966; Ballard et al., 1970; Samuels et al., 1973) but were only considered representing novel species with the application of species delineation standards as currently defined ( Wayne et al., 1987). Until now, strains from genomovars VI (B. dolosa), VII (B. ambifaria) and VIII (B. anthina) have not been found in historical collections.

Table 1. . Overview of the Burkholderia cepacia complex.
NameOther designationReferences
  • a

    . The taxonomic position of B. ubonensis within the B. cepacia complex has not been fully established

B. cepaciaB. cepacia genomovar IVandamme et al. (1997)
B. multivoransB. cepacia genomovar IIVandamme et al. (1997)
B. cenocepaciaB. cepacia genomovar IIIVandamme et al. (1997; 2003)
B. stabilisB. cepacia genomovar IVVandamme et al. (1997; 2000)
B. vietnamiensisB. cepacia genomovar VGillis et al. (1995); Vandamme et al. (1997)
B. dolosaB. cepacia genomovar VICoenye et al. (2001d); Vermis et al. (2003)
B. ambifariaB. cepacia genomovar VIICoenye et al. (2001e)
B. anthinaB. cepacia genomovar VIIIVandamme et al. (2002)
B. pyrrociniaB. cepacia genomovar IXVandamme et al. (2002)
B. ubonensisa Yabuuchi et al. (2000)

All of these B. cepacia complex species have been isolated from environmental and human clinical sources. Animal infections caused by B. cepacia complex have been reported (Berriatua et al., 2001) but in general, its distribution in animal species and infection therein, is not well-documented. Burkholderia cepacia complex bacteria are universal contaminants of cosmetic and pharmaceutical solutions (for reviews see Pankhurst and Philpott-Howard, 1996; Jimenez, 2001) and can cause contamination of water supplies (see for example Koenig and Pierson, 1997), sterile solutions (like disinfectant solutions, mouth wash and anaesthetics), and disposable equipment leading to nosocomial infections and pseudoepidemics (for a review see Coenye and LiPuma, 2003). Burkholderia cepacia complex organisms have also been recovered from unpasteurised bovine milk (Moore et al., 2001) and from gelatine, a product of animal origin often used in the food industry (De Clerck and De Vos, 2002). In humans B. cepacia complex bacteria have been associated with a wide variety of infections, most often in patients with an underlying disabling illness (including persons with chronic granulomatous disease). However, severe community acquired infections (including endocarditic and pneumonia) have been reported as well (for a review see Coenye and LiPuma, 2003). More importantly, patients with cystic fibrosis (CF) are particularly susceptible to infections caused by these bacteria.

B. cepacia complex and CF

Cystic fibrosis is the most frequent hereditary disease in Caucasian populations, affecting approximately 1 in 2570 live births (Rosenstein and Zeitlin, 1998). Cystic fibrosis symptoms are caused by a mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) gene, a cAMP dependent transepithelial chloride channel. The consequences of mutant CFTR protein are complex but the resultant altered epithelial surface liquid in some way predisposes the respiratory tract of affected persons to bacterial infections (Dinwiddie, 2000). Airway infections in CF patients are characterized by intercurrent acute exacerbations with fever, weight loss, increased cough, change in volume, colour or appearance of sputum, increased respiratory rate, dyspnea and appearance of infiltrates on chest radiographs (Gilligan, 1991). Typically, periods of relative well-being are followed by episodes of these pulmonary exacerbations, which result in a progressive deterioration of the lung function. The improvement of antimicrobial therapy has contributed significantly to the increased median survival age but chronic pulmonary infection is still the leading cause of death among CF patients (Liou et al., 2001). Common bacterial pathogens in young CF patients include Staphylococcus aureus and Haemophilus influenzae. Pseudomonas aeruginosa infection becomes more common during adolescence and by adulthood almost 80% of CF patients are infected with this organism (Lyczak et al., 2002).

Several reports on the recovery of bacteria now known as B. cepacia complex from CF patients appeared in the late 1970s and early 1980s (LiPuma, 1998). The first detailed description of its clinical significance was published by Isles et al. (1984), who also described the so-called ‘cepacia syndrome’, a progressive respiratory failure with necrotising pneumonia that affected some but not all patients. Similar increases in incidence were subsequently noted in other North-American and European CF centres (reviewed in Coenye and LiPuma, 2003). From these initial studies, several risk factors for B. cepacia complex acquisition could be identified, including age, severity of underlying lung disease, use of aminoglycoside antibiotics, previous hospitalization and the presence of a B. cepacia complex colonized sibling. Clustering of cases in some centres and the decrease of colonization following segregation of colonized and non-colonized patients in other centres suggested that B. cepacia complex could be transmitted between CF patients; since then numerous reports have provided epidemiologic and genotyping evidence for interpatient spread (reviewed in Coenye and LiPuma, 2003). Compared to P. aeruginosa, B. cepacia complex infects only a small proportion of CF patients but nevertheless has a significant impact on survival (Liou et al., 2001). In addition, it has been shown that B. cepacia complex colonized CF patients have a much poorer outcome following lung transplantation (the only therapeutic option for succesfull intermediate term survival for CF patients with end-stage pulmonary disease) than their non-infected counterparts (Aris et al., 2001; LiPuma, 2001).

Biocontrol, bioremediation and plant-growth-promotion applications

The biological control of plant diseases, insects and nematodes by microorganisms (both bacteria and fungi) has been proposed as an alternative or a supplement to chemical pesticides and the use of introduced biological control could have enormous ecological and economical benefits (Parke and Gurian-Sherman, 2001; Gerhardson, 2002). The two traditional approaches used for biological control of soil-borne plant pathogens in the field have been (i) crop rotation, to allow time for residential antagonists to ‘sanitize’ the soil or for propagules of specialized pathogens to die, and (ii) the addition of organic amendments to soil to stimulate residential antagonists. However, the greatest progress towards biological control of soil-borne plant pathogens has been made with antagonists introduced with the planting material, i.e. biological control with plant-associated microorganisms (Cook, 1990). Burkholderia cepacia complex strains can be used to control seedling and root diseases in vitro and in field tests, and can replace chemical alternatives. Field tests have shown that such strains can colonize the rhizosphere of several crops, including corn, maize, rice, pea, sunflower and radish, and thereby significantly increases the crop yield, even in the absence of pathogens (Parke et al., 1991; McLoughlin et al., 1992; Bowers and Parke, 1993; Hebbar et al., 1998; Tran Van et al., 2000). One of the most-studied biocontrol strains is B. ambifaria strain AMMDT. It was isolated from the rhizosphere of peas (Pisum sativum) grown in a Aphanomyces root rot nursery in Wisconsin (USA) in 1985. It has activity against Pythium aphanidermatum (responsible for pre- and post-emergence damping-off in peas) and Aphanomyces euteiches (responsible for root rot in peas) (Parke, 1990; Parke et al., 1991; Bowers and Parke, 1993; King and Parke, 1993; Heungens and Parke, 2000; 2001; Parke and Gurian-Sherman, 2001).

The exceptional metabolic versatility of B. cepacia complex bacteria and strains belonging to other Burkholderia species can also be used for bioremediation purposes. Constituents of crude oils [including polycyclic aromatic hydrocarbon (PHA) compounds], herbicides (including 2,4-dichlorophenoxyacetic acid and 2,4,5-trichlorophenoxyacetic acid, the principal component of Agent Orange), TCE and ether derivatives used as gasoline additives can be degraded by Burkholderia isolates (Kilbane et al., 1982; Folsom et al., 1990; Krumme et al., 1993; Bhat et al., 1994). Well-characterized biodegradative strains include G4 (Nelson et al., 1987; Folsom et al., 1990; Shields et al., 1991; Leahy et al., 1996; McClay et al., 1996; Massol-Deya et al., 1997) and CRE-7 (Mueller et al., 1996), which both belong to the B. cepacia complex. Several Burkholderia strains with potential applications in bioremediation are extremely well-studied with regard to their biotechnological potential but are taxonomically poorly characterized. For example, strain NF100 (Hayatsu et al., 2000) and strain LB400 (Haddock et al., 1993; Seeger et al., 1995; Billingsley et al., 1997; Master and Mohn, 1998) are phylogenetically closely related to B. glathei and B. fungorum, respectively, but their exact taxonomic status remains to be determined. An overview of Burkholderia strains capable of degrading recalcitrant xenobiotics can be found in the Biodegradative Strain Database (Urbance et al., 2003; http://www.bsd.cme.msu.edu/bsd/index.html).

Identification and misidentification of B. cepacia complex isolates

Identification of members of the B. cepacia complex is not straighforward (see for example Shelly et al., 2000; Henry et al., 2001; Brisse et al., 2002). In a large-scale study, McMenamin et al. (2000) recovered 1051 isolates from 608 CF patients receiving care in 115 treatment cities in 91 centres in the USA. Of the 770 isolates referred as B. cepacia complex, 11% were misidentified. On the other hand, of the 281 isolates received as unidentified or ‘not B. cepacia complex’, 36% actually were B. cepacia complex. Organisms that have been misidentified as B. cepacia complex include P. aeruginosa, Stenotrophomonas maltophilia, Ralstonia pickettii, Achromobacter xylosoxidans, Bordetella hinzii, Brevundimonas sp., Chryseobacterium sp. and members of the Enterobacteriaceae. In addition, among isolates misidentified as B. cepacia complex, a variety of novel species mainly belonging to the β-Proteobacteria have been described. These comprise several novel Burkholderia species (including B. fungorum, B. caledonica, B. terricola and B. hospita), novel Ralstonia species (including Ralstonia gilardii, Ralstonia mannitolilytica and Ralstonia taiwanensis) and members of the novel genera Pandoraea (Pandoraea apista, Pandoraea pnomenusa, Pandoraea norimbergensis, Pandoraea pulmonicola and Pandoraea sputorum) and Inquilinus (Inquilinus limosus). The clinical significance of these novel species is mostly unclear. Anecdotal evidence indicates that several are capable of chronic colonization of the CF lung and/or person-to-person transmission. The erroneous identification of an organism as B. cepacia complex has serious medical, social and psychological implications for infected CF patients and misidentification can have serious implications for potential biotechnological applications.

Prevalence of members of the B. cepacia complex in CF patients and in the environment

All nine B. cepacia complex species have been identified in CF sputum cultures, but several recent studies indicate that they are unequally represented. Surveys form the US (LiPuma et al., 2001), Canada (Speert et al., 2002), Italy (Agodi et al., 2001) and Belgium (De Boeck and P. Vandamme, unpubl. obs.) indicated that, although there are national differences, most patients are infected with B. cenocepacia or B. multivorans(Table 2).

Table 2. . Prevalence of Burkholderia cepacia complex organisms in CF specimens (in percentage of infected patients).
GenomovarUSA
(n = 606)
Canada
(n = 445)
Italy
(n = 62)
Belgium
(n = 33)
I 2.6 0.2 4.8
II37.8 9.7 4.857.8
III50.082.972.627.3
IV 0.2 3.8 3.212.1
V 5.1 1.6 3.0
VI 2.0
VII 0.7
VIII 0.2
IX 0.2

There are much less systematic studies regarding the distribution of the B. cepacia complex species in environmental samples. Balandreau et al. (2001) recovered 22 B. cepacia complex isolates from the rhizosphere of wheat and maize and from roots and shoots of wheat and lupine, grown in France and Australia. Of these, 21 belonged to B. cenocepacia, whereas a single isolate belonged to B. ambifaria. Fiore et al. (2001) recovered 120 B. cepacia complex isolates from maize rhizosphere. The majority belonged to B. cepacia, B. cenocepacia and B. ambifaria. In a similar study, Bevivino et al. (2002) recovered 75 isolates from maize rhizosphere. Only B. cenocepacia (53.4%), B. ambifaria (37.3%), B. pyrrocinia (8.0%) and B. cepacia (1.3%) were isolated. During a study of urban and suburban soils in three large cities in the USA, Miller and co-workers isolated B. cepacia complex isolates from 14 (15%) of 91 sampled sites. Sixty-eight B. cepacia complex isolates were recovered, mainly belonging to B. ambifaria and B. pyrrocinia (Miller et al., 2001; 2002). 127 B. cepacia complex isolates were recovered from agricultural soils in the USA by Gonzalez et al. (2001). The majority belonged to B. ambifaria (42.5%), B. cepacia (24.4%), B. cenocepacia (22.8%) and B. pyrrocinia (10.2%). Thus far, overall, B. cepacia, B. cenocepacia, B. ambifaria and B. pyrrocinia dominate in soil samples, whereas the other genomovars are rarely found. Although the knowledge about the distribution of the different genomovars in environmental samples is restricted and although misidentification is common, it is obvious that the same species that occur in the environment, colonize and infect CF patients, albeit to different degrees.

Is the natural environment the reservoir for B. cepacia complex infections in CF patients?

Indirect evidence for the speculation that the environment serves as a reservoir for acquisition of novel B. cepacia complex species comes from the observation that infection control measures (including segregation of patients) have reduced but not eliminated new infections (often with isolates showing novel fingerprint types). Direct evidence was obtained from genotyping studies, in which clinical and environmental isolates were compared using state-of-the-art molecular fingerprinting techniques. Govan et al. (2000) showed that the B. cepacia type strain ATCC 25416T.(isolated from rotting onions in the 1940s) was isolated from sputum of a CF patient in the UK. LiPuma et al (2002) reported that B. cenocepacia strain PHDC recovered from most CF patients in the mid-Atlantic region of the USA could also be isolated from agricultural soils. These findings, although anecdotal, show that human isolates are not necessarily distinct from environmental ones.

Concluding remarks

Although phylogenetically well-defined, the genus Burkholderia is functionally a remarkably diverse genus. The molecular and physiological background of this diversity and adaptability are largely unknown. In addition, little is known about factors that determine virulence, which implies that only correct identification provides a first basis for risk assessment and infection control. The genomes of most Burkholderia species (including B. cepacia complex species, B. glumae, B. glathei, B. gladioli and Burkholderia sp. LB400) consist of two to four circular replicons (Cheng and Lessie, 1994; Rodley et al., 1995; Lessie et al., 1996; Wigley and Burton, 2000; Parke and Gurian-Sherman, 2001). There appears to be a tremendous variation in genome size among these phylogenetically closely related organisms ranging from 4.7 Mb to more than 9 Mb. In addition, several Burkholderia species harbour an extensive array of insertion sequences (Lessie et al., 1996; Tyler et al., 1996) which may be involved in genomic rearrangements and regulation of gene expression. The genomes of strains of several Burkholderia species (including B. cenocepacia, B. pseudomallei and B. mallei) have been completely sequenced and other sequencing projects are in progress (Burkholderia sp. LB400) or are planned for the near future (B. vietnamiensis G4) (for an overview see http://www.igweb.integratedgenomics.com/GOLD/). It can be expected that knowledge derived from these genome sequencing projects will allow us to gain further insights into functional diversity, evolution and pathogenicity mechanism. This may provide the scientific basis for important decisions regarding infection control and the biotechnological use of Burkholderia strains.

Acknowledgements

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. T.C. also acknowledges the support from the Belgian Federal Government (Federal Office for Scientific, Technical and Cultural Affairs).

Ancillary