Functional Analysis of Burkholderia cepacia Genes bceD and bceF, Encoding a Phosphotyrosine Phosphatase and a Tyrosine Autokinase, Respectively: Role in Exopolysaccharide Biosynthesis and Biofilm Formation

Abstract

The biosynthesis of the exopolysaccharide (EPS) cepacian by Burkholderia cepacia complex strains requires the 16.2-kb bce cluster of genes. Two of the clustered genes, bceD and bceF, code for two proteins homologous to phosphotyrosine phosphatases and tyrosine kinases, respectively. We show experimental evidence indicating that BceF is phosphorylated on tyrosine and that the conserved lysine residue present at position 563 in the Walker A ATP-binding motif is required for this autophosphorylation. It was also proved that BceD is capable of dephosphorylating the phosphorylated BceF. Using the artificial substrate p-nitrophenyl phosphate (PNPP), BceD exhibited a V_max of 8.8 μmol of PNPP min⁻¹ mg⁻¹ and a K_m of 3.7 mM PNPP at 30°C. The disruption of bceF resulted in the abolishment of cepacian accumulation in the culture medium, but 75% of the parental strain's EPS production yield was still registered for the bceD mutant. The exopolysaccharide produced by the bceD mutant led to less viscous solutions and exhibited the same degree of acetylation as the wild-type cepacian, suggesting a lower molecular mass for this mutant biopolymer. The size of the biofilm produced in vitro by bceD and bceF mutant strains is smaller than the size of the biofilm formed by the parental strain, and this phenotype was confirmed by complementation assays, indicating that BceD and BceF play a role in the establishment of biofilms of maximal size.

Bacteria of the Burkholderia cepacia complex (Bcc) have emerged as opportunistic pathogens in patients with cystic fibrosis (CF) and immunocompromised individuals (23, 24, 27, 36). Furthermore, the number of human infections caused by Bcc strains has increased over the last 2 decades (24, 35, 36). Approximately 80% of the Bcc isolates recovered from the sputum of CF patients produce large amounts of exopolysaccharide (EPS) (16, 50), suggesting a possible role for this EPS in Bcc pathogenesis, as described for Pseudomonas aeruginosa alginate (24). In fact, the Bcc mucoid phenotype was found to affect cell-surface interactions and clearance in an animal model, which could enhance the persistence and virulence of Bcc in CF (13). Moreover, the EPS produced by Bcc interferes with the function of human neutrophils in vitro, inhibiting chemotaxis and the production of reactive oxygen species (8). Moreover, the size of the biofilm formed in vitro by mutants derived from a mucoid CF isolate correlates with their ability to produce exopolysaccharide (17).

Cepacian is the main exopolysaccharide produced by Bcc isolates, and it is composed of a branched acetylated heptasaccharide repeat unit made of d-glucose, d-rhamnose, d-mannose, d-galactose, and d-glucuronic acid at the molar ratio 1:1:1:3:1 (9-11, 55). The pathway leading to the nucleotide sugar precursors necessary for cepacian biosynthesis was proposed previously (51), and the bce cluster of genes directing its biosynthesis was identified (41). The bce genes also encode proteins putatively involved in repeat unit assembly, polymerization, and export. Among these genes are the bceD and bceF genes coding for proteins homologous to protein tyrosine phosphatases (PTPs) and protein tyrosine kinases (PTKs), respectively. The present study is focused on their functional analysis.

Protein phosphorylation on tyrosine has long been considered specific to eukaryotes, but in the last two decades, several evidences clearly indicate that it also occurs in bacteria (15). One of the best-studied examples of tyrosine phosphorylation/dephosphorylation in bacteria is related to polysaccharide production. Indeed, many clusters of genes directing the synthesis and regulation of exopolysaccharides or capsules encode a PTP and a PTK (15). In general, PTK homologues from gram-negative bacteria are integral membrane proteins harboring two transmembrane domains flanking a large periplasmic loop and have a cytoplasmic C-terminal region with Walker A and Walker B ATP-binding motifs and a tyrosine-rich C terminus where tyrosine phosphorylation occurs (2, 19, 20, 44, 46, 59, 62). In contrast, in gram-positive bacteria, the PTK homologues are present in two separate proteins, exhibiting significant sequence similarity to the two halves of the single peptides from gram-negative bacteria, and both proteins are required for phosphorylation (38, 43). This is also the case for the two proteins GelC and GelE, characterized in the gellan-producing Sphingomonas elodea ATCC 31461, which constitute the exception to the single-peptide PTK from gram-negative bacteria (40). The precise role of protein tyrosine autokinases in polysaccharide biosynthesis is not known. However, their presence in both gram-positive and gram-negative bacteria suggests that they may play a role in a conserved step of this biosynthetic process, presumably in the regulation of polymer chain length (61). A positive correlation between tyrosine phosphorylation and high-molecular-mass polysaccharide synthesis has been observed for Sinorhizobium meliloti succinoglycan and for capsular polysaccharides from Escherichia coli K30 or Streptococcus pneumoniae D39 (3, 46, 47). However, in other bacterial systems, phosphorylated tyrosine kinases are believed to act differently as negative regulators in colanic acid biosynthesis by E. coli K-12, in emulsan biosynthesis by Acinetobacter lwoffii RAG-1, and in capsular polysaccharide biosynthesis by S. pneumoniae Rx1 (42-44, 59).

PTP homologues are low-molecular-weight proteins with a conserved CX₄CR active motif in the phosphate-binding loop, flanked at some distance by an essential aspartate residue (30). In several bacterial species, PTKs are specific endogenous substrates for the corresponding PTPs. This is the case for the pairs PTP/PTK, Wzb/Wzc_CA, and Etk/Etp from E. coli, Ptp/Ptk from Acinetobacter johnsonii, and AmsI/AmsA from Erwinia amylovora, among others (6, 29, 59).

In the present study, we report the sequence analysis of a chromosomal region inside the bce gene cluster from the mucoid clinical isolate B. cepacia IST408. This region is required for cepacian biosynthesis and is where the bceD and bceF genes map. Results on the functional analysis of these genes, encoding a PTP and a PTK, respectively, are shown, as well as their involvement in cepacian biosynthesis. The hypothesized involvement of BceD and BceF on the formation of biofilms of maximal size was also examined.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The strains and plasmids used in this work are listed in Table 1. The cystic fibrosis isolate Burkholderia cepacia IST408, previously described as a high exopolysaccharide producer (50), was used as the parental strain for the mutants constructed in this work. E. coli strains were grown in Lennox broth (53) at 37°C. B. cepacia strains were grown in Pseudomonas isolation agar (Difco) plates at 37°C. Mannitol solid medium (12) or S medium (50) was used to quantify EPS production at 30°C by the B. cepacia strains. Growth media were supplemented with antibiotics when required to maintain the selective pressure at the following concentrations (in μg ml⁻¹): for B. cepacia, trimethoprim, 100; chloramphenicol, 300; for E. coli, ampicillin, 100; kanamycin, 50; trimethoprim, 100; chloramphenicol, 25.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Genotype or description	Reference or source
Escherichia coli strains
XL1-Blue	recA1 lac [F′ proAB lacIq ZαM15 Tn10 (Tcr)] thi	7
Sure	e14− (mcrA) Δ(mcrCB-hsdSMR-mrr)171 endA1 supE44 thi-1 gyrA96 relA1 lac recB recJ sbsC umuC:Tn5 (Kmr) uvrC [F′ proAB lacIq ZαM15 Tn10 (Tcr)]	Stratagene
Burkholderia cepacia strains
IST408	Cystic fibrosis isolate, genomovar I, EPS+	50
IST408 bceD::Tp	IST408 derivative with bceD gene interrupted by a Tpr gene cassette	This work
IST408 bceF::Tp	IST408 derivative with bceF gene interrupted by a Tpr gene cassette	This work
Plasmids
pUC-TP	pUC-GM derivative with a 1.1-kb Tpr gene cassette, Apr Tpr	56
pWH844	pQE9 derivative, lacIq Apr	54
pBBR1MCS	4,717-bp broad-host-range cloning vector, Cmr	31
pMLBAD	pBBR1 ori, araC-PBAD, Tprmob+	34
pMLBAD-Cm	pMLBAD derivative containing the dhfr gene replaced by the cat gene from pBBR1MCS	This work
pBCSK	Phagemid derived from pUC19, Apr	Stratagene
pBCSKΔBamHI	pBCSK derivative without the BamHI site	This work
Plasmid D	TnMod-KmO derivative carrying a 6,063-bp EcoRI fragment encompassing the bceD, bceE, bceF, and bceG genes, Kmr	41
pLM211-1	pWH844 derivative containing the bceD gene	This work
pLM45-1	pWH844 derivative containing the bceF gene	This work
pLM45-1.K563A	pLM45-1 derivative expressing His6-BceFK563A	This work
pLM411-4	pBCSK derivative containing a 2,056-bp SacII fragment from plasmid D encompassing the bceD gene and flanking regions	This work
pSF412-1	pLM411-4 derivative containing a Tpr gene cassette at the SphI site disrupting the bceD gene	This work
pSF54-1	pBCSKΔBamHI derivative containing a 2,218-bp HindIII fragment from plasmid D encompassing the bceF gene	This work
pSF57-5	pSF54-1 derivative containing a Tpr gene cassette at the BamHI site disrupting the bceF gene	This work
pLM63-1	pMLBAD-Cm derivative carrying a 2,242-bp KpnI/HindIII fragment with the coding region of the bceF gene	This work
pLM69-1	pMLBAD-Cm derivative carrying a 501-bp KpnI/HindIII fragment with the coding region of the bceD gene	This work

DNA manipulation techniques.

Total DNA was extracted from bacterial cells, harvested from liquid cultures grown overnight at 37°C, using a cell and tissue kit (Gentra Systems) following the manufacturer's instructions. Plasmid DNA isolation and purification, DNA restriction/modification, agarose gel electrophoresis, Southern blotting experiments, and E. coli transformation were carried out using standard procedures (53). A specific mutation in one amino acid from BceF was constructed in vitro using the QuikChange site-directed mutagenesis method (Stratagene). Briefly, complementary oligonucleotides K11A-1 (5′-CGGGCATCGGCGCGAGCTTCCTGACGG; substitutions in the primers are underlined) and K11A-2 (5′-CCGTCAGGAAGCTCGCGCCGATGCCCG) were designed to contain the desired codon change. The template consisted of plasmid pLM45-1 coding for the BceF protein. Following PCR amplification, the reaction products were digested with DpnI to eliminate the template, and the remaining DNA was introduced into E. coli XL1-Blue by electroporation. The mutated derivatives were sequenced to confirm the mutation and verify that no other changes were introduced.

B. cepacia electrocompetent cells, prepared as described by Sá-Correia and Fialho (52), were transformed by electroporation using Bio-Rad Gene Pulser II (200 Ω, 25 μF, 2.5 kV) and grown overnight before being plated in selective medium. Triparental conjugation to B. cepacia was performed using the helper plasmid pRK2013 (22).

Construction of bceD and bceF insertion mutations.

The 2,056-bp SacII fragment from plasmid D containing the bceD gene coding region and flanking regions was cloned into the multiple cloning site of pBCSK. The plasmid obtained, pLM411-4, has an SphI site within the bceD coding region, and it was used to insert an SphI fragment containing the trimethoprim (Tp) resistance cassette obtained from plasmid pUC-TP (56). The Tp cassette is transcribed in the same orientation as the bceD gene, and the resulting plasmid with the insertion mutation in the bceD gene was designated pSF412-1.

To obtain a construction for bceF disruption, a 2,218-kb HindIII fragment containing the bceF coding region was amplified using primers BceF-up (5′-CCCAAGCTTGAACACGCAAGCGAAAC; restriction sites are in italics) and BceF-low (5′-GGGAAGCTTGGATCAGGCGCTCAGGT) and cloned into the HindIII site of the previously modified plasmid pBCSK lacking the BamHI site to allow subsequent manipulations. The plasmid obtained (pSF54-1) was further digested with BamHI, which has a single recognition site within the bceF coding region. The Tp resistance cassette was obtained from pUC-TP by restriction with XbaI. After fill-in, the Tp cassette was cloned into pSF54-1, originating plasmid pSF57-5, which transcribes the cassette in the opposite orientation of the bceF gene. Each of the bceD::Tp and bceF::Tp insertion constructs in plasmids pSF412-1 and pSF57-5, respectively, was introduced into B. cepacia IST408 by electroporation, and transformants were selected by growth on Pseudomonas isolation agar medium with trimethoprim. The colonies obtained were then screened in the presence of chloramphenicol. Colonies that did not grow in the presence of chloramphenicol but were trimethoprim resistant were considered as candidates for having an allelic exchange of bceD or bceF by the bceD::Tp or bceF::Tp construct, respectively. The candidate insertion mutants were further characterized by Southern hybridization or PCR amplification.

Construction of plasmids for complementation experiments.

In order to complement the EPS⁻ phenotype and the biofilm formation ability of the B. cepacia bceF::Tp mutant, the recombinant plasmid pLM63-1 was constructed. For this, the bceF coding region was amplified by PCR using primers bceFBAD-up (5′-ACGGGTACCGAACACGCAAGCGA) and bceFBAD-low (5′-GTGAAGCTTGGATCAGGCGCTCA) and IST408 genomic DNA as the template. The amplified fragment was restricted by KpnI/HindIII and inserted into the same sites of the pMLBAD-Cm vector, obtained by replacing the dhfr gene, coding for trimethoprim resistance of pMLBAD, with the cat gene from pBBR1MCS, coding for chloramphenicol resistance. To complement the biofilm formation ability of the B. cepacia bceD::Tp mutant, the recombinant plasmid pLM69-1 was constructed. For this, the bceD coding region was amplified by PCR with primers bceDBAD-up (5′-AGAGGTACCGTTCCGGAACATCC) and bceDBAD-low (5′-CAGAAGCTTCGTCAGCGCGAC), using IST408 genomic DNA as the template. The amplified fragment was restricted by KpnI/HindIII and inserted into pMLBAD-Cm. The nucleotide sequences of the cloned genes were confirmed by sequencing.

In vivo complementation of the EPS⁻ phenotype of IST408 bceF::Tp.

Plasmid pLM63-1, encoding the parental bceF gene, was mobilized into the B. cepacia IST408 bceF::Tp mutant strain by triparental conjugation. To determine the ability of this plasmid to restore cepacian synthesis, the complemented strain IST408 bceF::Tp/pLM63-1 was grown in mannitol solid medium, supplemented with 1% of l-arabinose, at 30°C for 5 days, and the mucoidy of the corresponding colonies was observed. Mannitol liquid medium supplemented with 1% of l-arabinose was also used in order to determine the amount of EPS produced based on the dry weight of the ethanol precipitates from cell-free culture supernatants.

Construction of bceD and bceF overexpression plasmids.

The genes bceD and bceF were PCR amplified with primers Pho-up (5′-AAAGGATCCTTCCGGAACATCCT) and Pho-low (5′-CCCAAGCTTGTTTCAGCATAGTT) or BceF-up and BceF-low, respectively. The amplified product of bceD was digested with BamHI/HindIII and cloned into the same sites of linearized pWH844, resulting in plasmid pLM211-1. The amplified product of bceF was digested with HindIII and cloned in the HindIII-linearized plasmid pWH844, and the plasmid obtained was named pLM45-1.

Expression and purification of His₆-BceD and His₆-BceF.

E. coli Sure cells harboring pLM211-1 or pLM45-1 were grown in Lennox medium supplemented with 0.1% glucose and containing 300 mg liter⁻¹ of ampicillin, at 37°C, to an optical density at 640 nm of 0.5. Induction was started by adding 0.4 mM IPTG (isopropyl-β-d-thiogalactopyranoside), followed by incubation for 4 h. The His-tagged BceD was purified from cell-free lysates of E. coli prepared in sonication buffer (50 mM Tris-HCl, 300 mM NaCl, 10% [vol/vol] glycerol, pH 7.5). After removal of cell debris by centrifugation at 19,000 × g for 30 min at 4°C, His₆-BceD was purified by Ni²⁺ affinity chromatography. The Ni²⁺-nitrilotriacetic acid (NTA) matrix was washed with 6 volumes of buffer A (50 mM Tris-HCl, 300 mM NaCl, 10% [vol/vol] glycerol, pH 7.5) containing 20 mM imidazole and 6 volumes of buffer B (100 mM Tris-HCl, 500 mM NaCl, pH 8.9) containing 50 mM imidazole. Bound proteins were eluted with buffer B containing 250 mM imidazole.

His₆-BceF purification was done as described for His₆-BceD, except that 1.0% (vol/vol) Triton X-100 was added to the lysate after sonication and elution was done with 300 mM imidazole. Both proteins were purified freshly before each experiment. The protein concentrations of the purified His-tagged protein solutions were determined by the method of Bradford (5), using bovine serum albumin as the standard.

Phosphatase activity.

Phosphatase activity was determined based on the continuous monitoring, at 405 nm, of the p-nitrophenol (PNP) formed from p-nitrophenol phosphate (PNPP) at 37°C in a Hitachi UV 2000 double-beam spectrophotometer. The reaction mixture contained, in a total volume of 1 ml, 100 mM sodium citrate buffer, pH 6.5, 1 mM EDTA, 0.1% (vol/vol) β-mercaptoethanol, and PNPP concentrations ranging from 0.5 to 40 mM. The concentration of PNP formed was estimated using a molar extinction coefficient of 18,000 M⁻¹ cm⁻¹ (48). In experiments carried out to determine the optimal pH for phosphatase activity, the citrate buffer pH was varied from 5.5 to 7.5. All reactions were initiated by the addition of 5 μl of purified His-tagged phosphatase. One unit of enzyme activity was defined as the amount of enzyme that catalyzes the formation of 1 μmol of PNP per minute under the assay conditions. The results presented for enzyme assays and for protein concentrations are the means from at least three independent experiments.

Dephosphorylation of His₆-BceF was monitored by Western immunoblot analysis. For that, 1 μg of His₆-BceF and 1 μg of His₆-BceD were incubated in 30 μl of buffer consisting of 100 mM sodium citrate (pH 6.5) and 1 mM EDTA at 37°C for 1 to 22 h. The reaction was stopped by the addition of an equal volume of 2× sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer, and the mixture was heated at 100°C for 5 min and subsequently analyzed by SDS-PAGE, followed by immunodetection using antiphosphotyrosine antibodies. The relative signal intensity of the His₆-BceF bands obtained by immunodetection was quantified using Quantity One software from Bio-Rad.

Western immunoblot analysis.

Protein samples were separated on 15% SDS-PAGE gels (33) prior to being transferred onto nitrocellulose membranes (58). The nitrocellulose filters were incubated with either antiphosphotyrosine antibodies (PT-66; Sigma), at a dilution of 1:2,000, or monoclonal antibodies against the His tag (QIAGEN), at a dilution of 1:2,000. The membranes were then incubated with alkaline phosphatase-conjugated rabbit anti-mouse or goat anti-rabbit (Promega) and reacted with ECL reagent (Amersham Biosciences). Broad-range prestained SDS-PAGE standard from Bio-Rad was used as a molecular weight marker.

Production and characterization of bacterial exopolysaccharides.

EPS production was assessed based on the dry weight of the ethanol-precipitated polysaccharide present in 2-ml culture samples of the different strains cultivated in S liquid medium at 30°C with orbital agitation. The viscosities of solutions prepared with the polysaccharides produced by the different strains under study were compared using EPS obtained after 72 h of culture at 30°C with orbital agitation. For this, bacterial cells present in the cultures were separated by centrifugation at 20,000 × g for 15 min. The EPSs were precipitated from cell-free supernatants by the addition of 2.5 volumes of cold ethanol and then air dried and redissolved in distilled water prior to dialysis (molecular mass cutoff, 12 kDa) against water for 3 days at 4°C, followed by EPS lyophilization. Aqueous solutions of the lyophilized polymers (5 g liter⁻¹) were prepared, and their viscosity was measured at 30°C using a Brookfield cone and plate viscometer, model LVIIT. Results are the mean values for at least two independently prepared EPS solutions. The acetyl content of the EPSs produced was determined as described by McComb and McCready (37), using glucose pentaacetate as the standard. Results are the mean values from at least three independent determinations. Glucose and galactose content of the polysaccharide fractions was quantified by enzymatic assays after hydrolysis with HCl and subsequent neutralization (4, 32).

Biofilm formation assay.

Biofilm formation assays were based on the method described by Cunha et al. (17). Overnight liquid cultures of the different B. cepacia strains were used to inoculate S liquid medium or mannitol medium supplemented with 1% of l-arabinose and grown at 30°C with orbital agitation until mid-exponential phase. The cultures were subsequently diluted to a standardized culture optical density at 640 nm of 0.1, and 100 μl of these cell suspensions was used to inoculate the wells of a 96-well polystyrene microtiter plate (Greiner Bio-One) containing 100 μl of S medium or mannitol medium supplemented with l-arabinose. Wells containing sterile growth medium were used as negative controls. Plates were incubated at 30°C for 24 or 48 h without agitation. The biofilm formed was quantified as follows. Culture media and unattached bacterial cells were removed from the wells by careful rinsing with water (three times, 200 μl for each rinse). Adherent bacteria were stained with 200 μl of a 1% crystal violet solution for 15 min at room temperature, and after three gentle rinses with 200 μl of water each time, the dye associated with the attached cells was solubilized in 200 μl of 95% ethanol and the biofilm was quantified by measuring the absorbance of the solution at 590 nm using a VersaMax tunable microplate reader (Molecular Devices). Results are the mean values for at least five repeats from three independent experiments.

Computational analysis of nucleotide and protein sequences.

DNA and protein data were analyzed using the open reading frame (ORF) finder tool resident at the National Center for Biotechnology Information (NCBI). The algorithm BLAST (1) was used to compare sequences of the deduced amino acids to database sequences available at the NCBI. Alignments were performed using the program CLUSTAL W (57). DNA sequences from the B. cenocepacia J2315 genome were extracted from the database at http://www.sanger.ac.uk.

Nucleotide sequence accession number.

The nucleotide sequence reported here for the bce genes was submitted to GenBank under accession no. DQ418767.

RESULTS

Sequence analysis of a B. cepacia chromosomal region where bceD and bceF genes map.

In a previous study, we used a mutagenesis strategy, based on the use of the plasposon pTnMod-KmO, to obtain mutants unable to produce exopolysaccharide from the mucoid CF isolate B. cepacia IST408. This manipulation led to the identification of three EPS-deficient strains (IST408-SS1, IST408-SS2, and IST408-SS3) (41). After recovery of the plasposons with the genomic flanking regions as recombinant plasmids, sequencing reactions were carried out to determine the place of insertion of the plasposon and the DNA region involved in EPS biosynthesis was identified by a homology search against the genome of B. cenocepacia J2315 (41). In this work, we report the sequence determination and characterization of plasmid D, carrying a chromosomal region from the B. cepacia IST408-SS1 mutant, obtained after digestion with EcoRI and self-ligation. This plasmid contains an EcoRI fragment of 6,063 bp where the plasposon insertion site was identified, followed by the sequencing in both strands of the DNA insert. Homology searches at the nucleotide level indicated that this chromosomal region of 6,063 bp is 97% identical to a Burkholderia sp. strain 383 genomic region ranging from nucleotides 2592997 to 2599051. This environmental strain is also known as ATCC 17660 and is closely related to the species B. cepacia. The DNA insertion of plasmid D also showed 94% identity to a genomic region of the cystic fibrosis isolate B. cenocepacia J2315 (nucleotides 944407 to 950464). Moreover, other Burkholderia strains such as B. cenocepacia AU1054, B. thailandensis E264, B. pseudomallei 1106b, and B. xenovorans LB400, with genome sequences deposited in GenBank, also have similar chromosomal regions.

Computer-assisted sequence analysis of the 6,063-bp chromosomal region of B. cepacia IST408 revealed four complete ORFs designated bceD, bceE, bceF, and bceG and two incomplete ORFs (bceC and bceH). The ORFs are in the same orientation, confirming the previous conclusion based on the inspection of the available B. cenocepacia J2315 sequence (41). The ORFs bceE and bceG are homologous to several polysaccharide export proteins and glycosyl transferases from family 2, respectively. They are putatively involved in cepacian biosynthesis but will not be further examined here. The two other complete ORFs, bceD and bceF, are homologous to genes encoding low-molecular-mass acid PTPs and PTKs, respectively, and are the focus of the present study.

B. cepacia IST408 BceD has over 94% identity, at the amino acid level, with other putative PTPs from several strains from Burkholderia species. This is the case for the environmental strain 383, the cystic fibrosis isolate B. cenocepacia PC184, two environmental strains from the species B. vietnamiensis and B. ambifaria, and the epidemic strain B. dolosa AU0158 (Table 2). For the clinical isolate B. pseudomallei 1106b and the two environmental strains B. thailandensis E264 and B. xenovorans LB400, the amino acid identity value with BceD is between 62 and 77% (Table 2). The degree of identity, at the amino acid level, between BceD and other experimentally characterized low-molecular-mass PTPs from different genera, such as Wzb from Escherichia or Salmonella, Ptp from Acinetobacter, AmsI from Erwinia, and EpsP from Ralstonia, is within the range of 33 to 37% (Table 2).

TABLE 2.

Features of the bceD and bceF genes from Burkholderia cepacia IST408

ORF	Position in sequence DQ418767 (nucleotides)	No. of amino acids	Putative function	% Identity/% similarity	Organism (protein name)	GenBank accession no.
bceD	985-1428	147	Low-molecular-mass	100/100	Burkholderia sp. strain 383	ABB12385
			phosphotyrosine	41/54	Burkholderia sp. strain 383	ABB12976
			phosphatase	35/55	Burkholderia sp. strain 383	ABB06443
				97/99	B. cenocepacia PC184	ZP_00978876
				95/97	B. vietnamiensis G4	ZP_00424416
				95/95	B. ambifaria AMMD	ZP_00685665
				94/95	B. dolosa AU0158	ZP_00985113
				77/85	B. pseudomallei 1106b	ABA51365
				75/83	B. thailandensis E264	YP_438745
				62/78	B. xenovorans LB400	ZP_00280319
				37/56	Acinetobacter johnsonii (Ptp)	CAA75430
				39/53	Escherichia coli (Wzb)	BAE76575
				39/52	Salmonella enterica serovar Typhimurium (Wzb)	AAL21021
				35/49	Erwinia amylovora (AmsI)	CAA54881
				33/56	Ralstonia solanacearum (EpsP)	ZP_00945864
bceF	2651-4876	741	Protein-tyrosine kinase	98/99	Burkholderia sp. strain 383	ABB12383
				37/56	Burkholderia sp. strain 383	ABB06442
				36/54	Burkholderia sp. strain 383	ABB12977
				33/53	Burkholderia sp. strain 383	ABB11930
				95/97	B. cenocepacia PC184	ZP_00978874
				93/97	B. dolosa AU0158	ZP_00985115
				92/97	B. vietnamiensis G4	ZP_00424418
				92/97	B. ambifaria AMMD	ZP_00685663
				80/88	B. pseudomallei 1106b	ZP_00490661
				80/88	B. thailandensis E264	YP_438747
				72/84	B. xenovorans LB400	ZP_00280321
				41/59	Ralstonia solanacearum (EpsB)	CAD18169
				35/57	Erwinia amylovora (AmsA)	CAA54882
				33/54	Salmonella enterica serovar Typhimurium (Wzc)	AAL21020
				35/54	Escherichia coli (Wzc)	BAA15913
				32/53	Acinetobacter johnsonii (Ptk)	CAA75431

The proteins most homologous to BceD within each Burkholderia genome analyzed are listed in Table 2. However, more than one gene coding for putative PTPs was found to be present in all of the genomes examined. For example, in Burkholderia sp. strain 383, besides the protein with GenBank accession number ABB12385 with 100% identity, at the amino acid level, to BceD, two other proteins (ABB12976 and ABB06443) with 41 and 35% identity, respectively, to BceD were found (Table 2). Similar results were obtained for the other Burkholderia species examined, with one of the putative PTPs being highly homologous to BceD while the other two or three encoding regions exhibited an identity below 40% (data not shown).

The homology between the protein BceF from B. cepacia IST408 and genomic regions of the sequenced Burkholderia strains showed that BceF is homologous to putative protein-tyrosine kinases from Burkholderia sp. strain 383, B. cenocepacia PC184, B. dolosa AU0158, B. vietnamiensis G4, and B. ambifaria AMMD, with more than 92% identity at the amino acid level (Table 2). The homologous putative PTKs from B. thailandensis E264, B. pseudomallei 1106b, and B. xenovorans LB400 are within the range of 72 to 80% identity, at the amino acid level, to BceF. Biochemically characterized PTKs, such as EpsP, AmsA, Wzc, and Ptk from Ralstonia, Erwinia, Salmonella and Escherichia, and Acinetobacter genera, respectively, presented an identity to B. cepacia IST408 PTK within the range of 32 to 41% (Table 2).

The search for other BceF homologues within the sequenced Burkholderia sp. strain 383 genome revealed the presence of a strong homologue with 98% identity (GenBank accession number ABB12383) and three others (ABB06442, ABB12977, and ABB11930) with 33 to 37% identity (Table 2). Similar conclusions were taken from the inspection of other sequenced Burkholderia genomes (data not shown).

With respect to the conserved sequence motifs characteristic of PTPs, BceD exhibits the major signature of this type of enzyme, C-X₄-C-R (¹⁰CHANVCR¹⁶), but has the tyrosine residue of the conserved motif DPY substituted by a histidine residue (Fig. 1a). The conserved motifs of PTKs are also present in BceF, which shows the Walker A (⁵⁵⁶GPTPGIGKS⁵⁶⁴) and Walker B (⁶⁶²VLID⁶⁶⁵) ATP-binding motifs and six tyrosine residues at positions 596, 651, 659, 728, 732, and 738 in its C-terminal region (Fig. 1b).

FIG. 1.

Comparison of the partial sequences of B. cepacia IST408 BceD and BceF protein homologues. (a) Alignment of B. cepacia (Bc) BceD with experimentally confirmed prokaryotic low-molecular-mass PTPs from Acinetobacter johnsonii (Aj) Ptp, Acinetobacter lwoffii (Al) Wzb, Ralstonia solanacearum (Rs) EpsP, Salmonella enterica serovar Typhimurium (St) Wzb, Escherichia coli (Ec) Wzb, and Erwinia amylovora (Ea) AmsI. (b) Alignment of B. cepacia (Bc) BceF with experimentally confirmed prokaryotic PTKs from Erwinia amylovora (Ea) AmsA, Klebsiella pneumoniae (Kb) Yco6, Salmonella enterica serovar Typhimurium (St) Wzc, Escherichia coli (Ec) Wzc, Acinetobacter johnsonii (Aj) Ptk, and Ralstonia solanacearum (Rs) EpsB. The conserved amino acids within the motifs and the C-terminal tyrosine-rich regions of PTKs are shown in bold. X indicates any amino acid, h indicates a hydrophobic amino acid, and alternative residues are enclosed in brackets. Asterisks indicate the amino acid residues that are identical in all proteins; one or two dots indicate semiconserved or conserved substitutions, respectively.

An insertion mutant for bceF, but not for bceD, is impaired in cepacian biosynthesis.

To investigate the hypothesized role of bceD and bceF genes in cepacian biosynthesis, we constructed insertion mutants by using a Tp gene resistance cassette based on the procedure described by Sokol et al. (56). The mutants obtained, designated B. cepacia IST408 bceD::Tp and B. cepacia IST408 bceF::Tp, showed growth curves similar to the wild-type B. cepacia IST408 growth curve (Fig. 2a). To assess cepacian production phenotype, strains were grown in S liquid medium and samples were taken during growth and ethanol precipitated to quantify the polysaccharide present in the supernatant. No precipitate was obtained with the mutant strain IST408 bceF::Tp, but the mutant strain IST408 bceD::Tp was mucoid and showed a reduction of only 25% of the ethanol-precipitated polysaccharide produced in the supernatant compared to the wild-type strain IST408 (Fig. 2b and d). Apparently, the polysaccharides produced by both IST408 and IST408 bceD::Tp strains are identical with respect to the neutral sugar composition since they contained glucose and galactose at a ratio of approximately 1:3 (Table 3). Moreover, both polymers exhibited similar degrees of acetylation in the range of 3.2 to 3.8 acetyl residues/repeat unit (Table 3).

FIG. 2.

(a) Growth curves and (b) cepacian production by B. cepacia IST408 (□), B. cepacia IST408 bceD::Tp (Δ), and B. cepacia IST408 bceF::Tp (•). The standard deviation in panel a is below 5%. In panel c, the viscosities of the aqueous solutions prepared with 5 g liter⁻¹ of the ethanol precipitate (considered equivalent to the EPS recovered) isolated from the culture supernatants of B. cepacia IST408 (□) and B. cepacia IST408 bceD::Tp (Δ) are shown. The data are based on mean values from the results of at least three independent cell cultivations. In panel d, the mucoid colony morphologies of the following strains are compared: (i) IST408, (ii) IST408 bceD::Tp, (iii) IST408 bceF::Tp, and (iv) IST408 bceF::Tp complemented in trans with pLM63-1, expressing the bceF gene, after 5 days of cultivation in mannitol medium supplemented with 1% arabinose. OD 640, optical density at 640 nm; cP, centipoise.

TABLE 3.

Neutral sugar composition and acetylation degree of the exopolysaccharides produced by B. cepacia strains^a

Strain	Amt of Glc (μg ml−1)	Amt of Gal (μg ml−1)	Glc/Gal ratio	Amt of acetyl (mg liter−1)	Total amt of sugar (mg liter−1)	Acetyl/sugar	No. of acetyl residues/RU
IST408	0.053 ± 0.005	0.156 ± 0.002	1.0:2.9	118.9 ± 8.3	902.1 ± 63.2	0.13	3.4
IST408 bceD::Tp	0.038 ± 0.002	0.115 ± 0.003	1.0:3.0	117.4 ± 4.0	909.1 ± 21.6	0.13	3.4

^aData represent means ± standard deviations. Values were calculated considering the following molecular masses: acetyl group, 43 g mol⁻¹; and heptasaccharide repeat unit 1,133 g mol⁻¹. Glc, glucose; Gal, galactose; RU, repeat unit.

The viscosities of aqueous solutions prepared with the purified EPSs produced by the two strains were different, with the solution prepared from the ethanol-precipitated material obtained from IST408 bceD::Tp exhibiting a slightly reduced viscosity for solutions with identical EPS concentrations (Fig. 2c). The protein contents present in the two EPS solutions derived from B. cepacia IST408 and B. cepacia IST408 bceD::Tp were 40 μg ml⁻¹ and 45 μg ml⁻¹, respectively. It is therefore unlikely that the registered differences in the viscosities of solutions prepared with the ethanol precipitate are due to different levels of protein contamination of the EPSs synthesized by the two strains. These observations, together with the previous characterization of the biopolymers, suggest a reduced molecular mass for the biopolymers produced by IST408 bceD::Tp.

To confirm that the insertion of the trimethoprim cassette into the bceF gene has no polar effect on the expression of the downstream bce genes, an in trans complementation experiment using the bceF gene cloned in the replicative plasmid pLM63-1 was performed. Since the expression of the bceF gene present in pLM63-1 is dependent on induction by arabinose and is repressed by glucose, the standard medium for EPS production (S medium) could not be used for the complementation experiment. As an alternative, mannitol medium supplemented with 1% of arabinose was used. Under these conditions, the introduction of pLM63-1 into the bceF mutant led to the recovery of cepacian biosynthesis, associated with the recovery of mucoidy of the colonies grown in solid medium (Fig. 2d). Despite the fact that all of the colonies obtained from the complementation of B. cepacia IST408 bceF::Tp with plasmid pLM63-1 were mucoid and that it was possible to recover EPS by ethanol precipitation of the cell-free culture supernatant, the EPS production yield was only 15% of the wild-type strain production level. This could be due to a deficient expression of the bceF gene under the control of the E. coli arabinose promoter that was tested in this work for the first time in B. cepacia. It can also be the result of the overexpression of the BceF protein leading to the alteration of the enzymatic activity or optimal relation between the several protein components involved in the biosynthesis of the polysaccharide.

BceD shows phosphatase activity.

In order to prove that BceD is a phosphatase, the protein was purified and the hypothesized enzymatic activity was determined. For that, the bceD gene from the IST408 strain lacking the start codon ATG was amplified and cloned in the expression vector pWH844. The resulting plasmid pLM211-1 allowed the production of the BceD protein with an N-terminal addition of 12 amino acid residues, including 6 histidines. Plasmid pLM211-1 was used to transform E. coli Sure cells, and after induction with 0.4 mM IPTG, a protein of approximately 18 kDa was overproduced, consistent with the calculated molecular mass of the fusion protein His₆-BceD (Fig. 3a, lanes 3 and 4).

FIG. 3.

(a) Coomassie blue-stained SDS-PAGE of the fractions obtained during purification of His₆-BceD from E. coli Sure cells harboring plasmid pLM211-1 containing the bceD gene. Lane 1, molecular mass standards; lane 2, protein cell extract obtained before IPTG induction; lane 3, protein cell extract obtained after IPTG induction; lane 4, His₆-BceD protein eluted from a Ni²⁺-NTA affinity chromatography column. (b) pH dependence of His₆-BceD phosphatase activity using 10 mM of PNPP as substrate. (c) Phosphatase activity of the His₆-BceD protein in the presence of increasing concentrations of the substrate PNPP at pH 6.5.

The phosphatase activity of His₆-BceD was assayed for its ability to cleave the artificial substrate PNPP. The purified fusion protein was able to efficiently hydrolyze this synthetic substrate at an optimum pH value of 6.5 (Fig. 3b). The corresponding kinetic constants, K_m and V_max, determined at 37°C, were 3.7 mM and 8.8 μmol min⁻¹ mg⁻¹, respectively. These values are in the same range as those reported for other prokaryotic low-molecular-mass phosphatases (44, 48, 59).

BceF autophosphorylation on tyrosine is prevented by a mutation in the Walker A motif.

To demonstrate the predicted activity of BceF as a tyrosine-phosphorylatable protein, the His tag fusion protein was purified to homogeneity by affinity chromatography. Efficient expression of an 82-kDa protein was consistent with the calculated molecular mass of the His₆-BceF fusion protein obtained in the soluble fraction of cell crude extracts following IPTG induction of E. coli Sure cells harboring plasmid pLM45-1 (Fig. 4a, lanes 2 and 3). The presence of phosphorylated tyrosine residues was detected in the purified His₆-BceF fusion protein by Western immunoblot analysis using antibodies against phosphotyrosine (Fig. 4b, lane 1, upper panel).

FIG. 4.

(a) Coomassie blue-stained SDS-PAGE of the fractions obtained during purification of His₆-BceF from crude cell extract of E. coli Sure cells harboring plasmid pLM45-1 expressing His₆-BceF. Lane 1, protein cell extract obtained before IPTG induction; lane 2, protein cell extract obtained after IPTG induction; lane 3, His₆-BceF protein eluted from a Ni²⁺-NTA affinity chromatography column. (b) Western blot analysis of His₆-BceF (lanes 1) or His₆-BceF_K563A (lanes 2) to detect the presence of phosphorylated tyrosine residues, using phosphotyrosine antibodies (upper panel), or the presence of the histidine tag, using antibodies against the His tag (lower panel).

The BceF protein has two conserved sequences at the C-terminal cytoplasmic domain (⁵⁵⁶GPTPGIGKS⁵⁶⁴ and ⁶⁶²VLID⁶⁶⁵) which are similar to the Walker A and Walker B motifs, respectively, described for many other organisms as being involved in ATP binding and hydrolysis (15). To assess the relevance of the ATP-binding motif in His₆-BceF with respect to protein tyrosine kinase activity, a modified fusion protein, with a specific substitution of the conserved lysine at position 563 from the Walker A motif, was constructed by site-directed mutagenesis. After overexpression and purification of His₆-BceF_K563A, this fusion protein was tested by immunoblotting for the presence of phosphorylated tyrosine residues. The results obtained indicate that the mutated protein did not react with the PT66 antibody (Fig. 4b, lane 2, upper panel). The stripping of the membrane and reprobing with the antibody against the His tag confirmed the presence of His₆-BceF and His₆-BceF_K563A (Fig. 4b, lanes 1 and 2, lower panel). These results firmly support the idea that the mutation in the ATP-binding motif blocked BceF autophosphorylation on tyrosine.

BceF is a substrate for BceD phosphatase activity.

In other bacterial systems, such as E. coli, Klebsiella pneumoniae, and Acinetobacter lwoffii, it was demonstrated that tyrosine autokinases are endogenous substrates for low-molecular-mass PTPs (44, 48, 59). To elucidate whether BceF is a substrate for BceD phosphatase activity, the purified proteins His₆-BceF and His₆-BceD were incubated together at 37°C in dephosphorylation buffer. Samples were removed at various incubation times and analyzed by Western immunoblotting using the PT66 antibody against phosphotyrosine to follow the phosphatase activity of BceD. In the presence of His₆-BceD, it was demonstrated that the tyrosine kinase His₆-BceF was dephosphorylated over time (Fig. 5, lanes 1 to 4), while in the absence of His₆-BceD, no His₆-BceF dephosphorylation occurred (Fig. 5, lanes 5 to 8). Estimation of the relative immunoblot signal intensity, using the lane 1 band as 100%, shows that after 22 h of incubation of the His₆-BceF protein with His₆-BceD, the band from lane 4 retained 63% of the signal, while the band from lane 8 showed similar signal intensity. This result suggests that the enzymatic activity of the tyrosine kinase His₆-BceF may be regulated in vivo by the dephosphorylating activity of His₆-BceD.

FIG. 5.

In vitro dephosphorylation of His₆-BceF by His₆-BceD. The two purified fusion proteins were incubated as described in Materials and Methods, and the reaction was stopped at 0, 1, 3, and 22 h. The samples were analyzed by immunodetection using phosphotyrosine antibodies. In lanes 1 to 4, His₆-BceF was incubated with His₆-BceD for the indicated times: lane 1, 0 h; lane 2, 1 h; lane 3, 3 h; lane 4, 22 h. In lanes 5 to 8, His₆-BceF was incubated under identical conditions but in the absence of His₆-BceD, for the same periods of time as in lanes 1 to 4, respectively, as a control experiment.

bceD and bceF expression is required for maximal biofilm size.

The sizes of the biofilms formed by the mucoid strain IST408 and the two mutant strains, IST408 bceD::Tp and IST408 bceF::Tp, were quantified after 24 and 48 h of incubation in S medium at 30°C. These assay conditions were previously shown to lead to thicker biofilms and the highest cepacian production yield (17). Although the amounts of the biofilms formed by the three strains were similar after 24 h of incubation, the amount of biofilm formed by the two mutant strains after 48 h was remarkably smaller than the amount of biofilm formed by IST408 (Fig. 6a). This result is particularly interesting when we keep in mind that both mutant strains were derived from the same strain, but while IST408 bceD::Tp still produces a significant amount of EPS, the isogenic IST408 bceF::Tp strain does not produce detectable levels of cepacian.

FIG. 6.

Size of the biofilm formed after 24 h (open bars) or 48 h (solid bars) of cultivation, without shaking, of (a) the CF isolate B. cepacia IST408 and the insertion mutants B. cepacia IST408 bceD::Tp and B. cepacia IST408 bceF::Tp in the wells of polystyrene microtiter dishes containing S medium at 30°C and (b) B. cepacia IST408/pMLBAD-Cm, B. cepacia IST408 bceD::Tp/pMLBAD-Cm, B. cepacia IST408 bceD::Tp/pLM69-1, B. cepacia IST408 bceF::Tp/pMLBAD-Cm, and B. cepacia IST408 bceF::Tp/pLM63-1 in the wells of polystyrene microtiter dishes containing mannitol medium supplemented with 1% of arabinose at 30°C. The microtiter dishes were rinsed to remove planktonic cells, and the biofilms were stained with crystal violet. Absorbance at 590 nm (A₅₉₀) quantifies the amount of crystal violet associated with the biofilm after staining.

Due to the reasons described before, the size of the biofilm from the bceD and bceF complemented mutants could not be assessed in S medium, and so, the alternative mannitol medium supplemented with 1% of arabinose was tested (Fig. 6b). Under these growth conditions, the maximal size of the biofilm registered was below 2 units of absorbance and the results obtained after 24 or 48 h of incubation were similar. Although this alternative growth medium was not optimized to assess biofilm formation in vitro, it was possible to observe an increase in the size of the biofilm formed by the IST408 bceD::Tp and IST408 bceF::Tp mutants when complemented with the lacking genes expressed from recombinant plasmids (Fig. 6b).

DISCUSSION

In this study, we report results on the functional analysis of the BceD and BceF proteins from the opportunistic human pathogen Burkholderia cepacia as well as their involvement in cepacian biosynthesis and in the size of the biofilms formed. The protein BceF was found to be tyrosine phosphorylated, and the importance of the ATP-binding site for tyrosine phosphorylation was demonstrated. The other protein under study, BceD, is homologous to bacterial low-molecular-mass acid phosphotyrosine phosphatases and dephosphorylates BceF, in agreement with the demonstrated activity of Wzb phosphatases from A. lwoffii and E. coli (44, 59, 60).

The disruption of the bceF gene resulted in an EPS-defective strain, as shown for the homologous proteins ExoP from Sinorhizobium meliloti, Wzc_CA from E. coli K-12, Wzc_cps from E. coli K30, and GelC/GelE from Sphingomonas elodea, whose lack of expression resulted in strong reduction of the amount of the polysaccharides produced (2, 21, 40, 46, 60). The disruption of the bceD gene reduced wild-type cepacian production by about 25% and led to a slight reduction of the polymer viscosity, while in E. coli K30, E. coli K-12, Erwinia amylovora Ea1/79, and Acinetobacter lwoffii RAG-1, the production of capsular polysaccharides or exopolysaccharides is dramatically abrogated following disruption of the wzb_CPS, wzb_CA, amsI, and wzb phosphatase genes, respectively (6, 44, 60, 62). However, in agreement with our observation, Minic et al. (39) recently described that the lack of the EpsB phosphatase from Streptococcus thermophilus implicates only a slight reduction in the amount of the exopolysaccharide produced.

A hypothetical model proposed before to explain the polymerization and export process of bacterial polysaccharide synthesis predicts the existence of a multienzyme complex, composed of several proteins acting together, including a flippase, a polymerase, an outer membrane lipoprotein complex, a tyrosine kinase, and a phosphotyrosine phosphatase (44, 45, 49). It has been hypothesized that the presence of all of these protein components, with the right stoichiometry, is required for the activity of this multienzyme complex. This model is consistent with the phenotypes registered with all of the mutants lacking the tyrosine kinase from all of the species tested (20, 25, 43, 44, 46, 59, 62), including the IST408 bceF mutant examined here. Indeed, this mutant does not produce cepacian. On the other hand, it appears that the mode of action of the phosphotyrosine phosphatase, another member of the polymerization/export complex in gram-negative bacteria, is strain specific. In fact, in Burkholderia cepacia, a mutant strain for the BceD phosphatase still retains 75% of wild-type cepacian production ability. However, the biopolymer produced by the mutant exhibits different rheological properties, suggestive of a lower molecular mass. The production of a significant amount of cepacian in the absence of the phosphatase BceD and the fact that BceF is phosphorylated on tyrosine and requires an active Walker A for autophosphorylation suggest that BceF has to be phosphorylated to allow cepacian production. This contrasts with the requirements for K30 capsular polysaccharide production in E. coli K30, where although the kinase Wzc_CPS has to be phosphorylated for CPS production, the mutation in the Wzb_CPS phosphatase resulted in trace amounts of CPS on the cell surface (62).

The ability of bacteria to form biofilms has been associated with their capacity to cause disease in the human host (14). It is well documented that the EPSs produced by E. coli and Vibrio cholerae are essential for the development of mature biofilms, as strains producing null or small amounts of EPS produce only thin biofilms that are devoid of normal architecture (18, 63). Alginate production in Pseudomonas aeruginosa has also been correlated with the ability of this bacteria to form thick and mature biofilms, where the bacteria exhibit higher resistance to antimicrobials (26, 28) and to host phagocyte killing. Therefore, it is considered that biofilm formation contributes to persistent infections that may lead to the characteristic chronic deterioration of CF patients' airways (24). In a previous work, we compared the abilities of the CF isolate B. cepacia IST408 and three isogenic mutants to form biofilms (17, 41). Two of these plasposon insertion mutants had the plasposon inserted in the tyrosine kinase bceF gene, and in the third mutant, the transposon was in the bceI gene, putatively coding for a polysaccharide polymerase. The amount of biofilm formed by the three EPS-defective mutants was strongly reduced compared with that formed by the mucoid wild-type strain (17). Since the two bceF plasposon mutants were polar, it was not possible to be sure whether the EPS-deficient phenotype is due to impaired expression of the bce downstream genes or to the absence of a functional tyrosine kinase. The results obtained in this work, using the nonpolar bceF mutant strain constructed here, allowed the confirmation of the crucial role that a functional tyrosine kinase has in EPS biosynthesis. The inability of the B. cepacia bceF mutant to form, in vitro, biofilms of the same size as those produced by the parental strain IST408 is consistent with the notion that EPS biosynthesis is an important factor in obtaining biofilms with the maximal size. However, although it produced about 75% of the EPS of the parental strain IST408, the bceD mutant also formed biofilms with a reduced size, close to the size of bceF mutant biofilms. Since it is believed that alterations in EPS structure significantly change its physicochemical properties, it is possible that the less viscous EPS produced by the bceD mutant strain may prevent the stable development of a biofilm of larger size. Although the development of thick biofilms in the B. cepacia complex certainly involves other strain-dependent factors besides the biosynthesis of EPS (17), this study reinforces the notion of EPS involvement in the development of biofilms with a potential role in the persistence and virulence of respiratory infections caused by these opportunistic bacteria.