University at Buffalo - The State University of New York
Skip to Content
Distribution and Expression of the ZmpA Metalloprotease in the Burkholderia cepacia Complex

Distribution and Expression of the ZmpA Metalloprotease in the Burkholderia cepacia Complex

  1. P. A. Sokol*
  1. Department of Microbiology & Infectious Diseases, University of Calgary, Calgary, Alberta T2N 4N1, Canada


The distribution of the metalloprotease gene zmpA was determined among strains of the Burkholderia cepacia complex (Bcc). The zmpA gene was present in B. cepacia, B. cenocepacia, B. stabilis, B. ambifaria and B. pyrrocinia but absent from B. multivorans, B. vietnamiensis, B. dolosa, and B. anthina. The presence of zmpA generally correlated with extracellular proteolytic activity with the exception of five strains, which had zmpA but had no detectable proteolytic activity when skim milk agar was used as a substrate (zmpA protease deficient). Western immunoblot experiments with anti-ZmpA antibodies suggest that the zmpA protease-deficient strains do not secrete or accumulate detectable ZmpA. Transcriptional zmpA::lacZ fusions were introduced in selected strains of the Bcc. zmpA::lacZ was expressed in all strains, but expression was generally lower in the zmpA protease-deficient strains than in the zmpA protease-proficient strains. Quantitative reverse transcriptase real-time PCR demonstrated that zmpA protease-deficient strains did express zmpA mRNA, although at various levels. ZmpA has previously been shown to be positively regulated by the CepIR quorum-sensing system. Addition of exogenous AHLs did not restore extracellular protease production to any of the zmpA protease-deficient strains; however, introduction of cepR in trans complemented protease activity in two of five strains. Extracellular proteolytic activity was restored by the presence of zmpA in trans in two of the five strains. These studies suggest that although some strains of the Bcc contain the zmpA gene, multiple factors may influence its expression.

The Burkholderia cepacia complex (Bcc) is a group of opportunistic pathogens that can cause severe respiratory infections in individuals with cystic fibrosis (CF) (25, 28, 41). These organisms are resistant to most available antibiotics, and some strains are highly transmissible. Chronic colonization with the Bcc is associated with a more rapid deterioration of lung function than either colonization with Pseudomonas aeruginosa alone or no colonization (10, 41). The Bcc presently consists of nine species: B. cepacia (formerly genomovar I) (45), B. multivorans (formerly genomovar II) (45), B. cenocepacia (formerly genomovar III) (44), B. stabilis (formerly genomovar IV) (46), B. vietnamiensis (formerly genomovar V) (15), B. dolosa (formerly genomovar VI) (6, 47), B. ambifaria (formerly genomovar VII) (7), B. anthina (formerly genomovar VIII) (43), and B. pyrrocinia (formerly genomovar IX) (43). Recent studies have revealed heterogeneity in the recA gene sequence among different isolates of B. cenocepacia, which led to the distinction of four recA lineages, referred to as IIIA, IIIB, IIIC, and IIID (26, 44).

Cell-cell signaling is used by certain bacteria to monitor their own population density and modulate gene expression accordingly once a critical cell density has been reached (12). Two sets of cell-cell signaling genes are present in B. cenocepacia. The cepIR genes are widely distributed in the Bcc (16, 23), whereas the cciIR genes are only found in B. cenocepacia strains containing the cci genomic island (1). The N-acyl homoserine lactones (AHLs) produced by CepI and CciI have been identified as N-octanoyl-homoserine lactone (OHL) and N-hexanoyl-homoserine lactone (HHL) (23, 30). Both the cepIR and cciIR genes influence protease production (30, 40).

A potential virulence trait that may contribute to the severity of Bcc infections is the ability to secrete extracellular proteases. A total of 69 to 88% of clinical isolates from the Bcc have been reported to produce extracellular proteases (13, 14, 31, 32). Variability in virulence has been shown between and within species of the Bcc (4, 5, 18, 22, 29, 37); thus, it is possible that the ability to produce extracellular proteases may account for some of these differences.

ZmpA (formerly PSCP) is a zinc metalloprotease originally described in B. cenocepacia Pc715j (31). We have shown that ZmpA is an important virulence factor in some strains, since a K56-2 zmpA mutant was less able to persist in a chronic lung infection model and caused less pathological damage to the lung than the parent strain, K56-2 (9). Interestingly, a zmpA mutant in strain Pc715j was similar in virulence to the parent strain, suggesting that other proteases in this strain may compensate for the zmpA defect (9). ZmpA is able to degrade several biologically important substrates, including neutrophil α-1 proteinase inhibitor, α2-macroglobulin, gamma interferon, type IV collagen, and fibronectin, which may account for its ability to cause tissue damage as well as modulate the host immune system (19). The objectives of the present study were to determine the distribution and expression of the zmpA gene within different species of the Bcc.


Strains, plasmids, and growth conditions.Bacterial strains and plasmids used in this study are described in Table 1. Escherichia coli DH5α and strains of the Bcc were routinely grown at 37°C in Luria-Bertani broth (LB) (Invitrogen, Burlington, Ontario, Canada) or on 1.5% LB agar plates. For protease assays, dialyzed brain heart infusion agar containing 1.5% skim milk (D-BHI milk) (39) was used. For AHL extraction and quantification, strains from the Bcc were grown in tryptic soy broth (BD Diagnostic Systems, Franklin Lakes, N.J.) at 30°C for 12 h. The following concentrations of antibiotics were used when necessary: 100 μg of ampicillin, 50 μg of kanamycin, and 800 μg of trimethoprim per ml for E. coli and 100 μg, 400 μg, or 800 μg of trimethoprim per ml for the Bcc.


Bacterial strains and plasmids used in this study

DNA manipulations.Molecular biology techniques were performed as generally described by Sambrook et al. (35). Genomic DNA was isolated according to the method of Walsh et al. (48). Recombinant plasmids were electroporated into E. coli DH5α or Bcc cells using a Gene Pulser (Bio-Rad, Richmond, CA) according to the manufacturer's instructions or transformed into E. coli DH5α cells made competent by treatment with CaCl2 (24). For Southern hybridization, approximately 200 ng of DNA was digested with PstI, separated by agarose gel electrophoresis, and transferred to GeneScreen II membrane (Perkin-Elmer). The blot was hybridized with a 250-bp PCR product amplified with primers 1-2RRP3 and 1-2FFP2 as described below and labeled with 32P with the Rediprime II Random Prime Labeling System (Amersham Biosciences).

PCR conditions.The zmpA gene was amplified using oligonucleotide primers 1-2RRP3 (CCAACCTGAACTATTCGG) and 1-2FFP2 (TGCGGATCGTGGCGC) or PNPF (ATCACCTTCTAGCATCC) and PNPR (CCTTCTCGTCATTCACC). PCRs were performed in the AB Applied Biosystems GeneAmp PCR System 2400 as follows: initial denaturation at 96°C for 4 min, denaturation at 96°C for 30 s, primer annealing at 56°C for the amplification of zmpA for 30 s, and extension at 72°C for 4.5 min for 30 cycles, with a final 10-min extension at 72°C. PCRs were routinely run on a 0.8 to 1.0% agarose gel, and bands were visualized by being stained with ethidium bromide. PCR products were purified from the agarose gel with the QIAquick gel extraction kit (Qiagen, Mississauga, Ontario, Canada) according to the manufacturer's instructions. Cloning of PCR products was performed using the TOPO-TA cloning kit into vector pCR2.1-TOPO, according to the manufacturer's directions (Invitrogen).

Real-time quantitative reverse transcriptase PCR (RT-PCR).Total RNA was isolated using RiboPure bacteria (Ambion, Austin, TX) from approximately 10 A600 U of culture incubated for 18 h at 37°C with shaking at 250 rpm. Residual DNA was removed by treatment with RNase-free DNase I using 8 U of TURBO DNase, followed by DNase Inactivation reagent (Ambion); the concentration was determined by measuring absorbance at 260 nm. To confirm that residual DNA was removed, control PCRs were performed using PlatinumTaq DNA polymerase (Invitrogen Canada, Inc.) and the Fndh and Rndh primers. Oligonucleotide primer sequences were designed with Primer Express software, version 2 (Applied Biosystems, Foster City, CA), based on the unpublished genome sequence of B. cenocepacia J2315 ( Primers EBS-E62 (5′-CGCGGCAGACATCGACTAC-3′) and RBS-E62 (5′-AGATGCCGTTGCGGTTGT-3′) were used to amplify zmpA. Primers Fndh (5′-GCGATCGGGCTGTACAAGTT-3′) and Rndh (5′-AGTGGCTCAGCGACTGGAA-3′) were used to amplify the NADH dehydrogenase gene (ndh) (BCAM0166) as a control gene, since ndh was expressed at a similar level in all strains examined.

Purified zmpA and ndh RNAs were used to construct standard curves. pTOPOZMPA (9) was digested with SpeI and purified with a Qiagen gel extraction kit (Qiagen, Mississauga, ON). The linear plasmid was used as a template to synthesize RNA using the T7 Ribomax Express Large Scale RNA Production system (Promega, Madison, WI). RNA was purified with a RNeasy Mini Kit (Qiagen), according to the manufacturer's recommendations. Primers (5′-ATGGACAACACCACGCCCAC-3′) and (5′-TCAGTGCAGCTTGATCGACG-3′) were used to amplify a 1.3-kb fragment containing ndh which was cloned into pCR2.1Topo (Invitrogen) downstream of the T7 promoter, generating pBS7. In vitro transcription of pBS7 was performed as above. The purified zmpA and ndh RNAs, in concentrations ranging from 20 pg to 0.2 fg, were used to construct standard curves for relative quantification with the TaqMan instrument and SYBR green RT-PCR reagents (Applied Biosystems). RT-PCR was performed with 25 μl (1× SYBR Green PCR Master Mix, 0.25-U/ml MultiScript reverse transcriptase, 0.4-U/ml RNase inhibitor, and 50 nM forward and reverse primers) using the following PCR cycles: 1 cycle at 48°C for 30 min; 1 cycle at 95°C for 10 min; and 45 cycles, each at 95°C for 15 s and 60°C for 1 min. Signal intensity was measured at the end of each elongation phase. A standard curve was generated from the cycle threshold of standard dilution series by the TaqMan 7500 system SDS software (version 1.3). All quantitative RT-PCRs (qRT-PCRs) were performed twice in triplicate.

β-Galactosidase assays.Expression of zmpA was quantified using a liquid β-galactosidase assay (34). Overnight starter cultures were subcultured 1/200 in 50 ml of medium and grown at 37°C for 20 h with shaking at 200 rpm. All assays were conducted in triplicate and repeated at least twice. Values shown represent the percentage of activity of the mean Miller units obtained with the K56-2 control cultures.

Protease assay on D-BHI skim milk plates.Overnight cultures were subcultured 1/50 in 50 ml of culture. Cells were grown until mid-log to late log phase (optical density at 600 nm [OD600] = 0.5 to 1.0) and normalized to an OD600 of 0.3; 3 μl was spotted in triplicate onto D-BHI milk plates (39). Plates were incubated at 37°C for up to 48 h and examined for zones of clearing around the colonies at 24 and 48 h by determining the length of the radius (from the outside of the colony to the edge of the zone). Each protease assay was repeated at least twice in triplicate.

AHL extraction and quantification.Agrobacterium tumefaciens A136 (pCF218)(pMV26) was grown at 30°C in LB or on LB solidified with 1.5% agar. Medium was supplemented with 25 μg of kanamycin per ml and 4.5 μg of tetracycline per ml or 3 μg tetracycline per ml in liquid broth as required. AHLs were extracted from the following strains: B. cepacia CEP 509; B. cenocepacia K56-2, ATCC 17765, CEP 511, J415, C5424, and J2315; B. stabilis LMG 14086 and LMG 14294; and B. ambifaria LMG 19182T as described previously (20). AHLs were quantitated with the luxCDABE reporter Agrobacterium tumefaciens A136(pCF218)(pMV26) as previously described (40). Briefly, AHL extracts were diluted 10 fold, and dilutions from 1/1,000 to 1/1,000,000 were assayed using an overnight culture of A. tumefaciens A136(pCF218)(pMV26). Ten microliters of extract, 10 μl of A. tumefaciens A136(pCF218)(pMV26), and 80 μl of LB were mixed in 96-well black plates with a clear flat bottom (Costar; Corning, Inc.). Luminescence was measured using a Wallac Trilux Luminescence counter (Perkin-Elmer Life Sciences) after 8 h of incubation at 30°C. Synthetic OHL (Sigma-Aldrich, Oakville, Ontario, Canada) was used to prepare a standard curve for comparison.

Cross-feeding assays.To determine if the addition of exogenous AHLs would restore protease production to zmpA protease-deficient strains, cross-feeding and AHL addback assays were performed. The following strains were tested: B. cepacia CEP 509 and B. cenocepacia J415, CEP 511, C5424, and J2315. B. cenocepacia K56-I2, a cepI::tp mutant shown to be protease negative but able to have protease production restored with the addition of exogenous AHLs (20), was used as a control. B. cenocepacia K56-2 was streaked perpendicularly to the test strains on D-BHI milk plates (39). The ability of zmpA protease-deficient strains to restore protease production to B. cenocepacia K56-I2 by the production of sufficient AHLs was also determined. Plates were incubated for up to 48 h and observed for a zone of clearing around the junction, which would indicate restoration of extracellular proteolytic activity by the AHLs produced by K56-2. Alternatively, 2 nmol of synthetic OHL and HHL (Fluka) were added to a filter disk (Becton Dickinson) and placed adjacent to the test strain, and the plates were observed for the restoration of protease production. Test strains were streaked perpendicularly to B. cenocepacia K56-I2 and observed for protease production at the junction.

Preparation of cell culture supernatants and extracts.B. cepacia CEP 509; B. cenocepacia Pc715j, K56-2, CEP 511, ATCC 17765, J2315, J415, and C5424; B. stabilis LMG 14086 and LMG 14294; and B. ambifaria LMG 19182T were grown overnight, subcultured to a final OD600 of 0.05, and grown for 18 h with shaking at 200 rpm. Cultures were centrifuged at 10,000 × g for 1 h at 4°C. The resultant supernatants were precipitated by the slow addition of cold acetone-trichloracetic acid (final concentration, 10% [vol/vol]) for 4 h at 4°C with stirring, followed by placement at −20°C for 2 h and then at 4°C for 1 h. The precipitated culture supernatants were centrifuged at 15,000 × g for 1 h, and the pellets were washed three times with 100% acetone. After being air dried, the resultant pellets were resuspended in 30 mM Tris-HCl, pH 7.5. The amount of protein present was quantified by the method of Bradford with a protein assay kit (Bio-Rad Laboratories). Ten micrograms of each precipitated protein preparation was separated by tricine-sodium dodecyl sulfate-14% polyacrylamide gel electrophoresis (36). For total cell extracts, cell pellets were resuspended in 30 mM Tris-HCl (pH 8.0), boiled for 20 min, and electrophoresed as above. Periplasmic extracts were prepared by resuspension of pellets from 100 ml of culture in 40 ml of 30 mM Tris-HCl (pH 8.0) containing 20% sucrose, followed by incubation at ambient temperatures for 20 min. The cells were centrifuged at 3,500 rpm for 20 min, resuspended in 2 ml of ice-cold 5 mM MgSO4, incubated on ice for 20 min, and centrifuged as above; the supernatant enriched for periplasmic protein was collected and electrophoresed. Gels were transferred to polyvinylidene difluoride membranes (Millipore), essentially as described by Towbin et al. (42). After being blocked with 5% bovine serum albumin, the blots were reacted with rat anti-ZmpA antibody (19).

Nucleotide sequencing.Nucleotide sequencing was performed by either Macrogen, Inc. (Seoul, South Korea), or the University of Calgary Core DNA Services (Calgary, Alberta, Canada). Sequence analysis was performed with DNAMAN software (Lynnon Biosoft, Vaudreuil, Quebec, Canada). The zmpA sequences from strains K56-2, C5424, J415, and ATCC 17765 have been deposited in GenBank and assigned the following accession numbers, respectively:DQ069247, DQ06948, DQ069249, and DQ06950.


Distribution of the zmpA gene in the Burkholderia cepacia complex.Oligonucleotide primers 1-2RRP3 and 1-2FFP2 amplified a 228-bp PCR product from all strains of B. cepacia, B. cenocepacia, B. ambifaria, and B. pyrrocinia (Fig. 1) but not from strains of B. multivorans, B. vietnamiensis, B. dolosa, and B. anthina. A weak PCR product was consistently obtained for B. cepacia ATCC 17759 (lane 3) and LMG 17997 (lane 5), B. cenocepacia PC184 (lane 22), and B. stabilis LMG 14294 (lane 26), C7322 (lane 27), LMG 14086 (lane 28), and LMG 18888 (lane 29). To confirm that the faint 228-bp product from these strains was truly an amplicon of zmpA, these PCR products were cloned and sequenced. The deduced amino acid sequence of the products amplified from each of these strains was 95 to 100% identical to the portion of Pc715j ZmpA amplified with this set of primers, indicating that zmpA is present in these strains (data not shown). These results suggest that zmpA is present in B. cepacia, B. cenocepacia, B. stabilis, B. ambifaria, and B. pyrrocinia but absent from B. multivorans, B. vietnamiensis, B. dolosa, and B. anthina (Fig. 1 and Table 2).

FIG. 1.

Detection of zmpA in the B. cepacia complex by PCR. Lanes: 1, 1-kb Plus DNA ladder (Invitrogen); 2, ATCC 25416T; 3, ATCC 17759; 4, CEP 509; 5, LMG 17997; 6, C5393; 7, LMG 13010; 8, C1576; 9, CF-A1-1; 10, JTC; 11, C1962; 12, ATCC 17616; 13, 249-2; 14, J2315; 15, Pc715j; 16, K56-2; 17, C5424; 18, C6433; 19, C1394; 20, 1-kb Plus DNA ladder (Invitrogen); 21, 1-kb Plus DNA ladder (Invitrogen); 22, PC184; 23, CEP 511; 24, J415; 25, ATCC 17765; 26, LMG 14294; 27, C7322; 28, LMG 14086; 29, LMG 18888; 30, PC259; 31, LMG 16232; 32, FC441; 33, LMG 10929T; 34, L06; 35, LMG 18943; 36, LMG 17828; 37, LMG 19182T; 38, LMG 19467; 39, H2O control; 40, 1-kb Plus DNA ladder (Invitrogen); 41, 1-kb Plus DNA ladder (Invitrogen); 42, LMG 19468T; 43, LMG 20980T; 44, LMG 20983; 45, LMG 16670; 46, LMG 20982; 47, LMG 21822; 48, LMG 14191T; and 49, H2O control. B. viet., B. vietnamiensis; B. dol., B. dolosa; B. ambif., B. ambifaria; B. anth., B. anthina; and B. pyrr., B. pyrrocinia.


Distribution of protease gene and extracellular protease activity in the Burkholderia cepacia complex

To confirm that the species negative for zmpA by PCR analysis lacked the gene rather than contained a zmpA gene with nucleotide sequence variation in the primer binding site, Southern hybridization analysis with an internal zmpA probe was also performed with selected strains of each species (Fig. 2 and Table 2). Strains negative for zmpA by PCR did not hybridize with the zmpA probe; however, the probe clearly hybridized with two B. stabilis strains that were only weakly positive by PCR.

FIG. 2.

Detection of zmpA in the B. cepacia complex by Southern hybridization. Southern blot of genomic DNA from representative strains digested with PstI, hybridized with a 250-bp fragment internal to zmpA, amplified by PCR, and labeled with 32P. Lane 1, ATCC 25416; lane 2, C5393; lane 3, LMG 13010; lane 4, J2315; lane 5, K56-2; lane 6, LMG 14294; lane 7, LMG 14086; lane 8, PC259; lane 9, LMG 16232; lane 10, LMG 18943; lane 11, LMG 19468T; lane 12, LMG 19467; lane 13, LMG 17828; lane 14, LMG 20980; lane 15, LMG 20982; lane 16, LMG 21822; lane 17, Pc715j.

Protease activity in the Burkholderia cepacia complex.Corbett et al. (9) reported that a K56-2 zmpA isogenic mutant produced little or no zone of clearing on D-BHI skim milk plates, suggesting that in K56-2 ZmpA is the major extracellular protease capable of cleaving casein. To determine if the presence of zmpA correlated with casein-degrading activity, protease assays were performed on 40 strains of the Bcc. Most isolates from B. cepacia, B. cenocepacia, B. stabilis, B. ambifaria, and B. pyrrocinia were positive for extracellular proteolytic activity, but no isolates from B. multivorans, B. vietnamiensis, B. dolosa, and B. anthina had detectable extracellular proteolytic activity (Table 2), suggesting that protease activity generally correlated with the presence of zmpA.

B. cenocepacia C1394 and Pc184 were protease negative at 24 h but had zones comparable to K56-2 and Pc715j at 48 h (Table 2). B. stabilis C7322, although protease negative at 24 h, had zones comparable to those of the other B. stabilis strains by 48 h. These data suggest that induction of protease activity may be slower in these three strains but that they produce similar amounts of final activity.

Interestingly, B. cepacia CEP 509; B. cenocepacia J2315, C5424, and J415; and B. stabilis LMG 14294 had little or no detectable proteolytic activity despite the presence of the zmpA gene. J415 had a zone of clearing of <1 mm at 48 h. To determine if ZmpA was secreted by these strains, protein profiles of culture supernatants of these strains and selected zmpA protease-proficient strains (Pc715j, K56-2, CEP 511, ATCC 17765, LMG 14086, and LMG19182T) were examined on Western blots reacted with anti-ZmpA. Anti-ZmpA reacted with an approximately 36-kDa protein in supernatants from B. cenocepacia Pc715j, K56-2, and ATCC 17765 and B. ambifaria LMG 19182T. The zmpA protease-deficient strains and two of the zmpA protease-proficient strains (Cep 511 and LMG14086) did not react with the antibody to ZmpA (Fig. 3). These data suggest that zmpA may not be properly transcribed, translated, or secreted in the zmpA protease-deficient strains. To determine if there was an accumulation of ZmpA intracellularly in these strains, whole-cell lysates and periplasmic extracts were also examined by Western blotting. No ZmpA was detected in either fraction from either the zmpA protease-proficient or the zmpA protease-deficient strains (data not shown).

FIG. 3.

Detection of ZmpA in culture supernatants. Western immunoblot of trichloroacetic acid-precipitated supernatants reacted with anti-ZmpA antibody. Lane 1, Pc715j; lane 2, K56-2; lane 3, Cep 511; lane 4, ATCC 17765; lane 5, LMG 14086; lane 6, LMG 19182; lane 7, Cep 509; lane 8, J2315; lane 9, J415; lane 10, C5424; lane 11, LMG 14294; lane 12, ZmpA.

Expression of zmpA in the Burkholderia cepacia complex.To determine why strains J2315, C5424, J415, LMG 14294, and CEP 509, which contain zmpA, lack extracellular protease activity, the ability of these strains to express zmpA was compared to selected zmpA protease-proficient strains. Plasmids pSG208, which contains a zmpA::lacZ transcriptional reporter fusion (40), and pUCP28T (vector control) were introduced in the zmpA protease-proficient strains K56-2, LMG 19182T, ATCC 17765, LMG 14086, and CEP 511 and the zmpA protease-deficient strains J2315, C5424, J415, LMG 14294, and CEP 509. Plasmid pSG208 was unstable in Pc715j. The expression of the zmpA::lacZ transcriptional fusion was determined by β-galactosidase assays (Fig. 4). β-Galactosidase activity was generally higher in the protease-proficient strains than in the protease-deficient strains, with the exception of strains Cep511 and J415.

FIG. 4.

Comparison of zmpA::lacZ expression in zmpA protease-proficient and zmpA protease-deficient strains. Expression of the zmpA::lacZ fusion contained on pSG208 was determined by measuring β-galactosidase assays. Assays were conducted in triplicate and repeated at least twice. Values shown are the mean percentages of K56-2 activity, which is represented as 100%, and are of a representative experiment. β-Galactosidase activity resulting from strains with the pUCP28T vector control was negligible and is not shown.

Expression of zmpA was also measured by quantitative real-time reverse transcriptase PCR. All of the strains expressed zmpA mRNA, but the level of expression did not correlate with activity on D-BHI milk plates (Fig. 5). Together, these experiments suggest that all of the strains that contain zmpA have the ability to express this gene, although the expression levels vary considerably among strains.

FIG. 5.

Comparison of zmpA expression in zmpA protease-proficient and zmpA protease-deficient strains by quantitative real-time RT-PCR. Expression of zmpA and ndh were determined by absolute quantification. Concentrations of the mRNA level of each strain are expressed as moles of target gene mRNA per mole compared to the concentration of the standard curve. The bars represent the means ± standard deviation of triplicate assays. Black bars, zmpA; gray bars, ndh.

Complementation of zmpA protease-negative strains.One possible explanation for the absence of detectable extracellular proteolytic activity in the zmpA protease-deficient strains is that zmpA contains a mutation that affects enzyme activity in these strains. Recently, the genomic sequence of B. cenocepacia J2315 was completed by the Wellcome Trust Sanger Institute ( The J2315 zmpA gene is 99.5% identical to the B. cenocepacia Pc715j zmpA gene (9). The only difference between the amino acid sequences is a substitution of a serine at residue 489 in Pc715j for the threonine in J2315. The ability of Pc715j zmpA to restore protease activity in strains LMG 14294, C5424, J415, J2315, and CEP 509 was determined by introducing pSG200, a high-copy-number plasmid, with zmpA. Extracellular protease activity was restored in J415 and C5424 when zmpA was provided in trans, but not in LMG 14294, J2315, or CEP 509 (Fig. 6).

FIG. 6.

Complementation of zmpA protease-deficient strains with Pc715j zmpA in trans. The assay shown is representative of two experiments performed in triplicate. Only strains J415 and C5424 had extracellular proteolytic activity when harboring pSG200. Strain CEP 509 produces a dark melanin-like pigment, which is visible around the colonies, but no zone of clearing was detectable.

The ability of Pc715j zmpA to restore protease activity in J415 and C5424 suggests that zmpA in these strains may contain a mutation that alters enzyme activity. The J415 and C5424 zmpA genes were amplified by PCR using primers PNPF and PNPR, and their nucleotide sequences were determined. C5424 ZmpA was 99.82% identical to Pc715j ZmpA and contained a threonine at residue 489. J415 ZmpA was 96.81% identical to Pc715j ZmpA and contained 18 amino acid changes, including a threonine at residue 489. To determine if this serine-threonine substitution correlated with the lack of extracellular proteolytic activity, we sequenced zmpA from two zmpA protease-proficient strains, K56-2 and ATCC 17765 (data not shown). Both strains also harbored a threonine at position 489, indicating that the absence of extracellular proteolytic activity in B. cenocepacia C5424, J2315, and J415 could not be explained by threonine in this position. There were no other consistent sequence differences that could account for the lack of protease activity in these strains.

Effect of cepI and cepR on zmpA expression.We have recently demonstrated that the cepIR cell-cell signaling system positively regulates the expression of zmpA at the transcriptional level (40). To determine if the lack of protease activity in the zmpA protease-deficient strains might be due to differences in cepI or cepR, the effects of exogenous AHL and the ability of cepR in trans to restore protease activity were determined. AHLs were extracted from the zmpA protease-proficient and zmpA protease-deficient strains and tested for activity as previously described (40). AHL activity was detected in all the strains tested (data not shown). Cross-feeding assays were also used to determine if the lack of protease activity was related to AHL production. The zmpA protease-deficient strains were streaked perpendicularly to B. cenocepacia K56-2 (which produces both OHL and HHL) or supplemented with synthetic OHL or HHL to determine if the availability of exogenous AHLs would restore extracellular protease activity (21). Neither cross-streaking with B. cenocepacia K56-2 nor the addition of synthetic OHL or HHL was able to restore protease production in the zmpA protease-deficient strains (data not shown), indicating that the lack of protease activity was not due to concentrations of AHLs insufficient to activate zmpA expression. We also tested for the ability of the zmpA protease-deficient strains to restore extracellular proteolytic activity in a cepI mutant, K56-I2. All zmpA protease-deficient strains were able to restore protease activity to K56-I2 in the cross-streaking assay, confirming that these strains produce sufficient functional AHLs to activate transcription of the zmpA gene by CepR. The ability of cepR, introduced on a high-copy-number plasmid (pSLR100) to restore protease activity to the zmpA protease-deficient strains, was also determined (Table 3). Introduction of pSLR100 increased protease activity in J2315 and C5424 compared to the same strains without the plasmid or with the vector control.


Complementation of protease activity by pSLR100 (cepR)a

The predicted promoter sequences of J2315 and Pc715j zmpA were compared and found to be identical in both the −35 (TTGTAA) and −10 (TTCTAGCAT) regions ( The predicted cep box sequence upstream of zmpA was also identical in these strains, suggesting that although an increase in cepR copy number results in protease activity in J2315, the CepR binding site is not altered in J2315. The J3215 cepR sequence is identical to that of K56-2, indicating that the difference in activity is not due to a cepR mutation.


The distribution of zmpA correlated with Bcc species. zmpA was detected in B. cepacia, B. cenocepacia, B. stabilis, B. ambifaria, and B. pyrrocinia but absent from B. multivorans, B. vietnamiensis, B. dolosa, and B. anthina. Strains possessing zmpA had proteolytic activity on D-BHI milk agar, whereas strains from which the gene could not be amplified had no protease activity, suggesting that ZmpA is responsible for most of the protease activity in this assay. Five strains (B. cepacia CEP 509; B. cenocepacia C5424, J2315, and J415; and B. stabilis LMG 14294) had little or no detectable proteolytic activity or detectable ZmpA, despite the presence of the zmpA gene. No intracellular accumulation of either ZmpA or its precursor form, preproZmpA, was detectable, which would be indicative of a defect in secretion; however, it is possible that ZmpA would be degraded if it were not secreted properly.

A functional zmpA gene present in trans was able to restore protease production in strains C5424 and J415, suggesting that the zmpA genes in these strains contained mutations that affected their activity or that the increase in copy number of zmpA due to the plasmid produced levels of expression sufficient to result in activity. J415 expressed both the zmpA::lacZ fusion and zmpA mRNA detectable by qRT-PCR at fairly high levels, suggesting that J415 zmpA may have a mutation that decreases activity, but this strain is capable of producing extracellular protease activity when carrying a functional copy of zmpA. J415 ZmpA has 18 amino acids that differ from Pc715j. It is possible that one or more of these alterations affect enzyme activity.

Interestingly, strain C5424 also produced detectable protease activity when cepR was introduced on a plasmid. Multiple copies of cepR might directly increase expression of either the zmpA gene or another gene involved in zmpA expression or secretion. Introduction of cepR also restored protease activity to J2315, which suggests that the zmpA gene is functional in this strain but that a regulatory factor is deficient, since zmpA on a high-copy-number plasmid did not result in protease activity. Exogenous OHL was not able to restore protease activity to either C5424 or J2315. These results suggest that the increased protease activity induced by cepR is not due to an increase in OHL production as a result of enhanced cepI expression but may be due to another as-yet-unidentified gene regulated by CepR.

Strains LMG 14294 and CEP509 were not proteolytically active with either a functional copy of zmpA or cepR in trans, suggesting that these strains have other defects in protease expression. These strains had lower expression of both zmpA mRNA detectable by qRT-PCR and the zmpA::lacZ fusion than most of the zmpA protease-proficient strains.

In Pseudomonas aeruginosa, several global regulators have been shown to contribute to the regulation of secondary metabolites and virulence determinants at the posttranscriptional level including Vfr (2, 3, 49), DksA (17), and RsmA (33). Vrf, a homologue of Escherichia coli cyclic AMP receptor protein, positively regulates elastase and pyocyanin production via activation of the transcriptional regulator LasR (2). Posttranscriptional control by DskA is required for full translation of the lasB elastase gene and the rhamnosyltransferase-encoding rhlAB (17). Interestingly, RsmA is a global posttranscriptional negative regulator of extracellular enzymes and N-AHL production (33). RsmA controls the temporal expression of the lasI gene, and lasI is induced earlier and at a higher level during the exponential growth phase in a rsmA mutant. It is possible that LMG 14294 and Cep509 lack a regulator of protease production. Studies are in progress to identify potential global regulators involved in zmpA expression in B. cenocepacia.

Western immunoblot experiments with anti-ZmpA antibodies suggest that the zmpA protease-deficient strains do not secrete detectable ZmpA or that the possibly inactive products are degraded. The antibody did not react with two of the zmpA protease-proficient strains, which may indicate that ZmpA is not secreted from LMG 14086 and CEP 511 or is produced in very small amounts. It is possible that the polyclonal antibodies to ZmpA do not recognize ZmpA from all species in the complex, although the antibodies did react with ZmpA in several strains of B. cenocepacia and B. ambifara LMG 19182. It is also possible that the extracellular proteolytic activity observed on D-BHI milk agar in these strains was due to other extracellular proteases.

We have determined that, although ZmpA may be present in many strains and species of the Bcc, there are likely multiple factors influencing its expression including the cepIR quorum-sensing system. ZmpA has been shown to be an important virulence factor in some strains of B. cenocepacia (9). The majority of B. cenocepacia strains are capable of producing protease, but expression varies and may be influenced by a number of factors. It is possible that J2315 and C5424 produce protease activity in vivo, due to increased expression of either cepR or zmpA. Strains C1394 and PC 184 have no protease activity at 24 h yet have significant activity on plates by 48 h, suggesting that expression of zmpA is delayed in these strains. Strains of B. multivorans, which have been shown to be generally less virulent than B. cenocepacia in both a chronic lung infection model and an alfalfa infection model (4), lack the zmpA gene and protease activity in the conditions examined. It is possible that the absence of zmpA accounts for some of the reduced virulence of B. multivorans.


This study was supported by a grants from Canadian Institutes of Health Research (CIHR) and the Canadian Cystic Fibrosis Foundation to P.A.S. S.G. was the recipient of a studentship award from the Natural Sciences and Engineering Research Council of Canada (NSERC).


    • Received 30 May 2005.
    • Accepted 19 September 2005.
  • *Corresponding author. Mailing address: Department of Microbiology and Infectious Diseases, Faculty of Medicine, University of Calgary Health Sciences Centre, Calgary, Alberta T2N 4N1, Canada. Phone: (403) 220-6037. Fax: (403) 270-2772. E-mail: psokol{at}