Transcriptional and Posttranscriptional Control of Cable Pilus Gene Expression in Burkholderia cenocepacia

Abstract

Burkholderia cenocepacia is an important member of the Burkholderia cepacia complex, a group of closely related bacteria that inhabits a wide variety of environmental niches in nature and that also colonizes the lungs of compromised humans. Certain strains of B. cenocepacia express peritrichous adherence organelles known as cable pili, thought to be important in the colonization of the lower respiratory tract. The genetic locus required for cable pilus biogenesis is comprised of at least five genes, designated cblB, cblA, cblC, cblD, and cblS. In this study a transcriptional analysis of cbl gene expression was undertaken. The principal promoter, located upstream of the cbl locus, was identified and characterized. By using lacZ transcriptional fusions, the effects of multiple environmental cues on cbl gene expression were examined. High osmolarity, temperature of 37°C, acidic pH, and low iron bioavailability were found to induce cbl gene expression. Northern hybridization analysis of the cbl locus identified a single, stable transcript corresponding to cblA, encoding the major pilin subunit. Transcriptional fusion studies combined with reverse transcription-PCR analysis indicated that the stable cblA transcript is the product of an mRNA processing event. This event may ensure high levels of expression of the major pilin, relative to other components of the assembly pathway. Our findings lend further insight into the control of cable pilus biogenesis in B. cenocepacia and provide evidence for regulation of cbl gene expression on both the transcriptional and posttranscriptional levels.

Burkholderia cepacia is a complex of gram-negative bacteria widespread in nature that consists of at least nine genomic species, or genomovars (3, 6, 41). Burkholderia cenocepacia, formerly genomovar III, has emerged as the dominant B. cepacia complex respiratory pathogen in compromised individuals and particularly cystic fibrosis (CF) patients (12, 19). B. cenocepacia colonization in individuals with CF generally leads to the establishment of chronic infection, causing a significant decrease in life expectancy. Colonization by B. cenocepacia can also lead to acute infections and fatal outcomes in CF patients (12, 19). Although several putative virulence factors have been identified, including the iron-scavenging siderophore ornibactin (35) and a type III secretion system (38), the molecular mechanisms facilitating colonization and pathogenesis of B. cenocepacia are still poorly understood (22).

B. cenocepacia adherence to host cells and mucosal surfaces likely plays an important role in the initiation and establishment of infection. The ability of B. cenocepacia to colonize the CF lung, as well as spread from patient to patient, has been associated with the expression of filamentous extracellular adherence organelles known as cable pili (4, 37). These peritrichously expressed structures derive their name from their unique cable-like intertwined morphology (29). Cable pili have been shown to facilitate bacterial binding to both mucin and CF respiratory epithelia, suggesting a direct role for cable pili in mediating colonization (27, 28). Aside from the role of cable pili in adhesion, little is known about the mechanisms governing the expression and assembly of these structures in B. cenocepacia.

The DNA locus required for B. cenocepacia cable pilus biogenesis is comprised of at least five genes, designated cblB, cblA, cblC, cblD, and cblS (30). The cblA gene encodes the major structural subunit of cable pili (29), while cblB, cblC, and cblD are predicted to encode the periplasmic chaperone, the outer membrane usher, and the minor pilus structural subunit, respectively. The fifth gene, designated cblS, is predicted to encode a new member of the sensor kinase superfamily of bacterial two-component systems. It has recently been demonstrated that the cblBACD locus is sufficient for heterologous expression of cable pili in Escherichia coli (30). The B. cenocepacia cblBACD gene products share high homology with the assembly machinery of the CS1 family of pili, elaborated by certain strains of human enterotoxigenic E. coli (ETEC) (30). The CS1 family includes CS1, CS2, CS4, and CFA/I pili (31), which have been implicated in colonization of the human small intestine and the establishment of infection by ETEC (9, 17).

The genes required for CS1 pilus biogenesis, as well as the biogenesis of other pilus types, are typically organized as operons (31). Expression of CS1 and other pilus operons is subject to both positive and negative regulation at the transcriptional level (15, 23). A number of studies have examined the environmental regulation of pilus expression in ETEC and other pathogenic E. coli. These studies have drawn a correlation between stimuli resembling those encountered in vivo, including pH, osmolarity, and temperature, and transcriptional activation of pilus gene expression (7, 10, 15, 20, 25, 42). In contrast, far less is known about the regulation of pilus gene expression in nonenteric bacteria, including respiratory pathogens such as B. cenocepacia.

In addition to transcriptional control, expression of some pilus operons has been shown to be regulated at the posttranscriptional level. Specifically, mRNA processing and the various stabilities of the resulting mRNAs have been proposed as mechanisms for facilitating differential expression of the various structural and assembly components of pilus biogenesis pathways. Posttranscriptional mRNA processing mechanisms have been shown to control CFA/I pilus expression in ETEC, as well as expression of the F1845, Pap, and S fimbriae of pathogenic E. coli (1, 2, 14, 24).

In this study we (i) undertook a transcriptional analysis of B. cenocepacia cbl gene expression, (ii) identified and characterized the principal promoter upstream of the cbl locus, (iii) examined the environmental modulation of cbl gene expression, and (iv) characterized an mRNA processing event, predicted to result in higher expression levels of the major structural subunit of cable pili, CblA, relative to the other components of the pilus biogenesis pathway. Our findings lend new insight into the regulation of cable pilus gene expression in B. cenocepacia and provide evidence for control at both transcriptional and posttranscriptional levels.

MATERIALS AND METHODS

Bacterial strains, plasmids, and media.

The bacterial strains and plasmids used in this study are listed in Table 1. B. cenocepacia (formerly B. cepacia complex genomovar III) strain BC7 was obtained from the Belgium Coordinated Collections of Microorganisms/Laboratorium Microbiologie Ghent (BCCM/LMG). Strain BC7 is a cable-piliated CF clinical isolate of B. cenocepacia (29). E. coli strains were grown with aeration at 37°C in Luria-Bertani (LB) broth (32) or on LB agar plates supplemented with ampicillin (100 μg/ml), tetracycline (12 μg/ml), or chloramphenicol (30 μg/ml) as necessary. B. cenocepacia strains were grown with aeration at 37°C in LB or in M9 minimal medium (32), supplemented with 0.2% glucose and 0.3% Casamino Acids (wt/vol). For propagation of B. cenocepacia strains harboring transcriptional fusion constructs, tetracycline was added to liquid media (25 μg/ml) and LB agar (500 μg/ml).

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant characteristics	Reference or source
E. coli strains
DH5α	supE44 lacU169 (φ80 lacZΔM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1	Bethesda Research Laboratories
XL-1 Blue	recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac (F′proAB lacIqZΔM15 Tn10)	Stratagene
S17-1	Integrated RP4-2, Tc::Mu, Km::Tn7	34
B. cenocepacia strains
BC7	Cystic fibrosis clinical isolate, cable piliated, formerly designated PC7	29
CM256	cblA::cat derivative of BC7	40
Plasmids
pBluescript SK(−)	Cloning and single-stranded phagemid; Apr	Stratagene
pGEM-T Easy	TA cloning vector; Apr	Promega
pRKlac290	lacZ transcriptional fusion vector, IncP1 replicon, mob+; Tcr	11
p3A4	Cosmid with a portion of the cbl locus, including the 3′ end of cblD and the entire cblS gene	This study
pMT55	0.6-kb EcoRI fragment cloned into pRKlac290, generating a cblB-lacZ transcriptional fusion construct; Tcr	This study
pMT58	1.1-kb EcoRI fragment cloned into pRKlac290, generating a cblB-lacZ transcriptional fusion construct; Tcr	This study
pMT59	1.0-kb BamHI/HindIII fragment cloned into pRKlac290, generating a cblA-lacZ transcriptional fusion construct; Tcr	This study
pMT62	0.4-kb EcoRI/HindIII fragment cloned into pRKlac290, generating a cblS-lacZ transcriptional fusion construct; Tcr	This study
pMT92	1.0-kb EcoRI/BamHI fragment cloned into pRKlac290, generating a cblC-lacZ transcriptional fusion construct; Tcr	This study
pMT93	1.1-kb BamHI/HindIII fragment cloned into pRKlac290, generating a cblD-lacZ transcriptional fusion construct; Tcr	This study
pMT95	0.4-kb PstI/HindIII fragment cloned into pRKlac290, generating a cblB-lacZ transcriptional fusion construct; Tcr	This study

DNA manipulations.

DNA-modifying enzymes, including restriction endonucleases, T4 polynucleotide kinase, T4 DNA ligase and T4 polymerase, were obtained from either Promega or New England Biolabs. Superscript II was obtained from Invitrogen, and Taq polymerase was obtained from Promega. Plasmid DNA was isolated by the boiling lysis method or by using the QIAprep Spin Miniprep kit (QIAGEN Inc.). Recombinant plasmids were introduced into E. coli by electroporation with a Gene Pulser II (Bio-Rad). Genomic DNA from B. cenocepacia was extracted by means of a PureGene kit (Gentra). Southern blot hybridizations were generally performed as described by Sambrook et al. (32) with Hybond N nitrocellulose membranes and probes labeled with [α-³²P]dCTP (Amersham Pharmacia Biotech) by the random primer method.

Cloning and sequencing of the B. cenocepacia cbl locus.

The B. cenocepacia cbl locus from strain BC7 was cloned and sequenced by two converging strategies. Initially, portions of the locus harboring cblB, cblA, cblC, and the first 1,074 nucleotides of cblD were cloned from strain CM256, a cblA::cat derivative of the parental strain BC7 (40). The last 90 nucleotides of the cblD gene and the entire cblS gene were cloned from a cblD-cross-hybridizing cosmid, designated p3A4, which was identified by probing a B. cenocepacia strain BC7 cosmid library, constructed as previously described by our laboratory (39). Multiple subclones of the cbl locus were generated, and their sequences were determined on both strands. Nucleotide sequencing was performed by the Advanced Genetic Analysis Center at the University of Minnesota by using the dideoxy chain termination method and an ABI 1371A DNA sequencer (Applied Biosystems). Oligonucleotide primers used for sequencing were standard forward and reverse (T3 and T7) pBluescript primers or custom oligonucleotides synthesized by Integrated DNA Technologies. Double-stranded sequences were aligned and assembled by the EditSeq and SeqMan components of a demonstration version of the Lasergene sequence analysis software package (DNASTAR Inc.). Nucleotide and amino acid sequence searches and analysis utilized the BLASTX and BLASTP programs at the National Center for Biotechnology Information.

Growth conditions and measurement of β-galactosidase activity.

Transcriptional fusion constructs were generated in the low-copy-number vector pRKlac290, harboring a promoterless β-galactosidase reporter gene, lacZ, and are described in Table 1 and Fig. 1. Transcriptional fusion constructs were introduced into B. cenocepacia strain BC7 by conjugation, as previously described, by using E. coli S17-1 as the donor strain (39). B. cenocepacia cultures were grown in the presence of tetracycline (25 μg/ml), in order to ensure maintenance of pRKlac290 and pRKlac290-derived constructs. For measurement of β-galactosidase activity, B. cenocepacia strains harboring the plasmid-borne transcriptional fusion constructs were grown to stationary phase for 17 h in 3 ml of LB or M9 medium, and aliquots were used to inoculate fresh 3-ml volumes of the corresponding medium. Cultures were grown for an additional 16 to 18 h, until an optical density at 600 nm (OD₆₀₀) of ∼0.2 was reached, at which point the first β-galactosidase measurements were taken. The β-galactosidase activities were assayed throughout the growth phase as described by Miller (21). Assays were performed in triplicate with a minimum of two independent experiments for each transcriptional fusion construct and/or growth condition.

FIG. 1.

Physical map of the B. cenocepacia cbl locus and transcriptional fusions generated in this study. The arrows indicate the direction of transcription. DNA fragments used to generate transcriptional fusion constructs to the β-galactosidase reporter gene are shown as black bars under the physical map of the cbl locus. Levels of β-galactosidase activity in B. cenocepacia strain BC7 harboring the various transcriptional fusions are shown to the right. Representative β-galactosidase activities (± standard error) measured in either LB or M9 medium, taken at 10 h, are presented in Miller units. Abbreviations: B, BamHI; H, HindIII; E, EcoRI; P, PstI; X, XhoI. Parentheses indicate the nonendogenous restriction endonuclease sites that were introduced during subcloning.

To examine the effect of pH on cbl gene expression, B. cenocepacia strains harboring either pRKlac290 (vector control) or the cblB transcriptional fusion construct pMT55 were grown in M9 medium (pH 7.0) or in M9 medium in which the pH was adjusted to either 5.0 or 6.0, by using concentrated HCl, or pH 8.0, by using 10 M NaOH. In order to determine the effect of osmolarity, M9 medium was prepared without NaCl or supplemented with 5 M NaCl to a final concentration of either 100 or 200 mM. To examine the effect of temperature, bacteria were grown with shaking at 250 rpm at either 30°C or 37°C. In order to examine the effect of iron, M9 minimal medium was supplemented with 50 μM FeCl₃. Conversely, iron was chelated in LB broth by the addition of 100 μg of ethylenediamine-N,N′-diacetic acid (EDDA) per ml.

Primer extension.

To determine the transcription initiation site of the cblB promoter, a synthetic oligonucleotide primer, cbl50, complementary to nucleotides −213 to −232 relative to the cblB start codon, was used in primer extension reactions (Table 2). In order to map the 5′ end of the stable cblA transcript, primer extension analysis was performed with primer cbl49, complementary to nucleotides +746 to +727, relative to the cblB translational start codon (Table 2). The cbl49 and cbl50 primers were 5′ end-labeled with [γ-³²P]ATP by using T4 polynucleotide kinase and were hybridized to 9 or 21 μg, respectively, of total RNA isolated from B. cenocepacia strain BC7, grown in M9 medium to an OD₆₀₀ of 1.0. Total bacterial RNA was isolated by using Trizol reagent (Invitrogen), according to the manufacturer's instructions. After a 5-min RNA denaturation at 70°C, primers were annealed at 45°C for 30 min, followed by reverse transcription (RT) with Superscript II (Invitrogen) at 37°C for 30 min. The primer extension products were precipitated with LiCl, extracted with phenol:chloroform (1:1), and reprecipitated with ethanol. To precisely determine the 5′ ends of transcripts, DNA sequencing reactions were carried out by means of a Thermo Sequenase cycle sequencing kit (Amersham Pharmacia Biotech) with the same 5′ end-labeled primers as used in the primer extension reactions. The primer extension products and sequencing ladders were analyzed by denaturing electrophoresis on 6.5% polyacrylamide sequencing gels. After electrophoresis, the gels were dried and exposed to X-ray film (Kodak).

TABLE 2.

Oligonucleotide primers

Oligonucleotide	Location	Oligonucleotide sequence (5′→3′)
cbl1	Nucleotides 663 to 681 of cblB; forward	AATGGCAGATGTGCAGCAG
cbl2	Nucleotides 51 to 68 of cblC; reverse	CGCGATGTCCATCACATAC
cbl8	Nucleotides 239 to 259 relative to the cblB transcriptional initiation site; reverse	ATATGGAATCCATTTCACGTG
cbl13	Nucleotides 178 to 195 of cblA; forward	GCAGCTGTAGTGAACACG
cbl15	Nucleotides 398 to 415 of cblA; reverse	TCT GACCGATCGACAGCG
cbl25	Nucleotides 2287 to 2304 of cblC; forward	AAGCGCACGCTGTTCATG
cbl27	Nucleotides −61 to −43 relative to the cblS start codon; forward	TTGCCGCTTGCGATGCC
cbl28	Nucleotides 705 to 722 of cblD; reverse	AGCTTGAGGTCGACGGTG
cbl49	Nucleotides 727 to 746 of cblB; reverse	GTCCGGGTCTCGTTATTCGC
cbl50	Nucleotides +72 to +91 relative to the cblB transcriptional initiation site; reverse	TGCAGCCAATCACTCAAGCG
cbl59	Nucleotides 325 to 341 of cblS; reverse	CGGTCCGGAACCAGCTC
cbl60	Nucleotides +72 to +90 relative to the cblB transcriptional initiation site; forward	CGCTTGAGTGATTGGCTGC

Northern hybridization analysis.

B. cenocepacia strain BC7 was grown in M9 medium to an OD₆₀₀ of 1.0, and total RNA was extracted by Trizol reagent (Invitrogen). Equivalent amounts of strain BC7 RNA were denatured, electrophoresed in formaldehyde agarose gels, and blotted onto Hybond N nitrocellulose membranes (Amersham Pharmacia Biotech). An RNA molecular size marker (Invitrogen) was also electrophoresed and used as a reference. Membranes were hybridized with DNA probes corresponding to each of the five genes in the cblBACDS locus. The following DNA fragments were used as probes: a 0.8-kb cblA EcoRI fragment; a 0.6-kb cblB PstI fragment; a 1.4-kb cblC BamHI/PstI fragment; a 0.7-kb cblD HindIII/XhoI fragment; and a 0.9-kb cblS EcoRV fragment. All probes were labeled with [α-³²P]dCTP (Amersham Pharmacia Biotech) by the random primer method (32).

RT-PCR.

Total bacterial RNA was isolated from B. cenocepacia strain BC7 grown in M9 minimal medium to an OD₆₀₀ of 1.0 by using Trizol reagent. To ensure that the RNA was devoid of contaminating DNA, the preparation was treated with RNase-free RQ1 DNase (Promega) for 1 h. The isolated RNA was used as a template in RT-PCRs, utilizing the SuperScript One-Step RT-PCR system (Invitrogen), generally according to the manufacturer's instructions. Briefly, the RT reaction was carried out at 45°C for 30 min in a thermal cycler (Hybaid), immediately followed by 40 cycles of PCR, consisting of 1 min at 94°C, 1 min at 52°C, and 2 min 20 s at 68°C, ending with a 10-min incubation at 72°C. RT-PCR was performed in 40-μl reaction mixtures, with 0.4 to 0.8 μg of total B. cenocepacia RNA and appropriate oligonucleotide primer pairs (see Fig. 7 and Table 2). For RT-PCR amplification of all transcripts, 0.4 μg of total RNA was used as a template, with the exception of the cblBA transcript, for which 0.8 μg of RNA was utilized. Reactions in which the reverse transcriptase Taq polymerase mix was replaced with Taq polymerase alone were also performed to confirm the absence of contaminating DNA in the RNA sample. RT-PCR amplifications were performed at least twice with total RNA preparations obtained from a minimum of two independent extractions. The RT-PCR and PCR products were analyzed by agarose gel electrophoresis.

FIG. 7.

RT-PCR analysis of cbl gene expression. Reactions were performed as described in Materials and Methods. The horizontal arrow (P) upstream of the cblBACDS genes indicates the position of the cblB-proximal promoter. The black bars and arrows below the physical map of the cbl locus indicate the designations and locations of the oligonucleotide primer pairs used and the predicted sizes of RT-PCR products. RT-PCRs corresponding to cblA (A), cblBA (B), cblBAC (C), cblCD (D), and cblDS (E) were analyzed by agarose gel electrophoresis. In each of the five panels A through E, the positions of molecular size markers (kb) are indicated to the left. The presence (+) or absence (−) of the RT enzyme in the amplification reactions is indicated. The arrows indicate the RT-PCR products obtained. X, XhoI.

Nucleotide sequence accession number.

The DNA sequence of the cbl locus has been deposited in the GenBank database under accession number AY114293.

RESULTS

Cloning and sequencing of the B. cenocepacia cbl locus.

Our efforts to clone and sequence the B. cenocepacia cbl locus were initiated prior to the recent publication of the sequence of the cbl locus of strain BC7 by Sajjan et al. (30). Our sequence of the cblBACDS locus is 99% identical to that published by Sajjan et al. and is 100% identical to the nucleotide sequence of the cbl locus in the closely related B. cenocepacia strain J2315, whose genome has recently been sequenced by the Sanger Centre (http://www.sanger.ac.uk/Projects/B_cenocepacia).

Transcriptional fusion analysis of the cbl locus.

The cable pilus biogenesis locus is comprised of at least five genes, designated cblB, cblA, cblC, cblD, and cblS. The tandem arrangement of the cblBACDS genes suggested an operonic structure with transcription initiating from an A+T-rich region identified upstream of cblB (Fig. 1). In order to delimit the upstream sequences required for expression of cblB and possibly other cbl genes, a series of transcriptional fusion constructs was generated. DNA fragments were inserted into the multiple cloning site of vector pRKlac290 to generate transcriptional fusions to the lacZ reporter gene. Constructs were introduced into B. cenocepacia strain BC7, and β-galactosidase activity was measured in either rich (LB) or minimal (M9) medium throughout growth phase. Preliminary studies showed that B. cenocepacia strain BC7 does not exhibit intrinsic β-galactosidase activity (data not shown) and that introduction of the pRKlac290 vector into B. cenocepacia strain BC7 results in only low-level β-galactosidase activity (Fig. 1).

Transcriptional fusion constructs pMT58 and pMT55, containing approximately 900 and 381 nucleotides upstream of the predicted cblB start codon, respectively, exhibited indistinguishable β-galactosidase activities in both LB and M9 and in all phases of growth (Fig. 1 and data not shown). In contrast, a third deletion derivative (pMT95), encompassing 127 nucleotides upstream of the cblB start codon, exhibited β-galactosidase activities similar to the levels of the pRKlac290 vector control in both LB and M9 media (Fig. 1). Together, these results demonstrate that the cis-acting DNA elements required for maximal expression of cblB are located between nucleotides −381 and −127, relative to the cblB start codon.

To determine whether additional promoters may be responsible for the transcription of genes downstream of cblB, transcriptional fusion constructs encompassing the intergenic regions between cblB and cblA (pMT59), cblA and cblC (pMT92), cblC and cblD (pMT93), and cblD and cblS (pMT62) were generated (Fig. 1). Only construct pMT93, encompassing the cblC-cblD intergenic region, exhibited β-galactosidase activity above the background level of the vector control (Fig. 1). The activity of this transcriptional fusion, however, was less than twofold higher than the activity of the vector control and significantly lower than that of the cblB transcriptional fusions pMT58 and pMT55 in both LB and M9 media.

Identification of the cblB promoter transcriptional initiation site.

Primer extension analysis was performed in order to precisely determine the transcriptional initiation site of the cblB-proximal promoter (Materials and Methods). A single primer extension product was consistently obtained, corresponding to a single transcriptional initiation site, located 303 nucleotides upstream of the predicted cblB translational start codon (Fig. 2). Total cellular RNA was also hybridized to a second primer, designated cbl8, complementary to nucleotides −49 to −69 with respect to the cblB translational start site. Analysis of the primer extension product obtained with this primer identified the same transcriptional initiation site that is identified with primer cbl50 (data not shown).

FIG. 2.

Primer extension analysis of the cblB promoter. (A) Lanes G, A, T, and C denote the corresponding sequencing reactions, and the primer extension product was loaded in lane 1. The arrow indicates the single primer extension product obtained with primer cbl50. (B) Nucleotide sequence of the cblB promoter region. The horizontal bars indicate the location of primers cbl8 and cbl50, used in the primer extension analysis. The transcriptional initiation site corresponding to the primer extension product is designated +1 and shown in bold. The numbers to the left of the sequence indicate the positions of the nucleotides relative to the cblB transcriptional initiation site. Putative −35 and −10 promoter elements, ribosomal binding site, and the deduced amino acid sequence of the N terminus of CblB are also indicated in bold letters.

Analysis of the DNA region upstream of the cblB transcriptional initiation site revealed the presence of several A+T-rich tracts, as well as partial direct and inverted repeats. The putative −35 (TATATT) and −10 (CAAAAT) promoter regions share only weak homology with known σ consensus sequences. Four out of six nucleotides in the −10 region match the conventional σ⁷⁰ consensus, while four out of seven match the E. coli consensus sequence for the stationary phase-specific σ factor, RpoS.

Regulation of cbl gene transcription in response to environmental cues.

To determine if expression of B. cenocepacia cbl genes is regulated in response to environmental signals, we utilized the transcriptional fusion construct pMT55. We initially examined the β-galactosidase activity throughout growth in rich (LB) or minimal (M9) medium, as these media are known to either repress or induce, respectively, pilus gene expression in E. coli (43). When B. cenocepacia strain BC7 harboring pMT55 was grown in either rich (LB) or minimal (M9) medium, transcriptional activity increased approximately twofold during the mid-late exponential phase, with peak activities observed in stationary phase (Fig. 3). No differences in growth were observed between strain BC7 harboring pRKlac290 and BC7 harboring pMT55 in either LB or M9 (data not shown). Overall, the cblB transcriptional fusion pMT55 exhibited four- to fivefold higher activity in the minimal medium, suggesting that the growth environment and growth phase can significantly influence cbl gene expression.

FIG. 3.

Regulation of cbl gene expression in rich (LB) and minimal (M9) media. B. cenocepacia strains harboring transcriptional fusion constructs were grown in LB (A) or M9 (B) medium. β-Galactosidase measurements were taken at 2-h intervals and are shown in Miller units. The bars indicate the standard errors of the measurements.

Given the known role of pH in regulating pilus gene expression in E. coli (33, 43), we tested the effect of pH on cbl gene transcription. B. cenocepacia strain BC7 harboring the cblB transcriptional fusion construct pMT55 was grown in M9 medium, ranging in pH from 5.0 to 8.0. Due to the inability of B. cenocepacia strain BC7 to grow in M9 at pH 9.0, the most basic pH in which β-galactosidase activity was measured was 8.0. The pH values of the cultures were monitored during growth and were found to remain constant throughout the experiment. The highest overall β-galactosidase activity was measured at pH 6.0, with an approximately twofold increase compared to levels of activity measured at pH 7.0 (Fig. 4A). The β-galactosidase activity was lowest at pH 5.0 and only slightly higher at pH 8.0. Based on these results, it appears that the optimal pH for cblB expression is between 6.0 and 7.0.

FIG. 4.

Analysis of the effects of environmental cues on cbl gene expression. B. cenocepacia strain BC7 harboring the cblB transcriptional fusion construct pMT55 was grown in standard or modified M9 medium, as indicated. β-Galactosidase measurements were taken at 2- or 4-h intervals and are shown in the graphs on the left in Miller units. The bars indicate the standard errors of the measurements. The corresponding growth curves are shown in the accompanying graphs on the right. (A) Analysis of the effect of pH on cbl gene expression. (B) Analysis of the effect of osmolarity on cbl gene expression. (C) Analysis of the effect of temperature on cbl gene expression.

There is evidence that the airway surface liquid (ASL) in the CF lung may have a higher concentration of Na⁺ and Cl⁻ ions than normal ASL (44). The high concentration of NaCl in the CF lung could create a hyperosmotic environment that may influence B. cenocepacia gene expression during the course of infection. The effect of osmolarity on cbl gene expression was examined in M9 medium with concentrations of NaCl ranging from 0 to 200 mM. An increase in osmolarity led to a corresponding increase in β-galactosidase activity of B. cenocepacia harboring the cblB transcriptional fusion construct pMT55 (Fig. 4B). A two- to threefold increase in activity was measured when M9 medium was supplemented with 200 mM NaCl, compared to medium without NaCl. At a concentration of 100 mM NaCl, the measured β-galactosidase activity was intermediate, demonstrating a dose-dependent induction of cblB expression in response to increased concentration of NaCl.

Temperature has been shown to play an important role in controlling CFA/I pilus gene expression in ETEC. CFA/I pili are expressed at 37°C, the physiological temperature of the human body, but not at 20°C, suggesting that temperature may be a cue sensed by E. coli to distinguish between the in vivo and ex vivo environments (15). The effect of temperature on B. cenocepacia cbl gene expression was examined by measuring the β-galactosidase activity of B. cenocepacia strain BC7 harboring pMT55 when grown in M9 medium at either 30°C or 37°C. Initially, the β-galactosidase activities were similar at both temperatures (Fig. 4C). However, the induction of β-galactosidase expression consistently observed at 37°C was absent at 30°C. The highest level of β-galactosidase activity measured at 37°C was twofold greater than that achieved at 30°C. These results suggest that transcription from the cblB-proximal promoter is a temperature-dependent process and that growth at 37°C is required for induction of cbl gene expression.

Iron is both an essential and limiting nutrient in vivo, and iron starvation has been shown to activate expression of a number of bacterial virulence factors, including the ETEC CFA/I fimbriae (16). In order to determine if iron availability plays a role in cbl gene expression, β-galactosidase activities were determined for B. cenocepacia harboring the cblB transcriptional fusion construct pMT55, grown under both iron-replete and iron-deplete conditions. To examine the effect of increasing iron concentration, M9 medium, normally containing only trace amounts of the metal, was supplemented with 50 μM FeCl₃. Initially, the measured β-galactosidase activities of the cblB transcriptional fusion in both M9 medium and M9 medium supplemented with iron were indistinguishable (Fig. 5A). The exponential-phase induction of cblB promoter activity was observed in both M9 medium and M9 medium supplemented with FeCl₃. However, the induction was delayed by approximately 4 h when the medium was supplemented with FeCl₃. Furthermore, the β-galactosidase activity did not reach the same level in M9 medium supplemented with iron as it did in M9 medium alone, suggesting that iron may lead to repression of the mid-exponential phase induction of cbl gene expression. To further examine the role of iron in cbl gene expression, B. cenocepacia strain BC7 harboring pMT55 was grown in LB medium, an iron-rich medium, or LB medium supplemented with 100 μg of the iron chelator EDDA per ml. The measured β-galactosidase activities in the presence of EDDA were approximately twofold higher than activities in LB medium alone (Fig. 5B), indicating that limiting iron bioavailability leads to induction of cbl gene expression. Together, our results suggest that iron starvation is a signal that leads to an increase in cbl gene expression.

FIG. 5.

Effect of iron bioavailability on cbl gene expression. (A) Analysis of the effect of iron supplementation on cbl gene expression. (B) Analysis of the effect of iron chelation on cbl gene expression. β-Galactosidase activity was monitored at 2-h intervals throughout growth. The bars indicate the standard errors of the measurements.

Northern hybridization analysis of cbl gene transcripts.

Northern hybridization analysis was performed to further characterize the transcriptional organization of the cblBACDS locus. The tandem arrangement of genes in the cbl locus indicated that the cbl genes are transcribed as an operon. Furthermore, the lack of detectable promoter activity from the intergenic regions downstream of the cblB-proximal promoter, with the exception of weak activity within or adjacent to the cblC-cblD intergenic region, supported the conclusion that the cbl genes are cotranscribed as a polycistronic operon. We therefore expected to detect a single polycistronic transcript, corresponding to the entire B. cenocepacia cblBACDS locus. Surprisingly, only one transcript of 0.7 kb consistently hybridized to a cblA-derived probe (Fig. 6). Furthermore, under the same conditions, cblB-, cblC-, cblD-, and cblS-derived probes did not reproducibly hybridize to any transcripts (Fig. 6), suggesting that mRNAs corresponding to these genes are of low abundance and/or may be unstable. However, as reported above, the cblB-proximal promoter was successfully mapped by primer extension, indicating that transcripts corresponding to the cblB gene are generated.

FIG. 6.

Northern hybridization analysis of cbl gene expression. Total bacterial RNA was extracted from B. cenocepacia strain BC7 grown in M9 minimal medium to an OD₆₀₀ of 1.0. Radiolabeled DNA fragments derived from the cblB, cblA, cblC, cblD, or cblS genes were used as probes in hybridizations, as described in Materials and Methods. The DNA fragments used as probes are shown as gray bars under the physical map of the B. cenocepacia cbl locus. The results of Northern hybridization analyses using probes corresponding to cblB, cblA, cblC, cblD, and cblS are shown below the gray bars. The positions of the bands in the RNA ladder are indicated on the left. The arrow indicates a 0.7-kb transcript hybridizing to the cblA probe. Abbreviations: B, BamHI; H, HindIII; P, PstI; V, EcoRV; X, XhoI.

RT-PCR analysis.

To further investigate the transcriptional organization of the cbl locus, we utilized RT-PCR analysis using total RNA extracted from B. cenocepacia strain BC7 grown under the same conditions as for the isolation of RNA used in the primer extension and Northern hybridization analyses. We initially aimed to use RT-PCR to amplify the stable cblA transcript, identified by Northern hybridization analysis. With primers cbl13 and cbl15, an abundant 0.2-kb product was obtained, corresponding to the predicted size of a portion of the cblA transcript (Fig. 7A).

Additional primer sets were used to determine whether polycistronic transcripts, corresponding to other genes in the locus, could also be amplified by RT-PCR. Portions of transcripts corresponding to cblBA, cblBAC, cblCD, and cblDS were successfully amplified (Fig. 7B through E), confirming the operonic organization of the cbl locus. The RT-PCR products obtained from these reactions, however, were significantly less abundant than the 0.2-kb cblA-amplified product. Repeated attempts to amplify transcripts corresponding to the entire cblBACDS gene cluster or to cblBACD were unsuccessful. This result may be due to low transcript abundance, high G+C content, mRNA secondary structure, transcript size limitation, or any combination thereof.

Lack of promoter activity immediately upstream of the cblA gene.

The identification of a single 0.7-kb transcript hybridizing to the cblA-derived probe in Northern hybridization analysis, potentially encompassing all 501 nucleotides of the cblA gene, suggested two possible explanations for its origin: (i) the cblA transcript is initiated from a promoter immediately upstream of the cblA gene, within the cblB coding region, or (ii) a posttranscriptional event leads to the processing of a larger transcript, initiated from the cblB-proximal promoter. To examine the former possibility, a DNA fragment encompassing the cblB-cblA intergenic region was cloned into vector pRKlac290, generating a transcriptional fusion to lacZ (pMT59) (Fig. 1). No significant difference in β-galactosidase activity was measured between B. cenocepacia harboring pMT59 or the vector control in either LB or M9 medium throughout growth (Fig. 1 and data not shown), suggesting that the cblA gene is not transcribed from an independent promoter located within the cblB coding region.

Mapping the cblA mRNA processing site.

A second hypothesis to account for the origin of the stable 0.7-kb cblA transcript is that it is generated by the processing of a larger mRNA, initiated at the cblB-proximal promoter. A region of dyad symmetry, predicted to form a stem-loop structure in the corresponding transcript, was identified immediately downstream of the cblA gene. The stem-loop, followed by six uracyl residues in the transcript, constitutes a strong Rho-independent transcriptional terminator and indicates the position of the 3′ end of the 0.7-kb cblA transcript. Furthermore, termination of transcription at the stem-loop would position the 5′ end of the 0.7-kb cblA transcript approximately 200 nucleotides upstream of the cblA start codon and within the cblB coding region. To investigate this further, primer extension was utilized to identify the 5′ end of the stable 0.7-kb cblA transcript. Total bacterial RNA was isolated from B. cenocepacia strain BC7 and hybridized to the 5′ end-labeled primer cbl49, complementary to the region immediately upstream of the cblA gene. Three predominant primer extension products were identified (Fig. 8), corresponding to nucleotides TAT (UAU in the corresponding mRNA), located at positions +952, +953, and +954 relative to the cblB transcriptional initiation site. Mapping of the 5′ end of the stable cblA transcript confirms that the 0.7-kb mRNA originates from within the cblB coding region and also indicates that the 3′ end of the 0.7-kb transcript is immediately downstream of the cblA translational stop codon. Together with the transcriptional fusion, Northern hybridization, and RT-PCR analyses, the primer extension results suggest that the cblA gene is cotranscribed with cblB on a dicistronic transcript, which may be posttranscriptionally cleaved to yield an abundant, stable 0.7-kb cblA transcript and an unstable, truncated cblB transcript.

FIG. 8.

Primer extension analysis of the 5′ end of the cblA mRNA. (A) Lanes G, A, T, and C denote the corresponding sequencing reactions, and the primer extension reaction was loaded in lane 1. The black arrows indicate the three predominant primer extension products obtained with primer cbl49. (B) Nucleotide sequence of the cblB-cblA intergenic region. The numbers to the left of the sequence indicate the positions of the nucleotides relative to the cblB transcriptional initiation site. The horizontal bar indicates the location of primer cbl49, used for the primer extension. The vertical arrows indicate nucleotides +952, +953, and +954, corresponding to the three predominant 5′ end nucleotides of the cblA mRNA. The putative ribosomal binding site and the deduced amino acid sequence of CblB and CblA are indicated in bold letters.

DISCUSSION

In this study we investigated the transcriptional organization of the cblBACDS locus, encoding components of the B. cenocepacia cable pilus biogenesis pathway. A promoter upstream of the cblB gene was identified and characterized, and the effects of multiple environmental cues on cbl gene expression were investigated. Our studies have also provided evidence for posttranscriptional control of cable pilus gene expression through differential stability of cbl transcripts. This mechanism may ensure a high level of expression, relative to the other components of the assembly pathway, of the major structural subunit of cable pili.

The cblB promoter was found to be four- to fivefold more active in minimal M9 medium than in rich LB medium. Growth in rich media has also been found to repress the transcription of the pap, daa, and fan operons, encoding Pap pili, F1845, and K99 fimbriae in E. coli, respectively (43). We then began to dissect the role of individual environmental stimuli in cbl gene expression. The activity of the cblB promoter was sensitive to pH, induced by acidic conditions (pH 6.0), and repressed in more acidic (pH 5.0) or basic (pH 8.0) environments. Our findings indicate that the expression of cbl genes is maximal under slightly acidic conditions, with the optimal pH being between 6.0 and 7.0. This range correlates well with the known pH of the human ASL, which has been determined to be 6.78 ± 0.2 (13). The proposed increased acidity of the CF ASL (5) may have an additional inducing effect on cbl gene expression.

Although the ionic content of the CF ASL has been a matter of debate, there is evidence for increased levels of Cl⁻ ions compared to normal ASL (44). We found that increasing the NaCl concentration had a positive effect on cbl gene expression, with the lowest levels of expression measured in the NaCl-free M9 medium. Several studies have determined the concentrations of both Na⁺ and Cl⁻ in the ASL to be approximately 100 mM each (13), which is in the range of the NaCl concentrations tested in this study. Growth temperature also had a significant effect on cbl gene expression, with up to twofold higher levels at 37°C compared to expression levels at 30°C. Our findings suggest that the cable pilus expression may be increased at the physiological temperature of the human body.

Iron limitation had an inducing effect on cbl gene expression. Iron is a scarce nutrient in the human body, with the majority of the metal sequestered inside host cells or by transport and storage proteins (26). Sokol and coworkers have demonstrated that secretion of ornibactin, an iron-scavenging siderophore, is essential for virulence of B. cenocepacia in both chronic and acute models of infection (35). Furthermore, the B. cenocepacia fur gene has recently been identified, encoding a homolog of the pleiotropic iron-responsive transcriptional repressor (18). The B. cenocepacia Fur protein may directly or indirectly lead to a partial repression of cbl genes under iron-rich conditions, which is counteracted by derepression in iron-limiting environments. Although the consensus Fur-binding sequence 5′-GATAATGATAATCATTATC-3′ (8) was not identified within the cblB promoter region, there are multiple tracts of A+T nucleotides proximal to the cblB transcriptional initiation site, which Fur may interact with to mediate repression of cbl genes under iron-replete conditions. Our results indicate that acidic pH, high osmolarity, temperature of 37°C, and iron limitation are all inducing conditions for cbl gene expression and may be sensed by B. cenocepacia in the CF lung, resulting in induction of cable pilus expression.

Several of the environmental conditions examined in this study had an effect on the growth rate of B. cenocepacia strain BC7. However, there was no direct correlation between growth rate and cbl gene expression. For example, incubation of B. cenocepacia strain BC7 both at 30°C or in the presence of EDDA resulted in a reduced growth rate. However, these conditions had opposite effects on cbl gene expression, repressing or inducing expression, respectively (Fig. 4C and Fig. 5B). These observations indicate that growth rate per se is not a direct indicator of the level of cbl gene expression.

In addition to the cblB-proximal promoter characterized in this study, only one other region of the cbl locus, located within or adjacent to the cblC-cblD intergenic region, gave rise to transcriptional activity above levels of the vector control (Fig. 1). While the measured activity was significantly lower than that of the cblB-proximal promoter, we cannot rule out the possibility that a weak promoter within this region also contributes to the expression of cblD and/or cblS. We also cannot exclude the formal possibility that additional promoters, which have yet to be identified, may be active under growth conditions other than those examined in this study.

By Northern hybridization analysis, we were unable to detect transcripts hybridizing to probes other than cblA. Similar findings have been reported for transcripts corresponding to the genes encoding the ETEC CFA/I usher and minor pilin, homologs of the B. cenocepacia cblC and cblD gene products, respectively (14). However, using RT-PCR, we were able to amplify transcripts corresponding to portions of the cblBA, cblBAC, cblCD, and cblDS genes, which along with the transcriptional fusion studies strongly argues that all five genes are expressed and cotranscribed.

Our deletion analysis of the cblB-proximal promoter revealed that 78 base pairs upstream of the cblB-proximal promoter transcriptional initiation site are both required and sufficient for full activity in both rich and minimal media (Fig. 1). This region of DNA may be responsible for binding transcriptional regulator(s) of cbl gene expression. Downstream of the cblS gene, we have recently identified an open reading frame predicted to encode a protein with high sequence homology to the DNA-binding response regulators of bacterial two-component signal transduction systems. The putative response regulator, designated CblR, along with the CblS putative sensor kinase and possibly additional components of the signal transduction pathway, may be involved in the transcriptional control of cbl gene expression.

Analysis of the DNA sequence immediately downstream of the cblA gene identified a region of dyad symmetry, predicted to form a stem-loop structure in the corresponding mRNA, through interactions between nine G+C base pairs (Fig. 9). The stem-loop structure is followed by a stretch of six uracyl nucleotides in the mRNA, which together may constitute a strong Rho-independent transcriptional terminator. It is likely that transcriptional termination preferentially occurs downstream of the cblA gene, resulting in a cblBA dicistronic transcript. Termination of transcription at the putative stem-loop structure is consistent with the size of the processed 0.7-kb cblA transcript, whose 5′ end was mapped by primer extension. Furthermore, under the same RT-PCR conditions, the molar amount of the amplified cblA transcript was significantly higher than that of the cblBAC product (Fig. 7A and C). Since the 3′ end oligonucleotide primer used to amplify the cblBAC product is positioned downstream of the transcriptional terminator, the amount of the RT-PCR product obtained is reflective of the relative efficiency of transcription continuing past the stem-loop. Our results suggest that termination at the stem-loop structure occurs in approximately 80% of transcription events. Although our RT-PCR analysis was semiquantitative, the significantly higher abundance of the amplified cblA transcript compared to the amount of cblBAC transcript suggests that termination of transcription at the stem-loop structure is highly efficient. This transcriptional termination mechanism would result in reduced transcription of the cblCDS genes relative to the cblBA genes. Additionally, stem-loop structures at the 3′ ends of mRNAs have been shown to stabilize transcripts, protecting them from 3′ to 5′ exonuclease activities of RNases (36). Therefore, the stem-loop structure may also act to stabilize the cblA transcript, generated by mRNA processing. A model for transcriptional and posttranscriptional control of cbl gene expression is presented in Fig. 9.

FIG. 9.

A model for transcriptional and posttranscriptional regulation of cbl gene expression. In response to environmental signals, the cbl genes are cotranscribed from the cblB-proximal promoter (P). Transcription is preferentially terminated downstream of cblA by a Rho-independent transcriptional termination mechanism, facilitated by the stem-loop structure (  |○) downstream of cblA. Thus, the stem-loop functions as an attenuator, reducing the expression of cblC, cblD, and cblS. A cblBA dicistronic transcript is processed within the cblB coding region by an as yet unknown mechanism, yielding a truncated cblB mRNA and the stable 0.7-kb cblA transcript. Since the truncated cblB mRNA does not encode a full-length CblB protein, the cblBA mRNA processing event effectively negatively regulates CblB expression. The 0.7-kb cblA mRNA is stabilized by the 3′ end stem-loop structure, leading to high-level expression of the major structural subunit of cable pili, relative to other components of the assembly pathway. In contrast, the truncated cblB mRNA is rapidly degraded. Low-level transcription through the terminator downstream of cblA allows transcription of the cblC, cblD, and cblS genes. A weak promoter within or adjacent to the cblC-cblD intergenic region may also contribute to the expression of cblD and/or cblS. Nucleotides in the cblBA transcript, shown in bold, indicate the mRNA processing site.

We have mapped the 5′ end of the 0.7-kb cblA transcript and have found that it originates from within the cblB coding region. The 5′ end of the stable 0.7-kb cblA transcript corresponds to the cblBA mRNA processing site. The pattern of three major products obtained by primer extension suggests imprecise cleavage of cblBA mRNA, which is consistent with the known activity of RNases (1, 14). RNA processing and differential stability are mechanisms known to control pilus gene expression in other systems (1, 2, 14, 24). Although we did not consistently detect cblB-hybridizing mRNAs by Northern hybridization analysis, the mapping of the transcriptional initiation site of the cblB-proximal promoter demonstrates the presence of cblB transcript(s). The transcript mapped by primer extension is either a dicistronic cblBA mRNA, a polycistronic transcript, a truncated cblB RNA, or any combination thereof. It is clear, however, that the cblB transcript(s) are significantly less abundant than the 0.7-kb cblA mRNA, as over twofold more total B. cenocepacia RNA was used to map the transcriptional initiation site of the cblB-proximal promoter than was used to map the 5′ end of the stable cblA transcript. Moreover, the RT-PCR product corresponding to cblBA mRNA was significantly less abundant than the product corresponding to the processed cblA mRNA. Together, our results strongly suggest that cblBA mRNA processing is a highly efficient event.

Pilus gene expression in E. coli is known to be regulated at the posttranscriptional level through mRNA processing and differential stability. Our study is the first to provide evidence for similar posttranscriptional control of a pilus operon in a nonenteric pathogen. The results presented here suggest that regulation of pilus gene expression on the posttranscriptional level may be more widespread in bacteria than previously appreciated. Studies are currently under way to identify elements, both cis and trans, controlling cbl gene expression on the transcriptional and posttranscriptional levels.