Positive Autoregulation of mrkHI by the Cyclic Di-GMP-Dependent MrkH Protein in the Biofilm Regulatory Circuit of Klebsiella pneumoniae

ABSTRACT

Klebsiella pneumoniae is an important cause of nosocomial infections, primarily through the formation of surface-associated biofilms to promote microbial colonization on host tissues. Expression of type 3 fimbriae by K. pneumoniae facilitates surface adherence, a process strongly activated by the cyclic di-GMP (c-di-GMP)-dependent transcriptional activator MrkH. In this study, we demonstrated the critical importance of MrkH in facilitating K. pneumoniae attachment on a variety of medically relevant materials and demonstrated the mechanism by which bacteria activate expression of type 3 fimbriae to colonize these materials. Sequence analysis revealed a putative MrkH recognition DNA sequence (“MrkH box”; TATCAA) located in the regulatory region of the mrkHI operon. Mutational analysis, electrophoretic mobility shift assay, and quantitative PCR experiments demonstrated that MrkH binds to the cognate DNA sequence to autoregulate mrkHI expression in a c-di-GMP-dependent manner. A half-turn deletion, but not a full-turn deletion, between the MrkH box and the −35 promoter element rendered MrkH ineffective in activating mrkHI expression, implying that a direct interaction between MrkH and RNA polymerase exists. In vivo analyses showed that residues L260, R265, N268, C269, E273, and I275 in the C-terminal domain of the RNA polymerase α subunit are involved in the positive control of mrkHI expression by MrkH and revealed the regions of MrkH required for DNA binding and transcriptional activation. Taken together, the data suggest a model whereby c-di-GMP-dependent MrkH recruits RNA polymerase to the mrkHI promoter to autoactivate mrkH expression. Increased MrkH production subsequently drives mrkABCDF expression when activated by c-di-GMP, leading to biosynthesis of type 3 fimbriae and biofilm formation.

IMPORTANCE Bacterial biofilms can cause persistent infections that are refractory to antimicrobial treatments. This study investigated how a commonly encountered hospital-acquired pathogen, Klebsiella pneumoniae, controls the expression of MrkH, the principal regulator of type 3 fimbriae and biofilm formation. We discovered a regulatory circuit whereby MrkH acts as a c-di-GMP-dependent transcriptional activator of both the gene cluster of type 3 fimbriae and the mrkHI operon. In this positive-feedback loop, whereby MrkH activates its own production, K. pneumoniae has evolved a mechanism to ensure rapid MrkH production, expression of type 3 fimbriae, and subsequent biofilm formation under favorable conditions. Deciphering the molecular mechanisms of biofilm formation by bacterial pathogens is important for the development of innovative treatment strategies for biofilm infections.

INTRODUCTION

Biofilm-related infections substantially contribute to patient morbidity and place a significant burden on health care systems. The evolution of high-technology medicine such as intensive and invasive hospital care practices and immunosuppressive therapy and the emergence of multiply antibiotic-resistant organisms have contributed to the rise in nosocomially acquired infections among susceptible patients. Klebsiella pneumoniae is a Gram-negative, opportunistic pathogen that is frequently associated with nosocomial infections such as pneumonia, urinary tract infections, and septicemia (1). An important clinical pathogenic mechanism of K. pneumoniae is its ability to form biofilms on inert surfaces such as medical devices, which can lead to rapid microbial colonization on host tissues (2–6).

Like some other bacteria that cause serious biofilm infections, adherence of K. pneumoniae cells on a surface requires a chemosensory system involving cyclic di-GMP (c-di-GMP) to transduce the biosynthesis of fimbriae (7–10). K. pneumoniae isolates express two different, well-characterized fimbrial structures that are involved in surface attachment—the type 1 and type 3 fimbriae (4, 5, 11–16). FimK regulates type 1 fimbriation through a phase-variation process (17), while type 3 fimbrial production is controlled by MrkH, a novel c-di-GMP-dependent transcriptional activator (7, 8, 10, 18) that binds to a nucleotide sequence known as the “MrkH box” that is located upstream of the type 3 fimbrial operon, mrkABCDF (9). A series of double-base mutations constructed within the MrkH box revealed that alterations in the 6-bp core of the MrkH box (TATCAA) exerted a major reduction in MrkH-mediated mrkABCDF transcriptional activation (9). The type 3 fimbriae mediate attachment to human endothelial and urinary bladder cell lines and also participate in biofilm formation on abiotic surfaces, including surfaces coated with human extracellular matrix proteins such as type IV and type V collagen (13, 15, 19–22).

The most widespread signal transduction system regulating bacterial biofilm formation involves the second messenger, c-di-GMP. This cyclic dinucleotide is a “lifestyle switch regulator” that controls the transition between sessility and motility. It acts to coordinate diverse physiological processes, including motility, adhesion, biofilm formation, fimbrial expression, exopolysaccharide synthesis, cell-cell communication, virulence, and cell cycle progression (23, 24). The best-studied c-di-GMP receptors are members of a family of proteins containing a PilZ domain (25). In K. pneumoniae, the MrkH regulator contains a well-conserved c-di-GMP binding PilZ domain and a point mutation in the PilZ domain completely ablates MrkH-mediated mrkABCDF operon transcriptional activity (7, 8, 10).

Given the central role of MrkH as a powerful transcriptional activator of the gene cluster of type 3 fimbriae, we were interested in investigating the regulation of MrkH expression in K. pneumoniae. MrkH is encoded within the mrkHIJ gene cluster, located adjacent to the mrkABCDF operon (26). MrkI, which contains a LuxR-like DNA binding domain at its C-terminal region, is also involved in regulating expression of type 3 fimbriae (7, 8, 10). MrkJ, a phosphodiesterase enzyme, degrades intracellular c-di-GMP to suppress expression of type 3 fimbriae and biofilm formation (8, 26). Transcriptional analysis suggests that mrkHI is transcribed as a single unit whereas mrkJ is activated independently (7, 8, 10).

Although the mechanism of MrkH-mediated mrkABCDF gene expression has been recently elucidated, there are conflicting reports on the regulation of the mrkHI operon. In two relevant studies, Johnson et al. (7) showed no autoregulatory effect of MrkI on mrkHI transcription, while Wu et al. (10) demonstrated that MrkI autoactivates mrkHI transcription in a ferric uptake regulator (Fur)-dependent manner. In this report, we show that MrkH binds directly to what we term a “MrkH box” located upstream of the mrkHI operon and activates its transcription in a c-di-GMP-dependent manner. Additionally, we show that the MrkH-mediated activation of mrkHI gene expression requires an interaction with the C-terminal domain of the α subunit (α-CTD) of the RNA polymerase (RNAP).

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth condition.

The bacterial strains and plasmids used in this study are listed in Table S1 in the supplemental material. K. pneumoniae AJ218 is a human urinary tract infection isolate (27). Unless otherwise stated, all bacterial strains were cultured at 37°C, with shaking, in Luria-Bertani (LB) broth overnight. When necessary, the growth medium was supplemented with appropriate antibiotics at the following concentrations: ampicillin, 100 μg/ml; trimethoprim, 40 μg/ml; chloramphenicol, 30 μg/ml; kanamycin, 50 μg/ml.

DNA manipulations and preparations.

PCR amplifications were performed with GoTaq Green master mix (Promega, Madison, WI) or Phusion High-Fidelity PCR master mix (New England BioLabs, Ipswich, MA) or VentR DNA polymerase (New England BioLabs) in accordance with the instructions of the manufacturers. Restriction endonucleases and T4 DNA ligase (New England BioLabs) were used for DNA cloning. Various lengths of wild-type and mutant mrkH gene and promoter regions were constructed by either standard or overlapping PCR techniques. The relevant forward and reverse primers were obtained from Geneworks (Hindmarsh, South Australia, Australia) (see Table S2 in the supplemental material), and K. pneumoniae AJ218 genomic DNA was used as the template. Mutations in the mrkHI promoter region and the mrkH gene were constructed by overlapping-extension PCR using mutagenic oligonucleotides (see Table S2) as described previously (9). Amplified fragments were cloned into pGEM-T Easy (Promega) and sequenced. Various lengths and mutations of the mrkHI promoter were excised from pGEM-T Easy and ligated into the pMU2385 lacZ fusion vector to generate single-copy transcriptional fusion vectors (28). The wild-type and mutant mrkH genes were cloned from pGEM-T Easy to pACYC184 (8, 29).

β-Galactosidase assay.

The β-galactosidase assay was assayed as described elsewhere (30). Overnight Escherichia coli MC4100 transformants were diluted 1:20 with LB medium supplemented with appropriate antibiotics. The bacterial cultures were grown at 37°C with shaking until the optical density at 600 nm (OD₆₀₀) of the culture reached 0.6 to 0.8, after which the production of β-galactosidase was assayed in triplicate for each sample.

RNA extraction.

E. coli MC4100 (31) and K. pneumoniae AJ218 carrying different plasmid derivatives were grown in LB medium with appropriate antibiotics to the mid-log phase at an OD₆₀₀ of 0.8. A 10-ml volume of culture was incubated with 20 ml of RNAProtect solution (Qiagen, Germany) at room temperature. The cells were pelleted, and RNA was extracted using a FastRNA Pro Blue kit (Q-Biogene, Canada). Residual DNA was removed using DNase I and an RNase-free DNase set (Qiagen) prior to RNA purification with an RNeasy MiniElute cleanup kit (Qiagen).

Real-time qPCR.

500 ng of extracted cellular RNA was used as the template to synthesize cDNA of interest using SuperScript II reverse transcriptase (Life Technologies, Carlsbad, CA) with random hexamers (Life Technologies) according to the manufacturer's instructions. The absence of residual genomic DNA was verified before quantification of the gene expression was carried out. Quantitative PCRs (qPCRs) were performed using a 20-μl reaction mixture with the following ingredients: 10 μl 2× SsoFast Evagreen Supermix (Bio-Rad Laboratories, Hercules, CA), a 1 μM concentration of primer pair qrtMrkHRev/qrtMrkHFor or primer pair qrpoDFor/qrpoDRev (see Table S2 in the supplemental material), and a 10-ng volume of the synthesized first-strand cDNA. Each reaction was performed in triplicate. Amplification and detection of the gene of interest were performed with a CFX96 reverse transcription-PCR (RT-PCR), C1000 thermal cycler (Bio-Rad) using the following protocol: 95°C for 30 s, followed by 40 cycles of 95°C for 5 s and 55°C for 20 s. The data were analyzed with CFX Manager Version 2.0 (Bio-Rad). The qPCR data were normalized against a K. pneumoniae housekeeping gene, rpoD, and the expression ratios of mrkH in strains with or without mrkJ and strains with or without yfiRNB were determined.

Expression and purification of wild-type MrkH-8×His protein.

Wild-type mrkH was amplified and cloned into pET11a to create pETmrkH-8His, as described elsewhere (8). For overexpression of mrkH, E. coli BL21(DE3) (32) carrying pETmrkH-8His was induced with 0.6 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 3 h at 16°C with shaking. Overexpressed MrkH-8×His was purified using nickel-nitrilotriacetic acid (Ni-NTA)–agarose (Qiagen).

Electrophoretic mobility shift assay (EMSA).

Primer mrkHRev was labeled at the 5′ end with [γ-³²P]ATP (PerkinElmer, Waltham, MA) and T4 polynucleotide kinase (Promega). The mrkH promoter region was generated by PCR using primer ³²P-mrkHRev and primer mrkH-184 with pGEM-T Easy carrying the mrkHI promoter region as a template. ³²P-labeled mrkH fragments (approximately 0.5 nM) were incubated with various amounts of purified MrkH-8×His and 200 μM c-di-GMP at 30°C for 20 min in binding buffer [10 mM Tris HCl (pH 7.4), 50 mM KCl, 1 mM dithiothreitol (DTT), 100 mg/ml bovine serum albumin (BSA), 5 ng/ml poly(dI-dC)]. Glycerol was added to reach a final concentration of 6.5%. Samples were analyzed by electrophoresis on 5% native polyacrylamide gels containing 50 μM c-di-GMP. Electrophoresis was carried out at room temperature for approximately 12 h at 10 V/cm.

Primer extension assay.

Total cellular RNA was extracted (as described above) from E. coli MC4100 strains containing pMrkH (8) with either pMU2385 (control) or pHI-lacZ1. Primer mrkHRev was labeled at the 5′ end with both [γ-³²P]ATP and T4 polynucleotide kinase and subsequently coprecipitated with 5 mg of total RNA isolated from E. coli MC4100 containing pMrkH with either pMU2385 (control) or pHI-lacZ1. Hybridization was performed at 45°C for 15 min using 10 ml of Tris-EDTA (TE) buffer containing 150 mM KCl. The primer extension reaction was started by the addition of 24 μl of extension solution (20 mM Tris HCl [pH 8.4], 10 mM MgCl₂, 10 mM DTT, 2 mM [each] deoxynucleoside triphosphates [dNTPs], 1 U/ml avian myeloblastosis virus [AMV] reverse transcriptase) and was carried out at 42°C for 1 h. Samples were then precipitated and analyzed on a sequencing gel.

Biofilm formation.

The following materials were used in biofilm formation experiments: Nunc 96-well nontreated polystyrene microplates (Thermo Scientific, Waltham, MA), Falcon 14-ml polypropylene tubes (BD, Franklin Lakes, NJ), 30-ml polypropylene containers (Technoplas, Australia), HDA 14-mm-diameter glass coverslips (Yancheng Huida Medical Instruments, China), 8-mm-diameter stainless steel balls (Livingstone International, Australia), Bardia silicone elastomer-coated latex Foley urinary catheters (Bard Medical, Covington, GA), and Millicoat 96-well human collagen IV cell adhesion strips (Millipore, Temecula, CA). Sterile glass coverslips, stainless steel balls, and urinary catheters (1-cm cross-sectioned lengths) were placed in 48-well microplates. Biofilm assays were performed according to a method previously described by O'Toole (8, 33), with minor modifications. Briefly, the different materials were incubated in M63B1-GCAA minimal medium (containing 1% glycerol and 0.3% Casamino Acids) subcultured with ∼10⁷ CFU/ml K. pneumoniae. Following 24 h of static incubation at 37°C (8 h for catheter sections), materials were washed three times in double-distilled water (dH₂O). Biofilms attached to surfaces were stained with 0.1% (wt/vol) crystal violet solution (Sigma-Aldrich, St. Louis, MO), solubilized with 33% acetic acid, and then quantified at OD₆₀₀. The data for each strain represent average values taken from five replicate samples in determinations performed in three independent experiments.

Statistical analysis.

Statistical analysis was conducted using one-way analysis of variance (ANOVA) with the Bonferroni posttest (GraphPad Prism Version 6; GraphPad, La Jolla, CA); a P value of <0.05 was considered statistically significant.

RESULTS

MrkH mediates adherence of K. pneumoniae to medically relevant substrates.

MrkH, a transcriptional activator of expression of type 3 fimbriae, has been previously shown to promote biofilm formation on polystyrene and polyvinyl chloride surfaces (8, 9). We sought to investigate the formation of K. pneumoniae biofilms on a wider range of the medically relevant materials that a nosocomial pathogen might encounter and also to determine the importance of MrkH and type 3 fimbriae in mediating attachment to these different surfaces. The materials studied here included polystyrene, polypropylene, polycarbonate, glass, stainless steel, urinary catheters, and type IV human collagen-coated surfaces. In a crystal violet-based static biofilm assay, all materials tested could support the attachment of wild-type K. pneumoniae (Fig. 1). The mrkA gene of the mrkABCDF operon encodes the major structural subunit (MrkA) of type 3 fimbriae. We observed a significant reduction in biofilm formation by the K. pneumoniae ΔmrkA and ΔmrkH mutants on each surface. These mutant phenotypes were successfully complemented with multicopy wild-type mrkA or mrkH (pMrkA and pMrkH, respectively). Thus, we observed that K. pneumoniae can form robust biofilms on materials with different surface properties and that MrkH-mediated control of type 3 fimbriae is the principal mechanism for initiating biofilm formation. A key step in this regulation of biofilms by MrkH is therefore the regulation of expression of MrkH itself.

FIG 1.

Biofilm formation by K. pneumoniae AJ218 on different materials. Levels of biofilm formation by the K. pneumoniae wild-type strain, type 3 fimbrial ΔmrkA mutant, ΔmrkH mutant, ΔmrkH mutant containing pMrkH, and ΔmrkH mutant containing pACYC184 (empty vector control) are shown. Biofilm formation was assessed after 24 h using a crystal violet-based static assay. All values represent the means of the results of analyses of five replicate samples for each strain performed in three independent experiments (n = 15). The error bars represent ± standard deviations.

Mapping the transcriptional start site of the mrkHI operon.

Sequence analysis and a genome comparison between K. pneumoniae and E. coli indicated that a MrkH homologue is not encoded by E. coli. Given the high degree (>99% identity) of RNAP subunit conservation between the two bacterial species, transcriptional analyses for K. pneumoniae can be conveniently performed in an E. coli system (8, 9).

Primer extension was carried out to locate the mrkHI transcription start site. Total cellular RNA was extracted from E. coli MC4100 carrying the pMrkH plasmid and either pMU2385 (control plasmid) or its derivative carrying the transcriptional fusion, pHI-lacZ1 (test plasmid) (see below). The RNA was hybridized with primer ³²P-mrkHRev and extended with AMV reverse transcriptase. As shown in Fig. 2A, a distinct extension band was seen only with the test strain. This mapped the mrkHI transcriptional start site to the adenine residue located 53 bp upstream of the putative mrkH translational start site (Fig. 2B). Based on the start site of transcription, the putative promoter core elements, which included a −35 region (TAAAAA), a −10 region (CAGAAT), and a 17-bp spacer, were identified (Fig. 2B).

FIG 2.

Analysis of the mrkHI regulatory region. (A) Primer extension to identify the mrkHI transcription start site. Total cellular RNA was purified from E. coli MC4100 containing pMrkH with either empty pMU2385 (control) or pHI-lacZ1. RNA samples were hybridized with ³²P-mrkHRev. Primer extension was performed using AMV reverse transcriptase in the presence of dNTPs. GA, GA sequence ladder prepared using the mrkH PCR fragment generated using primer pair ³²P-mrkHRev and mrkH-184. Lane 1, experiment using RNA from E. coli MC4100 containing pMrkH and pHI-lacZ1. Lane 2, control experiment using RNA from E. coli MC4100 containing pMrkH and pMU2385. The positions corresponding to the ³²P-mrkHRev primer and the extension product are marked. (B) Nucleotide sequence of the mrkHI regulatory region. The mrkHI transcriptional start site is marked with an angled arrow. The numbering to the left of the sequences is relative to the start site of transcription. The −35, −10, and spacer sequences are marked. The MrkH box and its center position (relative to the start site of transcription) are marked above the sequence. The putative Fur box is underlined. The genetic changes of the various mutations are shown below or above the sequence.

Transcriptional analysis of the mrkHI regulatory region.

To characterize the mrkHI regulatory region, we generated four PCR fragments which spanned positions −184, −88, −66, and −28 to +126, relative to the transcriptional start site of the mrkHI operon (Table 1), and inserted each fragment into pMU2385 to generate single-copy lacZ transcriptional fusion plasmids (pHI-lacZ1, -2, -3, and -4, respectively). These pMU2385 derivatives were then transformed into E. coli MC4100 carrying either the pACYC184 vector (MrkH⁻ background) or pMrkH (MrkH⁺ background). The mrkHI transcriptional activation was analyzed by measuring the production of β-galactosidase.

TABLE 1.

Transcriptional analysis of the mrkHI regulatory region

lacZ reporter construct	mrkHI promoter activity (Miller units)a
	MrkH−	MrkH+
pHI-lacZ1 (−184 to +126)	5 ± 0.3	518 ± 80 (98)
pHI-lacZ2 (−88 to +126)	4 ± 0.3	453 ± 66 (109)
pHI-lacZ3 (−66 to +126)	4 ± 1.3	6 ± 1.1 (1)
pHI-lacZ4 (−28 to +126)	6 ± 0.3	7 ± 1.1 (1)

^aβ-Galactosidase assays were performed with E. coli MC4100 derivatives containing different mrkHI promoter-lacZ truncations. The values are the means ± standard deviations of the results from at least three biological replicates. The values of fold activation, equal to the specific activity of β-galactosidase of the MrkH⁺ strain divided by the activity of the MrkH⁻ strain, are shown in parentheses.

As shown in Table 1, in the MrkH⁻ background, the four constructs (pHI-lacZ1 to -4) exhibited very weak transcriptional activity (4 to 6 U of β-galactosidase activity). Strikingly, in the MrkH⁺ background, the transcriptional levels of E. coli strains carrying pHI-lacZ1 and pHI-lacZ2 increased 98- and 109-fold to 518 and 453 U, respectively; however, the promoter activities of E. coli strains carrying pHI-lacZ3 and pHI-lacZ4 were largely unchanged (Table 1). These results demonstrated that expression of the mrkHI promoter is strongly activated by the MrkH protein and indicated that a cis-acting element located between or overlapping positions −88 and −66 is responsible for MrkH-mediated activation of mrkHI transcription.

Identification of the MrkH binding site.

Having previously identified a MrkH recognition site (i.e., the MrkH box) at the mrkA promoter (CATCTATCAATG) (9), we were interested in scanning the K. pneumoniae genome sequence for other putative MrkH binding sites through nucleotide homology. Sequence analysis of the region upstream of the mrkHI promoter revealed the presence of a putative MrkH recognition site (MrkH box [TATCAA]) located between −78 and −73 relative to the mrkHI transcriptional start site (Fig. 2B). To examine whether the putative MrkH box is important for MrkH-mediated activation, the nucleotide sequences within this region were scrambled to create a mutant construct, pHBox-mut1 (Fig. 2B). The results of a β-galactosidase assay showed that this mutation caused a severe reduction of MrkH-mediated activation of the mrkHI promoter (Table 2).

TABLE 2.

The effects of various MrkH box mutations in the MrkH⁻ and MrkH⁺ backgrounds

lacZ reporter construct	Sequence	mrkHI promoter activity (Miller units)a
		MrkH−	MrkH+
pHI-lacZ2 (wild type)	TATCAA	8 ± 2.5	794 ± 36 (99)
pHBox-mut1	-CCAG-	6 ± 2.1	13 ± 0.8 (2)
pHBox-mut2	G-----	8 ± 2.7	58 ± 4.1 (7)
pHBox-mut3	-C----	8 ± 2.6	222 ± 32 (28)
pHBox-mut4	--G---	8 ± 1.1	29 ± 4.9 (3)
pHBox-mut5	---A--	11 ± 0.7	277 ± 47 (27)
pHBox-mut6	----C-	8 ± 1.3	129 ± 28 (17)
pHBox-mut7	-----C	11 ± 0.6	169 ± 9.4 (16)

^aβ-Galactosidase assays were performed with E. coli MC4100 derivatives containing pHI-lacZ2 mutant variants within the MrkH box. The values are the means ± standard deviations of the results from three biological replicates. The values of fold activation, equal to the specific activity of β-galactosidase of the MrkH⁺ strain divided by the activity of the MrkH⁻ strain, are shown in parentheses.

To further characterize the MrkH box, we introduced a single nucleotide change into each of its six positions (Table 2). All six mutant constructs (pHbox-mut2 to -7) displayed a significant reduction in MrkH-mediated mrkHI activation, thus demonstrating an important contribution of each nucleotide within the MrkH box to the optimal interaction with the MrkH protein.

To demonstrate direct binding of the MrkH protein to DNA containing the MrkH box sequence, an electrophoretic mobility shift assay (EMSA) was carried out. The DNA fragment spanning −184 to +126 relative to the mrkHI transcriptional start site was labeled with ³²P and incubated with histidine-tagged MrkH in the presence or absence of cofactor c-di-GMP and analyzed by native polyacrylamide gel electrophoresis. In the presence of c-di-GMP, a MrkH-DNA complex was observed at MrkH concentrations of 200 nM and 400 nM (Fig. 3). In the absence of c-di-GMP, little or no protein-DNA complex was observed. Taken together, the results from the mutational assays and EMSAs demonstrate that MrkH activates the expression of the mrkHI operon by binding to the MrkH box which is centered at −75.5 relative to its start site of transcription in a c-di-GMP-dependent manner.

FIG 3.

EMSA of MrkH-8×His binding to the mrkHI regulatory region. The ³²P-labeled PCR fragment containing the mrkHI regulatory region and putative MrkH box was generated using primer pair ³²P-mrkHRev and mrkH-184. The mrkH fragment was mixed with various amounts (from 0 to 400 nM) of purified MrkH-8×His in the absence (−) or presence (+) of c-di-GMP (200 μM). Following incubation at 30°C for 20 min, the samples were analyzed on native polyacrylamide gels. The migration positions for the free DNA (F) and protein-DNA complex (C) are marked.

Expression of mrkHI is influenced by MrkJ- and YfiN-mediated control of c-di-GMP expression.

The mrkJ gene (located immediately upstream of mrkHI) encodes a phosphodiesterase enzyme that degrades intracellular c-di-GMP (8, 26), while the yfiN gene (located in the yfiRNB operon) encodes a diguanylate cyclase that synthesizes c-di-GMP (8, 34). To investigate whether changes in the concentration of intracellular c-di-GMP affect mrkH expression, we compared the levels of MrkH-mediated mrkH gene expression in ΔmrkJ and ΔyfiRNB K. pneumoniae mutants and their complemented plasmid strains by performing a real-time quantitative PCR assay. The relative mrkH mRNA expression ratios were determined for two strain combinations. The mean expression ratios of mrkH transcripts in the YfiRNB⁺-overexpressing strain relative to the ΔyfiRNB mutant and in the ΔmrkJ mutant relative to the mrkJ-overexpressing (MrkJ⁺) strain were evaluated. The relative expression ratios of K. pneumoniae YfiRNB⁺ and ΔyfiRNB strains and K. pneumoniae ΔmrkJ and MrkJ⁺ strains were 18.5 ± 0.5 and 17.4 ± 0.6 (means ± standard deviations of the results from three biological replicates), respectively. Consistent with our observation that the interaction of purified MrkH with the mrkH promoter is c-di-GMP dependent, overexpression of a diguanylate cyclase promotes mrkH expression whereas overexpression of a phosphodiesterase decreases mrkH expression in K. pneumoniae.

The effect of mrkHI promoter mutations on MrkH-mediated activation.

To probe the mechanism by which MrkH activates mrkHI transcription, we constructed two pHI-lacZ2 variants (pHI-35UP and pHI-10UP) in which the −35 region or the −10 region of the mrkHI promoter was modified to fit perfectly the consensus sequence for docking RNA polymerase (Fig. 2B). Transcriptional analysis in E. coli showed that in the absence of MrkH (MrkH⁻), expression from pHI-35UP and pHI-10UP led to 1,233 and 61 U of β-galactosidase activity, respectively, representing 247-fold and 12-fold increases in the promoter strength compared to the level seen with the wild-type MrkH box (Table 3). When MrkH was provided in multiple copies on pMrkH (MrkH⁺), both mutant promoters produced approximately 2,400 U of β-galactosidase activity, giving a 5-fold increase in promoter activity compared to the wild-type MrkH box. The observation that modifying the −35 region led to (i) a major enhancement in transcriptional activity in the absence of MrkH and (ii) diminished activation of the mrkHI promoter by MrkH in the presence of MrkH suggests that the MrkH protein acts primarily to overcome an intrinsic defect incurred by the presence of the nonconsensus −35 sequence in this promoter.

TABLE 3.

The effects of various mutations on MrkH-dependent and -independent transcription from the mrkHI promoter

lacZ reporter construct	mrkHI promoter activity (Miller units)a
	MrkH−	MrkH+
pHI-lacZ1 (wild type)	5 ± 0.3	518 ± 80 (98)
pHI-35UP (consensus −35)	1,233 ± 57	2,366 ± 512 (2)
pHI-10UP (consensus −10)	61 ± 5.0	2,428 ± 332 (40)
pHI-5bp (Δ5 bp)	11 ± 1.8	17 ± 4.7 (1.6)
pHI-10bp (Δ10 bp)	12 ± 2.8	2,623 ± 263 (215)

^aβ-Galactosidase assays were performed with E. coli MC4100 derivatives containing pHI-lacZ1 variants. The values are the means ± standard deviations of the results from three biological replicates. The values of fold activation, equal to the specific activity of β-galactosidase of the MrkH⁺ strain divided by the activity of the MrkH⁻ strain, are shown in parentheses.

The involvement of the α-CTD of the RNA polymerase in MrkH-mediated activation of mrkHI expression.

The upstream location of the MrkH binding site relative to the mrkHI promoter suggests a direct interaction between the MrkH protein and the C-terminal domain of the α subunit of RNA polymerase (α-CTD of RNAP). To test this possibility, we first constructed two pHI-lacZ1 variants (pHI-5bp and pHI-10bp) that had either a 5-bp or a 10-bp deletion between the MrkH box and the −35 promoter (Fig. 2B) and analyzed their effect on MrkH-mediated activation. A half-turn deletion (Δ5-bp) almost completely abolished MrkH-mediated activation of mrkHI expression (Table 3). However, a full-turn deletion (Δ10-bp) increased the level of expression by 215-fold, thereby enhancing by 5-fold the ability of MrkH to activate the transcription of the wild-type mrkHI promoter. These results indicate that proper helical alignment of the DNA-bound MrkH and RNAP is required in order for MrkH to mediate efficient mrkHI activation. This arrangement may allow more-stable protein-protein interactions.

The observations described suggest a role for the α-CTD of RNAP in interacting with MrkH at the mrkHI promoter. To test this hypothesis and to identify residues in the α-CTD of RNAP involved in the positive control of mrkHI expression, a set of pLAW2 (35) derivatives in which the rpoA gene contained an alanine substitution at position 258 to position 275 were transformed into E. coli MC4100 carrying pMrkH and pHI-lacZ1. The results of β-galactosidase assays showed that when L260A, R265A, N268A, C269A, E273A, and I275A derivatives were overexpressed in the E. coli system, MrkH-mediated mrkHI expression was significantly reduced to between 40% to 80% of the level exhibited by the E. coli strain overexpressing the wild-type rpoA gene (Fig. 4A). Based on the crystal structure of the α-CTD of RNAP (36, 37), five of the six residues are exposed on the surface of the α-CTD, while the other (I275) is buried inside the structure (Fig. 4B). These results demonstrate the importance of these α-CTD residues in the (either direct or indirect) activation of mrkHI expression by MrkH.

FIG 4.

Identification of RNAP α-CTD amino acid residues important for MrkH-mediated mrkHI transcription. (A) Each of the plasmids containing rpoA mutations with alanine substitutions between positions 258 and 275 was introduced into E. coli MC4100 containing plasmids pHI-lacZ1 and pMrkH. The β-galactosidase activities of samples from three independent experiments are presented relative to the activity of the strain harboring pLAW2 containing wild-type (WT) rpoA. Data from alanine replacements that caused a significant reduction in β-galactosidase activity are shown in black. Data are shown as means ± standard deviations, with n = 3 per group. Each rpoA mutant was compared to the wild-type control using one-way ANOVA with the Bonferroni posttest, *, P < 0.05; **, P < 0.01; ***, P < 0.0001. (B) The structure of the α-CTD of RNAP. Surface-exposed residues L260, R265, N268, C269, and E273 implicated in MrkH-mediated mrkHI expression are marked.

Analysis of MrkH mutations and MrkH-mediated mrkHI transcription activation.

MrkH contains a PilZ domain with a conserved c-di-GMP binding site, which is critical for its function as a transcriptional activator of the mrkABCDF operon (7, 8). To examine whether c-di-GMP binding by the PilZ domain is also critical for the function of MrkH with respect to activation of the mrkHI operon, E. coli MC4100 carrying pHI-lacZ1 was transformed with either wild-type pMrkH or the MrkH mutant construct, pMrkH(113R-A). The mutant construct contains a single alanine substitution mutation within the c-di-GMP binding site of MrkH, which we have shown previously to destroy the ability of MrkH to bind to the mrkA regulatory region and activate mrkABCDF transcription (8). In vivo transcriptional analysis showed that the mutation in the c-di-GMP binding site also completely inhibited MrkH-mediated activation of mrkHI expression (Fig. 5). This further confirmed that the positive control of mrkHI transcription by MrkH is c-di-GMP dependent.

FIG 5.

Investigation of MrkH regions required for mrkHI transcription activation. β-Galactosidase assays were performed with E. coli MC4100 containing plasmid pHI-lacZ1 and wild-type pMrkH (MrkH⁺) or empty pACYC184 (MrkH⁻) or a mutant pMrkH construct containing one of the four AS insertion mutations in α1, β2, α2, or β10 or a point mutation in the c-di-GMP binding domain (R113A). Values represent the means of the results from three replicate samples. The error bars represent ± standard deviations.

We have previously attempted to functionally define the various structural folds of MrkH while studying the regulatory mechanism of MrkH at the mrkABCDF promoter (9). Using several algorithms to predict the MrkH secondary structure, we made alanine-serine (AS) substitution mutations within predicted α-helices and β-strands at the N terminus and C terminus of the protein (9). We generated mutations in the N-terminal α1 (positioned between residue 15 and residue 16 and denoted 15AS16) and β2 (39AS40) regions, as well as mutations in the C-terminal β10 (202AS203) and α2 (217AS218) regions. The mutant MrkH constructs encoded by pACYC184 derivatives were transformed into E. coli MC4100 containing a mrkA promoter-lacZ transcriptional fusion vector. We showed that each AS insertion within the α1, α2, and β10 regions resulted in the complete inability of MrkH to activate the mrkA promoter. However, AS insertion within the β2 region had little effect on MrkH-mediated activation of the mrkA promoter. We could show by Western blotting that these four mutant MrkH proteins are stably expressed (9). In the present study, we tested the ability of these mutant proteins to activate the expression of the mrkHI promoter. The plasmid pMrkH (wild-type MrkH control) and each of its derivatives encoding the mutant MrkH proteins were transformed into E. coli MC4100 containing pHI-lacZ1. The results of β-galactosidase assays showed that mutation of the N-terminal β2 region had no effect on mrkH transcription activation; however, mutation of the N-terminal α1 region and C-terminal α2 and β10 regions completely abrogated transcriptional activity of the mrkH promoter (Fig. 5). Thus, some functional similarities appear to exist in certain regions of MrkH that participate in transcription activation of both the mrkHI and mrkABCDF operons.

DISCUSSION

The MrkH protein is a major virulence regulator of K. pneumoniae whose expression is required for transcriptional initiation of the type 3 fimbrial operon (7–9) and for the formation of biofilms on a variety of materials (Fig. 1). Hence, the transcriptional regulation of mrkH expression per se could represent an important process in controlling the switch between planktonic and biofilm lifestyles. In this study, we investigated the expression and regulation of the mrkHI operon and demonstrated that MrkH strongly activates its own transcription in response to a secondary messenger, c-di-GMP. We are able to reveal the regulatory mechanism by which MrkH activates the mrkHI promoter through interactions with DNA and RNAP and to establish a MrkH/c-di-GMP-mediated biofilm regulatory circuit operating in K. pneumoniae (Fig. 6).

FIG 6.

Model for a K. pneumoniae biofilm regulatory circuit coordinated by MrkH and c-di-GMP. c-di-GMP is produced from GTP by diguanylate cyclases (DGCs) and is degraded by phosphodiesterases (PDEs). Upon binding c-di-GMP, MrkH can specifically bind to the MrkH box, facilitating the recruitment of RNAP (via the interaction of MrkH with the α-CTD) to both the mrkHI and mrkABCDF promoter regions. The autoactivation of mrkH expression subsequently leads to increased MrkH production and, once activated by c-di-GMP, drives mrkABCDF expression, biosynthesis of type 3 fimbriae, and rapid biofilm formation.

Primer extension and mutational experiments showed that transcription of the mrkHI operon is driven by a promoter that contains a −35 region (TAAAAA versus the consensus TTGACA), a −10 sequence (CAGAAT versus the consensus TATAAT), and a 17-bp spacer. The poor conservation of the promoter core elements correlates with the very weak basal levels of transcription from the mrkHI promoter in the MrkH⁻ background (Tables 1, 2, and 3). The sigma subunit of RNAP is known to make direct contacts with conserved nucleotides within the −35 and −10 regions during transcription initiation (38); therefore, the lack of conservation in several key positions of the core elements should considerably weaken the ability of RNAP to recognize the mrkHI promoter and form a stable initiation complex. Indeed, replacing the −35 or −10 region with the consensus sequences greatly enhanced the basal levels of mrkHI transcription in the MrkH⁻ background (Table 3).

The ability of MrkH to switch on its own transcription relies on the binding of the regulatory protein to a 6-bp sequence (TATCAA) which we named the MrkH box. This nucleotide hexamer is the core sequence of the 12-bp operator (CATCTATCAATG) responsible for MrkH-mediated transcriptional activation of the type 3 fimbrial operon mrkABCDF (9). Deletion of a half-turn (5 bp) of DNA between the MrkH box and the −35 region of the mrkHI promoter completely destroyed the activation by MrkH, whereas deletion of a full turn (10 bp) of DNA resulted in an enhancement of activation that was 5 times greater than that of the wild-type promoter (Table 3). This “face-of-the-helix” effect indicates that a direct contact between MrkH and RNAP is likely to be important for mrkHI activation. Furthermore, the strong positive effect of the 10-bp deletion indicates that the MrkH box is not positioned for maximal activation in the wild-type mrkHI regulatory region. Presumably, this promoter-operator arrangement was selected for to avoid the production of unnecessary amounts of MrkH and MrkI proteins in K. pneumoniae under inducible conditions. Alternatively, other situations could exist where the 10-bp deletion arrangement negatively impacts the fitness of cells.

The upstream location and the relative distance between the MrkH box and the mrkHI promoter core sequence pointed to the possibility that MrkH acts as a class I activator which contacts the α-CTD of RNAP to stimulate transcription initiation of the mrkHI operon. By testing a series of rpoA derivatives with single alanine substitutions, we identified six residues (L260, R265, N268, C269, E273, and I275) that are important for MrkH-mediated activation (Fig. 5). Structural and functional analyses of interactions between α-CTD and various bacterial promoters have shown that E273 is involved in an interaction with the E. coli transcriptional regulator Fis at the propP2 promoter (39) and that L260, R265, N268, and C269 residues are responsible for binding to the upstream promoter (UP) elements of several bacterial promoters, including rrnBP1, lacP1, and T7D (40–46). The L260 residue is also essential for transcription activation at the lacP1 promoter by the cyclic AMP (cAMP) receptor protein (CRP) (45) and the mtr promoter by TyrR (47). We have previously shown that three (R265, E273, and I275) of the six residues are also important for MrkH-mediated activation of the mrkABCDF operon (9), indicating that similar but not identical regulatory mechanisms are used by MrkH to control transcriptional initiation from the mrkHI and mrkABCDF promoters.

In the mrkHI regulatory region, the sequence between −63 and −48 is AT rich and could serve as an UP element for an interaction with the α-CTD of RNAP. This region partially overlaps with the “Fur box” (Fig. 2), which is the binding site of the Fur regulatory protein (48). Fur is able to fine-tune the transcription of the mrkHI operon in response to the presence of iron (10). It is known that the α-CTD of RNAP interacts with UP elements in the minor groove of DNA (44, 49) whereas Fur binds to the major DNA groove of the Fur box (50). It is therefore possible that the regulatory region of the mrkHI operon is able to accommodate both the α-CTD of RNAP and Fur.

Several lines of evidence suggest that c-di-GMP plays a central role in assisting MrkH to activate its own transcriptional expression: (i) a mutation in the c-di-GMP binding motif in the PilZ domain (R113A) abolished the activation of the mrkHI promoter by MrkH (Fig. 5); (ii) the presence of c-di-GMP enhanced the affinity of MrkH for its target DNA (Fig. 3); (iii) deleting the mrkJ gene, which encodes a phosphodiesterase responsible for c-di-GMP degradation, increased the production of the mrkH transcript (see above); and (iv) deleting the yfiN gene, which encodes an integral-membrane diguanylate cyclase involved in c-di-GMP synthesis, reduced the production of the mrkH transcript (see above). From these results, it is apparent that c-di-GMP is an important intracellular signal that controls biofilm formation by regulating the transcription of the mrkHI and mrkABCDF operons through the MrkH protein (Fig. 6). The external signals that induce biofilm formation in K. pneumoniae remain largely unknown; however, a range of environmental factors such as oxygen, bile salts, bicarbonate ions, zinc, and quorum-sensing autoinducers have been shown to influence intracellular c-di-GMP turnover of other bacterial pathogens by directly regulating the activities of their diguanylate cyclase and phosphodiesterase proteins (51).

The critical importance of MrkH in the induction of expression of type 3 fimbriae and the fact that this protein contains multiple functional motifs that interact with c-di-GMP, DNA, and RNA polymerase make it an excellent target for developing chemical compounds that prevent biofilm formation by K. pneumoniae. A high-throughput screening system similar to that described by Yang et al. (52) has been established, and experiments are under way to identify small-molecule inhibitors of the MrkH protein from a commercial chemical library.