Systematic Mutational Analysis of Histidine Kinase Genes in the Nosocomial Pathogen Stenotrophomonas maltophilia Identifies BfmAK System Control of Biofilm Development

Abstract

The Gram-negative bacterium Stenotrophomonas maltophilia lives in diverse ecological niches. As a result of its formidable capabilities of forming biofilm and its resistance to multiple antibiotic agents, the bacterium is also a nosocomial pathogen of serious threat to the health of patients whose immune systems are suppressed or compromised. Besides the histidine kinase RpfC, the two-component signal transduction system (TCS), which is the canonical regulatory machinery used by most bacterial pathogens, has never been experimentally investigated in S. maltophilia. Here, we annotated 62 putative histidine kinase genes in the S. maltophilia genome and successfully obtained 51 mutants by systematical insertional inactivation. Phenotypic characterization identified a series of mutants with deficiencies in bacterial growth, swimming motility, and biofilm development. A TCS, named here BfmA-BfmK (Smlt4209-Smlt4208), was genetically confirmed to regulate biofilm formation in S. maltophilia. Together with interacting partner prediction and chromatin immunoprecipitation screens, six candidate promoter regions bound by BfmA in vivo were identified. We demonstrated that, among them, BfmA acts as a transcription factor that binds directly to the promoter regions of bfmA-bfmK and Smlt0800 (acoT), a gene encoding an acyl coenzyme A thioesterase that is associated with biofilm development, and positively controls their transcription. Genome-scale mutational analyses of histidine kinase genes and functional dissection of BfmK-BfmA regulation in biofilm provide genetic information to support more in-depth studies on cellular signaling in S. maltophilia, in the context of developing novel approaches to fight this important bacterial pathogen.

INTRODUCTION

Histidine kinases (HKs) are the cellular sensors of the two-component signal transduction systems (TCSs) employed by most bacteria to detect and respond to environmental stimuli (1). With the exception of a few bacterial species, such as Mycoplasma spp., the bacterial cell generally encodes several to hundreds of TCSs, and the total number of TCSs possessed is a metaphor for the bacterial intelligence quotient (IQ) in terms of a bacterium's ability to adapt to various ecological niches (2). The prototypical TCS consists of a membrane-bound HK and a cytoplasmic response regulator (RR). After monitoring a specific stimulus, HK autophosphorylates itself by hydrolyzing ATP and transfers the phosphoryl group to a conserved histidine residue within the dimerization and histidine phosphotransfer (DHp) domain. Next, the phosphoryl group is transferred to a conserved aspartate residue within the N-terminal receiver domain of its cognate RR (3). Thereafter, the activated RR controls downstream gene expression, cellular behavior, or enzymatic activity, depending on the biochemical property of its C-terminal output domain. Therefore, a bacterium's ability to live in complex environments depends largely on the number, structure, and regulatory function of its HK repertoire (4). In the context of infection, some HKs (e.g., PhoQ, QseC, and SaeS) act as canonical regulators of the expression of virulence factors in bacterial pathogens (5). TCSs are encoded by the genomes of prokaryotes, slime molds, fungi, and plants. However, TCSs have not been identified in any proteome of animals (6). Hence, HK is potentially an ideal molecular target for the development of novel antibiotics defending bacterial pathogens (7, 8).

During the past 3 decades, the Gram-negative, pathogenic bacterium Stenotrophomonas maltophilia has gradually become a serious threat to people whose immune systems are compromised or suppressed by immunodeficiency disease or the prolonged antibiotic therapy used in hospitals (9, 10). This pathogen causes fatal diseases such as pneumonia, urethritis, meningitis, and bloodstream infections in such immunocompromised patients. One difficulty in treating S. maltophilia infections is that most isolates of this bacterium have the strong ability to form biofilms on the surfaces of, for example, perfusion tubes, indwelling intravascular devices, medical devices, water pipes, and host tissues (11). These bacterial aggregates are sources of infection, and it is hard to efficiently remove them. In addition, S. maltophilia genomes encode a diverse range of multidrug efflux pumps and other molecular machineries that enhance the bacterium's resistance to various antibiotics (12). For these reasons, S. maltophilia has become a serious threat to human health. Some studies have estimated that a mortality rate of up to 77% is likely attributable to S. maltophilia infection in reported nosocomial infection cases (10, 13). To fight S. maltophilia, the development of novel therapies or antibacterial agents based on targeting the regulation of biofilm formation and virulence factor expression is desirable. However, TCS regulation, one of the central molecular mechanisms potentially targetable by novel antibacterial agents, lacks proper investigation in S. maltophilia. Indeed, only HK RpfC, which is responsible for regulating cell-cell communication and virulence factor expression in the close-relative bacteria Xanthomonas spp., has been shown to control diffusible signaling factor (DSF)-triggered modulation in S. maltophilia (14–16). Our previous study identified FsnR, an orphan RR that controls transcription of flagellar genes and biofilm formation (17), but the cognate HK for FsnR remains unclear.

In the present study, we conducted a systematic mutational investigation of all of the HK genes of S. maltophilia involved in controlling biofilm formation. Among 62 putative HK genes, we obtained 51 insertion mutants using vector integration methodology by homologous, single-crossover events. Phenotypic characterization identified a series of HK genes involved in bacterial growth, swimming motility, and biofilm development. Among them, a TCS encoded by a bicistronic operon, designated BfmA-BfmK (Smlt4209-Smlt4208), was genetically confirmed to regulate bacterial cell aggregation. BfmK is a bona fide HK that has autokinase activity, and BfmA acts as a transcription factor that binds the promoter regions of the bfmA-bfmK operon to autoregulate its own expression. In addition, BfmA binds the promoter region of Smlt0800 to positively regulate its transcription. This gene encodes an acyl coenzyme A (acyl-CoA) thioesterase that is involved in biofilm formation. Our results provide essential genetic information about the HKs of S. maltophilia to facilitate future investigation on the cellular regulation of this bacterial pathogen.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The bacterial strains and plasmids used in the present study are listed in Table 1. S. maltophilia ATCC 13637 was cultured at 28°C in NYG medium (5 g/liter tryptone, 3 g/liter yeast extract, 20 g/liter glycerol; pH 7.0). S. maltophilia competent cells were prepared by culturing the bacterium in 210 medium (4 g/liter yeast extract, 8 g/liter casein enzymatic hydrolysate, 5 g/liter sucrose, 3 g/liter K₂HPO₄, 0.3 g/liter MgSO₄·7H₂O; pH 7.0); cells were collected by centrifugation (12,000 × g) and washed three times with ice-cold glycerol (10%). The Escherichia coli DH5α cells used for molecular cloning were usually cultured at 37°C in Luria-Bertani medium. Antibiotics were used at the following concentrations: kanamycin, 50 μg/ml; ampicillin, 100 μg/ml; spectinomycin, 100 μg/ml; and streptomycin, 200 μg/ml. Transformation of S. maltophilia and E. coli was achieved by electroporation (18 kV cm⁻¹, 25 μF, and 200 Ω) using a Bio-Rad pulser Xcell electroporation system (Bio-Rad, USA).

TABLE 1.

Bacterial strains and plasmids used in this study^a

Strain or plasmid	Genotype or description	Resource or reference
Strains
S. maltophilia ATCC 13637	Wild-type strain	CGMCC
E. coli DH5α	Host strain used for molecular cloning	Lab collection
E. coli BL21(DE3)	Host strain used for protein expression	Lab collection
SM004–SM055 (a total of 51 strains)	Δ0107–Δ4627 (a total of 51 strains), mutants with insertions of histidine kinase genes of S. maltophilia ATCC 13637, constructed by pK18mob vector integration; Kanr	This study
SM056	IFD-bfmA, in-frame deletion mutant of bfmA (Smlt4209)	This study
SM057	IFD-bfmK, in-frame deletion mutant of bfmK (Smlt4208)	This study
SM058	IFD-bfmA-bfmA, complementary strain of IFD-bfmA, containing a pBBRMCS2::bfmA vector; Kanr	This study
SM059	IFD-bfmK-bfmK, complementary strain of IFD-bfmK, containing a pBBRMCS2::bfmK vector; Kanr	This study
SM060	IFD-bfmA-bfmA*, similar to SM059, but with a His6 coding sequence at the 3′ end of bfmA, for ChIP assay; Kanr	This study
SM061	IFD-acoT, in-frame deletion mutant of acoT (Smlt0800)	This study
SM062	IFD-acoT-acoT, complementary strain of IFD-acoT, containing a pBBRMCS2::acoT vector; Kanr	This study
SM063	IFD-bfmA-acoT, strain IFD-bfmA containing a pBBRMCS2::acoT vector, for epistatic analysis; Kanr	This study
Plasmids
pK18mob	Suicide vector to create a mutant by a single crossover; Kanr	Schäfer et al. (40)
pK18mobsacB	Suicide vector to create a mutant by a double crossover; Kanr	Schäfer et al. (40)
pET30a	Protein expression vector; Kanr	Novagen
pBBR1MCS2	Broad-host-range vector used for genetic complementation	Lab collection
pBBR-4208	pBBR1MCS2::Smlt4208, genetic complementation vector; Kanr for bfmK mutant	This study
pBBR-4209	pBBR1MCS2::Smlt4209, genetic complementation vector; Kanr for bfmA mutant	This study
pBBR-0800	pBBR1MCS2::Smlt0800, genetic complementation vector; Kanr for acoT mutant	This study
pET30a-4209	pET30a::Smlt4209, for BfmA expression; Kanr	This study
pET28a-4208	pET28a::Smlt4208, for BfmK expression; Kanr	This study

^aKan^r, kanamycin resistant; ChIP, chromatin immunoprecipitation; CGMCC, China General Microbiological Culture Collection Center.

Molecular cloning and genetic manipulation of bacterial strains.

The insertional inactivation mutants (by homologous, single-crossover methodology) and in-frame deletion (by homologous, double-crossover methodology) S. maltophilia mutants were constructed using the suicide vectors pK18mob and pK18mobsacB, respectively, as previously reported (17, 18). The primers used to amplify the relevant DNA sequences are listed in Table 2. Insertional inactivation mutants were selected on NYG plates containing kanamycin, and in-frame deletion mutants were screened on NYG plates containing 10% sucrose. To construct the point mutations, the Fast Mutagenesis System (Transgene, China) was used according to the product manual. Commonly used molecular cloning methods, such as PCR, ligation, and DNA restriction, were performed as instructed in Molecular Cloning: a Laboratory Manual (19).

TABLE 2.

Primers used in this study^a

Primer	Sequence (5′–3′)	Purpose
IFD4208	GAATTCTGCGGAAGCCCTGGTGCA	Construction of bfmK in-frame deletion mutants
	GGATCCGAGCCAGTCACTGGCACTGC
	GGATCCATCGCCGAGGGCGACCGA
	AAGCTTGATCAGGGCATGGTGCTTGG
C4208	AAGCTTATGCTGGCCCTGGCCAGC	For bfmK complementation
	GAATTCTCAGGTGAACTTTCCCAGCAG
P4208	CCATGGATCTGGCCCTGGCCAGCCTG	For BfmK expression
	AAGCTTGGTGAACTTTCCCAGCAGCAG
H237A	GCTGCAGTGGCGGCAGACCTGCGC	Construction of BfmK(H237A)
	TGCCGCCACTGCAGCCAGCATATG
IFD4209	GAATTCGGAGAATCGGACGCCAGC	Construction of bfmA in-frame deletion mutant
	GGATCCCAGCAGCACGTCCAGCGC
	GGATCCGGGCAGGCCTCACCCGGT
	AAGCTTTGCCCTGTGGTGGCGCAT
C4209	GGTACCATGACAACTCCCGCCCGT	For bfmA complementation
	AAGCTTTCACAGCACCTGTACCTCGGC
P4209	CATATGACAACTCCCGCCCGTGTC	For BfmA expression
	AAGCTTCAGCACCTGTACCTCGGCG
D55A	GACGTGATCATCCTCGCGTGGATGATG	Construction of BfmA(D55A)
	CGCGAGGATGATCACGTCCGGGCGATC
His4209	GGTACCATGACAACTCCCGCCCGT	Construction of the strain used for ChIP assays
	AAGCTTTCAGTGGTGGTGGTGGTGGTGCAGCACCTGTACCTCGGC
PR4209	CGGCAGGATGTGTTCGGGAG	For EMSA, PSmlt4209 probe
	GCAGGCAGTATCAGGCAGG
IFD0800	AAGCTTATGCGCAAGACGGGATGG	Construction of acoT in-frame deletion mutant
	GGATCCGGTTTTCGACTGCTGCCT
	GGATCCGCCTATCGCCAGCGCCTGG
	GAATTCTCAGAGCAGCGGCCTGAG
C0800	AAGCTTATGCGCAAGACGGGATGG	For acoT mutant complementation
	GAATTCTCAGAGCAGCGGCCTGAG
P0800	CGATGGCCACGCTGGTGCCT	For EMSA, PSmlt0800 probe
	TGGAGTCTCCTCATCGACAG
Q0570	GAAGTGGAGCTGGGCGAAAG	For RT-PCR, Smlt0570
	GGACGCATCGTCTGGATGTAC
Q0800	AAAAGCCGACCGTGGTAGTGA	For RT-PCR, Smlt0800
	GTTGGGCGGCACGTCAAT
Q0978	AACGCTGACCCGCGAAGT	For RT-PCR, Smlt0978
	TGGTCGTTGGCGAACACG
Q3568	GAGCTGCCGGACAAGATTGC	For RT-PCR, Smlt3568
	ATCCAGCACCAGATCGCCCACC
Q3949	ACCTGGGCAAACCATTCGA	For RT-PCR, Smlt3949
	TCCAGCAGCCGGTATTCG
Q4208	GCGGGTCGGTGTACTGAAGG	For RT-PCR, bmfK
	ATCGGCCAAGCGTCGTAG
Q4209	ATGACGACTCCCGCCCGT	For RT-PCR, bmfA
	CGACCTGCAGCAGCACGTC

^aChIP, chromatin immunoprecipitation; RT-PCR, reverse transcriptase PCR.

Phenotypic characterization.

The colony morphologies of the bacterial strains cultured on NYG plates for 24 h were observed. Bacterial growth under rich NYG medium and aerobic conditions was measured by an automated microbiology growth curve analysis system, Bioscreen C (Oy Growth Curves Ab, Ltd., USA). For the bacterial swimming motility assays, the strains were inoculated into semisolid 0.1% NYG agar with a toothpick and cultured for 12 h at 28°C, and the diameters of the swimming zones were measured. For biofilm quantification, the crystal violet staining method used was reported previously (17). S. maltophilia cultures were inoculated quantitatively into NYG medium within 96-well polystyrene plates, and the plates were incubated at 28°C for 5 h without shaking. Meanwhile, bacterial growth in 96-well plates was measured via the optical density at 600 nm (OD₆₀₀) readings of a microplate reader (Tecan Infinite 200 Pro). The wells in the plate were washed with water prior to being stained with crystal violet (1%) for 20 min before being washed. Next, the crystal violet stain was solubilized in absolute ethanol, and the biofilm quantities were measured at OD₅₉₀ on a microplate reader (n = 8).

RT-PCR and real-time quantitative PCR assays.

Bacterial total RNA, extracted using TRIzol (Invitrogen, USA), was quantified with a NanoDrop spectrophotometer (Thermo Fisher, USA). Next, the extracted RNA was purified with DNA-free DNase (Life Technologies) to remove any contaminating DNA. cDNA was synthesized from the purified RNA using random primers (Promega, USA) and Superscript III reverse transcriptase (RT) (Invitrogen, USA). S. maltophilia DNA templates were included as positive controls for the PCRs, and transfer-messenger RNA (tmRNA) amplification was used as the loading control for the RT-PCRs; samples that lacked reverse transcriptase during cDNA synthesis were included as negative controls to evaluate potential DNA contamination. The primers used to amplify sample genes are listed in Table 2.

Protein expression, purification, and phosphorylation assays.

Recombinant BfmA-His₆ and BfmK-His₆ proteins were expressed in E. coli BL21(DE3) cells using pET28a and pET30a (Novagen) expression vectors, respectively. The PCR primers used to construct the recombinant vector are listed in Table 2. BfmA-His₆ was purified by Ni-nitrilotriacetic acid (NTA) affinity chromatography according to the manufacturer's manual (Novagen). Inverted membrane vesicles of the BfmK HK were prepared according to the method used in our previous study (20). The purified protein and inverted membrane vesicles were stored in storage buffer (50 mM Tris-HCl, 0.5 mM EDTA, 50 mM NaCl, 5% glycerol; pH 8.0).

We performed in vitro phosphorylation assays according to the method used previously (20). Briefly, for the autophosphorylation assays, BfmK inverted membrane vesicles were incubated with 100 μM ATP containing 10 μCi [γ-³²P]ATP (PerkinElmer) in a 20-μl volume of autophosphorylation buffer (50 mM Tris-HCl [pH 7.8], 2 mM dithiothreitol [DTT], 25 mM NaCl, 25 mM KCl, 5 mM MgCl₂) for 10 min (28°C). To detect phosphotransferase between BfmK and BfmA, 20 μM purified, soluble BfmA was added to the reaction mixture, after which it was incubated at 28°C for the times indicated in Fig. 4. The reactions were stopped by 6× sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer. The phosphorylated proteins were then separated on a 12% SDS-PAGE gel. After electrophoresis, the gels were exposed to a phosphorimaging screen for 1 h, and the radioactive signals from the gels were recorded by a PhosphorImager system (Amersham Biosciences).

FIG 4.

Operon organization of the Smlt4208-Smlt4209 locus and its role in biofilm development. (A) Genomic localization of the Smlt4204-Smlt4209 locus. Arrows indicate genes and their transcriptional directions. Gene names are listed, and the primers (P1 to P12) used for RT-PCR are indicated around the arrows. (B) Dissection of the operon organization of the Smlt4204-Smlt4209 locus by RT-PCR. cDNA was reverse transcribed with random primers using total RNA from S. maltophilia ATCC 13637 grown overnight in NYG medium at 28°C. RT denotes PCR amplification using cDNA transcribed from RNA as the template; DNA denotes the positive control, using bacterial total DNA as the PCR template; −RT denotes the negative control, for which reverse transcriptase was absent during cDNA synthesis. (C) In vitro phosphorylation assay of the BfmA-BfmK system. BfmK inverted membrane vesicles (50 μg) and soluble BfmA protein (20 μM) were used. [γ-³²P]ATP was added to the reaction mix for the times indicated. The reactions were stopped with SDS loading buffer, and the samples were separated by 12% SDS-PAGE. Int., band intensity that was estimated by Quantity One software. (D) bfmK (Smlt4208) and bfmA (Smlt4209) control biofilm formation in S. maltophilia ATCC 13637. Bacterial biofilms were quantified by OD₅₉₀ measurements using the crystal violet staining method. Bars represent standard deviations (n = 8). *, P < 0.01, calculated by Student's t test. WT, wild-type strain; IFD-bfmK and IFD-bfmA, bfmK and bfmA in-frame deletion mutants, respectively; IFD-bfmK-bfmK and IFD-bfmA-bfmA, complementary strains. All of the strains, including the wild-type strain, contained a corresponding blank or recombinant pBBRMCS2 vector. (E) Growth curves of the bacterial strains in rich NYG medium. Bars represent standard deviations (n = 3).

ChIP.

The protocol used for chromatin immunoprecipitation (ChIP) followed that of a previous study (20). In brief, S. maltophilia was grown in NYG medium until the OD₆₀₀ was 0.4. The cells were cross-linked with 1% formaldehyde and subsequently quenched by 0.5 M glycine for 5 min. Bacterial cultures (4 ml) harvested by centrifugation were washed twice in 10 ml of cold Tris-buffered saline (TBS) buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl) and then resuspended in 1 ml of lysis buffer (10 mM Tris [pH 8.0], 20% sucrose, 50 mM NaCl, 10 mM EDTA, 10 mg/ml lysozyme). IP buffer (50 mM HEPES-KOH [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 1 mM phenylmethylsulfonyl fluoride) was added to the bacterial cell suspension, and the cells were sonicated with a Diagenode Bioruptor UCD-300 sonicator (Diagenode, USA) to generate DNA fragments of about 200 bp (on average). After centrifugation, the solution was precleared with 20 μl of protein A at 4°C for 10 min on a slow rotator, and a 100-μl aliquot was kept as the loading control DNA (input sample). For the ChIP assays, 20 μl of protein A-Sepharose (50% slurry) and 2 μl of an anti-His₆ antibody were added to an 800-μl aliquot, and the mixture was incubated at 4°C overnight with slow rotation. The next day, the beads were collected by centrifugation and washed with IP buffer and wash buffer (10 mM Tris-HCl [pH 8.0], 250 mM LiCl, 1 mM EDTA, 0.5% Nonidet P-40 [equivalent to Triton X-114], 0.5% sodium deoxycholate). The immunoprecipitated chromatin was removed from the beads by adding 100 μl of elution buffer (50 mM Tris [pH 7.5], 10 mM EDTA, 1% SDS), and the solution was incubated for 10 min at 65°C. RNase A was added to remove RNA contamination. Proteinase K was used for cross-linking reversal. The DNA was purified using a PCR purification kit (Qiagen, USA). For semiquantitative PCR, 1 μl of the eluted DNA and 1 μl of the control DNA were used as the templates for PCR with the appropriate primers. The quantities of the captured DNAs were normalized to the control input DNA.

EMSA.

For the electrophoresis mobility shift assay (EMSA), promoter regions of the bfmA and acoT operons were PCR amplified using the primers shown in Table 2. PCR products were end labeled with [γ-³²P]ATP using T4 polynucleotide kinase (New England Biolabs [NEB], USA). BfmA proteins (1 μg) and probes (4 fmol) were incubated in reaction buffer [10 mM Tris (pH 7.0), 50 mM KCl, 1 mM DTT (pH 7.5), 2.5% glycerol, 5 μl MgCl₂, 50 ng/μl poly(dI·dC), 0.05% NP-40, and 10 mM EDTA] for 20 min. DNA loading buffer (0.25% bromophenol blue, 40% sucrose) was added to stop the reaction, and the samples were separated by 8% native PAGE. Phosphorimaging screens were used to detect the radioactive signals. Different concentrations of unlabeled DNA probes were used as competitors.

RESULTS

S. maltophilia encodes multiple histidine kinases with diverse domain compositions.

Because the complete genome of S. maltophilia ATCC 13637 has not been sequenced, the S. maltophilia K279a genome was used in this study as a reference genome (12). According to the prokaryotic two-component systems database (21), this genome contains 60 HK genes, and 16 of them are hybrid HK types that each contain an additional receiver domain as a phosphorylation acceptor. Independent database searches also revealed that Smlt2710 is an HWE HK (an HK group with signature His and Trp-Glu residues within the presumed N and G1 boxes, respectively) (22). Furthermore, Smlt4131 contains an ATPase domain with conserved N, G1, and G2 boxes but lacks the typical H box of an HK, suggesting that it is a degenerated HK. Smlt2710 and Smlt4131 were also included in the analysis, so a total of 62 putative HK genes of S. maltophilia were examined in the present study (Fig. 1).

FIG 1.

Putative secondary structures of all annotated sensor histidine kinases of S. maltophilia K279a. Protein domains were predicted by searching the Pfam database (E value, ≤1.0). Domain names are the same as those in the Pfam database: HK (HisKA domain), HA (HATPase C domain), REC (response regulator domain), Hpt (Hpt domain), PAS (PAS fold), CHASE (CHASE domain), GAF (GAF domain), RP (two-component regulator propeller), YYY (Y_Y_Y domain), CheW (CheW-like domain, PF01584), HAMP (HAMP domain), SBP (SBP domain), KDPD (KdpD domain), and 2CSK (2CSK domain). Gray vertical rectangular bars on the N-terminal regions of each protein represent predicted transmembrane helices.

A comparison with the HK genes of Xanthomonas species members (23), which are close relatives of S. maltophilia, revealed that all of the orthologs of virulence-associated HKs of Xanthomonas spp., including RpfC (Smlt2234), VgrS (or ColS, Smlt3765), PhoQ (Smlt0278), RavS (Smlt2324), RavA (Smlt2322), RaxH (Smlt3567), and HpaS (Smlt4106), were also present in S. maltophilia K279a (Fig. 1). Because S. maltophilia is not a plant pathogen, these virulence-associated HKs might exhibit subtle functional differentiation in signal detection between the two bacterial taxa (Stenotrophomonas and Xanthomonas). The majority of these HK genes (56 genes) contain at least an RR gene in close vicinity in the genome. The exceptions are Smlt0107, Smlt0570, Smlt3019, Smlt3625, Smlt3730, and Smlt4540; these six HKs are orphans, whose cognate RRs might be absent or carried on other genomic loci.

Systematic insertional inactivation of HK genes and phenotypic characterization of mutants.

To obtain a mutational library of HK genes in S. maltophilia ATCC 13637, we initially aimed to construct recombinant suicide vectors (pK18mob) containing truncated DNA sequences of HK genes as inserts (300 to 450 bp). Among the 62 genes predicted to be HKs in the S. maltophilia K279a genome, 60 recombinant vectors were obtained (see Fig. S1 in the supplemental material). Two genes, Smlt1219 and Smlt4540, failed to be PCR amplified by the various primers, suggesting that they are absent from the S. maltophilia ATCC 13637 genome. Starting from these suicide vectors, a total of 51 insertional mutants were obtained, and they were verified by PCR (Table 1). Despite great efforts, 11 genes (Smlt1219, Smlt1437, Smlt1541, Smlt2322 [ravA], Smlt2382, Smlt2646, Smlt3765 [vgrS], Smlt3948, Smlt4131, Smlt4540, and Smlt4627) proved to be quite difficult to knock out. These genes might be essential in S. maltophilia ATCC 13637, or their inactivation might have led to serious growth deficiencies under the culture condition used in this study.

The morphologies of the colonies and the swimming motilities of the HK gene mutants were observed on rich NYG plates (1.5% agar) and semisolid NYG plates (0.1% agar), respectively (Fig. 2), and the growth curves in the NYG medium of all the mutants were measured (see Fig. S2 in the supplemental material). The mutants with mutations in five genes, Smlt0769, Smlt0882, Smlt0977, Smlt1636, and Smlt1785, grew substantially slower than the wild-type strain or the other mutants (see Fig. S2). Smlt0882 is the ortholog of RcsC in enterobacteria. The RcsC HK modulates bacterial capsule synthesis, and mutation of it usually impacts the cell envelope stress response (24). Orthologs of Smlt1636 in other bacteria have not been investigated. However, inactivation of the two genes (Smlt0882 and Smlt1636) also decreased the swimming motilities of the bacteria (Fig. 2C), which might be caused by their low growth rates. Figure 2C and Table S1 in the supplemental material show, besides these two gene mutants, seven other mutants that also exhibited decreased swimming abilities; they were Smlt1785, Smlt2260 (cheA), Smlt2267 (cheA2), Smlt2324 (ravS), Smlt2380, Smlt2710, and Smlt3944 mutants. Among the gene products, Smlt2260 and Smlt2267 are orthologs of the biological sensors CheA and CheA2 of E. coli, respectively, suggesting that they are involved in chemotaxis regulation. Smlt2324 (RavS), Smlt2380, Smlt2710, and Smlt3944 are PAS domain-containing HKs that are responsible for detecting oxygen, redox potential, or light (25), thereby indicating that these signals are critical to the motility of S. maltophilia.

FIG 2.

Colony morphology and swimming motility of insertional mutants with mutations in genes encoding histidine kinases of S. maltophilia ATCC 13637. (A) Positions of the mutants on the plates. The order of the columns and rows are indicated numerically and alphabetically, respectively. (B) Morphologies of the bacterial strains grown on a rich NYG plate. A total of 1 μl of each bacterial culture was inoculated on the plate, and the bacteria were grown for 24 h at 28°C. (C) Swimming motility of the bacterial strains grown on a semisolid NYG plate containing 0.1% agar. After inoculation by toothpick stabbing, the strains were grown for 12 h at 28°C. The diameters of bacterial swimming zones were measured and are listed in Table S1 in the supplemental material.

Systematic screening for biofilm-deficient mutants of S. maltophilia.

To investigate the biofilm formation capabilities of the HK gene mutants, bacterial cells from cultures grown overnight in rich NYG medium were collected and washed, and their individual concentrations were each adjusted to an OD₆₀₀ of 0.4. Equal volumes of bacterial cultures (20 μl) were inoculated into 96-well plates containing 180 μl of fresh NYG medium in each well. The plates were then incubated for 5 h at 28°C, and the biofilm quantities were calculated using the traditional crystal violet staining method (Fig. 3A). Meanwhile, the bacterial growth in the wells was estimated by determining the OD₆₀₀ absorbance values of each culture (Fig. 3B). To estimate the biofilm formation relative to bacterial growth, the ratio of biofilm quantity versus planktonic bacteria (OD₆₀₀) was calculated (Fig. 3C). It showed that the biofilm-to-growth ratios of four mutants (the ΔSmlt0596, ΔSmlt0884, ΔSmlt3625, and ΔSmlt4477 mutants) were not altered significantly from that of the wild-type strain. The biofilm-to-growth ratios of nine mutants (the ΔSmlt0158, ΔSmlt0595, ΔSmlt0882, ΔSmlt1421, ΔSmlt1636, ΔSmlt1785, ΔSmlt2380, ΔSmlt2710, and ΔSmlt3944 mutants) were significantly increased, whereas the other 38 mutants, including the ΔSmlt0278 (phoQ), ΔSmlt2324 (ravS), and ΔSmlt4106 (hpaS) mutants, exhibited decreased biofilm-to-growth ratios. Of these, the Smlt2324 (ravS) mutation led to a decrease in biofilm that was also observed in our previous study using random, large-scale transposon mutagenesis (17). It is interesting to note that, although insertional inactivation of the rpfC ortholog (Smlt2234) resulted in a significant decrease in the biofilm-to-growth ratio (Fig. 3C), its biofilm quantity regardless of growth was not decreased (Fig. 3A). This is quite different from that of the Xanthomonas species member whose rpfC gene mutation has been shown to cause a conspicuous deficiency in cell-cell communication and biofilm development (26).

FIG 3.

Biofilm development of histidine kinase gene mutants of S. maltophilia ATCC 13637. (A) Quantification of biofilm levels in the bacterial strains. Biofilms were quantified by a crystal violet staining methodology. (B) Bacterial strain growth. The OD₆₀₀ absorbance (ABS) of the bacterial cultures in 96-well plates was measured before biofilm quantification. (C) Ratios of biofilm quantity versus bacterial growth (OD₆₀₀). The error bars represent standard deviations (n = 8). *, statistically significantly different from the value for the wild-type strain (P ≤ 0.05, calculated by Student's t test).

Smlt4209-Smlt4208 is a TCS that controls biofilm formation.

We next focused on further investigation of Smlt4208, because its mutant was stably impaired in biofilm formation and not in growth. Furthermore, no other orthologs of this HK gene have been investigated before. As Fig. 4A shows, Smlt4208 is located in a genomic locus that contains an RR gene (Smlt4209), a gene that encodes a putative MarR family transcription factor (Smlt4207), and several genes with unknown functions. The operon predictions that we conducted using different algorithms are controversial: ProOpDB predicted that Smlt4205-Smlt4209 forms an operon containing five genes, from Smlt4205 to Smlt4209 (27), whereas both theoretical prediction by the DOOR database and the MicrobesOnline operon predictions tool proposed an operon that contains only Smlt4208 and Smlt4209 (28, 29). To experimentally decipher the potential operon structure of this locus, a series of primers was designed, and RT-PCR was used to check for intergenic transcription between these genes. The RT-PCR analysis revealed that Smlt4208 and Smlt4209 do indeed constitute a bicistronic operon after we observed amplification of a PCR product between the two genes, which suggests that they form a transcription unit and form an intact TCS (Fig. 4B). However, this analysis also revealed that the two genes were not cotranscribed with other nearby genes in the genome. This experimental result thus supports the theoretical predictions made by the DOOR database and the MicrobesOnline operon predictions tool. Consequently, we named Smlt4209 and Smlt4208 bfmA (biofilm activator) and bfmK (biofilm activator cognate kinase), respectively.

To investigate the biochemical properties of BfmA and BfmK in protein phosphorylation, the prokaryotic expression vectors pET30a and pET28a were employed to express full-length BfmA and BfmK proteins, respectively, using E. coli BL21(DE3) cells as the host. In addition, point mutations were also created in the genetic codes corresponding to the phosphorylation sites of BfmA and BfmK proteins, the modified genes of which were used to express recombinant BfmA(D55A) and BfmK(H237A) proteins, respectively. Soluble BfmA protein (purity, >98%), the inverted membrane vesicles of BfmK (BfmK proportion, >20%), and their recombinant mutants were expressed and successfully purified by Ni-NTA affinity chromatography (Fig. 4C). The in vitro phosphorylation assay conducted on the recombinant proteins revealed that the full-length BfmK inverted membrane vesicle was able to autophosphorylate its conserved His-237 site, whereas its recombinant form, BfmK(H237A), failed to be phosphorylated (Fig. 4C). This result demonstrates that BfmK is a bona fide HK with autokinase activity. When BfmA was added to the reaction mixture, although no band representing the phosphorylated BfmA (BfmA-P) was observed, the phosphorylation level of BfmK-P decreased to about 80% of the control's. Addition of BfmA(D55A) protein did not decrease the level of BfmK-P (Fig. 4C). This result suggests that BfmK can phosphorylate its cognate BfmA, but BfmA-P is probably unstable, with a half-life too short to be detected.

Because bfmA and bfmK are organized in an operon, insertional inactivation of bfmK (ΔSmlt4208) may have a polar effect in terms of influencing transcription of the whole operon. To avoid this defect, in-frame bfmA and bfmK deletion mutants (IFD-bfmA and IFD-bfmK, respectively) were constructed by the homologous, double-crossover methodology using the suicide vector pK18mobsacB. Also, the full-length bfmA and bfmK sequences were PCR amplified and inserted into the multiple-cloning site of the broad-host-range vector pBBRMCS2 to construct two recombinant vectors. The recombinant vectors were then transformed accordingly into the in-frame bfmA and bfmK deletion mutants to genetically complement the mutation (strains IFD-bfmA-bfmA and IFD-bfmK-bfmK, respectively). In these two constructs, the vector-borne bfmA and bfmK mutants were under the control of a PlacZ promoter. The biofilm formation abilities of these strains were then tested. As Fig. 4D shows, the biofilm quantities of the IFD-bfmA and IFD-bfmK mutants decreased significantly to 61% and 53% of the levels of the wild-type strain, respectively, whereas genetic complementation almost completely suppressed the effects caused by the gene deletion. The bacterial growth of these strains was also determined (Fig. 4E), and the results show that the in-frame bfmA and bfmK deletion mutants and the complementary IFD-bfmK-bfmK strain grew the same as the wild-type strain. The complementary IFD-bfmA-bfmA strain grew more slowly than the other strains, which might be caused by overexpression of bfmA, since the Plac promoter driving its transcription had high activity under the tested growth condition (20). The result again excludes the possibility that decreased biofilm formation in the mutants was caused by slow bacterial growth. In addition, the flagella, activities of extracellular proteases and amylases, and susceptibilities to rifamycin, streptomycin, and tetracycline were not remarkably changed for bfmA and bfmK mutants (data not shown). Collectively, the above analyses show that the TCS BfmA-BfmK controls biofilm development in S. maltophilia.

Screen of downstream genes regulated by BfmA using ChIP.

The regulon of the BfmA-BfmK TCS has not been investigated previously. We used the STRING database (30) to predict the possible functional partners of BfmA and BfmK. Judging by the information available from the genomic neighborhood, or from co-occurrence or gene fusion events in different species, BfmK and BfmA each potentially interact with 10 proteins, the scores of which are high (scores > 0.850) (see Fig. S3A and B in the supplemental material), with three of them being other HKs (Smlt0977 [PhoR], RpfC [Smlt2234], and SmeS [Smlt4477]). Six pairs of these proteins, namely, PhoB-PhoR, Smlt1218-Smlt1219, Smlt1540-Smlt1541, Smlt3944-Smlt3949, SmeS-SmeR, and BfmA-BfmK, are themselves clustered in the genome, suggesting a strong functional correlation with the TCS BfmA-BfmK.

Because BfmA is a typical OmpR family RR with a C-terminal transcription factor (TF) region (151 to 228 amino acids [aa]) as the output domain, we proposed that some of these related protein coding genes are subject to direct regulation by BfmA at the transcriptional level. Therefore, 15 operon structures of the coding genes of these interacting proteins were predicted by the DOOR database, and PCR primers were designed according to the predicted promoter regions of these operons (see Fig. S3C in the supplemental material). ChIP, together with semiquantitative PCR, was then used to screen for possible DNA binding to the TF BfmA. To perform this experiment, we constructed another bfmA complementary strain similar to the IFD-bfmA-bfmA strain but with a His₆ epitope tag coding sequence added to the 5′ region of the stop codon of the bfmA sequence. This DNA insert, cloned into the pBBRMCS2 vector, was used to complement the bfmA mutant (strain IFD-bfmA-bfmA*). Using a monoclonal antibody against BfmA-His₆, ChIP was conducted based on this recombinant strain (IFD-bfmA-bfmA*). As Fig. 5A shows, the promoter regions of PSmlt1216, PSmlt1423, PSmlt3944, and PSmlt4477 failed to be amplified by PCR even when the input DNA samples were used as DNA loading controls, thereby suggesting that substantial genetic polymorphisms were present in these cis-regulatory sequences. Additionally, PSmlt1436, PSmlt1540, PSmlt2233, PSmlt2700, and PSmlt2801 PCR products were absent in the samples, suggesting that these promoters did not bind BfmA. The above nine promoter regions were then excluded from the following analyses. However, compared with the background control that used coimmunoprecipitated DNA from the His₆ tag-lacking IFD-bfmA-bfmA control strain for the PCR amplifications, the amounts of the PSmlt0570, PSmlt0800, PSmlt0978, PSmlt3568, PSmlt3949, and PSmlt4209 PCR products from the IFD-bfmA-bfmA* ChIP samples increased substantially (Fig. 5A), suggesting that BfmA-His₆ binds to these promoter regions directly in vivo and that the corresponding genes or operons are under the control of BfmA.

FIG 5.

BfmA regulates transcriptions of its own operon and Smlt0800. (A) Screen to identify BfmA binding sequences using chromatin immunoprecipitation (ChIP). Bacterial strains were cultured in NYG medium to an OD₆₀₀ of 0.4 each, and ChIP was conducted to collect BfmA-bound double-stranded DNA. Semiquantitative PCR was used to quantify the amount of DNA coimmunoprecipitated with BfmA. No, ChIP DNA sample from untagged strain (IFD-bfmA-bfmA); His₆, ChIP DNA sample from His₆-tagged strain (IFD-bfmA-bfmA*); Sample, PCR amplification using ChIP DNA; Input, PCR amplification using bacterial DNA extracted from the cell lysate before ChIP and subsequently used as a loading control. (B) Semiquantitative RT-PCR analysis of transcription levels of candidate genes. cDNA was reverse transcribed using total RNA from S. maltophilia ATCC 13637. −RT denotes the negative control, and reverse transcriptase was absent during cDNA synthesis. All of the bacterial strains tested contained a pBBRMCS2 vector for comparability. The experiment was repeated three times independently. (C) Transcription levels of the genes regulated by BfmA, as quantified by quantitative real-time PCR. cDNA amplified from tmRNA was used as an internal standard. Each assay was conducted using biological and technical triplicates. The results from one representative experiment are shown. Vertical bars represent standard deviations. *, statistically significant differences (P < 0.05), as calculated by Student's t tests.

bfmA positively regulates transcription of its own operon and Smlt0800.

To evaluate the regulatory effect of BfmA on the candidate genes downstream of it, the six operons whose promoter regions could bind to BfmA were selected, and the mRNA amounts of the wild-type strain and the IFD-bfmA mutant were compared by semiquantitative RT-PCR. As Fig. 5B shows, no positive amplification bands for the representative genes, Smlt3949 and Smlt3568, were observed, suggesting that their transcription levels were too low to be detected under the growth conditions used in this study. The quantity of mRNA from Smlt0570 and Smlt0978 remained stable in the strains tested (Fig. 5B), and quantitative real-time PCR also showed that their mRNA levels were similar between the wild-type strain and the bfmA mutant (see Fig. S4 in the supplemental material). These results suggested that the transcription of the two genes is disassociated from the control of BfmA, despite the existence of TF-promoter binding events. The transcription levels of Smlt0800 and bfmK were substantially lower in the IFD-bfmA mutant, and genetic complementation of bfmA restored their transcription levels to near or above the wild-type level, suggesting that BfmA acts as a positive regulator to modulate their expression.

To verify the above result, quantitative real-time PCR was employed to compare the mRNA quantities of Smlt0800 and bfmK. The result obtained is in good agreement with the result obtained for the semiquantitative RT-PCR assay (Fig. 5C); the in-frame deletion of bfmA caused 60% and 96% decreases in the mRNA quantities of Smlt0800 and bfmK, respectively, and these decreases were almost completely suppressed by genetic complementation of bfmA in the mutant (Fig. 5C). Taken together, the results show that bfmA positively autoregulates the transcription of its own operon. Furthermore, BfmA is also a positive regulator that controls the transcription of Smlt0800, which encodes an acyl-CoA thioesterase (named AcoT) that takes part in the hydrolysis of acyl-CoA.

BfmA binds the promoter regions of acoT to regulate its expression and biofilm.

To verify the in vivo ChIP-PCR results and determine whether BfmA is a TF with double-stranded DNA binding activity, 200-bp sequences from the promoter regions of the bfmA-bfmK operon and acoT were PCR amplified, end labeled with [γ-³²P]ATP by T4 polynucleotide kinase, and used as probes for in vitro EMSAs. As Fig. 6A and B show, the BfmA recombinant protein formed stable protein-DNA complexes with the labeled DNA probes of PbfmA and PSmlt0800. The binding events between BfmA and the two probes were specific, because addition of increasing concentrations of unlabeled DNA probes gradually competed with the ³²P-labeled probes, thereby resulting in signal fade or complete disappearance of the isotopic signals. Collectively, the in vitro and in vivo evidence supports the assertion that BfmA binds directly to the promoter region of acoT and the bfmA-bfmK operon and positively controls their transcription.

FIG 6.

BfmA directly binds to the promoter region of bfmA and acoT and regulates biofilm formation. EMSAs reveal that BfmA binds directly to the promoter region of bfmA itself (A) and acoT (B). A total of 4 fmol of [γ-³²P]ATP-labeled DNA sequences corresponding to the bfmA and acoT promoter regions was used in the assay. Increasing amounts of unlabeled probes were used as competitors (10× to 1,000×) for binding to BfmA protein. The black triangle on the right side of each panel represents the BfmA-DNA binding complex. (C) bfmA is epistatic to acoT, which contributes to biofilm formation. Biofilm quantity was measured in different bacterial strains. IFD-acoT, acoT in-frame deletion mutant; IFD-acoT-acoT and IFD-bfmA-acoT, acoT and bfmA mutants containing a full-length acoT gene subjected to the control of the Plac promoter. All of the strains, including the wild-type strain, contained a corresponding blank or recombinant pBBRMCS2 vector. Bars represent standard deviations. *, significant difference (P < 0.05, by Student's t test; n = 8).

To evaluate whether BfmA-regulated acoT expression is associated with biofilm, an in-frame deletion mutant of acoT was constructed (IFD-acoT). As Fig. 6C shows, acoT deletion caused a remarkable decrease (38%) in bacterial cell aggregation. Although genetic complementation (strain IFD-acoT-acoT) by a full-length acoT gene did not restore the deficiency to the wild-type level, it significantly increased the bacterial biofilm formation compared to that of the mutant. Furthermore, when acoT was overexpressed in the bfmA mutant (strain IFD-bfmA-acoT), the biofilm deficiency caused by the bfmA mutation was significantly suppressed, albeit not completely restored. These genetic and epistatic analyses suggest that acoT is also involved in biofilm development and that the decrease of the acoT transcription level in the bfmA mutant is one of the causes of the biofilm deficiency.

DISCUSSION

HKs are the dominant cellular sensors used by prokaryotes to detect environmental stimuli; hence, they are proposed to be candidate molecular targets for developing novel antibiotic chemicals (7, 11). Based on the bioinformatics predictions of the characteristics of HK genes (62 genes) (Fig. 1), the present study successfully constructed 51 HK gene mutants of the nosocomial pathogen S. maltophilia and identified a number of mutants with phenotypic deficiencies in the following: colony morphology, bacterial swimming on semisolid media, and biofilm formation on the surface of polystyrene (Fig. 2 and 3). Among these HK genes, molecular analyses revealed that a TCS, BfmA-BfmK (Smlt4209-Smlt4208), controls biofilm development (Fig. 4D). Based on the biochemical evidence collected in this study, BfmK is an HK with autokinase and phosphotransferase activities (Fig. 4C), and BfmA is a transcription factor capable of binding double-stranded DNA in vitro and in vivo (Fig. 5 and 6). After prediction of the possible functional partners of the BfmA-BfmK system by STRING database searches, subsequent ChIP-PCR screening identified six promoters that were bound by BfmA. Quantitative real-time PCR and EMSA showed that BfmA binds directly to the 5′ cis-regulatory sequences of the bfmA-bfmK operon and an acyl-CoA thioesterase-coding gene (acoT, Smlt0800) and positively modulates their transcription. The results suggest that the BfmA-BfmK TCS has autoregulation and controls the mRNA level of acoT, which was shown to take part in biofilm formation (Fig. 6C). To the best of our knowledge, this is the first study to systematically investigate the biological function of the TCS in S. maltophilia. It will facilitate our future studies aimed at combating the fatal infections caused by this bacterial pathogen.

With a genome size of approximately 5.0 Mb, S. maltophilia contains a large number of HK genes (62 genes), including 44 orthodox HK genes, 16 hybrid HKs, an HWE family HK, and an atypical HK that lacks a representative H box. The number of S. maltophilia HKs is larger than the number of HKs of its close relatives with similar genome sizes. For example, the plant pathogens Xanthomonas spp. generally contain approximately 38 to 60 HK genes (23). In addition, as Fig. 1 indicates, the signal sensor domains encoded by S. maltophilia are diverse and include PAS, SBP, KdPD, 2CSK, GAF, and CHASE domains. This suggests that S. maltophilia potentially possesses strong abilities to survive a wide range of environmental challenges. By systematic inactivation of the HK genes in S. maltophilia and by characterizing the resulting phenotypic alterations, we found that the HK genes of this bacterium differ substantially from their orthologs. For example, the HK RpfC and its cognate RR, RpfG, were shown to be a major TCS that controls biofilm formation, extracellular polysaccharide and enzyme production, and virulence against the host plant of Xanthomonas spp. (31–33). Interestingly, insertional inactivation of the rpfC ortholog, Smlt2234, did not significantly impact biofilm development regardless of bacterial growth (Fig. 3A), which is quite different from that of Xanthomonas spp. Protein sequence alignments indicated that the two HKs are conserved (identity, 65%). However, according to its genomic annotation, Smlt2234 (883 aa) has a longer signal input region, with nine putative transmembrane helices, than RpfC (726 aa), which contains five predicted transmembrane helices. It remains unknown whether this difference is caused by genomic annotation or represents the real protein length. As for another HK, RavS, in Xanthomonas campestris pv. campestris, the causative agent of black rot disease in cruciferous plants, inactivation of ravS resulted in an increase in bacterial swimming (34). However, as our previous study and Fig. 2C show (17), inactivation of the ravS ortholog Smlt2324 caused a remarkable decrease in swimming motility, with the mutant exhibiting a phenotypic change completely opposite of that of the wild-type bacterium. The reason behind this discrepancy is unknown. In addition, in X. campestris pv. campestris and Xanthomonas oryzae pv. oryzae, both of which cause rice bacterial blight disease, although the vgrS (or colS) mutant grew very slowly, this gene was shown to be mutable (18, 35, 36). However, our repeated efforts to inactivate the vgrS ortholog (Smlt3765) failed, suggesting that vgrS might be an essential gene in S. maltophilia or that it has a much greater impact on bacterial growth. Taken together, the results of our study suggest that the functional differences in HK regulation between S. maltophilia and Xanthomonas spp. make them ideal experimental models for investigating the evolution of signaling cascades in close-relative organisms.

The newly identified TCS, BfmA-BfmK, is a regulatory system used by S. maltophilia for biofilm manipulation. BfmK is an orthodox HK with C-terminal DHp and CA domains. Its N-terminal input region has an unknown function, while its 89-aa-length periplasmic sensor is surrounded by two putative transmembrane helices. A HAMP linker is present between the N and C termini of the HK (Fig. 1). Therefore, a specific environmental stimulus detected by the BfmK HK is hardly predictable by the secondary structure. As Fig. 4C shows, BfmK autophosphorylates its own conserved histidine residue, and the BfmK phosphorylation level was decreased in the presence of BfmA, suggesting that BfmK can transfer the phosphoryl group to BfmA and that the half-life of the phosphorylated BfmA is probably very short (Fig. 4C); hence, it is a biochemically verified TCS. BfmK and BfmA orthologs are distributed widely in proteobacteria, but their functions have not been studied before. Based on the theoretical predictions and subsequent molecular analyses, we identified two downstream operons that were directly controlled by BfmA-promoter binding (Fig. 5 and 6). Autoregulation of bfmA-bfmK was verified, suggesting that a positive regulatory feedback of bfmA-bfmK transcription exists to amplify the efficiency of this TCS, as found in other TCSs (37, 38). Furthermore, BfmA is likely to control the cellular concentration of CoA by positively modulating the transcription of acoT and is thus involved in degradation of acyl-CoA into free fatty acid and CoA, suggesting that the TCS is involved in multiple cellular metabolisms, especially the β-oxidation of fatty acid (39). Future studies on the phosphorylation process of BfmK-BfmA, signal detection, and the transcriptional regulation of BfmA are needed to understand the role of this TCS in regulating bacterial physiology.