An Inducible Fusaric Acid Tripartite Efflux Pump Contributes to the Fusaric Acid Resistance in Stenotrophomonas maltophilia

Background Fusaric acid (5-butylpicolinic acid), a mycotoxin, is noxious to some microorganisms. Stenotrophomonas maltophilia displays an intrinsic resistance to fusaric acid. This study aims to elucidate the mechanism responsible for the intrinsic fusaric acid resistance in S. maltophilia. Methodology A putative fusaric acid resistance-involved regulon fuaR-fuaABC was identified by the survey of the whole genome sequence of S. maltophilia K279a. The fuaABC operon was verified by reverse transcriptase-PCR. The contribution of the fuaABC operon to the antimicrobial resistance was evaluated by comparing the antimicrobials susceptibility between the wild-type strain and fuaABC knock-out mutant. The regulatory role of fuaR in the expression of the fuaABC operon was assessed by promoter transcription fusion assay. Results The fuaABC operon was inducibly expressed by fusaric acid and the inducibility was fuaR dependent. FuaR functioned as a repressor of the fuaABC operon in absence of a fusaric acid inducer and as an activator in its presence. Overexpression of the fuaABC operon contributed to the fusaric acid resistance. Significance A novel tripartite fusaric acid efflux pump, FuaABC, was identified in this study. Distinct from the formally classification, the FuaABC may constitute a new type of subfamily of the tripartite efflux pump.


Introduction
Fusarium is a large genus of filamentous fungi widely distributed in soil. The genus includes a number of economically important plant pathogenic species, such as Fusarium graminearum and Fusarium oxysporum f.sp. cubense. Fusaric acid (5-butylpicolinic acid), a mycotoxin produced by several Fusarium species [1], is firstly known to decrease plant cell variability [2,3]. Fusaric acid is reported to be toxic to some microorganisms such as Pseudomonas fluorescens and mycobacteria [4], but is not generally used as an antimicrobial in clinic. Little is known about fusaric acid resistance in microorganisms. A fusaric acid-resistance associated operon has been reported from Burkholderia cepacia [5].
Stenotrophomonas maltophilia, an aerobic, nonfermentative, Gramnegative bacterium, is ubiquitous in nature, including soil, water, plants, and animals [6]. It is a member of endophytic bacteria, which can be isolated from plant rhizosphere, roots, and stems. S. maltophilia can generate antifungal organic compounds [7], plant growth factors [8], hydrolytic enzymes [9], and has been used for microorganism-based biological control in agriculture. Owing to the same habitats of endophytic S. maltophilia and Fusarium, the fusaric acid produced by Fusarium is a challenge for the survival of S. maltophilia.
In addition to its role in biocontrol, S. maltophilia is also an opportunistic human pathogen, causing nosocomial infection and community acquired [10]. Treatment of S. maltophilia infection is difficult since this pathogen is characterized by intrinsic and acquired resistance to a variety of antibiotics. The known mechanisms to combat the antimicrobial compounds in S. maltophilia include antibiotic hydrolysis or modification, target genes modification, membrane permeability alteration, and efflux pump overexpression [11]. Among them, the efflux pump is a crucial mechanism to remove a variety of toxic compounds and help bacteria to escape the attacks either from medical treatment or from natural environmental compounds. According to the structural characteristics, the efflux pumps are divided into five families: the resistance-nodulation-cell division (RND), the major facilitator superfamily (MFS), the small multidrug resistance (SMR), the multidrug and toxic compound extrusion (MATE), and the ATP-binding cassette (ABC) [12][13][14]. In Gram-negative bacteria, the RND, MFS, and ABC pumps may form a tripartite system to extrude the substance directly from cytoplasm to extracelluar environment. Tripartite pumps consist of an inner membrane protein (IMP), an outer membrane protein (OMP), and a membrane fusion protein (MFP) to link the IMP and the OMP. The genes encoding the IMP, OMP and MFP are generally organized into an operon. The known MDR pumps are generally chromosomally encoded, evolutionarily conserved, and may play a critical role in making bacteria to adapt the stresses occurring in their own habitats. With respective to the energetics of these efflux pumps, RND-, MFS-, SMR-, and MATE-type pumps use an electrochemical potential of protons across the cytoplasmic membrane as the energy source to extrude substrates. ATP hydrolysis is the energy source for the ATP-type pump to function [15].
A fusaric acid-resistance-involved tripartite FuaABC pump was identified in S. maltophilia in this study. The involvement of an AraC-type transcriptional regulator, FuaR, in its expression was investigated. Furthermore, the impact of the overexpressed FuaABC pump on antibiotic resistance was evaluated.

Materials and Methods
Bacterial Strains, Media, and Growth Conditions Table 1 lists the bacterial strains, the plasmids, and the primers used in this study. All cultures were grown in LB broth at 37uC with shaking. PCR primer design was based on sequence data obtained from the S. maltophilia K279a genome [16].

Construction of the fuaABC-xylE Single-copy Fusion Strain, KJFua23
A 755-bp DNA fragment (labeled as amplicon 5 in Fig. 1) containing the C-terminus of the fuaC gene and downstream of the fuaC gene was obtained by PCR using the primer sets of 5F/5R and cloned into pEX18Tc, yielding plasmid p5. A xylE cassette was retrieved from pTxylE [19] and inserted into the PstI site of p5, generating plasmid pFua23. The orientation of xylE, confirmed by sequencing, was the same as that of the fuaABC gene. The conjugation between S. maltophilia KJ and E. coli S17-1(pFua23) was performed as previously described [17]. The xylE gene in pFua23 was inserted onto the intergenic region (IG) downstream the fuaC gene, without disrupting any gene. This construction yielded a fuaABC-xylE transcription fusion in KJFua23 chromo-some and the expression of the xylE gene can represent the expression fuaABC operon.

Construction of Promoter-xylE Transcriptional Fusion Plasmids, pFuaA xylE , pFuaB xylE
To determine the possible promoter regions of fuaABC cluster, two xylE transcriptional fusions, pFuaA xylE and pFuaB xylE , were constructed using the pRK415 vector. PCR amplicon 2 and 3 ( Fig. 1) were cloned into pRK415 and the xylE gene was then inserted following the amplicon fragment to yield plasmid pFuaA xylE and pFuaB xylE , respectively. The orientation of the xylE gene in constructs was opposite to that of P lacZ of the pRK415 vector. The promoter fragments assayed contained 108 bp upstream the fuaA gene in pFuaA xylE and 237 bp upstream the fuaB gene in pFuaB xylE .

Induction Assay
The strains tested were cultured overnight at 37uC with continuous shaking. The overnight culture was added to the fresh LB broth with an initial OD 450 nm of 0.2 and incubated for a further 30 min. The inducer was added as indicated. Control group without inducer was established. After incubation for a further 6 h, the cells were harvested for the catechol 2,3dioxygenase (C23O) activity assay.

Determination of Catechol 2,3-dioxygenase (C23O) Activity
Catechol-2,3-dioxygenase is encoded by the xylE gene and its activity was measured as described previously [20]. The rate of hydrolysis was calculated by using 44,000 M 21 cm 21 as the extinction coefficient. One unit of enzyme activity (Uc) was defined as the amount of enzyme that converts 1 nmole substrate per minute. The specific activity was expressed as Uc/OD 450 nm.

Susceptibility Test
The MICs of the antimicrobials were determined with the standard agar dilution method on Mueller-Hinton agar as recommended by the CLSI [21]. After incubation of the paltes at 37uC for 18 h, the MIC was determined by obsreving the lowest concentration of the antimicrobial in which bacterial growth was inhibited. The MICs were determined alone or in combination with 20 mg/L of fusaric acid or 10 mg/L of carbonyl cyanide 3chlorophenylhydrazone (CCCP).

Reverse Transcriptase-PCR (RT-PCR)
Total RNA from S. maltophilia was isolated with a PureLinkTM Total RNA Purification System (Invirtogen, Carisbad, CA, USA) according to the manufacturer's instruction. RNase-free DNaseI (Invitrogen, Carisbad, CA, USA) was used to eliminate DNA contamination. RT-PCR was carried out to firstly amplify the firststrand cDNA using the MMLV Reverse Transcriptase 1 st Strand cDNA Synthesis Kit (Epicentre Biotechnologies, Taiwan), and then PCR amplification of the cDNA was performed with Taq DNA polymerase. The primers used are listed in Table 1.

Quantitative Real-Time PCR (QRT-PCR)
DNA-free RNA was prepared as aforementioned protocol. cDNA was synthesized from DNA-free RNA with a random 6mer primer using the MMLV Reverse Transcriptase 1 st Strand cDNA Synthesis Kit (Epicentre Biotechnologies, Taiwan). Quantitative real-time PCR (QRT-PCR) was, then performed in the ABI Prism 7000 Sequence Detection System (Applied Biosystems) using the Smart Quant Green Master Mix (Protech Technology Enterprise Co., Ltd.) according to the manufacturer's protocols. Relative quantities of mRNA from each gene of interest were determined by the comparative cycle threshold method [22]. The mRNA of 16S rDNA was chosen as the internal control to normalize the relative expression level. The individual target gene was amplified with the primers listed in Table 1 (FuaAQ-F/ FuaAQ-R, FuaBQ-F/FuaBQ-R, FuaCQ-F/FuaCQ-R and rDNA-F/rDNA-R for fuaA, fuaB, and fuaC genes and 16S rDNA, respectively). Each experiment was performed at least three times.

Bioinformatics Analysis
Multiple sequence alignments among the assayed proteins were constructed using the ClustalX program. Phylogenetic analysis was performed using phylip package3.69. DNA distances were calculated by DNADist using the Kimura model. Phylogenetic trees were constructed using the Neighbor-Joining methods. The bootstrap number was obtained in 1000 replications.

Nucleotide Sequence Accession Number
The nucleotide sequences of S. maltophilia KJ fuaR-fuaABC cluster have been deposited in GenBank under accession no. JX524207.

Results
Fusaric Acid-resistance-like Gene Cluster of S. maltophilia The MIC value of S. maltophilia KJ for fusaric acid was 512 mg/ L, as established by susceptibility test ( Table 2), indicating that S. maltophilia KJ has an intrinsic fusaric acid resistance. In an attempt to identify the fusaric acid resistance-related gene(s), a genomewide search was performed on the S. maltophilia K279a genome [16]. An ORF, tagged as Smlt2796 and annotated as a putative transmembrane fusaric acid resistance efflux protein, attracted our attention. Two ORFs downstream Smlt2796, Smlt2797 and Smlt2789, had the signature sequences of periplasmic membrane fusion protein (MFP) and outer membrane protein (OMP). These features suggest that the proteins encoded by the Smlt2796-2797-2798 cluster constitute a tripartite efflux pump. A 795-bp orf (Smlt2795) encoding a putative AraC-type transcriptional regulator was located upstream from the Smlt2796-Smlt2797-Smlt2798 cluster and transcribed in the opposite direction, with a 70-bp Smlt2795-Smlt2796 intergenic region. This genomic organization suggests that Smlt2796-2797-2798 form an operon and Smlt2795 plays a regulator role in the expression of this operon. Therefore, the homologues of Smlt2795-2798 cluster in S. maltophilia KJ were named as fuaR, fuaA, fuaB, and fuaC, respectively, in this study (Fig. 1).
To assess whether the fuaABC cluster is indeed responsible for intrinsic fusaric acid resistance in the wild-type S. maltophilia KJ, a chromosomal fuaABC deletion mutant was constructed. The resulting mutant KJDABC had a four-fold decrease in the MIC of fusaric acid compared with the wild-type KJ (Table 2), supporting the hypothesis that the fuaABC cluster contributes to fusaric acid resistance.

FuaA, fuaB, and fuaC Form an Operon and the FuaABC Operon is Inducibly Expressed by Fusaric Acid
To evaluate the expression of fuaABC cluster, reverse transcriptase-PCR (RT-PCR) on the wild-type strain KJ, cultured in LB medium without fusaric acid, was performed. Primer sets FuaAQ-F/FuaAQ-R, FuaBQ-F/FuaBQ-R, and FuaCQ-F/FuaCQ-R were used for amplification of fuaA, fuaB, and fuaC transcripts, respectively. No significant fuaA, fuaB, and fuaC transcripts were detected. The MIC difference in fusaric acid between KJ and KJDABC was demonstrated by the susceptibility test (Table 2). In some instances, the inducers to trigger the efflux pump expression are the extruded substrates of the efflux pump [23]. Based on these observations, fusaric acid is likely an inducer of expression of the fuaABC cluster. To test this possibility, RT-PCR was used to examine the fuaABC expression of strain KJ without or with the treatment of fusaric acid. Indeed, fuaA, fuaB, and fuaC transcripts were observed in the fusaric acid-treated S. maltophilia KJ. Furthermore, we attempted to verify the possibility of the fuaABC operon using the transcript of the fusaric acid-treated KJ strain and RT-PCR. Primer C1st was used for the first-strand cDNA synthesis and primer pairs across fuaAB (primers AB-F and AB-R) and fuaBC (primer BC-F and BC-F) were used for PCR amplification. The amplicons with expected sizes were detected (data not shown), indicating that fuaA, fuaB, and fuaC are transcribed as a single unit in the fusaric acid-treated KJ strain.
To further test whether the promoter of the fuaABC operon is inducible by fusaric acid, a 108-bp DNA fragment upstream fuaA gene was used to construct the P fuaA ::xylE transcription fusion plasmid, pFuaA xylE . As shown in Table 3, an insignificant C23O activity was observed in KJ(pFuaA xylE ) when cells were grown without fusaric acid. However, the C23O activity was increased approximately 22-fold with the addition of fusaric acid (Table 3). Since a 215-bp intergenic region exists between fuaA and fuaB genes (Fig. 1), we tested the possibility that fuaB has its own promoter. Insignificant C23O activity was detected in KJ(pFuaB-xylE ) either in the presence or absence of fusaric acid (Table 3), indicating that there is no detectable promoter activity in the 237bp region upstream fuaB gene under the conditions tested. The promoter of the fuaABC operon is located in the intergenic region of fuaA and fuaR genes and induced by fusaric acid.

Construction of a Single-copy fuaABC-xylE Transcription Fusion Strain, KJFua23
For the convenience of monitoring the fuaABC operon expression in following assays, a chromosomal fuaABC transcription fusion construct, KJFua23, was constructed by inserting a xylE reporter gene downstream of the fuaC gene without disruption of any gene (Fig. 1). The C23O activity of KJFua23 was increased more than 28-fold under fusaric acid-treated condition (460.8 v.s. 113615 Uc/OD 450 nm ). To confirm this result, qRT-PCR was performed to analyze the transcripts of fuaA, fuaB, and fuaC genes of KJFua23 strain with or without fusaric acid. As expected, the transcripts of the three genes were increased to a similar extent in the presence of fusaric acid. Construct KJFus23 is adequate for monitoring the fuaABC operon expression.

Inducibility of the fuaABC Operon
To investigate whether other compounds, in addition to fusaric acid, will trigger the expression of the fuaABC operon, the C23O activities of KJFua23 with and without the treatment of compounds were measured. The compounds tested included chloramphenicol, nalidixic acid, tetracycline, kanamycin, gentamicin, and erythromycin at a concentration of 1/4 MIC ( Table 2). Among the six compounds tested, none was demonstrated to be potent inducers for the fuaABC expression.
To further elucidate whether the inducibility of the fuaABC operon is fusaric acid concentration dependent, the C23O activities of KJFua23 treated with fusaric acid of different concentrations (20,30,60, 90 mg/L respectively) were measured. The C23O activities of KJFua23 did not significantly change at the concentrations tested.
The induction course of fuaABC was monitored by recording the C23O activity at 1 h intervals for strain KJFua23 after the addition of 20 mg/L fusaric acid. The C23O activity was detectable at the first sampling without any apparent lag phase and gradually increased with time. Maximum C23O activity was obtained after 8 h of induction and lasted for at least 3 h (Fig. 2).

Role of FuaR in Expression of the fuaABC Operon
The fuaABC operon is separated by 70 bp from an upstream fuaR gene, an AraC-type transcriptional regulator transcribed divergently from fuaABC. To assess the role of fuaR in the fusABC expression, fuaR deletion was engineered into the wild-type KJ, yielding mutant KJDFuaR. The promoter activities of P fuaABC in the wild-type and DfuaR background strains were assessed by determining the C23O activities of KJ(pFuaA xylE ) and  KJDFuaR(pFuaA xylE ). KJ(pFuaA xylE ) had an inducible C23O activity phenotype in the presence of fusaric acid (22-fold increase), while KJDR(pFuaAxylE) showed an increase in the basal-level expression (5-fold increase) along with a loss of inducibility (Table 3). Consequently, FuaR functions as a repressor of the fuaABC operon in the absence of a fusaric acid inducer and as an activator in its presence. FuaR is essential for the inducibility of the fuaABC operon. The fusaric acid MIC of KJDFuaR was evaluated. KJDFuaR exhibited a fusaric acid MIC of 128 mg/L as low as strain KJDABC, further supporting the hypothesis that FuaR is essential for fuaABC operon induction.

Substrates Profile of the FuaABC Pump
It is well known that several tripartite pumps in Gram-negative bacteria can extrude a variety of substrates [12]. Therefore, it is of interest to decipher whether the FuaABC pump can extrude other antimicrobials, in addition to fusaric acid. However, the fuaABC operon is intrinsically quiescent and inducibly expressed by fusaric acid, but not by other antibiotics tested. Consequently, it is unreasonable to evaluate the substrates extruded by the FuaABC pump using the standard susceptibility test. As a result, the susceptibility test was comparatively assayed in the absence and presence of 20 mg/L fusaric acid. The wild-type KJ and KJDABC failed to demonstrate differences in MIC value for the antimicrobials tested in the fusaric acid-addition counterpart ( Table 2), indicating that the FuaABC pump cannot extrude the antimicrobials tested except fusaric acid.

Inactivation of FuaABC Pump Components
To examine the role of each component of the FuaABC pump in fusaric acid resistance, the fuaA, fuaB, and fuaC genes were independently deleted, yielding deletion mutants KJDFuaA, KJDFuaB, and KJDFuaC, respectively. KJDFuaA, KJDFuaB, and KJDFuaC reduced the MIC of fusaric acid to the values of KJDABC ( Table 2), indicating that each component of the FuaABC pump is essential for pump function and cannot be substituted by other proteins.

The Effect of CCCP on FuaABC Pump Activity
The proton potential of cytoplasmic membrane and ATP hydrolysis have been reported to provide energy for efflux pumps [15]. FuaA displayed ten transmembrane segments (TMS) predicted by the TMHMM tool (http://www.cbs.dtu.dk/ services/TMHMM/), but did not have any nucleotide binding domain (NBD), a critical domain involved in the ATP hydrolysis. This observation suggests that the energy source for the FuaABC pump is likely the transmembrane proton gradient, and not ATP hydrolysis. To test this hypothesis, the MIC of fusaric acid was determined in the presence of the proton uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP, 15 mg/L). Table 2 shows that the presence of CCCP reduced the fusaric acid resistance of the wild-strain KJ to the same level as that of KJDABC, supporting that the CCCP inhibits the activity.

Phylogenetic Analysis of FuaA
The bacterial efflux pumps are classified as five families, RND, MFS, ABC, MATE, and SMR [14]. The inner membrane transporter protein FuaA investigated in this study cannot be clearly classified as a member of one of these families. A conserved domain of fusaric acid resistance protein family was identified in amino acid residues 40-180 of FuaA protein. Therefore, the phylogeny between FuaA and the known inner membrane transporters of the five efflux pump families is of interest. Five, three, three, four, and two inner membrane transporters of RND-, MFS-, ABC-, MATE-, and SMR-type of gram-negative bacteria, respectively, were selected for phylogenetic analysis. The representatives of inner membrane transporters mainly focus on E. coli, P. aeruginosa, and S. maltophilia, and are known to be a component of tripartite efflux pumps except MATE-and SMR-type. A phylogenetic tree was constructed from FuaA and the 17 proteins (Fig. 3). As expected, each type transporter formed its own phylogenetic cluster, labeled as RND, MFS, SMR, MATE, and ABC clusters in Fig. 3. FuaA formed a separate branch close to the ABC-type transporter cluster.

Discussion
The classification of bacterial efflux pumps has been well described and includes the RND, MFS, ABC, MATE, and SMR families [14]. The transporter system for bacteria to adapt the environmental stress is diverse and complex. Phylogenetic analysis (Fig. 3) revealed that the FuaABC pump proposed in this article should be classified as a subfamily of ABC-type family or a new family, and its pump activity relies on the membrane proton gradient.
Although there are increasingly fusaric acid resistance genes annotated in the finished bacterial genome sequences, the actual role of these genes has been barely described. The sole actual demonstration is the fusABCDE operon of Burkholderia cepacia [5]. The FusA and FusE proteins of B. cepacia, which are an outer membrane protein and a membrane fusion protein, show 32% and 35% protein identity to FuaC and FuaB proteins of S. maltophilia, respectively. The most interesting finding is that the 142-aa FusB, 346-aa FusC, and 208-aa FusD proteins of B. cepacia exhibit 45%, 22%, and 39% protein identity to the N-terminus, central region, and C-terminus of the 656-aa FuaA protein of S. maltophilia, respectively. This suggests that the FusB, FusC, and FusD in B. cepacia can assemble to form an inner membrane transporter, like the FuaA in S. maltophilia.
The FuaR transcriptional regulator proposed in this study belongs to the AraC-type family. Most characterized proteins of the AraC family are positive transcriptional regulators [24]. However, some AraC-type regulators can function as a repressor or a positive regulator depending on the presence or absence of appropriate effectors, for example, the AraC protein from E. coli [25,26], the YbtA protein from Y. pestis [27], and the FuaR protein from S. maltophilia described in this article. According to the protein length, the AraC family regulators can be distinguished into two types. The first type AraC regulator generally consists of more than 250 amino acids with an approximately 99-amino-acid helix-turn-helix (HTH) DNA-binding motif at its C terminus. The N terminus of this type AraC regulator functions as the signal receptor which can bind to the effectors. Upon the effectors binding, the AraC regulator can act as a switch with its activity being radically altered. The AraC, XylS, and Rob of E. coli belong to the first type AraC regulator [26,28]. The second type AraC regulator is shorter in length, roughly 100-150 amino acids which mainly conserve the HTH motif and are devoid of the effectors binding domain. In general, the transcription of this type AraC regulator is controlled by another regulator. The SoxR and MarR of E. coli are representatives of the second type AraC regulator [29,30]. The FusR proposed in this article is 264 amino acids in length and has a HTH DNA-binding motif between residues 180 and 255. Furthermore, the regulatory role of FusR toward the fusABC operon can switch depending on the presence or absence of fusaric acid and the inducibility of the fuaABC operon is FusR protein dependent (Table 3). These observations strongly support the hypothesis that the FusR of S. maltophilia should be classified into the first type AraC regulator and its N terminus may contain a fusaric acid-responsible domain to play a regulatory role in the fuaABC expression. Based on the results of this study, we propose a possible transcriptional regulation mechanism for the fusABC operon expression in S. maltophilia. In the absence of fusaric acid, FusR binds onto the fusR-fusA intergenic region, forming a closed configuration, which is inaccessible by RNA polymerase, repressing fuaABC transcription. In the presence of fusaric acid, fusaric acid binds with FusR, changing the DNA-FuaR-fusaric acid configuration into an open state that is accessible by RNA polymerase, which, in turn, induces fuaABC expression. It is worthily mentioned that in the absence of FuaR, fusaric acid does not function as an inducer, even represses the expression of fuaABC operon (Table 3, strain KJDFuaR(pFuaA xylE )). Whether there is another transcriptional regulator, other than FuaR, involved in the expression of fuaABC operon remains to be elucidated.
A variety of S. maltophilia genome sequences have been released, including the clinical isolate strain K279a [16] and D457 [31] as well as plant isolate strains R553-1 and RR-10 [32]. All of these S. maltophilia genomes contain a repertoire of compounds resistanceassociated genes to resist the environmental pressures. In this study, a fusaric acid-resistance efflux pump, fuaABC, which appears to be responsible for endophytic fitness has been identified in clinical isolate S. maltophilia KJ and its function is well conserved. Sometimes, the xenobiotics-extrusion efflux systems also extrude the antibiotics used in clinic. In this study, we assessed the clinical significance of the FuaABC pump. Results conclude that the fuaABC operon cannot be induced by the antibiotics tested and FuaABC pump does not function in extrusion of antibiotics tested.
At this point, the FuaABC efflux pump has made little contribution to antibiotics resistance so far.

Author Contributions
Conceived and designed the experiments: TC-Y. Performed the experiments: S-TL C-CH Y-WH. Analyzed the data: R-MH T-CY. Contributed reagents/materials/analysis tools: R-MH. Wrote the paper: T-CY.