|Home | About | Journals | Submit | Contact Us | Français|
Streptococcus mutans, a Gram-positive organism, is the primary causative agent in the formation of dental caries in humans. To persist in the oral cavity, S. mutans must be able to tolerate rapid environmental fluctuations and exposure to various toxic chemicals. However, the mechanisms underlying the ability of this cariogenic pathogen to survive and proliferate under harsh environmental conditions remain largely unknown. Here, we wanted to understand the mechanisms by which S. mutans withstands exposure to methyl viologen (MV), a quaternary ammonium compound (QAC) that generates superoxide radicals in the cell. To elucidate the essential genes for MV tolerance, screening of ~3,500 mutants generated by ISS1 mutagenesis, revealed 15 MV-sensitive mutants. Among them, five and four independent insertions had occurred in SMU.905 and SMU.906 genes, respectively. These two genes are appeared to be organized in an operon and encode a putative ABC transporter complex; we designated the genes as vltA and vltB, for viologen transporter. To verify our results, vltA was deleted by using an antibiotic resistance marker; the mutant was just as sensitive to MV as the ISS1 insertion mutants. Furthermore, vltA and vltB mutants were also sensitive to other viologen compounds such as benzyl and ethyl viologens. Complementation assays were also carried out to confirm the role of VltA and VltB in viologen tolerance. Sensitivity to various drugs, including a wide range of QACs, was evaluated. It appears that a functional VltA is also required for full resistance toward acriflavin, ethidium bromide, and safranin; all are well-known QACs. These results indicate that VltA/B constitute a heterodimeric multidrug efflux pump of the ABC family. BLAST-P analysis suggests that homologs of VltA/B are widely present in streptococci, enterococci, and other important Gram-positive pathogens.
Bacteria have developed multiple strategies to resist action of toxic chemicals, including antibiotics (29). Some of these strategies can be specific for a drug or closely related compounds, such as inactivation of the β-lactam ring of antibiotics by the β-lactamase family enzymes (17). The other mechanism of resistance is by the modification of drug target to reduce the target's affinity for the drug (38). For many organisms, the frontline defense is to reduce the permeability of the cell envelope. The cytoplasmic membrane acts as a barrier to prevent toxic chemical influx into the cell. In Gram-negative bacteria an outer membrane that is less permeable to various toxic compounds gives further protection. However, Gram-positive bacteria lack the outer membrane but are surrounded by thick peptidoglycan cell wall that offers very little resistance to the diffusion of toxic chemicals. The most important mechanism of drug resistance is probably the active efflux of chemicals from the cell (29). These active extrusion mechanisms involve integral membrane proteins that utilize metabolic energy to expel drugs across the membrane against the concentration gradient (29). These effluxes can be specific for a given drug or group-specific chemicals or may recognize a wide range of structurally and chemically unrelated compounds. The latter efflux system, which is known as multidrug efflux or multidrug resistance (MDR) transporter, is very important for the emergence of antibiotic resistance among pathogens. MDR transporters are also important for bacteria to survive under complex environment and facilitate biofilm formation (27, 29).
Based on the bioenergetic criteria, MDR transporters can be classified into two major groups, ATP-dependent transporters and protein-motive-force (PMF)-dependent transporter. ATP-dependent transporters are the primary active transporters that belong to the ATP binding cassette (ABC) superfamily and utilize the free energy of ATP hydrolysis to extrude chemicals from the cells against the concentration gradient. The basic structure of ABC transporters consists of four domains: two integral membrane domains and two ATPase subunits. The ATPase subunits of ABC transporters include a characteristic ABC signature motif (16). Bacterial ABC transporters involved in the uptake also require an additional solute-binding domain, which provides specificity and maintains the direction of transport into the cell (16). PMF-dependent transporters are the secondary transporters that utilize PMF or sodium motive force for drug expulsion.
Streptococcus mutans, which persists in the oral cavity and maintains a biofilm lifestyle in the dental plaque, is the primary etiological agent of dental caries. The dental plaque, which contains more than 600 different microorganisms (1), is a very dynamic environment that often undergoes rapid changes in pH, nutrient availability, and oxygen tension. The mechanisms by which S. mutans copes with these dynamic changes are relatively well studied (2, 28). However, dental plaque is also constantly exposed to a number of compounds that are toxic to the resident microorganisms such as S. mutans. Oral healthcare products, tobacco products, and food additives are a significant source of toxic compounds. Degradation by-products of dental composite resins are another source of toxic chemicals that interfere with bacterial growth in the plaque. Bacteria in the biofilm community also generate various toxic compounds, such as methyl mercaptan and dimethyl sulfide, that can interfere the growth of other competing bacteria (18). During the growth in the biofilm, cells also maintain a balance of metabolism that involves production and detoxification of toxic by-products such that the levels accumulated are well within the capacity of the cell to adapt. Two such examples are the production of methyl glyoxal, a by-product of glycolysis, and peroxides, a by-product of the redox reaction (18, 20, 22). Furthermore, various phytochemicals present in the plant-based diet could be a potential source for toxic compounds related to viologen (21, 37).
How S. mutans tolerates exposure to various toxic substances during its growth in the oral biofilm is poorly understood. The goal of the present study was to elucidate the mechanism of tolerance to quaternary ammonium compounds (QACs). QACs are a group of compounds in which a central nitrogen atom is joined to four organic radicals and one hydrophobic alkyl chain. These compounds are widely used as broad-spectrum bactericides in antiseptics and disinfectants and also used as surfactants and dyes. The representative of QAC that we chose is methyl viologen (MV), also known as paraquat, because this molecule can also generate reactive oxygen species (ROS) in the cell by cyclic univalent reduction and reoxidation. Our study provides the first evidence of an ABC transporter complex, VltAB, which acts as a major multidrug efflux system for several QACs in S. mutans.
Escherichia coli strains DH5α[F− 80lacZΔM15 Δ(lacZYA-argF)U169 deoR recA1 endA1 hsdR17 (rK− mK+) phoA supE44 thi-1 gyrA96 relA1 λ−] (NEB) and TG1Rep+ [supE hsd-5 thi Δ(lac-proAB) F′(traD6 proAB+ lacIq lacZΔM15; repA from pWVO1] was used for routine cloning and for propagation of pGhost9::ISS1 at 37°C, respectively. These strains were grown in Luria-Bertani medium and, when necessary, ampicillin (Ap; 100 μg ml−1), erythromycin (Em; 400 μg ml−1), and spectinomycin (Sp; 100 μg ml−1) were included. S. mutans UA159, a standard laboratory strain which belongs to Bratthall serotype c, was originally isolated by Page Caufield (University of Alabama, Birmingham), and its whole genome has been sequenced recently (3). Twelve other S. mutans strains, including some clinical isolates, were also used for genomic analysis. S. mutans strains were routinely grown in Todd-Hewitt medium (BBL/Becton Dickinson) supplemented with 0.2% yeast extract (THY). When necessary, Em (10 μg ml−1) and Sp (300 μg ml−1) were included.
Insertional mutagenesis was performed with the plasmid pGh9:ISS1, according to the method described by Maguin et al. (9, 31), which was later adapted for S. mutans (41). Briefly, S. mutans was transformed with pGh9:ISS1, and transformants were selected on THY agar containing Em and incubated at 30°C. An overnight-grown liquid culture was made from a single transformed colony. Cultures were diluted 100-fold in the same medium, grown for 2 h at 30°C, and then shifted to growth at 37°C for 2.5 h to select for transposition events. This culture represents the starting transposon library. Cultures were then plated on THY-Em plates to obtain individual mutants, which were then inoculated into 96-well plates containing THY broth supplemented with Em, followed by growth overnight in a CO2 incubator containing 5% CO2. Using a 48-pin replicator (Sigma), the cultures were then spotted onto THY agar plates with or without MV (5 mM), and the plates were incubated at 37°C in the CO2 incubator. Colonies that grew on THY plates but failed to grow on THY+MV plates, were identified, cultured overnight in THY-Em at 37°C, and processed for analysis. Integration of ISS1 on S. mutans chromosome was first verified by Southern hybridization with an ISS1 specific probe (~900 bp) was amplified by PCR using the ISS1-For and ISS1-Rev primer set and pGhost9::ISS1 as a template (for primer sequences, please refer to Table Table1).1). The location of the inserted ISS1 element was then identified by inverse PCR as described previously (41). Chromosomal DNA, isolated from the selected mutants, was digested by HindIII or EcoRI enzymes, followed by heat inactivation at 65°C. About 2 μg of digested DNA was self-ligated by T4 DNA ligase, and the ligated samples were subjected to inverse PCR by using the primers ISS1Rout1 and ISS1Fout1. The PCR products were purified from agarose gel and sequenced with the primer ISS1-Rout2. The flanking sequences obtained from sequencing analysis were mapped on the genome of S. mutans UA159 by a BLAST search.
S. mutans cells carrying chromosomally inserted pGh9:ISS1 were subjected to multiple growth cycles in liquid THY medium at permissive and nonpermissive temperatures, in the absence of antibiotic, to induce plasmid DNA excision. For each growth cycle, a saturated culture grown at 37°C was diluted 1,000-fold in fresh THY medium, followed by incubation at 30°C for 16 h. After 16 h the cells were diluted and plated on THY agar. Colonies were then replicated on THY agar with or without Em to determine the efficiency of plasmid excision and to isolate pGh9:ISS1-cured strain.
To delete the SMU.905 (vltA) locus, an 1.7-kb fragment spanning the entire SMU.905 region was PCR-amplified from UA159 genomic DNA, using the primers 905CF and 905R. This fragment was cloned into the pGEM-T Easy vector (Promega, Wisconsin) to create pIBA21. A 0.87-kb Spr gene (aad9) was amplified from pUCSpec (23) by using the primers Spec-P-For and Spec-Rev (14). Plasmid pIBA21 was restricted with XcmI and blunted by treatment with T4 polymerase. The PCR-amplified aad9 gene was then cloned into this blunted plasmid to generate pIBA20. The orientation of the aad9 insert in pIBA20 was verified by PCR and found to be the same direction as the SMU.905 (vltA) locus. Plasmid pIBA20 was then linearized by NotI and used for UA159 transformation. Spr transformants were selected, and the deletion of the SMU.905 (vltA) locus was verified by PCR. A successful representative transformant was chosen and named IBSA26.
We used allelic exchange to precisely replace the deleted SMU.905 chromosomal locus in IBSA26 with the wild-type SMU.905 gene to construct strain IBSA50, essentially as previously described (8). Briefly, 200 ng of pGhost4 plasmid/ml (used as a helper for selection) with 1 μg of a second unselected target locus DNA/ml (PCR amplified with the primers Smu.CI2.Bam905F and Xho-C905-6R) was used. The transformed cells were plated on Em-containing medium, followed by incubation for 48 h at 30°C under microaerophilic conditions. Transformants that appeared on the THY-Em plates were checked for unselected chromosomal recombination events by patching on THY-Sp plates. An Sps transformant was selected, named IBSA50, and verified by PCR. A similar approach was used to replace the ISS1 interrupted SMU.906 locus in IBSA43 with the wild-type SMU.906 gene to generate strain IBSA51. The PCR fragment used for IBSA43 transformation was amplified with the primers Lnk905-906F and Lnk-906MR.
S. mutans UA159 and its derivatives were grown in THY medium with appropriate antibiotics to the mid-exponential phase (70 Klett units), and the cultures were harvested by centrifugation. The cell pellets were then suspended in 10 ml of RNAprotect bacterial reagent (Qiagen) and incubated at room temperature for 10 min. Total RNA was extracted by using an RNeasy minikit (Qiagen) according to the manufacturer's instructions, with a modified bacterial-lysis step (15). The purified RNA samples were further treated with Turbo DNase (Ambion) according to the manufacturer's instructions to remove residual DNA contamination. The quality and integrity of the purified RNA samples were ascertained by using 1.2% agarose gel electrophoresis. Total RNA was quantitated in a UV spectrophotometer (Shimadzu) according to the optical density at 260 nm (OD260) (1 OD260 unit = 40 μg/ml).
Reverse transcription-PCR (RT-PCR) was used to determine the transcriptional organization of the vltA-vltB locus according to a protocol described by Chong et al. (15). A 5-μg portion of DNA-free RNA was used for the synthesis of cDNA using SuperScript II reverse transcriptase (Invitrogen) according to the manufacturer's protocol. PCR was then performed on RNA (as a negative control), cDNA, and chromosomal DNA (as a positive control), using the primer pairs described in Table Table11 and depicted in Fig. 2, to determine which of the genes in the vltA-vltB locus were cotranscribed.
Semiquantitative RT-PCR (sqRT-PCR) was used to quantify the level of expression of vltAB in UA159. S. mutans cultures were grown to mid-exponential phase (KU70), followed by RNA extraction. sqRT-PCR was performed as previously described (13) using the Titan One-Tube RT PCR system (Roche). The gyrA gene was included as control to ensure that equivalent amounts of RNA were being used for each reaction.
To evaluate the sensitivity of the S. mutans mutants to various ROS-producing chemicals, cultures were exposed to reagents either through disk diffusion assays or growth on THY agar plates containing the chemical of interest. For disk diffusion assays, cultures were grown overnight in THY broth (with or without antibiotics as required) and then overlaid on THY agar plates, followed by the placement of filter paper disks (6 mm in diameter) containing various chemicals onto the inoculated agar. After overnight incubation at 37°C under microaerophilic conditions, the diameters of the zones of bacterial growth inhibition were measured.
For growth on THY agar plates containing chemicals, cultures were grown to exponential phase in THY broth with appropriate antibiotics at 37°C. Cultures were centrifuged, washed twice with 0.85% NaCl, and resuspended in 0.85% NaCl. The cultures were adjusted to an optical density (A600) of 5.0 and 10-fold serially diluted, and 7.5 μl of each dilution was spotted onto THY agar containing the chemicals of interest. The plates were incubated overnight at 37°C, under microaerophilic conditions, and the bacterial growth was evaluated as previously described (11). The following chemicals (obtained from Sigma) were used as indicated. MV (paraquat) was added to THY agar medium to a final concentration of 5 mM (or 10 mM), while 10 μl from a 1 M stock was added to each disk for the disk diffusion assay. Ethyl viologen (1.5 M), benzyl viologen (50 mg/ml), and diquat dibromide (100 or 500 mg/ml) were added to each disk. Menadione (1.0%), pyrogallol (400 mg/ml), and plumbagin (100 mM) were used in a disk diffusion assay for generating superoxide radicals. Hydrogen peroxide (1.5% [wt/vol]), cumene hydroperoxide (10%) and t-butyl-hydroxyperoxide (70%) were also used in disk diffusion assays.
Disk diffusion assays were performed to evaluate the antibiotic susceptibility of S. mutans UA159 and its derivatives as described previously (7). Briefly, antibiotic disks (6 mm in diameter; Becton Dickinson) were placed on THY agar plates that were overlaid with 10 ml of THY soft agar containing 200 μl of the S. mutans strain of choice. The plates were incubated overnight at 37°C under microaerophilic condition, and the zones of inhibition were measured. A list of the antibiotics used for the present study is given in Table S1 in the supplemental material.
Sensitivity of the S. mutans mutant strains to various chemicals was evaluated primarily by using the disk diffusion method. THY agar plates inoculated with the wild-type or the mutant cultures were overlaid with disks containing various toxic QACs or other stress-inducing chemicals. The chemicals (all from Sigma) tested included benzidine (4 mM), benzalkonium chloride (10 mg/ml), tetraethylammonium bromide (10 mg/ml), chlorhexidine gluconate (20%), crystal violet (1.5%), diamide (1.0 M), 2,2′-dipyridyl (50 mM), 4,4′-dipyridyl (1 M), ethidium bromide (1.5%), EDTA (EDTA, 0.5 M), CTAB (cetyltrimethylammonium bromide; 0.5%), hexadecylpyridinium chloride (1%), hydrazine (0.5 M), 1,4′-napthaquinone (0.1%), malachite green (1.5%), mitomycin C (12.5 μg/ml), 1,10′-phenanthroline (200 mM), potassium telurite (1.0%), puromycin (6.0 μg/ml), pyronin B (5 mg/ml), safranin (1.5%), and streptonigrin (5 mg/ml). A complete list of the chemicals used here is provided in Table S1 in the supplemental material.
To identify genes that are potentially involved in MV tolerance, we used ISS1 transposition mutagenesis since ISS1 appears to integrate itself randomly into the genome of streptococci (9, 39), including S. mutans (41). Furthermore, ISS1 rarely inserts more than once into the genome of the same cell (9, 39, 41). We used plasmid pGhost9::ISS1 (30, 31) to introduce the insertion element into wild-type strain UA159. The insertion frequency was found to be less than 0.5%, which is consistent from one independent experiment to the next and was similar to the frequencies reported for other streptococci (9, 39). A collection of approximately 3,500 mutants were grown in 96-well plates and replica patched onto THY, with or without MV. We obtained 15 mutants were that displayed an MV-sensitive growth phenotype. We performed Southern hybridization assay to examine how ISS1 insertion had occurred in each of the 15 mutants. We obtained 13 mutants into which ISS1 was inserted at a single location on the chromosome and two mutants in which ISS1 was integrated at more than one location (possibly at two sites [data not shown]). Of the 13 mutants that contained insertion at a single location, 4 contained multiple-ISS1 insertion sequence at the same location, and the rest contained a single-copy ISS1 insertion.
For majority of the clones, the site of the ISS1 insertion was identified by inverse PCR method as described in Materials and Methods. Of the 15 MV-sensitive mutants, the location of the insertion sites could be determined unambiguously for 13 mutants. Five independent insertions (IBSA27, IBSA28, IBSA34, IBSA35, and IBSA40) occurred in SMU.905, which encodes a putative ABC transporter complex. Four independent insertions (IBSA32, IBSA36 to IBSA38) had occurred in SMU.906, which is just downstream of SMU.905 and encodes a putative ABC transporter. For the remaining four mutants, insertions were mapped to SMU.1128 (encoding the histidine kinase ciaH; two insertions), SMU.902 (encoding a putative ABC transporter), and SMU.283 (encoding a small hypothetical protein). One mutant, IBSA41, had two insertions, one in ciaH and one in SMU.1149 that encodes a putative transporter for bacteriocin. Because CiaH was already known to be involved in superoxide stress tolerance in streptococci and other bacteria (36), we did not consider this protein for further characterization. We also precluded the small hypothetical protein encoded by SMU.283 since it shows similarity to bacteriocin-like peptide. Our repeated attempt to identify the ISS1 insertion sites in two of the mutants (IBSA30 and IBSA33) by inverse PCR and cloning did not generate any fruitful sequence. Since we obtained multiple insertions in SMU.905 and SMU.906 loci, we designated these genes as vltA and vltB, respectively, for viologen transporter and elected to focus our studies on these genes.
The vltA gene encodes a polypeptide of 579 residues with high homology to an ABC transporter protein; the ISS1 insertion occurred at five times in this gene at codon positions 58 (IBSA40), 75 (IBSA35), 187 (IBSA34), 539 (IBSA28), and 573 (IBSA27) (Fig. (Fig.1).1). The vltB gene, which lies just 11 bp downstream of vltA, appears to be organized as an operon with vltA and encodes a polypeptide of 591 residues, also with high homology to an ABC transporter protein. The ISS1 insertion occurred in vltB at positions 13 (IBSA36), 181 (IBSA32), 483 (IBSA37), and 564 (IBSA38) (Fig. (Fig.1).1). SMU.902 (623 residues) in which a single ISS1 insertion occurred at codon position 619 is found upstream of vltA (Fig. (Fig.1),1), while SMU.909, which encodes a malate permease, is found downstream of vltB. An intergenic region of 456 bp is present between the SMU.902 and vltA loci. In silico analysis by BPROM software (Softberry, Inc.) indicates that this region contains a weak promoter-like sequence (−10 box [TATATT] at position 362), indicating that vltAB may be transcribed separately from SMU.902. However, analysis of the same 456-bp intergenic region by FindTerm software (Softberry) failed to show any strong rho-independent terminator. Thus, to determine whether vltAB are transcriptionally linked to SMU.902, a linkage PCR analysis was performed using RNA isolated from exponentially grown cultures of UA159. As shown in Fig. Fig.2,2, it appears that both vltA and vltB are transcriptionally linked to SMU.902, whereas vltB is not linked to the downstream gene SMU.909 (Fig. (Fig.22 and data not shown). A pentacistronic operon encodes SMU.902, and it appears that SMU.902 is the terminal gene in that operon. About 3.2-kb upstream of SMU.902 lies another intergenic region of 213 bp that also contains a strong promoter-like structure (−10 box [TGCTATAAT]) 37 bp upstream of the first ATG codon. Thus, SMU.902, as well as vltAB, may be transcribed from this promoter. Our linkage analysis also indicated that the downstream gene SMU.909 is not transcriptionally linked to vltB. The intergenic region (114 bp) between vltB and SMU.909 contains a strong promoter-like sequence with a perfect −10 box (TATAAT); SMU.909 is probably transcribed from this promoter sequence independently of the upstream genes.
To confirm that the observed phenotype of the vltA ISS1 insertion mutation did not result from additional spontaneous mutations elsewhere in the genome, a deletion mutation of vltA was constructed in strain UA159 by a gene replacement system using a nonpolar antibiotic marker (aad) as described previously (10). When the growth of the mutant, IBSA26, was examined in THY broth, no obvious growth defects were observed (data not shown), indicating that vltAB locus does not influence overall growth. IBSA26 was then tested for its ability to withstand exposure to MV. As shown in Fig. Fig.3A,3A, IBSA26 displayed sensitivity to MV exposure as the original ISS1 insertion mutant, IBSA34, and its derivative IBSA42, which is cured of the ISS1 delivery plasmid (Ems mutant). However, the sensitivity of IBSA26 was more pronounced than the original ISS1 insertion mutant, IBSA34. The ability of IBSA26 to tolerate MV exposure was restored by reverse complementation with the wild-type SMU.905 gene (IBSA50, Fig. Fig.3B3B).
To confirm that vltB is also responsible for viologen tolerance, we used IBSA43, a plasmid cured derivative of IBSA32 (Fig. (Fig.1).1). As expected, IBSA43 displayed same degree of sensitivity to MV exposure as the vltA mutant strain (IBSA26). The ability of IBSA43 to resist MV exposure was also restored by reverse complementation with the wild-type vtlB (SMU.906) gene (IBSA51, Fig. Fig.3B3B).
A recent microarray based hybridization analysis suggests that as much as 20% of the UA159 genes are absent in one or more S. mutans clinical isolates (40). We used PCR analysis to determine whether the vltAB locus is present in various S. mutans strains. Based on analysis of the UA159 genome sequence, we designed two internal primers for the genes SMU.902, vltA, and vltB and used them for PCR amplification, along with chromosomal DNA isolated from 15 different S. mutans clinical isolates as a template. Of the 15 strains chosen for the analysis, 11 belong to serotype c, including three commonly used lab strains (UA159, NG-8, and GS-5), and the remainder beloged to serotype e (V100) and serotype f (OMZ175). We found that these three genes were present in all 15 S. mutans strains (data not shown). We also tested whether the intergenic region between SMU.902 and vltA loci is conserved among the isolates and found that the length of the intergenic region is also well conserved among various isolates (data not shown).
To test whether the vltA mutant (IBSA26) is sensitive to reagents that generate superoxide but chemically and structurally are distinct from MV, we used menadione, pryrogallol, and plumbagin. We observed that IBSA26 did not have an increased sensitivity to those superoxide generators (see Table S1 in the supplemental material). This observation suggests that vltA is not involved in general superoxide stress tolerance; rather, this ABC transporter is specific for MV-mediated toxicity tolerance. To confirm our hypothesis that vltA is not involved in general superoxide stress tolerance response, we measured the superoxide dismutase (SOD) activity in IBSA26. S. mutans, like other streptococci, encodes a single Mn-type SOD (34). SOD activity was measured in the crude cellular extract isolated from mid-exponentially grown cultures on a native polyacrylamide gel by negative staining using the nitroblue tetrazolium method (5). Only a single active band, in approximately equal intensities, was seen in the wild-type (UA159), the vltA mutant (IBSA26), and the vltB ISS1 insertion (IBSA42) strains (data not shown). Thus, vltAB is not involved in the general superoxide stress response.
We also examined the role of vltAB in oxidative stress responses. We used hydrogen peroxide to generate intracellular oxidative stress and used cumene hydroperoxide and t-butyl hydroperoxide to mimic lipid hydroperoxide stress (4). The zones of growth inhibition of IBSA26 by these peroxides were similar to the wild-type UA159 strain (see Table S1 in the supplemental material). Thus, vltA is not involved in the oxidative stress tolerance response.
The vltA mutant, IBSA26, while sensitive to MV, did not display an increased sensitivity to the structurally distinct superoxide generators. To test whether VltAB is involved in tolerance to other viologens, we used ethyl viologen (EV) and benzyl viologen (BV). We tested the original ISS1 insertion mutants (IBSA34 and IBSA37), as well as the vltA deletion mutant IBSA26. As shown in Fig. Fig.44 (and listed in Table 3), IBSA26 showed enhanced sensitivity to both EV and BV, with a zone of inhibition of 24 ± 2 mm and 27 ± 2 mm, respectively. As expected, we observed no obvious growth inhibition in UA159 with EV and BV (Fig. (Fig.4).4). The original ISS1 mutants of vltA also displayed similar zone of inhibition as the vltA deletion mutant (data not shown). We also tested the plasmid-cured ISS1 mutant of vltB (IBSA43) for its sensitivity to EV, BV, and MV. This strain also showed increased sensitivity to these compounds compared to the wild-type strain (Table 3). However, the zones of growth inhibition were ca. 70 to 80% of the IBSA26 strain.
We then tested the sensitivity of IBSA26 to diquat (DQ) because it represents the 2,2′-dipyridyl derivative that is structurally similar to viologen. However, when fresh overnight cultures were used, the effect of DQ on the growth inhibition of IBSA26 was not so obvious. Although the diameter of the zone of inhibition for both the wild type and the mutant was similar (~17 mm) with DQ (100 mg/ml), the wild type produced a cloudy zone of halo with residual bacterial growth present within the inhibited zone, while the mutant produced clear zone of halo. In contrast, when the overlay cultures were 48 h or older, the zone of inhibition was significantly larger for IBSA26 compared to UA159, and both the strains produced a clear zone of halo (see Table 4).
The viologens and DQ are structurally similar, charged molecules, and generate superoxide radicals. We wanted to examine whether vltA is also involved in the tolerance to compounds that contain similar dipyridyl structures but do not generate superoxide radicals. Therefore, we tested the sensitivity of the vltA deletion mutant to 4,4′-dipyridyl (4DP), benzidine (BZD), and 1,10′-phenanthroline (PHEN) in a disk diffusion assay. We observed no zone of inhibition with 4DP and BZD for both the wild-type and the vltA mutant strains, while the halo sizes for both the strains were similar when PHEN was used (~22 mm, see Table S1 in the supplemental material). Therefore, our results, taken together, suggest that VltAB is specifically involved in the tolerance of viologen and related charged compounds that generate superoxide radicals.
To examine whether prior exposure of S. mutans cells to sublethal concentration of MV can induce the transcription, we measured vltA transcripts by sqRT-PCR. We found that prior MV exposure did not induce vltA transcription, suggesting that expression of the vltAB operon may not be inducible (data not shown).
Since VltA is required for viologen tolerance, we wanted to know whether VltA is also necessary for resistance to other QACs. We tested the vltA deletion mutant (IBSA26) in disk diffusion assays for sensitivity toward commonly used biological dyes (malachite green, crystal violet, pyronin B, and safranin), compounds commonly used in disinfectant or in mouthwash (acriflavin, benzalkonium chloride, cetrimonium bromide, cetylpyridinium chloride, and tetraethylammonium bromide), and ethidium bromide. As shown in Fig. Fig.5,5, the vltA mutant revealed an enhanced sensitivity to acriflavin, ethidium bromide, and safranin O, with zones of growth inhibition ca. 130% larger than those of the wild-type UA159 strain. However, there were no significant differences in the zones of growth inhibition between the wild type and the mutant for the other QACs (see Table S1 in the supplemental material). To confirm our result that VltA/B is indeed involved in the efflux of these the QACs, we tested the plasmid cured ISS1 insertion mutant of vltB (IBSA43, Table Table2)2) for its sensitivity. To our surprise, IBSA43 did not show increased zones of growth inhibition to these drugs as the vltA mutant strain (data not shown). Thus, while the role of VltA in the tolerance to these QACs is very certain, the role of VltB in the efflux of other QACs is unknown.
Since many ABC transporters are known to be involved in both QAC and quinolone resistance (35), we examined the role of VltA in the resistance to quinolone antibiotics such as ciprofloxacin, levofloxacin, and nalidixic acid. We also included vancomycin (contains complex heterocyclic rings); lincosamide drugs such as clindamycin and chloramphenicol; trimethoprime (contains a diamine-pyrimidine group); rifampin (contains a complex ring structure similar to quat compound); and bacitracin, a cyclic polypeptide. These antibiotics were used in disk diffusion assays with IBSA26 and UA159 strains. We did not find any significant differences in the growth inhibition zones between the SMU.905 mutant and the wild-type strains (see Table S1 in the supplemental material).
We also tested sensitivity of IBSA26 to various toxic chemicals such as chlorhexamide, diamide, EDTA, hydrazine, mytomycin C, potassium tellurite, puromycin, and streptonigrin; all of these reagents are structurally very different. We also did not observe any significant difference between wild-type and mutant strains (see Table S1 in the supplemental material). Thus, VltA/B appears to have very restricted substrate specificity.
Phenotypic microarray (PM) assay is a relatively new method that allows testing for a large number of phenotypes simultaneously for a given strain (12). PM assays were performed on IBSA26 and UA159 for about 1,900 cellular phenotypes. In the metabolic panels (PM1 to PM8), overall signals were either sporadic or very low (see Fig. S1 in the supplemental material). PM analysis in the osmotic panel (PM10) generated good signal but no significant differences between the strains. Surprisingly, for the toxic chemical panels (PM11 to PM20), despite reasonably good growth in most of the conditions, we observed no significant differences between wild-type and mutant strains. This suggests that VltA has a very restricted substrate specificity that includes viologens and some QACs.
We obtained a single insertion (IBSA31) in the SMU.902 locus that lies just upstream of the vltA and vltB loci. SMU.902 encodes a polypeptide of 623 residues that shows high homology with an ABC transporter protein. The insertion that we obtained in SMU.902 was mapped at position 619. To gain a better insight into the function of SMU.902, we used IBSA31 to generate a strain, IBSA44, which is devoid of the delivery plasmid pGhost9. Both IBSA31 and IBSA44 were tested for MV sensitivity. As shown in Fig. Fig.6,6, both of the strains showed increased sensitivity toward MV. Furthermore, although both of the strains were highly sensitive to MV exposure, we observed that the IBSA44 strain (plasmid cured, Tables Tables33 and and4)4) was at least 10-fold less resistant to MV than the original insertion mutant IBSA31 that contains the plasmid. We also tested IBSA44 for its sensitivity toward acriflavin and ethidium bromide in a disk diffusion assay. As expected, IBSA44 was more sensitive to both of these reagents compared to the wild-type strain, suggesting that this insertion in SMU.902 somehow disrupted the wild-type function (data not shown).
For successful colonization and maintenance of a dominant presence in the oral cavity, S. mutans has developed multiple strategies. These strategies help this organism to grow under nutrition-limiting conditions and protect it from various environmental insults (28). Although most of the previous studies were focused on understanding the mechanisms of acid tolerance and oxidative-stress responses, our knowledge of the mechanisms of tolerance to various toxic chemicals remains limited. To obtain further insight into this process, a collection of random insertion mutants of S. mutans UA159 was screened to select clones with high sensitivity to MV, a charged dipyridyl-ring-containing QAC that also generates ROS. This approach allowed us to identify genes that may be responsible for defense against QAC, without prior knowledge of the genes' function(s). In the present study, we only screened approximately 3,500 such mutants; therefore, the screening process was not particularly exhaustive since S. mutans genome encodes about 1,900 genes. Of the five loci that were identified in our analysis, at least one gene, ciaH, was previously reported as an important player in the oxidative-stress response in streptococcus and other bacteria (2, 36, 41), signifying that the screening method used here is a viable approach. Four unique loci were identified by our search: three ABC transporter-encoding genes (SMU.902, vltA, and vltB) and a hypothetical protein (SMU.283). In the present study, we further characterized the vltAB loci in order to better understand the mechanism and substrate specificity of these ABC transporters.
Sequence analysis and genome organization strongly suggest that VltA and VltB encode a heterodimeric ABC-type exporter pump. Our linkage analysis also demonstrated that these two genes are transcriptionally linked. S. mutans genome analysis indicates that this organism encodes several ABC transporters, of which at least 42 are putative exporter pumps (3, 33). Since we specifically obtained multiple insertions in vltA and vltB (five in vltA and four in vltB), this ABC transporter appears to be the most important for viologen tolerance in S. mutans. Analysis of the sequences by a transmembrane (TM) helices prediction program, TMHMM (www.cbs.dtu.dk/services/TMHMM), of both VltA and VltB revealed that these two proteins contain 6 (residues 1 to 294) or 5 (residues 1 to 382) TM helices, respectively. Both VltA and VltB also contain putative nucleotide-binding domains, Walker A and Walker B motifs, and ABC signature sequences (16, 24) (Fig. (Fig.1).1). We also found that both VltA and VltB were necessary for viologen resistance, supporting our notion that VltA and VltB is a heterodimeric ABC-type multidrug efflux pump.
The ABC transporter that we identified export, in addition to viologen compounds, some other QACs such as acriflavin, ethidium bromide, and safranin. Analysis of the structures of these compounds failed to identify any common structural moiety that could easily explain the substrate specificity (Table (Table3).3). However, all of the compounds are charged heterocyclic molecules. Definitely charge plays a role in the substrate recognition since the ABC transporter complex did not recognize dipyridyl and benzidine, which are structurally very similar to viologen but uncharged. Similarly, VltA expelled DQ, which is structurally very similar to phenanthroline but charged, out from the cell, whereas phenanthroline was not recognized. On the other hand, charged alone is not sufficient to explain the substrate specificity as well. This is because we also tested several QACs ranging from compounds that contain simple structures such as tetraethylammonium bromide to compounds that contain heterocyclic rings such as malachite green and crystal violet; VltA recognized none of these QACs. Thus, in addition to charge, other physical characteristics such as hydrophobicity or amphiphilicity may be also very important.
Although IBSA43 (vltB mutant) displayed increased sensitivity to viologen compounds, this mutant, surprisingly, did not show any significant differences compared to the wild type when tested for sensitivity to acriflavine, ethidium bromide, and safranin. Thus, it appears that VltB is not involved in the resistance of these compounds. Although VltA and VltB are expected to interact with each other to form a functional heterodimeric ABC transporter, it is possible that VltA and VltB can each form homodimers and that these homodimers have different substrate specificities. For example, while VltA homodimer is involved in QAC resistance, VltB homodimer does not take part in QAC resistance. Future studies will address the question of whether VltA and VltB can also form homodimers and, if so, to what extent their substrate specificities differ from that of the VltA/B heterodimer.
BLAST searches using protein sequences as a query against the nonredundant database at the National Center for Biotechnology Information showed that homologues of VltA/B are widely present in streptococci, enterococci, and clostridia (see Fig. S2 in the supplemental material). In all cases, two open reading frames were located in tandem, and many genes seemed to encode multidrug resistance ABC-type proteins. The closest homologues (>90% identity), SAG1338 and SAG1337, are found in group B streptococci (GBS), and all of the sequenced GBS strains encode these ABC transporter genes. In contrast to other organisms in which VltA/B homologues are found, the genomic locus for this ABC transporter is somewhat conserved in GBS. Specifically, the four upstream genes are highly conserved, including the SMU.902 homologue, SAG1340. On the other hand, two genes immediately downstream of SAG1337 are homologues of SMU.911 and SMU.913. However, homologues of SMU.909 that encode a malate permease and SMU.910, which encodes glucosyltransferase, are absent in GBS. Our BLAST search also identified two ABC transporter proteins from Enterococcus faecalis, EfrA (EF2920) and EfrB (EF2919), which showed >80% identity to SMU.905 and SMU.906, respectively (26). It has been demonstrated that expression of both efrAB genes confers resistance to many drugs, including acriflavin, ethidium bromide, and safranin (26). Unfortunately, MV or other quat compounds were not tested in that study; thus, whether EfrAB is involved in viologen efflux remains to be seen.
Although we have identified VltA/B as an ABC transporter involved in the efflux of viologens and QACs, the organism may not encounter these chemicals frequently during its growth in natural habitat in the dental plaque. Dental plaque is a polymicrobial community that harbors more than 500 species or phylotypes (1), and the cell density can reach as high as 1011 CFU/ml (19). The oral biofilm is continuously challenged by changes in the environmental conditions and, as a response to such challenges, the bacterial community evolved, with individual members with specific functions such as primary or secondary colonizers, including the ability to metabolize or tolerate toxic excreted products produced by other species (25). About 20% of the oral bacteria are streptococci (32), and these organisms with their specific spatial and temporal distribution determine the development of the biofilms. When present in high numbers, the pioneer colonizers can antagonize S. mutans, as suggested by clinical studies (6). However, S. mutans can become dominant in oral biofilms, which leads to dental caries development. This dominance depends on competition with other organisms and is influenced by various factors. We speculate that the presence of numerous transporters, such as VltA/B, allows S. mutans to withstand toxic compounds produced by competing species or present in the plaque environment.
We thank Wolf Zückert for helpful suggestions.
This study was supported by an NIDCR grant (DE016686) and an NCRR COBRE project grant (P20RR01644) to I.B.
Published ahead of print on 31 January 2011.
†Supplemental material for this article may be found at http://aac.asm.org/.