ABSTRACT
Major facilitator superfamily (MFS) efflux pumps have been shown to be important for bacterial cells to cope with biocides such as chlorhexidine (CHX), a widely used molecule in hospital settings. In this work, we evaluated the role of two genes, smvA and smvR, in CHX resistance in Enterobacter cloacae complex (ECC). smvA encodes an MFS pump whereas smvR, located upstream of smvA, codes for a TetR-type transcriptional repressor. To this aim, we constructed corresponding deletion mutants from the ATCC 13047 strain (CHX MIC, 2 mg/liter) as well as strains overexpressing smvA or smvR in both ATCC 13047 and three clinical isolates exhibiting elevated CHX MICs (16 to 32 mg/liter). Determination of MICs revealed that smvA played a modest role in CHX resistance, in contrast to smvR that modulated the ability of ECC to survive in the presence of CHX. In clinical isolates, the overexpression of smvR significantly reduced MICs of CHX (2 to 8 mg/liter). Sequence analyses of smvR and promoter regions pointed out substitutions in conserved regions. Moreover, transcriptional studies revealed that SmvR acted as a repressor of smvA expression even if no quantitative correlation between the level of smvA mRNA and MICs of CHX could be observed. On the other hand, overproduction of smvA was able to complement the lack of the major resistance-nodulation-cell division (RND) superfamily efflux pump AcrB and restored resistance to ethidium bromide and acriflavine. Although SmvA could expel biocides such as CHX, other actors, whose expression is under SmvR control, should play a critical role in ECC.
INTRODUCTION
Biocidal agents are wildly used to limit the spread of antibiotic-resistant bacteria in hospitals. However, numerous compounds used for disinfection in health care facilities may enhance the risk of the emergence of antibiotic resistance, especially during a long period of low-level exposure (1). The bisguanide chlorhexidine (CHX) is frequently used in health care environments for disinfection of hospitalized patients’ skin (catheter maintenance) or mucous membranes (mouthwash) (2). The decreased susceptibility of bacteria to biocides involves one or more mechanisms such as activation of efflux pumps and modification or overexpression of targets (3). For example, some efflux pumps, such as AceI in Acinetobacter baumannii and CepA in Klebsiella pneumoniae, have been described to be involved in decreased susceptibility to CHX (4, 5). Moreover, Wand et al. demonstrated that increased tolerance to CHX in K. pneumoniae was associated with overexpression of the smvA gene encoding an efflux pump belonging to the major facilitator superfamily (MFS) and that this gene was under the control of the TetR family transcriptional regulator SmvR (6).
The species of the Enterobacter genus, mainly Enterobacter cloacae complex (ECC) and Enterobacter aerogenes, are responsible for 7% of infections among patients hospitalized in intensive care units (ICUs) (7). Actually, ECC is composed of 14 clusters (designated C-I to C-XIV), among which three (C-III, C-VI, and C-VIII) are the most frequently recovered from human clinical specimens (8–10). In addition, ECC members are capable of escaping the antibacterial action of several antibiotics and were classified among the multidrug-resistant (MDR) ESKAPE pathogens (Enterococcus faecium, Staphylococcus aureus, K. pneumoniae, A. baumannii, Pseudomonas aeruginosa, and Enterobacter spp.) (11).
The purpose of the study was (i) to evaluate the in vitro antibiotic and biocide susceptibilities of the ATCC 13047 reference strain and clinical strains belonging to ECC and (ii) to characterize the role of the ECL_01895/ECL_01896 locus (homologous to smvAR) by constructing the corresponding knockout mutants and overexpressing strains. Our data revealed that the SmvR-like regulator played an important role in CHX resistance, and no correlation was observed between MICs of CHX and transcriptional levels of smvA.
RESULTS AND DISCUSSION
ECL_01896 (smvA) is a putative efflux pump involved in decreased CHX susceptibility.One of the main mechanisms underlying decreased CHX susceptibility in Enterobacteriaceae is related to the expression of efflux pumps allowing the expulsion of antiseptic molecules (12). The gene coding for the quaternary ammonium (qac) compound efflux small multidrug resistance (SMR) transporter QacE delta 1 has been recovered among 68% of extended-spectrum β-lactamase-producing (EBSL) E. cloacae clinical isolates with decreased CHX susceptibility (13). However, the prevalence of this gene, located in an integron, did not seem to correlate with the decreased susceptibility to biocides in MDR enterobacterial species including E. cloacae, and no qac gene was recovered in any strain in the study of Kücken et al. (14). Analysis of the genome of E. cloacae ATCC 13047 and that of clinical isolates ECLJO_031, ECLJO_32, and ECLJO_38 allowed us to identify a locus (the gene ECL_01896; henceforth named smvA) encoding a major facilitator transporter of the MFS family (15). ECL_01896 showed 91% and 72% amino acid identities with the SmvA protein of Klebsiella aerogenes and of K. pneumoniae, respectively (data not shown). To evaluate the role of this SmvA-like protein, we constructed an smvA deletion mutant strain and showed that the MIC of CHX was reduced 4-fold compared to that for the ATCC 13047 parental strain (0.5 mg/liter and 2 mg/liter, respectively) (Table 1), indicating that smvA is intrinsically able to expel CHX in ECC.
MICs of biocides and antimicrobials for the different strains
ECL_01895 (smvR) regulates the transcription of smvA.Interestingly, the ECL_01895 gene (henceforth named smvR), located upstream of smvA and divergently transcribed, coded for a TetR family transcriptional regulator homologous to smvR of K. aerogenes and K. pneumoniae. Recently, Wand et al. showed that SmvR acted as a repressor of smvA transcription, and mutation in the sequence of the regulator was correlated with the CHX adaptation (16). Unexpectedly, in the ECC ATCC 13047 strain that exhibited a low MIC of CHX (2 mg/liter), the total deletion of smvR led to overexpression of smvA by 6-fold but only modestly increased the CHX MIC (4 mg/liter) (P < 0.05) (Fig. 1 and Table 1). On the other hand, the CHX susceptibility of the ECL_JO04 strain, which is genetically very close to the susceptible strain ECL_JO40 but has a total deletion of both smvA and smvR, was reduced to a greater extent (0.25 mg/liter) than that of the single smvA or smvR mutant (Table 1). In order to further analyze the role of the SmvR regulator, three clinical isolates (ECL_JO31, ECL_JO38, and ECL_JO32) with higher MICs of CHX (16 to 32 mg/liter) were also analyzed and complemented with smvR cloned with its own promoter into the pBAD202 vector (Table 1). ECL_JO32 was categorized as having decreased susceptibility because the MIC for the strain was <64 mg/liter. This cutoff value also corresponded to the minimal bactericidal concentration (MBC) and was determined for Enterobacter spp. based on the analysis of 54 clinical isolates (17). As shown in Table 1, all of these strains reverted to a CHX-susceptible phenotype with MICs of 2 to 4 mg/liter. In parallel, the transcriptional levels of smvA decreased greatly (nondetectable) in these complemented strains (Fig. 1). These results strongly pointed out the role of SmvR in CHX susceptibility in E. cloacae and are in agreement with those observed in K. pneumoniae (16). Analysis of SmvR sequences revealed that 14 to 15 amino acid substitutions were recovered in clinical isolates compared to the sequences of the two susceptible ATCC and ECL_JO40 strains. Among these substitutions, 11 were at the same positions (Fig. 2A). Note that one and three substitutions were specific for ECL_JO38 and ECL_JO32, respectively. These data contrast with the absence of SmvR mutations in Enterobacter sp. upon CHX exposure while mutations have been observed in Klebsiella, Citrobacter, and Salmonella (6). Because one substitution (G52A) was present into one of the two highly conserved regions of the SmvR sequence in clinical isolates, it is tempting to speculate that it could play a role in high CHX MICs. We also compared the shared promoter region for smvAR especially into the palindromic DYAD motifs which are typical for Tet repressors. Two interesting nucleotide changes were identified in the sequences from clinical isolates: one in the −35 box located in the well-conserved P3 motif and another in the P2 motif (Fig. 2B). The impact of such modifications on decreased CHX susceptibility needs to be further studied.
Expression ratios of smvA in ATCC 13047, the smvA mutant, the smvR mutant, clinical isolates, and strains complemented by pBAD202ΩsmvR. *, P < 0.05 (for results compared to data obtained with the ATCC 13047 strain).
Sequence alignments of SmvR (A) and the smvR-smvA intergenic region (B). In panel A, the amino acid substitutions compared to the sequences of the two susceptible strains ATCC 13047 and ECL_JO40 are in boldface. Conserved regions in Enterobacteriaceae in SmvR proteins are boxed. In panel B, potential conserved motifs P2 and P3 are boxed. Gray boxes indicate start codons of smvR and smvA ORFs. Putative −10 and −35 boxes are indicated in italics. Putative interesting nucleotide changes are in bold.
In order to verify whether CHX may constitute an inducer of smvR expression, we measured the transcriptional level of smvR in the presence of a sub-MIC of CHX (1/4 the MIC). As shown in Fig. 3, the expression of smvR was induced from 2.5- to 9-fold after 20, 40, and 60 min in the presence of CHX in the ATCC 13047, ECL_JO32, and ECL_JO38 strains. As for other Tet family repressors, it appears that when a signal molecule (here, CHX) binds to SmvR, a conformational change takes place that renders the repressor protein unable to bind DNA (18). As a consequence, smvR (likely under its own regulation) was overexpressed (Fig. 3). Concomitantly, increased expression of smvA was observed in the ECL_JO32 and ECL_JO38 strains but not in ATCC 13047 (Fig. 4). It may be suggested that, in the ATCC 13047 strain, the SmvR is structurally modified by CHX but may still interact with the promoter region of smvA.
Expression ratios of the smvR gene in ATCC 13047 and clinical isolates without and with CHX at 1/4 the MIC for 20, 40, and 60 min. *, P < 0.05 (for results compared to the data obtained with the ATCC 13047 strain without CHX).
Expression ratios of the smvA gene in ATCC 13047 and clinical isolates without and with CHX at 1/4 the MIC for 20, 40, and 60 min. *, P < 0.05 (for results compared to the data obtained with the ATCC 13047 strain without CHX).
The CHX susceptibility level is not quantitatively correlated to the level of smvA transcription.We showed that an almost undetectable level of smvA transcription led to a low MIC of CHX and that overexpression of smvA in all Enterobacter strains was coupled with a 2-fold increase of the MICs, arguing for a role of SmvA (Fig. 1 and Table 1). However, no strict quantitative correlation between the transcriptional level of smvA and the MIC was observed. Indeed, the transcription of smvA in ECL_JO31 (with a MIC of 16 mg/liter) was 4.3-fold higher than that in ECL_JO38 ( also with a MIC of 16 mg/liter) and 2-fold lower than that in an smvR mutant of the ATCC 13047 strain, for which the MIC is 4 mg/liter (Fig. 1). Similarly, we measured 3-fold more smvA transcript in the smvR mutant of ATCC 13047 than in ECL_JO32, whereas the MIC of CHX was 8-fold lower (Fig. 1 and Table 1). Recently, it has been shown that heterologous expression of smvA from K. pneumoniae or S. enterica in Escherichia coli (which does not possess smvAR homologs) modestly increased the MICs of CHX (from 0.5 to 2 to 4 mg/liter) (6). It is noteworthy that, in contrast to observations in these E. coli strains, we did not observe any growth defect for our ECC transconjugants (data not shown).
Thus, in Enterobacter, it may be hypothesized that SmvR may have targets other than smvA whose products are involved in reduced CHX susceptibility. Such targets should be other(s) MFS family efflux pumps because the addition of 20 mg/liter reserpine (an MFS pump inhibitor) dramatically reduced the MICs of CHX (2 to 4 mg/liter) in clinical isolates and strains overexpressing smvA (see Table S4 in the supplemental material).
Role of SmvA and SmvR in resistance to other biocide compounds.We previously showed that the ArcB efflux pump of the RND family was involved in the resistance of E. cloacae to biocides such as benzalkonium chloride (BC) and tetraphenylphosphonium and to the dyes acriflavine (ACR), crystal violet (CV), ethidium bromide (EtBr) and rhodamine (19). On the other hand, the lack of AcrB did not impact resistance to CHX or cetyltrimethylammonium bromide (CTAB) (Table 1) (19). To test whether SmvA and/or SmvR can play a role in the resistance to such molecules, MICs were determined for the different constructs of ECC previously used as well as for transcomplemented strains of an acrB mutant of Enterobacter. No differences in the MICs of CV, CTAB, and CB, were observed for the smvA and smvR mutants or transcomplemented strains compared to the MICs for the wild-type strain or parental counterparts (Table 1). On the other hand, a 4-fold increases in the MIC to EtBr and ACR was observed when smvA was overexpressed in the ATCC 13047 strain, and a 4-fold decrease was observed when smvR was overproduced in the ECL_JO38 clinical isolate (Table 1). However, the most important effects were obtained when smvA was expressed in the acrB mutant strain susceptible to EtBr and ACR. Indeed, complementation by smvA in this mutant almost completely restored levels of resistance to EtBr and ACR to those of the wild-type ECC strain (Table 1). This clearly shows that, depending on the biocide, the MFS family efflux pump SmvA was able to compensate the lack of the major RND family pump AcrB.
SmvA and SmvR are not involved in antibiotic resistance.The question as to whether biocides select for antibiotic resistance has been asked for a long time (3). This possible link may be due to the selective pressure created by the presence of antiseptics, leading to the emergence of mechanisms of resistance that are common with other antimicrobial compounds. In this context, the overexpression of efflux pumps may be of great interest. In E. cloacae, the acrB mutant is more susceptible to several biocides as well as to fluoroquinolones, tetracyclines, cotrimoxazole, chloramphenicol, fusidic acid, and erythromycin (19). In order to evaluate the putative role of SmvA and SmvR in antibiotic resistance, MICs of gentamicin (GEN), amikacin (AMK), cefotaxime (CTX), chloramphenicol (CHL), erythromycin (ERY), norfloxacin (NOR), and tetracycline (TET) for the different strains used in this study were determined. As shown in Table 1, no significant difference was observed for either the smvA or smvR mutant strain or for cells overproducing SmvA or SmvR compared to the MICs for their isogenic parental strains. In contrast to results with AcrB, these results argue for a minor involvement of the SmvA efflux pump in the acquisition of antibiotic resistance.
Concluding remarks.The transcriptional regulator SmvR of Enterobacter acted as repressor of the expression of smvA (encoding an MFS family efflux pump). However, (an)other member(s) of its regulon may be involved in the resistance to CHX observed in some clinical isolates. In addition, the overproduction of an MFS-type SmvA was able to partially complement the lack of the AcrB RND family efflux pump by restoring the resistance to EtBr and ACR biocides. Even though the smvRA locus did not seem to be implicated in the antimicrobial resistance of ECC, it appears as an important opportunistic trait. The characterization and surveying of such mechanisms enabling this pathogen to cope with biocides are important, especially in a hospital environment, to avoid the emergence and persistence of virulent isolates.
MATERIALS AND METHODS
Strains, medium, and growth conditions.Bacterial strains and plasmids used in this study are listed in Table S1 in the supplemental material. The reference strain was E. cloacae subsp. cloacae ATCC 13047 (GenBank accession number ECL13047), for which the genome sequence is available (GenBank accession numbers CP001918, CP001919, and CP001920); this strain belongs to cluster XI (15). Escherichia coli and E. cloacae strains were cultured with shaking (200 rpm) at 37°C in Luria-Bertani (LB) medium. In addition to the reference strain, three unrelated ECC clinical isolates (belonging to clusters III, VI, and VIII), with different CHX resistance levels, were collected from the University Hospital of Caen (France) and included in the study. Identification at the species level was performed by using the matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry (Microflex; Bruker Daltonik, Bremen, Germany) while determination of the cluster membership was obtained using the hsp60 gene sequence, as previously described (8).
Drug susceptibility testing.MICs of CHX, crystal violet (CV), cetyltrimethylammonium bromide (CTAB), benzalkonium chloride (BC), ethidium bromide (EtBr), acriflavine (ACR), cefotaxime (CTX), gentamicin (GEN), amikacin (AMK), chloramphenicol (CHL), erythromycin (ERY), norfloxacin (NOR), and tetracycline (TET) were determined by the broth microdilution (BMD) reference method using cation-adjusted Mueller-Hinton broth (CA-MHB) (Becton, Dickinson, Le Pont de Claix, France), in accordance with EUCAST guidelines (http://www.eucast.org/). MICs of CHX were also determined by BMD with or without reserpine (20 mg/liter). The MBC was determined only for CHX by subculturing the last tube showing visible growth and all the tubes in which there was no growth on already prepared plates containing Mueller-Hinton agar medium. The plates were then incubated at 37°C for 24 h, and the lowest concentration showing no growth was taken as the MBC.
Construction of the knockout mutants.Disruption of the genes encoding the MFS transporter (smvA) and its putative TetR regulator (smvR) was performed using a method previously described using the Red helper plasmid pKOBEG (20–22). The primers used in the study are reported in Table S2.
Construction of a multicopy plasmid library containing putative efflux pump or regulator open reading frames (ORFs).The regulator and the MFS efflux pump-encoding genes (smvA and smvR) including their own promoters were amplified by PCR using primers listed in Table S2. Each amplicon was then TA cloned into the pBAD202 Directional TOPO overexpression plasmid (low-copy-number plasmid, ∼20 copies/cell; Invitrogen, Villebon sur Yvette, France). E. coli TOP-10 cells (Invitrogen) carrying pBAD202 recombinants containing correctly oriented inserts were selected on LB plates with 40 mg/liter of kanamycin. After purification, each plasmid carrying the regulator or MFS efflux pump-encoding genes was used to transform the ECC ATCC 13047 strain and clinical isolates (Table S1).
RNA extraction and quantitative reverse transcription-PCR (qRT-PCR).Total RNAs were extracted from bacterial cells grown to the late exponential phase with or without CHX (1/4 the MIC) by using a Direct-Zol RNA miniprep kit (Zymo Research, Irvine, CA). Residual chromosomal DNA was removed by treating samples with a Turbo DNA-free kit (Life Technologies, Saint-Aubin, France). Samples were quantified using a NanoDrop One spectrophotometer (Thermo Fisher Scientific, Courtaboeuf, France). cDNA was synthesized from total RNAs (∼1 μg) using a QuantiTect reverse transcription kit (Qiagen) according to the manufacturer’s instructions. For each RNA sample, transcript levels were determined by the ΔCT (where CT is threshold cycle) method using the expression of the corresponding rpoB gene as a housekeeping control gene from the same RNA extraction (Table S2). The expression ratios were determined by comparison with the level of transcription measured for the ATCC 13047 strain. Experiments were performed at least three times, and Student's t test was performed. P values of less than 0.05 were considered statistically significant.
Whole-genome sequencing (WGS) and bioinformatic analysis.Genomic DNA was isolated from ECC isolates using a Quick-DNA fungal/bacterial miniprep kit (Zymo Research, Irvine, CA) according to the manufacturer’s recommendations. After DNA shearing, the DNA libraries were prepared using an NEBNext Ultra DNA library prep kit for Illumina (New England Biolabs, Ipswich, MA) and sequenced as paired-end reads (2 by 300 bp) using an Illumina MiSeq platform and a MiSeq reagent kit, version 3. The Illumina reads were trimmed using Trimmomatic (Bolger Bioinformatics, 2014), quality filtered with a FASTX-toolkit (http://hannonlab.cshl.edu/fastx_toolkit/), and then assembled using the SPAdes and plasmidSPAdes software programs (23, 24). The nucleotide sequences were also submitted to the ResFinder server (www.genomicepidemiology.org/) to identify acquired antimicrobial resistance genes (25). The presence of the qac and teh genes was examined using a BLAST search on a homemade database containing 59 genes recovered from the NCBI database. These genes are listed in Table S3.
ACKNOWLEDGMENTS
We warmly thank Michel Auzou, Sebastien Galopin, and Mamadou Godet for technical assistance.
François Gravey and Marion Aubourg are supported by grants from the Ministère de l’Enseignement Supérieur et de la Recherche and the Region Normandy, respectively.
We have no relevant financial disclosures or funding to declare.
FOOTNOTES
- Received 10 September 2019.
- Returned for modification 25 September 2019.
- Accepted 24 October 2019.
- Accepted manuscript posted online 4 November 2019.
Supplemental material is available online only.
- Copyright © 2019 American Society for Microbiology.
