The DmeRF System Is Involved in Maintaining Cobalt Homeostasis in Vibrio parahaemolyticus

Although cobalt (Co) is indispensable for life, it is toxic to cells when accumulated in excess. The DmeRF system is a well-characterized metal-response system that contributes to Co and nickel resistance in certain bacterial species. The Vibrio parahaemolyticus RIMD 2210633 genome also harbors a dmeRF operon that encodes a multiple antibiotic resistance regulator family transcriptional regulator and a cation diffusion facilitator family protein. Quantitative real-time PCR, growth curves analysis, inductively coupled plasma-mass spectrometry, β-galactosidase activity assays, electrophoretic mobility shift assays, and a mouse infection experiment were performed to characterize the function of the DmeRF system in V. parahaemolyticus. Zinc, copper, and Co significantly increase dmeF expression, with Co inducing the greatest increase. DmeF promotes V. parahaemolyticus growth under high-Co conditions. Additionally, increased accumulation of cellular Co in the ΔdmeF mutant indicates that DmeF is potentially involved in Co efflux. Moreover, DmeR represses the dmeRF operon by binding directly to its promoter in the absence of Co. Finally, the DmeRF system was not required for V. parahaemolyticus virulence in mice. Collectively, our data indicate that the DmeRF system is involved in maintaining Co homeostasis in V. parahaemolyticus and DmeR functioning as a repressor of the operon.


Introduction
Metals, such as cobalt (Co) and zinc (Zn), are indispensable for almost all organisms. Indeed, many proteins require metals as structural components or enzymatic cofactors [1]. Vertebrate hosts have evolved a strategy, termed nutritional immunity, to decrease the availability of metals to invading bacteria [2]. Consequently, bacteria have developed numerous and varied countermeasures such as the utilization of high-affinity metal acquisition systems [3]. Despite their essential roles, metals are harmful to bacteria when accumulated in excess [4]. Remarkably, increasing evidence shows that hosts exploit metal toxicity to defend against bacterial pathogens [5,6]. Metal homeostasis disturbance can lead to mismetalation (i.e., protein metalation with a non preferred metal) or the generation of reactive oxygen species (ROS) via the Fenton reaction; both are detrimental to cells [7]. Therefore, metal homeostasis should be tightly regulated. Since metals cannot be synthesized or degraded, bacteria maintain their homeostasis mainly through the modulation of metal import and export [4].
Co is a micronutrient that participates in various metabolic processes. Most notably, it is required by vitamin B 12 and certain proteins [8,9]. Nonetheless, excessive amounts of Co lead to toxicity. It can cause cell damage by catalyzing the production of ROS [10]. Moreover, Co competes with iron (Fe) in a variety of metabolic processes [11][12][13]. To prevent Co toxicity, efflux systems, such as P 1B-4 -type ATPase and cation diffusion facilitator

Identification of the DmeRF System in V. parahaemolyticus
In the genome of V. parahaemolyticus RIMD 2210633, the locus VP_RS21330 is annotated as dmeF, whose product is the CDF family Co/Ni efflux transporter DmeF. This protein exhibits 50.72%, 49.68%, 46.15%, and 44.68% amino acid sequence identity with DmeF from C. metallidurans, S. meliloti, A. fabrum, and R. leguminosarum, respectively. The locus VP_RS21325, immediately adjacent to and co-transcribed with dmeF ( Figure S1), encodes a multiple antibiotic resistance regulator (MarR) family transcriptional regulator. No significant similarity was found between this protein and DmeR from the Rhizobiaceae species. Despite this, VP_RS21325 was designated as dmeR based on its regulatory role in dmeF expression (see Section 2.5). In the Rhizobiaceae species, dmeR is located upstream of dmeF [14,[19][20][21], but in V. parahaemolyticus, the reverse applies ( Figure 1). Multiple sequence alignment also showed that the DmeF proteins from V. parahaemolyticus, C. metallidurans, and the Rhizobiaceae species are highly conserved, while the DmeR protein from V. parahaemolyticus shows little homology with those of the Rhizobiaceae species ( Figure 2). Together, these results revealed that the V. parahaemolyticus DmeF may have functions similar to its homologues from C. metallidurans and Rhizobiaceae.    To evaluate the involvement of DmeF in the response of V. parahaemolyticus to metal toxicity, dmeF expression in RIMD 2,210,633 incubated with elevated levels of various metals was determined by quantitative real-time PCR (qRT-PCR) analysis. As seen in Figure 3, dmeF expression increased approximately 14-fold after the Co treatment compared to the H 2 O treatment. Furthermore, the Zn and Cu treatments increased dmeF expression by approximately 4-fold and 5-fold, respectively, while Fe(II), Mn, and Ni had no significant effect on dmeF expression ( Figure 3). To evaluate the involvement of DmeF in the response of V. parahaemolyticus to metal toxicity, dmeF expression in RIMD 2,210,633 incubated with elevated levels of various metals was determined by quantitative real-time PCR (qRT-PCR) analysis. As seen in Figure  3, dmeF expression increased approximately 14-fold after the Co treatment compared to the H2O treatment. Furthermore, the Zn and Cu treatments increased dmeF expression by approximately 4-fold and 5-fold, respectively, while Fe(II), Mn, and Ni had no significant effect on dmeF expression ( Figure 3). Expression of dmeF is reported relative to the H2O treatment. Results represent the means and standard deviations (SD) from three biological replicates. The data were analyzed using one-way analysis of variance along with Bonferroni's post-test. **, p < 0.01; ***, p < 0.001.

DmeF Contributes to V. parahaemolyticus Growth under High-Cobalt Conditions
To explore the role of DmeRF in V. parahaemolyticus physiology, we generated gene deletion mutants (including the single mutants ΔdmeR and ΔdmeF, and the double mutant ΔdmeRF) and overexpression strains in the corresponding mutant's background (OEdmeR, OEdmeF, and OEdmeRF). These strains were verified by PCR analysis and DNA sequencing.
The sensitivities of the wild type (WT), mutants, and overexpression strains to various Co concentrations were determined by growth curves analysis. In the absence of Co, all the strains displayed similar growth ( Figure 4A). However, in the presence of elevated concentrations of Co, the ΔdmeF and ΔdmeRF mutants exhibited severely decreased growth compared to the WT strain ( Figure 4B−D). The ΔdmeR mutant exhibited similar growth compared to the WT strain, while the OEdmeF and OEdmeRF strains, in which dmeF expression is significantly upregulated ( Figure 5), grew much better than the WT strain ( Figure 4B−D). Surprisingly, decreased OEdmeR growth was also observed in the presence of elevated Co concentrations ( Figure 4B  Results represent the means and standard deviations (SD) from three biological replicates. The data were analyzed using one-way analysis of variance along with Bonferroni's post-test. **, p < 0.01; ***, p < 0.001.

DmeF Contributes to V. parahaemolyticus Growth under High-Cobalt Conditions
To explore the role of DmeRF in V. parahaemolyticus physiology, we generated gene deletion mutants (including the single mutants ∆dmeR and ∆dmeF, and the double mutant ∆dmeRF) and overexpression strains in the corresponding mutant's background (OEdmeR, OEdmeF, and OEdmeRF). These strains were verified by PCR analysis and DNA sequencing.
The sensitivities of the wild type (WT), mutants, and overexpression strains to various Co concentrations were determined by growth curves analysis. In the absence of Co, all the strains displayed similar growth ( Figure 4A). However, in the presence of elevated concentrations of Co, the ∆dmeF and ∆dmeRF mutants exhibited severely decreased growth compared to the WT strain ( Figure 4B−D). The ∆dmeR mutant exhibited similar growth compared to the WT strain, while the OEdmeF and OEdmeRF strains, in which dmeF expression is significantly upregulated ( Figure 5), grew much better than the WT strain ( Figure 4B−D). Surprisingly, decreased OEdmeR growth was also observed in the presence of elevated Co concentrations ( Figure 4B   The sensitivities of the WT, ΔdmeF, and OEdmeF strains to various other metals were also evaluated. Upon supplementation of Fe(II), Mn, Zn, Cu, and Ni, ΔdmeF displayed almost identical growth compared to the WT strain ( Figure S2). OEdmeF exhibited either slight or moderate growth decreases in the presence of trisodium citrate dihydrate (TCD, as a control for the Fe[II] treatment), Fe(II), Mn, Zn, and Cu, whereas it showed increased growth compared to the WT strain in the presence of Ni ( Figure S2A-F). Taken together, these results indicate that DmeF is involved in the resistance of V. parahaemolyticus to Co toxicity.

The ΔdmeF Mutant Accumulated Increased Levels of Cellular Cobalt Content
To better understand the mechanism underlying DmeF-mediated Co resistance, the WT, ΔdmeF, and OEdmeF strains grown in the presence of 0.1 mM Co were collected and   The sensitivities of the WT, ΔdmeF, and OEdmeF strains to various other metals were also evaluated. Upon supplementation of Fe(II), Mn, Zn, Cu, and Ni, ΔdmeF displayed almost identical growth compared to the WT strain ( Figure S2). OEdmeF exhibited either slight or moderate growth decreases in the presence of trisodium citrate dihydrate (TCD, as a control for the Fe[II] treatment), Fe(II), Mn, Zn, and Cu, whereas it showed increased growth compared to the WT strain in the presence of Ni ( Figure S2A-F). Taken together, these results indicate that DmeF is involved in the resistance of V. parahaemolyticus to Co toxicity.

The ΔdmeF Mutant Accumulated Increased Levels of Cellular Cobalt Content
To better understand the mechanism underlying DmeF-mediated Co resistance, the WT, ΔdmeF, and OEdmeF strains grown in the presence of 0.1 mM Co were collected and The sensitivities of the WT, ∆dmeF, and OEdmeF strains to various other metals were also evaluated. Upon supplementation of Fe(II), Mn, Zn, Cu, and Ni, ∆dmeF displayed almost identical growth compared to the WT strain ( Figure S2). OEdmeF exhibited either slight or moderate growth decreases in the presence of trisodium citrate dihydrate (TCD, as a control for the Fe[II] treatment), Fe(II), Mn, Zn, and Cu, whereas it showed increased growth compared to the WT strain in the presence of Ni ( Figure S2A-F). Taken together, these results indicate that DmeF is involved in the resistance of V. parahaemolyticus to Co toxicity.

The ∆dmeF Mutant Accumulated Increased Levels of Cellular Cobalt Content
To better understand the mechanism underlying DmeF-mediated Co resistance, the WT, ∆dmeF, and OEdmeF strains grown in the presence of 0.1 mM Co were collected and analyzed for Co content using inductively coupled plasma-mass spectrometry (ICP-MS).
As seen in Figure 6, the WT and ∆dmeF strains accumulated approximately 65 and 87 µg Co per g cells (dry weight), respectively. The OEdmeF strain, which had a higher expression of dmeF, accumulated only 6 µg Co per g cells (dry weight) ( Figure 6). The cellular Co content in ∆dmeF was significantly higher than that in the WT and OEdmeF strains. Moreover, when growing in the presence of 1 mM Ni, ∆dmeF accumulated higher Ni content than the WT strain (although not significantly) and OEdemF (significantly) ( Figure S3). analyzed for Co content using inductively coupled plasma-mass spectrometry (ICP-MS). As seen in Figure 6, the WT and ΔdmeF strains accumulated approximately 65 and 87 μg Co per g cells (dry weight), respectively. The OEdmeF strain, which had a higher expression of dmeF, accumulated only 6 μg Co per g cells (dry weight) ( Figure 6). The cellular Co content in ΔdmeF was significantly higher than that in the WT and OEdmeF strains. Moreover, when growing in the presence of 1 mM Ni, ΔdmeF accumulated higher Ni content than the WT strain (although not significantly) and OEdemF (significantly) ( Figure  S3). Figure 6. Cellular Co content in the WT, ΔdmeF, and OEdmeF strains. These strains were grown in the presence of 0.1 mM CoSO4 for 6 h. Cellular Co content was analyzed by inductively coupled plasma-mass spectrometry. Results represent the means and SD from five biological replicates. The data were analyzed using one-way analysis of variance along with Bonferroni's post-test. ***, p < 0.001.

DmeR Negatively Regulates the dmeRF Operon by Binding Directly to the Promoter While Cobalt Inhibits the Interaction
To test whether DmeR regulates the dmeRF operon, dmeF expression in the WT, ΔdmeR, and OEdmeR strains was determined by qRT-PCR analysis. In ΔdmeR, dmeF expression was approximately 5-fold and 31-fold higher than that of the WT and OEdmeR strains, respectively ( Figure 7A).  . Cellular Co content in the WT, ∆dmeF, and OEdmeF strains. These strains were grown in the presence of 0.1 mM CoSO 4 for 6 h. Cellular Co content was analyzed by inductively coupled plasmamass spectrometry. Results represent the means and SD from five biological replicates. The data were analyzed using one-way analysis of variance along with Bonferroni's post-test. ***, p < 0.001.

DmeR Negatively Regulates the dmeRF Operon by Binding Directly to the Promoter While Cobalt Inhibits the Interaction
To test whether DmeR regulates the dmeRF operon, dmeF expression in the WT, ∆dmeR, and OEdmeR strains was determined by qRT-PCR analysis. In ∆dmeR, dmeF expression was approximately 5-fold and 31-fold higher than that of the WT and OEdmeR strains, respectively ( Figure 7A).
The promoter region of dmeRF contains a predicted binding site for DmeR that consists of a DNA sequence flanked by two inverted repeats ( Figure S4A). The promotion of dmeRF was mutated by replacing the putative binding site with an unrelated DNA sequence, to generate P dmeRF ' ( Figure S4B). The WT and ∆dmeR strains harboring either pDM8 (a plasmid carrying a promoterless lacZ gene), P dmeRF -pDM8 (pDM8 carrying the promoter of dmeRF), or P dmeRF '-pDM8 (pDM8 carrying the mutant promoter of dmeRF) were measured for β-galactosidase activity. As shown in Figure 7B, the β-galactosidase activity produced by ∆dmeR harboring P dmeRF -pDM8 (2039 Miller units) was significantly higher than that produced by the WT strain harboring the same plasmid (743 Miller units). In addition, the strain (either WT or ∆dmeR) harboring P dmeRF '-pDM8 produced β-galactosidase activity comparable to that harboring pDM8 ( Figure 7B).
To further elucidate the regulatory mechanism of DmeR, electrophoretic mobility shift assays (EMSAs) were performed. Purified recombinant DmeR (rDmeR) was incubated with the promoter probes of either dmeRF (WT or mutant) or gyrB (as a negative control) in buffers containing EDTA or Co. As seen in Figure 8A, when 0.1 mM EDTA was present in the reactions, a single shifted band became visible as the concentration of rDmeR increased. In contrast, no shifted band was observed in reactions containing 0.5 mM Co ( Figure 8B). Moreover, when incubated with the mutant promoter probe of dmeRF, no shifted band was observed in reactions containing EDTA ( Figure 8C). Regardless of the presence of EDTA or Co, rDmeR was unable to shift the negative control probes ( Figure 8A-C).

DmeR Negatively Regulates the dmeRF Operon by Binding Directly to the Promoter While Cobalt Inhibits the Interaction
To test whether DmeR regulates the dmeRF operon, dmeF expression in the WT, ΔdmeR, and OEdmeR strains was determined by qRT-PCR analysis. In ΔdmeR, dmeF expression was approximately 5-fold and 31-fold higher than that of the WT and OEdmeR strains, respectively ( Figure 7A).  (B) Assays measuring the β-galactosidase activities of early-exponential phase cells of the WT and ∆dmeR strains harboring either pDM8 (a plasmid carrying a promoterless lacZ gene), P dmeRF -pDM8 (pDM8 carrying the promoter of dmeRF), or P dmeRF '-pDM8 (pDM8 carrying the mutant promoter of dmeRF). Results represent the means and SD from three independent experiments performed in duplicate. The data were analyzed using one-way analysis of variance along with Bonferroni's post-test. ns, not significant; *, p < 0.05; ***, p < 0.001. ΔdmeR strains harboring either pDM8 (a plasmid carrying a promoterless lacZ gene), PdmeRF-pDM8 (pDM8 carrying the promoter of dmeRF), or PdmeRF'-pDM8 (pDM8 carrying the mutant promoter of dmeRF). Results represent the means and SD from three independent experiments performed in duplicate. The data were analyzed using one-way analysis of variance along with Bonferroni's posttest. ns, not significant; *, p < 0.05; ***, p < 0.001.
The promoter region of dmeRF contains a predicted binding site for DmeR that consists of a DNA sequence flanked by two inverted repeats ( Figure S4A). The promotion of dmeRF was mutated by replacing the putative binding site with an unrelated DNA sequence, to generate PdmeRF' (Figure S4B). The WT and ΔdmeR strains harboring either pDM8 (a plasmid carrying a promoterless lacZ gene), PdmeRF-pDM8 (pDM8 carrying the promoter of dmeRF), or PdmeRF'-pDM8 (pDM8 carrying the mutant promoter of dmeRF) were measured for β-galactosidase activity. As shown in Figure 7B, the β-galactosidase activity produced by ΔdmeR harboring PdmeRF-pDM8 (2039 Miller units) was significantly higher than that produced by the WT strain harboring the same plasmid (743 Miller units). In addition, the strain (either WT or ΔdmeR) harboring PdmeRF'-pDM8 produced β-galactosidase activity comparable to that harboring pDM8 ( Figure 7B).
To further elucidate the regulatory mechanism of DmeR, electrophoretic mobility shift assays (EMSAs) were performed. Purified recombinant DmeR (rDmeR) was incubated with the promoter probes of either dmeRF (WT or mutant) or gyrB (as a negative control) in buffers containing EDTA or Co. As seen in Figure 8A, when 0.1 mM EDTA was present in the reactions, a single shifted band became visible as the concentration of rDmeR increased. In contrast, no shifted band was observed in reactions containing 0.5 mM Co ( Figure 8B). Moreover, when incubated with the mutant promoter probe of dmeRF, no shifted band was observed in reactions containing EDTA ( Figure 8C). Regardless of the presence of EDTA or Co, rDmeR was unable to shift the negative control probes (Figures 8A-C). promoter probes of either dmeRF or gyrB (negative control) in buffers containing EDTA (A) or Co (B). (C) Purified rDmeR was incubated with the promoter probes in buffers containing EDTA. The mutant demRF promoter probe was generated by replacing the putative binding site for DmeR with an unrelated DNA sequence. The rDmeR was added to each reaction mixture in the amounts indicated. The images are representative of at least three independent experiments.
Together, these results indicate that DmeR represses the dmeRF operon by binding directly to its promoter in the absence of Co.

The DmeRF System Is Not Required for V. parahaemolyticus Virulence in Mice
To investigate the role of DmeRF in V. parahaemolyticus virulence, an experimental infection of C57BL/6 mice was conducted. Ten mice per treatment were intraperitoneally infected with either phosphate-buffered saline (PBS) or one of four V. parahaemolyticus strains: the WT, ∆dmeR, ∆dmeF, or ∆dmeRF strains. At 12 h post infection, the survival rates for mice in the WT, ∆dmeR, ∆dmeF, and ∆dmeRF groups were 40%, 50%, 20%, and 20%, respectively ( Figure 9). The remaining mice in the WT and ∆dmeRF groups died during the following 12 h, and 10% of the mice infected with the ∆dmeR or ∆dmeF strains survived over the course of the experiment (Figure 9). In contrast, all mice injected with PBS survived. These results suggest that the DmeRF system has no significant role in V. parahaemolyticus virulence in mice.  Together, these results indicate that DmeR represses the dmeRF operon by binding directly to its promoter in the absence of Co.

The DmeRF System Is Not Required for V. parahaemolyticus Virulence in Mice
To investigate the role of DmeRF in V. parahaemolyticus virulence, an experimental infection of C57BL/6 mice was conducted. Ten mice per treatment were intraperitoneally infected with either phosphate-buffered saline (PBS) or one of four V. parahaemolyticus strains: the WT, ΔdmeR, ΔdmeF, or ΔdmeRF strains. At 12 h post infection, the survival rates for mice in the WT, ΔdmeR, ΔdmeF, and ΔdmeRF groups were 40%, 50%, 20%, and 20%, respectively (Figure 9). The remaining mice in the WT and ΔdmeRF groups died during the following 12 h, and 10% of the mice infected with the ΔdmeR or ΔdmeF strains survived over the course of the experiment (Figure 9). In contrast, all mice injected with PBS survived. These results suggest that the DmeRF system has no significant role in V. parahaemolyticus virulence in mice. Figure 9. Survival curves of mice infected with one of four V. parahaemolyticus strains or PBS. Ten mice per treatment were intraperitoneally infected with 1 × 10 8 CFU of the WT, ΔdmeR, ΔdmeF, or ΔdmeRF strains, or injected with 100 μL of PBS as the control. The data were analyzed using the logrank test.

Discussion
Even though the maintenance of metal homeostasis is essential for bacterial physiology and pathogenesis, not much is known about the mechanisms by which V. parahaemolyticus responds to metal overload. Metal efflux is one of the most important mechanisms employed by bacteria to reduce the damage caused by metal influxes [4]. The DmeRF system is a well-characterized metal-response system that is involved in Co resistance in C. metallidurans and Co/Ni resistance in certain Rhizobiaceae species [14,[18][19][20][21]. The DmeRF system was demonstrated to contribute to the maintenance of Co homeostasis in V. parahaemolyticus by providing the following lines of evidence: (i) V. parahaemolyticus DmeF shares a high level of homology (approximately 44% to 51% amino acid sequence identity) with its homologues, all of which are involved in Co (and Ni) resistance; (ii) dmeF expression is significantly upregulated in response to Zn, Cu, and Co exposure; (iii) the dmeF deletion mutants exhibit increased sensitivity to Co stress, while the overexpression strains that have higher dmeF expression exhibit decreased sensitivity to Co stress; (iv) when cultured in a medium supplemented with Co, the cellular Co content in the ΔdmeF mutant is significantly higher than that in the WT and OEdmeF strains; and (v) in the Figure 9. Survival curves of mice infected with one of four V. parahaemolyticus strains or PBS. Ten mice per treatment were intraperitoneally infected with 1 × 10 8 CFU of the WT, ∆dmeR, ∆dmeF, or ∆dmeRF strains, or injected with 100 µL of PBS as the control. The data were analyzed using the log-rank test.

Discussion
Even though the maintenance of metal homeostasis is essential for bacterial physiology and pathogenesis, not much is known about the mechanisms by which V. parahaemolyticus responds to metal overload. Metal efflux is one of the most important mechanisms employed by bacteria to reduce the damage caused by metal influxes [4]. The DmeRF system is a well-characterized metal-response system that is involved in Co resistance in C. metallidurans and Co/Ni resistance in certain Rhizobiaceae species [14,[18][19][20][21]. The DmeRF system was demonstrated to contribute to the maintenance of Co homeostasis in V. parahaemolyticus by providing the following lines of evidence: (i) V. parahaemolyticus DmeF shares a high level of homology (approximately 44% to 51% amino acid sequence identity) with its homologues, all of which are involved in Co (and Ni) resistance; (ii) dmeF expression is significantly upregulated in response to Zn, Cu, and Co exposure; (iii) the dmeF deletion mutants exhibit increased sensitivity to Co stress, while the overexpression strains that have higher dmeF expression exhibit decreased sensitivity to Co stress; (iv) when cultured in a medium supplemented with Co, the cellular Co content in the ∆dmeF mutant is significantly higher than that in the WT and OEdmeF strains; and (v) in the presence of Co, DmeR dissociates from the dmeRF promoter, allowing the transcription of dmeF.
The role of the DmeRF system in the metal tolerances of C. metallidurans and certain Rhizobiaceae species is well studied [14,[18][19][20][21]. Therefore, BlastP analyses and multiple sequence alignments were performed to examine the level of homology exhibited between the DmeRF systems of V. parahaemolyticus and their homologues in these other species. While the DmeF proteins share high levels of homology, no significant similarity was observed between the DmeR of V. parahaemolyticus and those of the Rhizobiaceae species. It is not surprising as V. parahaemolyticus DmeR is a MarR family regulator while the others belong to the RcnR/CsoR family. Moreover, the organization of the V. parahaemolyticus demRF operon is different from those of the Rhizobiaceae species. Consequently, it is worthwhile to explore the role of DmeRF in V. parahaemolyticus.
Typically, bacteria respond to metal excess by expressing specific genes. Therefore, dmeF expression in the presence of elevated levels of various metals was measured using qRT-PCR analysis. dmeF expression is induced by Zn, Cu, and Co, with Co serving as the most potent inducer. The expression of dmeF differs among V. parahaemolyticus, C. metallidurans, and the Rhizobiaceae species. In C. metallidurans, dmeF expression is constitutive and cannot be induced by metals [18]. By contrast, dmeF expression is strongly induced by Co, Ni, and Cu in S. meliloti [19], and specifically induced by Co and Ni in A. fabrum and R. leguminosarum, with Co being a more potent inducer [14,20]. Although Cu induces demF expression in S. meliloti, the ∆dmeF mutant exhibited no difference in growth compared with the WT strain under high-Cu conditions [19]. Similarly, in V. parahaemolyticus, there was no major difference in the growth of the ∆dmeF and WT strains under high-Zn or -Cu conditions. A similar observation has also been made for the Fe(II) and Co efflux pump PmtA in S. suis. While pmtA expression is induced by Fe(II), Co, and Ni, the ∆pmtA mutant displayed no growth decrease under Ni stress [15]. We speculate that excessive amounts of Zn or Cu may change DmeR conformation, resulting in partial derepression of the dmeRF operon.
In line with the induction of dmeF expression by Co, the dmeF deletion mutants exhibited obvious growth inhibition under high-Co conditions, whereas, the dmeF overexpression strains grew better than the WT strain in the presence of elevated Co levels. These results clearly suggest that DmeF is involved in Co resistance in V. parahaemolyticus, which is consistent with the observations in C. metallidurans and certain Rhizobiaceae species [14,[18][19][20][21]. In A. fabrum, the ∆dmeF mutant accumulated significantly higher levels of cellular Co content than the WT strain when cultured in a Co-rich medium [14]. A similar observation was made in our study, indicating that DmeF potentially mediates Co resistance by Co efflux. Yet, growth of the ∆dmeF and ∆dmeRF strains was not completely inhibited by high concentrations of Co, suggesting that additional Co resistance systems likely exist in V. parahaemolyticus. Surprisingly, the OEdmeR strain, which showed very low dmeF expression, exhibited more severe growth inhibition than ∆dmeF and ∆dmeRF under Co stress. We speculate that DmeR may play a role in the regulation of other Co resistance systems. Interestingly, the OEdmeF strain exhibited increased growth compared to the WT strain in the presence of high concentrations of Ni. Consistent with their growth, OEdmeF accumulated significantly lower levels of cellular Ni content under Ni conditions. These results indicate that in V. parahaemolyticus, DmeF may be involved in Ni resistance, albeit in a less prominent role than in Co resistance.
qRT-PCR analysis, β-galactosidase activity assays, and EMSAs showed that DmeR, a MarR family regulator, represses the dmeRF operon by binding directly to the promoter in the absence of Co. In several studied Rhizobiaceae species, dmeF is co-transcribed with a gene encoding the RcnR/CsoR family regulator, whose repression of dmeRF transcription has been confirmed in A. fabrum [14,19,20]. Accordingly, we speculate that V. parahaemolyticus and Rhizobiaceae adopt regulators belonging to different families to modulate the conserved Co resistance system. Furthermore, the mechanisms of the DmeRF system in V. parahaemolyticus were proposed according to our results and the findings in A. fabrum [14]. In the presence of limited or normal concentrations of Co, DmeR binds to the promoter region of dmeRF, resulting in transcriptional repression, whereas under high-Co conditions, excessive amounts of Co result in DmeR dissociating from the promoter, probably by causing a conformational change, and consequently, the repression is relieved.
In A. fabrum, the inactivation of dmeF has no effect on bacterial virulence in Nicotiana benthamiana [14]. Likewise, the DmeRF system plays no significant role in V. parahaemolyticus virulence in mice. We speculate that the Co concentration is low in mouse tissues, hence the DmeRF system does not serve an important function during the infectious process.
In conclusion, a metal-response system, DmeRF, composed of the metal efflux pump DmeF and the regulator DmeR, has been identified and characterized in V. parahaemolyticus. The DmeRF system contributes to the maintenance of Co homeostasis in V. parahaemolyticus, and DmeR functions as a transcriptional repressor of the dmeRF operon in the absence of Co.

Bacterial Strains, Culture Conditions, Plasmids, and Primers
The bacterial strains and plasmids used in this study are listed in Table 1. All Escherichia coli strains, V. parahaemolyticus RIMD 2210633, and its derivatives were routinely grown at 37 • C in Luria-Bertani (LB) broth or on LB agar. When required, carbenicillin, chloramphenicol, and isopropyl β-D-1-thiogalactopyranoside (IPTG) were supplemented at 50 µg/mL, 25 µg/mL, and 1 mM, respectively. The primers used in this study are listed in Table 2.  The promoter of gyrB PgyrB-R TCTATCCTGCCATGTTCCAC 1 The underlined sequences are restriction sites.

RNA Extraction and qRT-PCR Analysis
An overnight culture of the RIMD 2,210,633 strain was diluted 1:100 in LB broth and grown to the early-exponential phase (OD 600 of~0.7). Seven 1 mL aliquots were removed, 2 µL of H 2 O was added to one and metal solutions were added to the rest, one to each, to create final concentrations of 1 mM FeSO 4 , 1 mM MnSO 4 , 0.5 mM ZnSO 4 , 1 mM CuSO 4 , 0.25 mM CoSO 4 , and 1 mM NiSO 4 . After further incubation at 37 • C for 15 min, bacterial cells were collected by centrifugation. Total RNA was isolated from the cell pellets using an Eastep Super Total RNA Isolation Kit (Promega, Shanghai, China). Three independent experiments were performed to obtain triplicate biological samples.
In another experiment, overnight cultures of the WT, ∆dmeR, and OEdmeR were separately diluted 1:100 in LB broth and grown to the early-exponential phase (OD 600 of 0.7). Then, bacterial cells were collected for RNA isolation as described above. Three independent experiments were performed to obtain triplicate biological samples. After evaluations of RNA integrity and measurements of RNA concentrations, the qualified RNAs were subjected to qRT-PCR analysis. cDNA was generated from approximately 200 ng of RNA per sample using ToloScript RT EasyMix for qPCR (with 2-step gDNA Erase-Out) (TOLOBIO, Shanghai, China). Quantitative PCR was performed on the StepOnePlus Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) using NovoStart SYBR qPCR SuperMix Plus (Novoprotein, Shanghai, China) and gene-specific primers ( Table 2). Gene expression levels were analyzed using the 2 − ∆∆CT method [32], with gyrB as the internal standard.

Growth Evaluation
The WT, ∆dmeR, ∆dmeF, ∆dmeRF, OEdmeR, OEdmeF, and OEdmeRF strains were grown to the mid-exponential phase (OD 600 of~2) and diluted 1:100 in LB broth Supplemented with one of four concentrations of CoSO 4 (0, 0.1, 0.2, or 0.3 mM) or specific high concentrations of another metal (2 mM FeSO 4 , 1 mM MnSO 4 , 1.25 mM ZnSO 4 , 2 mM CuSO 4 , or 1 mM NiSO 4 ). To alleviate Fe(II) oxidation, 1 g/L of TCD was supplemented to the medium containing FeSO 4 [34]. The cultures were transferred into 96-well plates with 200 µL per well and three wells per treatment. The plates were incubated in a shaker at 37 • C and 120 rpm, and the OD 595 values were measured hourly using a CMax Plus plate reader (Molecular Devices, San Jose, CA, USA).

Intracellular Metal Content Analysis
The WT, ∆dmeF, and OEdmeF strains were grown to the mid-exponential phase (OD 600 of~2) and diluted 1:100 in LB broth supplemented with either 0.1 mM CoSO 4 or 1 mM NiSO 4 . After incubation for another 6 h, bacterial cells were collected by centrifugation. Five independent experiments were performed to obtain five biological samples for each strain. Sample washing, drying, digestion, and dilution were performed as previously described [35]. Co/Ni content in these samples was analyzed by ICP-MS at Yangzhou University. The metal content was expressed as µg of Co/Ni per g of cells (dry weight).

Construction of LacZ Fusion Strains and β-galactosidase Activity Assays
The promoter of the dmeRF operon was mutated by replacing the putative binding site for DmeR with an unrelated DNA sequence. The DNA carrying the mutant promoter of dmeRF (P dmeRF ') was synthesized by Tsingke Biotechnology Co., Ltd. (Beijing, China).
The promoter of dmeRF (either WT or mutant) was cloned into pDM8, a plasmid carrying a promoterless lacZ gene [31]. The resulting plasmid P dmeRF -pDM8/P dmeRF '-pDM8 was transformed into E. coli S17-1 λpir and then conjugated into the WT strain and ∆dmeR. The WT strain and ∆dmeR harboring the empty pDM8 plasmid served as the control strains.
β-galactosidase activity assays were performed as previously described [36,37], with some modifications. Overnight cultures of the LacZ fusion strains were diluted 1:100 in LB broth and grown to the early-exponential phase (OD 600 of~0.7). Bacterial cells were collected from 1 mL of each culture by centrifugation. The cell pellets were resuspended in 1 mL PM buffer (60 mM Na 2 HPO 4 , 40 mM NaH 2 PO 4 , 10 mM KCl, 1 mM MgSO 4 , and 50 mM β-mercaptoethanol, pH 7.0). The A 600 values of the bacterial suspensions were measured. For each suspension, 200 µL of suspension, 30 µL of chloroform, and 30 µL of 0.1% SDS were added to 500 µL of PM buffer, and then vortexed vigorously to lyse bacterial cells. The reaction was started by adding 200 µL of o-nitrophenyl-β-galactopyranoside (4 mg/mL in PM buffer). When the mixture turned yellowish, the reaction was stopped by adding 400 µL of 1 M Na 2 CO 3 . The mixture was centrifuged, and the A 420 value of the supernatant was measured. β-galactosidase activity, in Miller units, was calculated as A 420 × 1000 × min −1 × mL −1 × A 600 −1 .

rDmeR Expression, Purification, and EMSAs
The dmeR gene was cloned into the pET-30a plasmid, and the resulting plasmid, pET30a-dmeR, was transformed into E. coli BL21(DE3). The strain was grown at 37 • C to the mid-exponential phase (OD 600 of~0.8). Then, 0.5 mM of isopropyl-β-D-thiogalactopyra noside (IPTG) was added to induce DmeR expression. After further growth at 28 • C for 4 h, bacterial cells were harvested by centrifugation. The cell pellets were resuspended in binding buffer (20 mM Tris-HCl, 500 mM NaCl, and 20 mM imidazole, pH 8.0) and lysed by sonication. rDmeR purification and assessment of its quality and concentration were performed as previously described [38].
EMSAs were performed as previously described [38]. DNA fragments containing the promoters of the dmeRF operon (WT or mutant) or gyrB (negative control) were PCR amplified and then purified. The 20 µL reaction mixtures contained 20 ng of the DNA fragments, varying amounts of rDmeR (0, 0.1, 0.2, 0.5, or 1.0 µg), and 200 ng of poly (dI:dC) in either EMSA buffer 1 (150 mM NaCl, 0.1 mM DTT, 0.1 mM EDTA, and 10 mM Tris, pH 7.4) or the Co containing EMSA buffer 2 (150 mM NaCl, 0.1 mM DTT, 0.5 mM CoSO 4 , and 10 mM Tris, pH 7.4). The mixtures were incubated at 25 • C for 30 min and then resolved by 6% native polyacrylamide gel electrophoresis in 0.5 × TBE buffer at 100 V for 2 h. After staining with SYBR Green I for 30 min, the gel was photographed.

Mouse Infection Experiment
A total of 50 female C57BL/6 mice (specific, pathogen-free, 8−12 weeks old) were randomly divided into five groups (10 mice per group). The WT, ∆dmeR, ∆dmeF, and ∆dmeRF strains were grown in LB broth at 30 • C for 12 h and adjusted to 1 × 10 9 CFU/mL in PBS. For groups I to IV, the mice were intraperitoneally infected with 100 µL of the corresponding strain. Mice in group V were injected with 100 µL of PBS and served as the control. Mouse survival was recorded twice daily for seven days.
GraphPad Prism 5 (San Diego, CA, USA) was used for statistical analysis. Gene expression, metal content, and β-galactosidase activity were analyzed by one-way analysis of variance with a Bonferroni's post-test. The log-rank test was used for analyzing mouse survival curves.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms24010414/s1, Figure S1: Co-transcription analysis of the dmeR and dmeF genes; Figure S2: Growth curves analysis of the V. parahaemolyticus strains in the presence of various metals. Figure S3: Cellular Ni content in the WT, ∆dmeF, and OEdmeF strains. Figure S4: Mutation of the promoter of the dmeRF operon.