Defining the regulatory mechanism of NikR, a nickel-responsive transcriptional regulator, in Brucella abortus

Abstract

Metals are essential micronutrients for virtually all forms of life, but metal acquisition is a double-edged sword, because high concentrations of divalent cations can be toxic to the cell. Therefore, the genes involved in metal acquisition, storage and efflux are tightly regulated. The present study characterizes a nickel-responsive transcriptional regulator in the intracellular mammalian pathogen, Brucella abortus. Deletion of bab2_0432 (nikR) in B. abortus led to alterations in the nickel-responsive expression of the genes encoding the putative nickel importer NikABCDE and, moreover, NikR binds directly to a specific DNA sequence within the promoter region of nikA in a metal-dependent manner to control gene expression. While NikR is involved in controlling the expression of nikA, nikR is not required for the infection of macrophages or mice by B. abortus. Overall, this work characterizes the role of NikR in nickel-responsive gene expression, as well as the dispensability of nikR for Brucella virulence.

Full-Text

Metals serve as essential cofactors for many cellular proteins, and the ability of a cell to acquire metals from the environment effectively is essential to their survival [1]. Although necessary for survival, metal acquisition can also cause serious harm to a cell if homeostasis is not maintained. If the concentration of intracellular metal is too low, proteins requiring those metals will not function properly; however, if the intracellular concentration of metals is too high, proteins can incorrectly bind to the wrong metal cofactor, leading to the disruption of protein function [2, 3]. Because of this, bacteria and other organisms have developed systems to tightly regulate intracellular metal homeostasis. One method for regulating metal homeostasis is to control the expression of metal import and export systems through the use of transcriptional regulatory proteins [4]. Indeed, transcriptional regulators are important for coordinating proper metal homeostasis in the intracellular mammalian pathogen Brucella.

Brucella spp. are Gram-negative intracellular bacteria belonging to the class α-proteobacteria [5]. Brucella spp. infect a variety of domesticated and wild animals, and several species have the ability to infect humans following contact with infectious tissue or consumption of contaminated unpasteurized dairy products [6]. Infection of animals can lead to abortions and infertility, while infection of humans leads to an undulating fever that can become chronic when left untreated [7]. Upon entering the host, Brucella invade its immune cells, where they form their replicative niche [8]. While the intracellular environment shields the brucellae from the host’s immune response, the brucellae encounter less than ideal conditions for survival, including acidic pH, diminished oxygen concentrations, increased levels of reactive oxygen species and limited availability of nutrients, including metals [9].

If ingested, Brucella spp. must survive passage through the stomach. In order to surpass the acidic environment of the stomach, Brucella spp. utilize the enzyme urease, which converts urea and water to carbon dioxide and ammonia to increase the dangerously low pH [10, 11]. The brucellae genome contains two urease-encoding loci (designated ure1 and ure2); however, only ure1 is necessary for virulence, and it is hypothesized that ure2 remains present in the brucellae genome due to the presence of nickel import genes located at the 3′ end of the ure2 transcript [10–12]. Interestingly, urease is one of the few known enzymes that requires nickel as a metal cofactor for proper function [12, 13].

Metal acquisition is critical for the survival of Brucella during infection and trafficking to its intracellular niche. Iron, magnesium, manganese and zinc have been the focus of much of the previous Brucella research on metals, primarily due to the necessity of their acquisition for virulence [2, 14–19]. While there is little experimental evidence demonstrating nickel homeostasis in Brucella, there is an abundance of bioinformatic analyses proposing putative genes related to nickel transport and regulation. Brucella spp. encode two putative nickel importers, NikABCDE and NikKMLQO, and a putative nickel-responsive regulator, NikR [20]. Interestingly, nikKMLQO is transcribed with the ure2 operon [11]. However, it should be noted that while nikABCDE is encoded in a gene cluster that is hypothesized to be transcribed by one promoter upstream of nikA, empirical data have not demonstrated that nikABCDE is a bona fide operon [13]. The objective of the presented work was to characterize the regulatory mechanism and functional activity of the putative B. abortus nickel-responsive transcriptional regulator, NikR, a well-studied metalloregulatory protein for metal homeostasis and virulence in other pathogenic bacteria [21, 22].

bab2_0432 is annotated as a nickel-responsive regulator in GenBank and has previously been hypothesized to encode the nickel-responsive regulator NikR in Brucella spp. [13]. nikR is encoded divergently from the nikABCDE gene cluster on chromosome II of B. abortus 2308 (officially known as B. melitensis biovar Abortus 2308) (Fig. 1a). It is important to note that in B. abortus, nikA is a pseudogene due to a nonsense mutation, separating the gene into two pseudogenes (denoted in Fig. 1a in grey) [23]. An isogenic, unmarked, in-frame gene deletion of nikR was constructed as described previously [15]. The primers used to generate the deletion construct of nikR include: nikR-Up-For, nikR-Up-Rev, nikR-Dn-For and nikR-Dn-Rev (Table S1, available in the online version of this article).

Fig. 1.

NikR is a repressor of nikA expression in B. abortus. (a) Genetic organization of the bab2_0432-bab2_0438 locus in B. abortus 2308. bab2_0432 encodes the putative nickel responsive regulator NikR. bab2_0433-bab2_0438 encodes the putative nickel ABC transport system nikABCDE. nikABCDE encodes a putative periplasmic nickel-binding protein (nikA), putative ATPases (nikD and nikE) and putative inner-membrane proteins (nikB and nikC). Grey arrows indicate pseudogenes. Nucleotide sequence directly upstream of nikA and underlined probes representing oligos utilized in Fig. 2(d). Bolded sequence represents imperfect inverted repeated sequences and putative NikR-binding box. (b) Quantitative RT-PCR was used to assess the relative expression levels of nikA in B. abortus 2308 or B. abortus ΔnikR grown in the absence (−) or presence (+) of 100 µM NiSO₄. 16S rRNA levels were assessed as a control, and all mRNA levels were normalized to the level of the parental strain 2308 grown in the absence of NiSO₄. The bars depict the average level of specific mRNA for each strain and condition analysed in triplicate and the error bars represent the standard deviations. Statistical significance when comparing B. abortus 2308 (−NiSO₄) to subsequent strains and conditions was determined using a t-test (*, P<0.05).

Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was utilized to establish NikR as a repressor of nikA gene expression. B. abortus 2308 or ΔnikR were grown overnight to late exponential phase in Brucella broth and subsequently exposed to fresh Brucella broth +/−100 µM NiSO₄ for 20 min. RNA was then isolated from bacterial cultures as described previously [13]. qRT-PCR was conducted using primers to probe for nikA expression (RT-nikA-For, RT-nikA-Rev) and 16S rRNA expression as a control (RT-16S-For, RT-16S-Rev) (Table S1).

In the parental strain B. abortus 2308, nikA expression was significantly reduced in the presence of excess nickel (Fig. 1b). Moreover, compared to the parental strain grown in the absence of NiSO₄, nikA expression was significantly elevated in the ΔnikR strain with or without the addition of 100 µM NiSO₄. Taken together, these results demonstrate that NikR is required for the nickel-responsive repression of nikA in B. abortus 2308.

Electrophoretic mobility shift assays (EMSAs) were utilized to test the hypothesis that NikR regulates the import of nickel by binding directly to the promoter region of nikA. In these experiments, the promoters of several other putative nickel-responsive genes were also tested for binding by NikR, due to previous studies demonstrating that NikR from Helicobacter pylori interacts with multiple promoters to control the expression of nickel homeostasis genes [24]. For these experiments, recombinant NikR (rNikR) was produced with the addition of a Strep-tagII from the pASK-IBA7 cloning vector in BL21 Escherichia coli cells and subsequently purified via affinity chromatography as previously described [25]. Protein purity was accessed via SDS-PAGE (data not shown). The primers to amplify nikR for cloning into pASK-IBA7 were rNikR-For and rNikR-Rev (Table S1). EMSAs were employed to assess rNikR binding to the promoters of nikA, nikK, nikR and ureA2, as described previously, albeit with the addition of 100 µM NiSO₄ to the electrophoresis buffer, binding reaction buffer and acrylamide gel (Fig. 2a) [22]. The DNA probes corresponding to the promoter region of each gene were amplified by PCR using B. abortus 2308 genomic DNA with the following primers: nikK-For, nikK-Rev, nikA-For, nikA-Rev, nikR-For, nikR-Rev, ureA2-For and ureA2-Rev (Table S1). The DNA probes were 5′-end-labelled with [γ-³²P]ATP and polynucleotide kinase as described [22]. In these experiments, rNikR bound to the promoter region of nikA, but rNikR did not bind to the promoters of nikK, nikR or ureA2 under the conditions tested (Fig. 2a). It should be noted that without the addition of 100 µM nickel to the electrophoresis buffer, binding reaction buffer and acrylamide gel, no binding of rNikR to the nikA promoter was observed (data not shown).

Fig. 2.

NikR binds to a specific sequence in the promoter of nikA in a concentration- and nickel-dependent manner. (a) Recombinant NikR (rNikR) binds to the promoter region of nikA. Electrophoretic mobility shift assays (EMSAs) were employed using the promoter regions of Brucella abortus nikR, nikA, nikK, ure2A that were amplified by PCR and labelled with ³²P. EMSAs were performed with radiolabelled promoter region DNA with and without the addition of 2000 nM rNikR. (b) rNikR binds specifically to the nikA promoter in a concentration-dependent manner. EMSAs were performed with increasing concentrations of rNikR incubated with radiolabelled nikABCDE promoter region DNA. Some binding reactions contained unlabelled nikA promoter DNA (specific) or unlabelled nikK promoter DNA (non-specific). (c) rNikR binds to the nikA promoter in a nickel-dependent manner. EMSA performed with radiolabelled nikA promoter in the presence (+) or absence (−) of 200 nM rNikR and in the presence (+) or absence (−) of 250 µM EDTA. (d) rNikR binds to a specific sequence in the nikA promoter. EMSA performed using oligos of specific regions (probes) of the nikA promoter DNA in the presence (+) or absence (−) of 250 nM rNikR. (e) Identification of the rNikR binding box in the nikA promoter. EMSA performed with oligonucleotides of probe 2 with and without mutations, highlighted in red, of the putative NikR binding box in the presence (+) or absence (−) of 250 nM rNikR.

EMSAs were also performed to assess the specificity of binding between rNikR and the nikA promoter (Fig. 2b). rNikR bound almost completely to the nikA promoter at a concentration of 200 nM rNikR and, moreover, binding was initiated at a concentration of 50 nM. The concentration of DNA probe utilized throughout the EMSA was 1 nM. To assess the specificity of NikR–nikA interactions, binding reactions were performed in the presence of specific and non-specific promoter region DNA. When unlabelled nikA promoter was added at a 100× concentration compared to the radiolabelled nikA DNA probe, the binding of rNikR to the labelled nikA promoter was severely diminished. Alternatively, the addition of excess unlabelled nikK promoter DNA did not affect binding between rNikR and the nikA promoter. Collectively, these data demonstrate that rNikR binds to the nikA promoter region in a specific, concentration-dependent manner. Again, it remains unknown whether nikA, nikB, nikC, nikD and nikE are components of a single operon, but bioinformatic approaches predict that the nikABCDE cluster is transcribed under one promoter upstream of nikA.

As previously stated, the absence of nickel in the EMSA results in no binding between rNikR and the promoter of nikA, as binding was only facilitated in the presence of 100 µM of NiSO₄ in all aspects of the EMSA experimental design. This indicated that nickel is required for rNikR to bind to the nikA promoter. To test this hypothesis, an EMSA was performed with and without the addition of the metal chelator EDTA during the binding reaction (Fig. 2c). Similar to the EMSA experiments performed above, 100 µM NiSO₄ was added to the electrophoresis buffer, binding reaction buffer and acrylamide gel. In the presence of EDTA, no interaction between rNikR and the nikA promoter was observed, and these data are consistent with the conclusion that NikR is a metal-dependent DNA-binding protein.

Previously, DNase I footprinting assays were used to identify a specific sequence in the nikABCDE promoter region bound by NikR in E. coli [26]. Moreover, similar putative NikR-binding sites were identified in the nickel import systems of Klebsiella pneumoniae, Salmonella typhimurium, H. pylori and Bradyrhizobium japonicum. While defining the nikABCDE gene cluster and its role in urease activity in Brucella suis, Jubier-Maurin et al. described a putative NikR-binding box in the promoter region of nikA [13]. This sequence contains a putative Pribnow box, 5′-AAATAAT-3′, flanked by two imperfect inverted repeat sequences that show high similarity to the NikR-binding box in E. coli [13, 26]. Importantly, the putative NikR-binding site in B. suis is almost identical to the sequences found in the nikA promoter of B. abortus. Therefore, EMSAs were employed to test the hypothesis that the identified sequence within the nikA promoter is the authentic NikR-binding box in B. abortus. For this, EMSAs using engineered DNA probes were performed as previously described [27]. The EMSA probes were generated using designed oligonucleotide probes of 25–35 nucleotides that encompassed different regions of the nikA promoter (Fig. 1a). This experiment revealed that rNikR bound specifically to probe 2 and not to probes 1, 3, or 4 (Fig. 2d). Interestingly, probe 2 contains the imperfect inverted sequence (bold) 5′-TATATGATGTTTTTAGCAAATAATCATACT-3′, which is the putative NikR-binding box described by Jubier-Maurin et al. [13]. EMSAs measuring NikR binding to probe 2 were repeated with DNA probes containing scrambled sequences, highlighted in red, of either the first half of the inverted repeat sequence (M1), the second half of the inverted repeat sequence (M2), or both halves of the inverted repeat sequence (M1/2) (Fig. 2e). Mutating motif 1 (M1), motif 2 (M2), or motif 1 and motif 2 (M1/2) abolished the formation of the rNikR–probe2 complex (Fig. 2e). This is evidence that NikR requires the fully intact inverted repeat sequence in order to bind directly to the nikA promoter. A bioinformatic search of the B. abortus 2308 genome for similar sequences to the NikR-binding box uncovered no other promoter regions containing the NikR-binding box sequence. Altogether, these data confirm that rNikR binds directly to the promoter of nikA in a concentration- and metal-dependent manner, and that rNikR binds to a specific sequence in the B. abortus nikA promoter that is highly similar to a previously identified NikR-binding box in other bacteria [26].

The current data illustrate that the B. abortus NikR protein is a repressor of nikA expression, and that nikA is over-expressed in a nikR deletion strain in both the presence and absence of excess nickel. Thus, a deletion of nikR could result in higher concentrations of intracellular nickel when the bacterium is in a nickel-rich environment, which could be toxic for the cell. This phenomenon has been demonstrated previously for H. pylori, as the disruption of nikR led to growth inhibition upon exposure to elevated concentrations of nickel [28]. To determine the effect of nikR deletion on metal sensitivity in B. abortus, a disk diffusion assay was employed in which B. abortus 2308 and B. abortus ΔnikR were exposed to nickel (NiSO₄), copper (CuCl₂), or zinc (ZnCl₂), as described previously (Fig. 3a) [15]. Each disk was overlaid with 7 µl of a 1M solution of nickel (NiSO₄), copper (CuCl₂), or zinc (ZnCl₂). Overall, there was no significant difference between the parental strain B. abortus 2308 and B. abortus ΔnikR with regard to metal toxicity when exposed to nickel, copper, or zinc. This result was not completely unexpected due to the presence of at least one putative nickel efflux permease, rcnA, in the B. abortus genome [23]. Hypothetically, in the presence of excess nickel, RcnA could detoxify the B. abortus intracellular environment by expelling nickel ions.

Fig. 3.

ΔnikR is not significantly different in metal sensitivity or virulence when compared to the parental strain. (a) Disk diffusion assay to assess the sensitivity of Brucella strains to metals. B. abortus 2308 and ΔnikR were tested in a disk diffusion assay for susceptibility to toxic levels of the divalent cations Ni²⁺, Cu²⁺ and Zn²⁺. The results are shown as the average diameter (±standard deviation) of the zone of inhibition around a disk containing 7 µl of a 1 M solution of NiSO₄, CuCl₂, or ZnCl₂. The results are from a single representative experiment using quadruplet samples. (b) Survival and replication of B. abortus ΔnikR in primary murine macrophages. B. abortus ΔnikR was able to survive and replicate in macrophages, similarly to B. abortus 2308. B. abortus 2308 or B. abortus ΔnikR were used to infect naïve peritoneal murine macrophages at a multiplicity of infection (m.o.i.) of 100. Macrophages were lysed with 0.1 % deoxycholate 2, 24 and 48 h post-infection and lysates were serial diluted to quantify the log₁₀ c.f.u. brucellae/well of macrophages. (c) Assessment of B. abortus ΔnikR in a mouse model of chronic Brucella infection. B. abortus ΔnikR was not significantly different in its ability to infect BALB/c mice via intraperitoneal (IP) infection. BALB/c mice were infected IP with 1×10⁵ c.f.u. of either B. abortus 2308 or B. abortus ΔnikR. Spleens were extracted and homogenized 4 and 8 weeks post-infection, and the number of brucellae was calculated following serial dilution of homogenates.

To test the requirement of nikR for the full virulence of B. abortus, the ΔnikR strain was compared to the parental strain 2308 for the ability to survive and replicate in macrophages, as well as for the ability to chronically infect BALB/c mice. Naïve peritoneal macrophages were infected at a multiplicity of infection (m.o.i.) of 100 with either B. abortus 2308 or ΔnikR, and the bacterial load was accessed 2, 24 and 48 h post-infection as described previously [15, 25, 29, 30]. This experiment revealed no significant difference in the ability of ΔnikR to infect macrophages when compared to the parental strain 2308 (Fig. 3b).

BALB/c mice were infected intraperitoneally with 10⁵ c.f.u. of either B. abortus 2308 or B. abortus ΔnikR to assess the ability of the brucellae to infect the mice chronically. Each strain was used to infect five mice. Spleens were removed 4 and 8 weeks post-infection and homogenized, and serial dilutions were plated on agar medium to calculate the brucellae/spleen. Compared to the parental strain 2308, the nikR deletion strain exhibited no significant difference in the ability to colonize the spleens of BALB/c mice (Fig. 3c). Overall, a deletion of nikR in B. abortus does not alter the ability of B. abortus to survive and replicate in peritoneally derived macrophages, or to colonize the spleens of BALB/c mice infected intraperitoneally.

In conclusion, this study demonstrates that NikR is a metal-dependent repressor of the expression of the putative nickel importer nikABCDE in B. abortus (Figs 1 and 2). While it is not understood why Brucella spp. contain two nickel importers or whether they serve redundant functions, this work sheds light on the regulation of these systems. Indeed, we determined that nikA, and potentially nikABCDE, regulation by NikR in B. abortus is similar to the regulation of nickel import systems in other organisms, but NikR does not directly regulate the nikKMLQO operon. Therefore, it is likely that expression of the nikKMLQO system is controlled by a regulator other than NikR, but more work is needed to identify and characterize this transcriptional regulatory protein. In addition to the regulatory aspects of NikR, we demonstrate that nikR is dispensable for B. abortus virulence in the models tested (Fig. 3). Given the link between nickel and urease function, it is possible that NikR is required during oral infection by Brucella, as proper nickel homeostasis is expected to be necessary in this environment to ensure the full activity of urease. However, the role of NikR during oral infection remains to be assessed empirically, and future work will be aimed at characterizing NikR and other systems involved in nickel homeostasis in oral infection models.