Zinc Homeostasis in Brucella ovis Under Zinc Scarcity: Characterization of the ZnuABC Zinc Import System and of a Newly Identified Gene of the Zur Regulon

Simple Summary

Brucella ovis causes non-zoonotic ovine brucellosis, a bacterial infection that is responsible for significant economic losses worldwide. Zinc homeostasis is an essential mechanism for pathogenic bacteria that must regulate zinc uptake and efflux to maintain the required levels for survival and to avoid the toxicity of zinc excess. Although a previous report evaluated zinc export in B. ovis, the mechanisms governing zinc uptake remained unknown. In this work, we constructed and characterized mutants for the zinc uptake system ZnuABC and for zigA, a gene uncharacterized in Brucella but associated in other bacteria with zinc homeostasis and histidine metabolism. We found that expression of these genes was regulated depending on zinc availability. A potential link between zinc and histidine homeostasis, as well as a crosstalk between the homeostasis of zinc and other metals, raised new directions for future research. Interestingly, removal of ZnuA, ZigAZnuA or ZnuABC did not impair the virulence of B. ovis, which contrasts with the reported attenuation for the ZnuA mutant of zoonotic Brucella species. Overall, these results provide new relevant insights about metal regulation in Brucella, which remains almost unexplored in B. ovis, and suggest new differences between zoonotic and non-zoonotic Brucella species, despite their high DNA homology.

Abstract

The zinc import system ZnuABC is essential for the virulence of smooth (bearing O-polysaccharide chains) zoonotic Brucella species but remains uncharacterized in the non-zoonotic rough (lacking O-chains) species Brucella ovis, the causative agent of a type of ovine brucellosis of veterinary and economic importance. In this study, we constructed a panel of B. ovis mutants for genes of the znuACB locus that were characterized by evaluating growth, gene expression, protein abundance, metal content, and virulence in mice. We identified a ZigA ortholog (uncharacterized in the genus Brucella but linked to zinc homeostasis and histidine metabolism in other bacteria) that belongs to the Zur regulon and is transcriptionally modulated in a zinc-dependent manner as the ZnuABC system. A potential link of B. ovis ZigA with histidine catabolism and a crosstalk between the homeostasis of zinc and other metals have been identified as areas for further exploration. Remarkably, the znuA, zigAznuA and znuACB mutants of B. ovis displayed wild-type virulence in mice, which contrasts with the reported attenuation for the znuA mutant of smooth zoonotic Brucella species sharing a high degree of DNA homology. These results significantly enlarge the knowledge about the metallobiology of the genus Brucella, providing new perspectives for related research.

1. Introduction

Brucella ovis is a Gram-negative bacterium belonging to the genus Brucella that causes a type of ovine brucellosis responsible for large economic losses in the livestock sector [1,2]. It is a non-zoonotic and rough species (lacking O-polysaccharide chains in the lipopolysaccharide [LPS]) that mainly produces epididymitis and other genital lesions in rams. This contrasts with smooth (bearing O-chains in the LPS) Brucella melitensis strains that have a high zoonotic potential and are highly abortifacient in sheep [3]. LPS O-polysaccharide chains are critical virulence determinants for smooth Brucella species (i.e., B. melitensis, B. abortus or B. suis, which preferentially infect sheep and goats, cattle and pigs, respectively) [4], but naturally rough B. ovis is virulent in sheep and in laboratory animals with its R-LPS being crucial for virulence [5].

B. melitensis Rev1 is a smooth attenuated vaccine that also confers cross-protection against rough B. ovis. However, it induces antibodies against O-polysaccharide chains that interfere with the serological diagnosis of infections caused by zoonotic smooth Brucella strains [1]. Consequently, its use is forbidden in B. melitensis-free areas, which results in an increase in B. ovis infection in these regions [2,6] and raises the need to develop a B. ovis-specific vaccine. Since vaccines commonly used in control and eradication campaigns targeting B. melitensis and B. abortus infections are homologous attenuated strains (B. melitensis Rev1 and B abortus S19 or RB51, respectively), a rough attenuated B. ovis strain would also be an interesting vaccine candidate against B. ovis infection in sheep. This approach requires a deeper understanding of the B. ovis virulence factors, which have been significantly less studied than those supporting the virulence of smooth B. melitensis and B. abortus.

Pathogenic smooth and rough Brucella strains are facultative intracellular bacteria able to establish a replicative niche inside phagocytes, a required characteristic to maintain a chronic infection [4,7]. Consequently, virulence attenuation is usually observed with Brucella spp. mutants exhibiting defective intracellular behavior. Metal homeostasis is crucial at the host–pathogen interface since metals are essential micronutrients for a wide variety of cellular processes but are toxic when present in excess [8,9,10]. Several mechanisms employed by the host to restrict metal availability or to intoxicate the bacterial pathogen, even in the intracellular environment, have been described. But bacteria have also developed strategies to overcome these challenges and maintain proper levels of metal ions [9,10].

Zinc ranks second in abundance among transition metal cofactors, following iron [11,12], and it is estimated that 5–6% of bacterial proteins are Zn-binding proteins intervening in essential metabolic pathways [13]. On the contrary, the Irving–Williams series, which categorizes first-row transition metals according to their relative competitiveness for ligand binding in a cellular environment, predicts high-affinity interactions of zinc with metal-binding sites [14]. Accordingly, zinc excess can affect metalloproteins that require metals other than zinc for their functionality, resulting in defective bacterial performance [11,12]. Vertebrate hosts possess specific transporters for zinc uptake and efflux that, along with additional mechanisms (e.g., calprotectin or metallothioneins functioning as extracellular and intracellular zinc-sequestering proteins), intervene in zinc intoxication or restriction processes affecting bacterial pathogens [10,12]. To counteract these defense strategies of the host, which are also detected in immune cells [10,12], bacteria have developed zinc homeostasis mechanisms generally relying on specific import and export systems that are switched on–off by transcriptional regulators able to sense zinc bioavailability inside the cell. However, other strategies, such as the presence of zinc metallochaperones that contribute to intracellular zinc mobilization–allocation, or the increase in proteins that do not require zinc and that are able to perform the same function as zinc-dependent proteins, have also been described [12,13,15].

The tripartite high-affinity Zn²⁺ uptake system ZnuABC is an ATP-binding cassette (ABC) transporter consisting of the periplasmic zinc-binding ZnuA protein, the ZnuB permease that spans the cytoplasmic membrane, and the cytoplasmic ATPase ZnuC [13,16]. This system, whose expression is controlled by the metalloregulator Zur in a zinc-dependent manner, together with the zinc exporter ZntA, which is transcriptionally regulated by another zinc-sensing protein called ZntR, have been studied in many bacteria as the main contributors to zinc uptake and efflux, respectively [13,16]. Extensive in vitro characterization has been conducted to understand their structure and function in bacterial zinc homeostasis, but reports evaluating their role in the interaction with the host are comparatively limited. Nevertheless, the involvement of the ZnuABC uptake system in virulence has been reported for several bacteria [11,12], but its dispensability for other bacterial pathogens has also been described [17]. Similarly, the ∆zntA mutants of Chromobacterium violaceum or Vibrio parahaemolyticus were reported to be attenuated [18,19], while ∆zntA mutants of other bacteria exhibit full virulence [20,21,22].

In the genus Brucella, the ZnuABC uptake system has only been studied in smooth B. abortus and B. melitensis where ZnuA was shown to be essential for the virulence in cellular models and mice [23,24,25]. While this system remains unexplored in rough B. ovis, a previous study of our team reported relevant differences between this species and smooth B. abortus 2308 regarding the ZntA-ZntR system, which is involved in Zn²⁺ efflux under zinc excess conditions [21]. Specifically, the ∆zntR mutant of B. abortus 2308, defective in the MerR-type transcriptional regulator that modulates expression of the Zn²⁺ exporter ZntA, showed virulence attenuation, which was linked to an associated increased transcription of zntA [20]. On the contrary, a B. ovis PA ∆zntR mutant showed reduced zntA transcription and was not attenuated in the mouse model [21]. An in-phase deletion of 300 bp removing 100 amino acids of the N-terminal cytoplasmic region of ZntA that contains a potential Zn²⁺ binding site was also detected in B. ovis [21]. However, the B. ovis ZntA exporter displayed equivalent activity to that of the B. abortus protein in response to toxic zinc levels and was revealed as a potent cadmium exporter also contributing to cobalt, copper, and nickel detoxification [21], characteristics that had not been previously described for the ZntA protein of B. abortus 2308 [20].

Considering the differences found between smooth B. abortus and rough B. ovis with respect to the ZntR-ZntA system and zinc efflux, as well as the knowledge gap concerning zinc uptake systems in rough Brucella species, the main goal of this work was to construct and characterize a panel of single and multiple mutants targeting the znuACBzur zinc import system of B. ovis PA and an adjacent gene encoding a protein annotated in the B. ovis genome as a zinc metallochaperone GTPase called ZigA. This gene has never been studied in the genus Brucella, and its characterization is a matter of interest considering that ZigA of Acinetobacter baumannii intervenes in zinc release from the histidine–zinc labile pool and is required for growth during zinc starvation and for virulence [26]. Moreover, the B. ovis genome was scanned searching for orthologs of additional genes associated with zinc homeostasis in other bacteria, and these genes were included in gene expression studies. The characterization of the thirteen genetically modified strains obtained in this work, including virulence assessment in the mouse model, offers new insights into the poorly studied zinc homeostasis in the genus Brucella and suggests new relevant differences with smooth Brucella species.

2. Materials and Methods

2.1. Cloning Vectors, Bacterial Strains and Culture Conditions

Plasmid pCR^®-Blunt (Life Technologies, Carlsbad, CA, USA) or pGEM^®-T Easy (Promega, Madison, WI, USA) were used to clone PCR-amplified fragments. To obtain the B. ovis deletion mutants, target genes (Figure 1) were inactivated by overlapping PCR and cloned into pCVD-KanD, a plasmid that does not replicate in Brucella spp. and confers sucrose sensitivity and kanamycin resistance when integrated in the genome [27]. For complementation of mutant strains, wild-type genes were cloned into pBBR1MCS4, a plasmid that replicates in Brucella spp. and confers carbenicillin resistance [28].

Figure 1.

Genetic organization of the znuACB-zur zinc uptake cluster, and adjacent zigA, in B. ovis PA and B. abortus 2308. The predicted DNA binding site of the Zur transcriptional regulator, with the palindromic sequence of consensus Zur-box in other bacteria [29,30], is marked. DNA deletions in each constructed mutant are indicated by two-point arrows. A ΔzntAΔzur double mutant was also constructed by deleting zur in a ΔzntA mutant, which is defective in a P_1B-type ATPase involved in Zn²⁺ efflux [21].

Plasmids were transformed into and maintained in Escherichia coli strains CC118, JM109 or NZY5α [21] that were cultured on Luria Bertani medium, supplemented with 50 µg/mL kanamycin or carbenicillin when required, and incubated at 37 °C. B. ovis PA was used as the parental strain for the construction of the genetically modified strains listed in Table 1. B. ovis strains were cultured at 37 °C under a 5% CO₂ atmosphere in trypticase soy broth or agar (TSB or TSA; Pronadisa-Laboratorios Conda, Torrejón de Ardoz, Spain) supplemented with 0.3% yeast extract (YE; Pronadisa-Laboratorios Conda, Torrejón de Ardoz, Spain) and 5% horse serum (HS; Gibco-Life Technologies, Grand Island, NY, USA). TSA-YE-HS or TSB-YE-HS culture media were supplemented with 5% sucrose or 50 µg/mL kanamycin or carbenicillin when required.

Table 1.

B. ovis bacterial strains used in this work.

Strain Name	Relevant Characteristics	Origin
B. ovis PA	Parental strain	BCCN 1
B. ovis PA-pBBR1MCS4	Parental strain bearing pBBR1MCS4	[21]
B. ovis PA-pNVzntAPAcom5	Parental strain bearing pNVzntAPAcom5	[21]
B. ovis PA-pNVzntA2308com5	Parental strain bearing pNVzntA2308com5	[21]
B. ovis ∆zntA	∆zntA deletion mutant of B. ovis PA	[21]
B. ovis ∆zntA-pBBR1MCS4	∆zntA mutant bearing pBBR1MCS4	[21]
B. ovis ∆zntA-pNVzntAPAcom5	∆zntA mutant complemented with pNVzntAPAcom5 2	[21]
B. ovis ∆zntA-pNVzntA2308com5	∆zntA mutant complemented with pNVzntA2308com5 2	[21]
B. ovis ∆znuA	∆znuA deletion mutant of B. ovis PA	This work
B. ovis ∆znuCB	∆znuCB deletion mutant of B. ovis PA 3	This work
B. ovis ∆znuACB	∆znuACB deletion mutant of B. ovis PA 3	This work
B. ovis ∆zigA	∆zigA deletion mutant of B. ovis PA	This work
B. ovis ∆zur	∆zur deletion mutant of B. ovis PA	This work
B. ovis ∆zur-pBBR1MCS4	∆zur mutant bearing pBBR1MCS4	This work
B. ovis ∆zur-pNVzurcom2	∆zur mutant complemented with pNVzurcom2 2	This work
B. ovis ∆zigAznuA	∆zigAznuA mutant of B. ovis PA 3	This work
B. ovis ∆zigAznuACB	∆zigAznuACB mutant of B. ovis PA 3	This work
B. ovis ∆zntA∆zur	Double ∆zntA∆zur mutant of B. ovis PA 4	This work
B. ovis ∆zntA∆zur-pBBR1MCS4	∆zntA∆zur mutant bearing pBBR1MCS4	This work
B. ovis ∆zntA∆zur-pNVzntAPAcom5	∆zntA∆zur complemented with pNVzntAPAcom5 2	This work
B. ovis ∆zntA∆zur-pNVzurcom2	∆zntA∆zur mutant complemented with pNVzurcom2 2	This work

¹ BCCN, Brucella Culture Collection Nouzilly (Institut National de la Recherche Agronomique, Nouzilly, France). ² Multicopy plasmids replicating in B. ovis bearing the wild-type gene specified in the plasmid name (zntA or zur from B. ovis PA or zntA from B. abortus 2308). Plasmids pNVzntA_PAcom5 and pNVzntA₂₃₀₈com5 were previously constructed [21]. ³ Single deletion that includes several adjacent genes and intergenic regions (Figure 1). ⁴ Double deletion mutant obtained by deleting zur in the ∆zntA mutant of B. ovis PA obtained previously [21].

Growth curves of B. ovis PA-derived strains were established in TSB-YE-HS, supplemented with 50 µg/mL carbenicillin for strains containing pBBR1MCS4 or its recombinant derivatives, with bacterial cultures starting at optical density values at 600 nm (OD₆₀₀) of 0.05. Incubation was performed at 37 °C with agitation at 120 rpm and under a 5% CO₂ atmosphere. OD₆₀₀ values were recorded for 3 days at several time points. Results presented are representative of three independent experiments. For procedures described below involving RNA extraction or proteomic and metal concentration evaluation, samples were collected and processed after 16 h of incubation (exponential growth phase). Since a minimal medium for B. ovis strains has not been developed, zinc deprivation conditions were established by incubation in the presence of 35 µM N,N,N′,N′-Tetrakis(2-pyridylmethyl)ethylenediamine (TPEN; Sigma-Aldrich, St. Louis, MO, USA), which is widely used as a Zn²⁺ chelator, or 0.5 mM of the more general metal chelator ethylenediaminetetraacetic acid (EDTA; Sigma-Aldrich, St. Louis, MO, USA). For zinc toxicity assays, ZnCl₂ (Sigma-Aldrich, St. Louis, MO, USA) concentrations of 0.5, 3, or 5 mM were used, depending on the assay and strain analyzed. For gene expression analysis, 0.15 mM CdCl₂ was added to the culture medium in some assays. Appropriate concentrations of these compounds (those slightly inhibiting growth of parental B. ovis PA or selected mutant strains) were previously determined [21] or established in this work (see below).

2.2. Primers, DNA Techniques and Analysis of Nucleotide and Protein Sequences

Primers used in this work, detailed in Supplementary Table S1, were purchased from IDT (Leuven, Belgium) and designed based on the published sequences of chromosome I and II (GenBank accession numbers NC_009505 and NC_009504, respectively) of B. ovis 63/290 (ATCC 25840) [31]. The published whole genome sequences of B. abortus 2308 (GenBank accession numbers AM040264 and AM040265) were also used to retrieve sequences of selected genes.

High-fidelity PCR reactions were performed with the NZYProof 2x Green Master Mix (NZYTech Lda., Lisboa, Portugal) and used for PCR amplification of the DNA fragments to be cloned into pCR^®-Blunt or pGEM^®-T Easy or to sequence the zigA-zur locus of B. ovis PA. Routine PCR reactions, employed for the verification of B. ovis recombinant strains, were done with the Red Taq DNA polymerase master mix (VWR, Leuven, Belgium). Sanger DNA sequencing, required for verification of plasmid constructs or to determine the nucleotide sequences of the B. ovis PA zigA-zur locus (Figure 1), was performed at the Nucleus Service of the University of Salamanca.

The KEGG database (Kyoto Encyclopedia of Genes and Genomes) was used to identify orthologs and paralogs of selected genes (https://www.kegg.jp/kegg/genes.html (accessed on 2 November 2025)). Protein subcellular localization and other protein characteristics were predicted with PSORTb v3.0.3 (Brinkman Laboratory, Simon Fraser University, Burnaby, BC, Canada) [32] and UniProt Knowledgebase [33] (https://www.uniprot.org (accessed on 7 March 2026)), respectively. The Basic Local Alignment Search Tool (BLAST^®, blastn suite) was used to perform DNA similarity searches (https://blast.ncbi.nlm.nih.gov (accessed on 7 March 2026)) and EMBOSS Needle (European Bioinformatics Institute, Hinxton, UK) was used for pairwise sequence alignments of DNA or protein sequences (https://www.ebi.ac.uk/jdispatcher/psa/emboss_needle (accessed on 7 March 2026)). Protein stability was evaluated at https://web.expasy.org/protparam/ (accessed on 22 June 2026).

2.3. Construction of B. ovis Genetically Modified Strains

B. ovis mutant strains were obtained through a double homologous recombination process leading to the substitution of the wild-type gene by an inactivated gene resulting in an in-frame overlapping PCR that deletes almost the entire gene. The procedure was previously described [21,34] and involves a first stage where the 5′ and 3′ ends of each gene, together with 500–800 bp of adjacent DNA, are PCR-amplified with primers “target gene”MUT-F + “target gene”OVL-R and “target gene”OVL-F + “target gene”MUT-R, respectively. Both ends are then merged in a second PCR stage, with primers “target gene”MUT-F + “target gene”MUT-R, through the overlapping region of the two OVL primers used in the first PCR stage. PCR primers used in the construction of each mutant are described in (Supplementary Table S1). After cloning into pCR^®-Blunt or pGEM^®-T Easy and sequencing of the DNA insert, to discard undesired mutations, each inactivated gene was excised by digestion with restriction enzymes (XbaI plus SacI or SphI plus SalI, depending on the gene) and cloned into pCVD-KanD to obtain the recombinant plasmids used to mutagenize B. ovis. Recombinant pCVD-KanD plasmids were introduced into B. ovis by electroporation and selection of intermediate, revertant and the desired mutant strains was performed, as described previously [21,34], according to their pattern of sucrose and/or kanamycin sensitivity/resistance and specific PCR reactions allowing their differentiation.

Recombinant pBBR1MCS4 plasmid bearing wild-type zur (pNVzurcom2) was constructed by cloning the PCR fragment amplified with primers zurcom-F + znuCBMUT-R (Supplementary Table S1) first into pCR^®-Blunt and then into pBBR1MCS4 after digestion with XbaI and SacI restriction enzymes. This plasmid was selected and maintained in culture medium containing carbenicillin. Plasmids pNVzntA_PAcom5 and pNVzntA₂₃₀₈com5 were previously obtained [21] and were used either to complement the ΔzntAΔzur mutant strains or to overexpress the cloned gene in selected strains previously obtained [21].

2.4. Relative Quantification of RNA Transcripts and End-Point RT-PCR

The procedure described previously was followed [21,35]. In brief, RNA was extracted, with the E.Z.N.A.^® Bacterial RNA Kit (Omega Bio-tek Inc., Norcross, GA, USA), from bacterial strains cultured for 16 h (exponential growth phase) in TSB-YE-HS, supplemented with 3 mM ZnCl₂, 0.15 mM CdCl₂, 35 µM TPEN or 0.5 mM EDTA where appropriate. After DNaseI treatment of RNA, cDNA was synthesized with the RevertAid H minus first strand cDNA synthesis kit (Thermo Scientific, Vilnius, Lithuania) and, as control of DNA absence, reactions lacking retrotranscriptase were also included. Real-time PCR was performed in a StepOnePlus™ apparatus (Applied Biosystems, Foster City, CA, USA) using NZYSpeedy qPCR Green Master Mix (2×), ROX plus (NZYTech Lda., Lisboa, Portugal). Primer pairs for each target gene are listed in Supplementary Table S1. Relative expression was evaluated with the 2^−ΔΔCt method using the StepOne™ Software v2.3. B. ovis PA cultured in TSB-YE-HS and 16S were used as the reference strain and gene, respectively [36]. Three technical replicates and three independent biological samples were used for each gene, strain, and culture condition. Results are expressed as mean ± SD of the log₂ of the relative quantity (log₂RQ), and ΔCt values were used for statistical analysis [37].

When necessary, cDNA was subjected to end-point PCR, and the amplified fragments were visualized on a 2% agarose gel after electrophoresis using an O’GeneRuler Low Range DNA ladder (Thermo Scientific, Vilnius, Lithuania) as a size marker.

2.5. Protein Techniques and Evaluation of Metal Concentration

Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) [38] was performed, on bacterial cells grown in TSB-YE-HS for 16 h, using 14% acrylamide/bisacrylamide gels and a Protein II xi cell (Bio-Rad, Hercules, CA, USA). Pre-stained protein marker VI (Applichem-Panreac, Barcelona, Spain) was used for protein molecular weight standards and separated protein bands were stained with Coomassie blue. A Chemidoc MP apparatus (Bio-Rad) was used for image acquisition.

The proteomic analysis of SDS-PAGE-selected protein bands was performed, as described previously [21], in the proteomics facility of Centro de Investigación del Cáncer, Salamanca, Spain, according to its standardized procedures. In brief, protein bands excised from the gels were digested with trypsin and submitted to reversed-phase LC-MS/MS. Identification of proteins was performed by searching the MS/MS spectra against the B. ovis Uniprot proteome database using the Andromeda algorithm [39] and MAXQUANT [40]. The abundance of proteins in the selected gel slices was evaluated using the Intensity-Based Absolute Quantification (iBAQ) score [41]. Results are represented as means ± SD (n = 2) of log₁₀ iBAQ scores in two SDS-PAGE gels.

For the determination of metal concentrations present in bacterial cells, strains were cultured on TSB-YE-HS for 16 h in the presence of 35 µM TPEN, 0.5 mM EDTA or 30 µM ZnCl₂ when desired. A volume equivalent to 2 mL of an OD₆₀₀ value of 1 was centrifuged and the sediment washed and resuspended in 1 mL of 0.9% NaCl. A volume of 100 µL was used to determine the protein concentration using a Bradford assay solution (Applichem GmbH, Darmstadt, Germany). The remaining 900 µL was centrifuged and sedimented bacteria were incubated at 80 °C for 1 h to kill the bacteria. Metal concentrations were determined, after resuspending the bacterial pellets in 10 mL of 1% HNO₃, by inductively coupled plasma mass spectrometry (ICP-MS) with an Agilent 7800 apparatus (Agilent Technologies, Santa Clara, CA, USA) at the Nucleus Service of the University of Salamanca following its standardized procedures. Results are represented as means ± SD (n = 3) of ng metal/mg protein. Metal concentrations were also assessed in 1 mL TSB-YE-HS with 9 mL of 1% HNO₃, and results are represented as means ± SD (n = 3) of ng metal/mL.

2.6. Virulence in the Mouse Model

To evaluate the virulence of the bacterial strains in mice, the procedure previously described was followed [42]. Briefly, 6-week-old BALB/c mice (Charles River Laboratories, Chatillon-sur-Chalaronne, France) were intraperitoneally inoculated with 10⁶ CFU of each bacterial strain. Bacterial counts in the spleen were determined at week 3 and 7 post-infection in five mice per group and time point. Results were expressed as means ± SD (n = 5) of log₁₀ CFU/spleen values at each time point. The genotype of bacterial strains recovered from spleens was checked by PCR with specific primers for each deleted gene/s. The Spanish and European legislation for research with animals (RD 53/2013 and directive 2010/63/EU) was followed to design and perform experiments with mice.

2.7. Statistical Analysis

The GraphPad Prism Software version 7.0e (GraphPad Software Inc., San Diego, CA, USA) was used for statistical comparisons with one-way ANOVA, followed by Tukey’s test, or t-test for multiple or single comparisons, respectively. Statistically significant differences (p ≤ 0.05) were established with a 95% confidence interval.

3. Results

3.1. Sequence Analysis of the znuACBzur Locus of B. ovis PA and Search for Orthologs of Genes Related to Zinc Import in Other Bacteria

Analysis of the znuACBzur locus revealed that the protein encoded by BOV_A1025, the gene located at the 3′-end of znuA (Figure 1), is annotated in the B. ovis genome as zinc metallochaperone GTPase ZigA. The ZigA protein of B. ovis shares 56.4% amino acid identity with ZigA (Zur-induced GTPase A) of Acinetobacter baumannii, encoded by A1S_3411, and 70.6% identity with the YciC zinc chaperone of Agrobacterium tumefaciens, encoded by Atu3181. Both of the latter proteins have been shown to have an important role in survival under conditions of zinc deprivation [26,43]. Since this gene and its encoded protein have not been studied in the genus Brucella, BOV_A1025 was also targeted for mutagenesis.

Based on its amino acid sequence (Figure 2), ZigA of B. ovis PA is predicted to be a cytoplasmic protein (PSORTb v3.0.3) that contains the characteristic motifs described for G3E P-loop small GTPases belonging to the Cluster of Orthologous Groups 0523 (COG0523), specifically for those grouped in cluster 1 [44]. This cluster mainly consists of bacterial GTPases regulated by Zur such as A. baumannii ZigA [44]. This protein has been described as a metallochaperone transferring zinc to acceptor enzymes intervening in the catabolic pathway of histidine under zinc-limiting conditions [26]. COG0523 proteins are structurally composed of two domains, with characteristic motifs, that are also present in the B. ovis ZigA ortholog (Figure 2): (i) the N-terminal G-domain, of about 250 amino acids, contains the GTPase domain, which is composed of five G (GTP-binding) loops and is responsible for metal interaction as well as for GTP binding and hydrolysis [44,45], and (ii) the C-terminal domain that displays more variability, although it contains several conserved tryptophan residues [26,44], and is associated with the recognition and release of the client protein for metalation [44,45]. Both domains are connected by a linker that in COG0523 proteins grouped in cluster 1 has the consensus sequence PETEEYG [44]. These linker residues are also present in the ZigA protein of B. ovis (Figure 2).

Figure 2.

Amino acid sequence of the B. ovis PA ZigA protein highlighting its conserved motifs corresponding to G3E P-loop GTPases grouped in cluster 1 of COG0523 proteins [44]. Conserved motifs of the five G-loops of the N-terminal G-domain (responsible for GTPase activity) are in bold and framed as follows: G1/Walker A (red), G2 (orange), G3/Walker B (yellow), divergent G4 motif NKIS, instead the canonical NKXD sequence, (blue), and G5 (purple). The putative CXCC metal-binding motif located between G2 and G3 loops is framed in green. The bold red lettering represents the linker connecting the G-domain with the C-terminal domain, the latter involved in the recognition and release of client proteins to be metallated. The characteristic tryptophan residues of the C-terminal region are highlighted in bold and two conserved H and C residues [44] are marked with black circles.

The nucleotide sequence of the zigA-zur cluster (Figure 1) was determined for parental B. ovis PA. It only shows 8 nt substitutions (99.8% identity) when compared to the corresponding locus of B. abortus 2308 (Supplementary Figure S1A). None of these differences generate premature translation stop codons in protein encoding regions and only one single-nucleotide difference between both species is observed in the intergenic znuA-znuC region (Supplementary Figure S1A).

Considering that the B. ovis ZigA displays significant identity with proteins involved in Zn homeostasis in other bacteria that contribute to histidine catabolism [26], orthologs of hisB (involved in the histidine biosynthesis pathway; BOV_2001) [46], hutH (part of the histidine utilization operon; BOV_A0872) and hutC (described as a transcriptional repressor of the hut operon in B. abortus [47]; BOV_A0869) were identified in the B. ovis genome and included in the transcriptomic analysis performed in this work.

The yciC gene of A. tumefaciens, the ortholog of the B. ovis zigA gene, is transcribed in the opposite orientation to the troCBA operon (Atu3178, Atu3179, Atu3180) that encodes an ABC-type zinc uptake system that is independent of the classical znuABC system [43]. Accordingly, the presence of troCBA orthologs in the B. ovis genome was investigated, but the closest genes in B. ovis were those of the znuACB operon and no additional related genes were detected. No orthologs for the A. tumefaciens ZinT zinc-binding periplasmic protein (Atu1049), which is also regulated by Zur [17], were detected in the B. ovis genome.

The znuD gene is part of the Zur regulon in Neisseria meningitidis (nmb0964 gene), Pseudomonas aeruginosa (PA0781 gene) and A. baumannii (A1S_2892) and is highly transcribed under zinc-limiting conditions [29,30,48,49]. ZnuD is a TonB-dependent outer membrane (OM) receptor that shows some sequence homology with B. ovis BOV_1306, which is also annotated in the genome as a putative TonB-dependent OM receptor. Accordingly, expression studies performed in this work included BOV_1306 together with genes for two other putative TonB-dependent OM receptors (BOV_A0941 and BOV_A1093) detected in the B. ovis genome and BOV_1612, which is the first gene of the exbB-exbD-tonB operon. Since TonB-mediated transport is associated with the uptake of bacterial siderophores and the siderophore yersiniabactin also functions as a zinc transporter [50,51], the BOV_A0013 gene of B. ovis PA was also included in the transcriptomic assays. This gene is the ortholog of the dhbC gene of B. abortus 2308, which is involved in the biosynthesis of the monocatechol siderophore 2,3-dihydroxybenzoic acid (DHBA) [52] and brucebactin, a second Brucella siderophore derived from DHBA and structurally more complex [53].

3.2. Growth of the B. ovis Strains with Mutations in Genes Related to Zinc Import Is Not Impaired When They Are Cultured in the Presence of the Zn²⁺ Chelator TPEN

Because of its fastidious nature and unknown nutritional requirements, a defined minimal medium is not available for culturing B. ovis [54]. Therefore, to mimic zinc-restricted conditions, the Zn²⁺ chelator TPEN [20,55,56] was added to the culture medium. To determine the optimal TPEN concentration for subsequent studies, growth curves of parental B. ovis PA were performed in TSB-YE-HS supplemented with different TPEN concentrations (Figure 3A).

Figure 3.

Growth curves of parental B. ovis PA in the presence of several concentrations of the zinc chelator TPEN (A) and of B. ovis PA mutant strains in the presence of 35 µM TPEN (B). In panel (A), incubation with 35 µM ZnCl₂ restores growth of B. ovis PA cultured with 35 µM TPEN (blue).

Based on the experimental results shown in Figure 3A, 35 µM TPEN was chosen as the preferred concentration to analyze the mutant strains, as it showed some degree of growth inhibition of the parental strain that was abolished by adding 35 µM ZnCl₂ to the culture medium. The range of TPEN tolerance susceptibility of B. ovis PA is tight since, while in the presence of 32.5 µM TPEN its growth is almost unaffected, a concentration of 37.5 µM TPEN clearly impairs its growth (Figure 3A). Surprisingly, growth curves of all mutant strains analyzed in this work were equivalent to those of the parental strain when cultured in the absence or in the presence of TPEN (Figure 3B and Section 3.5). These results suggest the existence of additional mechanisms, other than the ZnuABC system, allowing zinc acquisition under these experimental conditions.

3.3. Growth in the Presence of TPEN Alters Expression of Genes Related to Zinc Homeostasis and Histidine Metabolism in B. ovis PA and Reveals Cotranscription of znuA and zigA

As expected, exposure of B. ovis PA to TPEN resulted in increased transcription of zinc import genes (znuA and znuC) and decreased transcription of zinc export genes (zntA and zntR). But notably, this exposure had no impact on the gene encoding the zinc-responsive regulator Zur and highly increased transcription of zigA (Figure 4A). However, these observations contrast with the absence of growth defects mentioned above for the mutant strains studied in this work in the presence of TPEN (Figure 3B and Section 3.5).

Figure 4.

Expression of genes related to zinc and histidine homeostasis in parental B. ovis PA cultured for 16 h in the presence of 35 µM TPEN (A), and end-point RT-PCR reactions targeting zigA, znuA or their intergenic region in cDNA obtained from B. ovis PA cultured in the presence of 35 µM TPEN for 16 h (B). In panel (A), statistically significant differences (p ≤ 0.05) when compared to B. ovis PA cultured in normal medium are marked with asterisks, * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001. In panel (B), regions amplified by end-point RT-PCR (from cDNA resulting from reactions with or without RT, or from DNA) in each lane of the electrophoresis gel are marked with the respective colored lines in the schematic of the upper part of the panel; primer pairs used were zigAF2-EP + zigART-R (1), znuART-F + znuAR2-EP (2) and zigAF-EP + znuAOVL-R (3); lanes M correspond to the O’GeneRuler Low Range DNA ladder (Thermo Scientific, Vilnius, Lithuania).

TPEN exposure did not modify the expression of hisB (gene of the biosynthetic pathway of histidine), but elevated transcription of the histidine utilization gene hutH and reduced transcription of the gene encoding HutC that is a repressor of hut genes (Figure 4A). These experimental findings align with the proposed link between zinc homeostasis and histidine metabolism that has been reported for other bacteria [26] and advance a new area for future research exploring whether this connection also exists in the genus Brucella.

Since zigA has not been studied previously in the genus Brucella and to determine whether it constitutes a monocistronic operon or is cotranscribed with znuA, end-point RT-PCR reactions were performed with primer pairs targeting intragenic regions of zigA (zigAF2-EP + zigART-R) and znuA (znuART-F + znuAR2-EP) or their intergenic region (zigAF-EP + znuAOVL-R) (Supplementary Table S1). The results of these analyses indicate that znuA and zigA constitute a polycistronic operon and display maximum transcription in response to zinc deprivation (Figure 4B, Supplementary Figure S1B).

3.4. B. ovis Zur Is a Transcriptional Repressor of zigA, znuA and the znuCBzur Operon, and Its Absence in the Δzur Mutant Significantly Increases Levels of the ZnuA Protein

Transcription levels in normal medium of genes related to zinc and histidine homeostasis were evaluated by RT-qPCR in the Δzur-pBBR1MCS4 mutant in comparison to those of B. ovis PA-pBBR1MCS4 and the complemented mutant Δzur-pNVzurcom2. The mutant strain showed highly increased levels of zigA, znuA, and znuC, while the other genes exhibited expression levels like those of B. ovis PA-pBBR1MCS4. The complemented mutant Δzur-pNVzurcom2 behaved as B. ovis PA-pBBR1MCS4 except for zur, whose expression level was significantly higher in the complemented strain in accordance with the multicopy nature of pBBR1MCS4 (Figure 5). Overexpression of zigA and znuA in the Δzur mutant, when compared to the parental strain, was also evident in the end-point RT-PCR assays, as both genes were detected in the Δzur mutant while their detection in the parental strain required the presence of TPEN in the culture medium (Figure 4B, Supplementary Figure S1B).

Figure 5.

Expression of genes related to zinc and histidine homeostasis in B. ovis PA-pBBR1MCS4, the Δzur-pBBR1MCS4 mutant and the same mutant complemented with wild-type zur. Expression of zur in the Δzur mutant is not represented (blue) and statistically significant differences (p ≤ 0.05) when compared to B. ovis PA-pBBR1MCS4 are marked with asterisks, *** p ≤ 0.001.

Selected mutant and complemented strains, B. ovis PA overexpressing the Zn²⁺ exporter ZntA (from B. ovis or B. abortus), and B. ovis PA cultured in the presence of 35 µM TPEN or 1 or 3 mM ZnCl₂ were submitted to SDS-PAGE to evaluate their protein profile (Figure 6A, Supplementary Figure S1C).

Figure 6.

Protein profile after SDS-PAGE of B. ovis PA and its derived mutants (A), and abundance of proteins ZntA (B) or ZnuA (C) determined, after reversed-phase LC–MS/MS, in selected strains and culture conditions. In panel (A), pre-stained protein marker VI (Applichem-Panreac, Barcelona, Spain) is shown and regions analyzed by reversed-phase LC–MS/MS are framed in blue (ZntA region) or orange (ZnuA region). In panels (B,C), statistically significant differences (p ≤ 0.05) when compared to B. ovis PA (B) or B. ovis PA-pBBR1MCS4 (C) are marked with asterisks, * p ≤ 0.05; ** p ≤ 0.01.

Although incubation of parental B. ovis PA in the presence of 1 mM ZnCl₂ did not provide a differential protein pattern in SDS-PAGE, which is in agreement with previous results [21], when the concentration of ZnCl₂ was increased to 3 mM, a protein band of approximately 70 kDa became apparent when compared to the results obtained in normal culture conditions (Figure 6A, Supplementary Figure S1C). The protein was identified by reversed-phase LC-MS/MS as the Zn²⁺ exporter ZntA and showed iBAQ values significantly higher than those observed in normal culture conditions (Figure 6B). On the contrary, iBAQ values of ZntA in B. ovis PA cultured with 35 µM TPEN were significantly lower than those measured in control culture conditions (Figure 6B).

Additionally, a markedly overrepresented protein of about 35 kDa was detected in the Δzur mutant and the ΔzntAΔzur double mutant, but not in the complemented Δzur mutant (Figure 6A). By reversed-phase LC-MS/MS, this protein was identified as ZnuA in the Δzur-pBBR1MCS4 mutant and showed iBAQ values significantly higher than those detected in B. ovis PA-pBBR1MCS4 (Figure 6C). ZnuA was also significantly overrepresented in B. ovis PA cultured with 35 µM TPEN and, aligning with previous results [21], in B. ovis PA-pNzntA₂₃₀₈com5 that overexpresses the ZntA Zn²⁺ exporter (Figure 6C).

Despite the highly increased transcription of zigA in the Δzur mutant (Figure 5), no protein was clearly overrepresented in this mutant in the region of 45 kDa predicted for ZigA (Figure 6A). This difference with ZnuA could be related to different protein stabilities, since the ProtParam tool (Expasy, Swiss Institute of Bioinformatics) classified ZigA as an unstable protein, contrarily to what was predicted for ZnuA.

3.5. B. ovis PA and the ΔzigAznuACB Mutant Cultured with TPEN Show Similar Zinc Content, While the ΔzntA and ΔzntAΔzur Mutants Contain Higher Zn Levels than the Parental Strain When Cultured with ZnCl₂

Growth in the presence of 35 µM TPEN was also analyzed for the ΔzntA mutant of B. ovis PA, which is defective in a powerful zinc exporter [21], for the Δzur mutant, which lacks the Zur repressor and consequently overexpresses genes of the zigAznuACB locus (Figure 5), and for the double ΔzntAΔzur mutant. These three mutants showed the same behavior as the parental strain (Figure 7A), as was also the case for the ΔzigAznuACB mutant (Figure 7B), which lacks the entire ZnuABC-predicted zinc uptake system together with the newly identified zigA gene (Figure 1). Since this mutant behaved like the parental strain in TPEN-containing culture medium, zinc levels in the parental strain and the ΔzigAznuACB mutant were assayed by ICP-MS, but no differences were detected between these strains under the conditions of the assay (Figure 7C).

Figure 7.

Growth curves of B. ovis PA and the ΔzntA, Δzur, ΔzntAΔzur (A) and ΔzigAznuACB (B) mutants with 35 µM TPEN, and evaluation by ICP-MS of Zn levels in parental B. ovis PA and the ΔzigAznuACB mutant (C). Growth curves of B. ovis PA-pBBR1MCS4, the ΔzntA-pBBR1MCS4 mutant and the ΔzntA mutant complemented with wild-type zntA (D,E) or B. ovis PA and the ΔzntAΔzur double mutant (G,H) in the presence of 0.5 mM (D,G) or 30 µM ZnCl₂ (E,H), and evaluation by ICP-MS of Zn levels in these strains cultured for 16 h in the presence of 30 µM ZnCl₂ (F,I). In panels (C,F,I), statistically significant differences (p ≤ 0.05) when compared to B. ovis PA (C,I) or B. ovis PA-pBBR1MCS4 (F) are marked with asterisks, * p ≤ 0.05.

Previous results demonstrated that the ΔzntA mutant of B. ovis PA tolerated ZnCl₂ concentrations about 100 times lower than the parental strain, a defect that was abolished by complementation with wild-type zntA [21]. This phenotype was confirmed in this work using 0.5 mM ZnCl₂, a concentration that prevented growth of the ΔzntA mutant but not that of the parental strain or the complemented mutant (Figure 7D). Considering these results, zinc levels were also evaluated by ICP-MS in B. ovis PA-pBBR1MCS4, the ΔzntA-pBBR1MCS4 mutant and the ΔzntA-pNVzntA_PAcom5 strain (mutant strain complemented with wild-type zntA of B. ovis PA) that were cultured for 16 h in the presence of 30 µM ZnCl₂, a concentration that did not affect the growth of the parental and complemented strains and that slightly affected the growth of the ΔzntA mutant (Figure 7E). The ΔzntA-pBBR1MCS4 displayed significantly higher levels of zinc than the parental strain and the ΔzntA mutant complemented with the wild-type gene (Figure 7F). Similar results to those obtained with the single ΔzntA mutant were obtained with the ΔzntAΔzur double mutant (Figure 7G–I). Complementation of this double mutant with wild-type zntA, but not with wild-type zur, restored the growth pattern of parental B. ovis PA (Supplementary Figure S2A,B). The mutants of the zigA-znuACBzur locus obtained in this work behaved as the parental strain in the presence of 0.5 mM ZnCl₂ (Supplementary Figure S2C,D).

3.6. The Siderophore Biosynthesis Gene BOV_A0013 Exhibits Significantly Reduced Transcription in Response to TPEN or Excess Cadmium, but Not in Response to Excess Zinc

Siderophores and their energized import system, frequently constituted by OM receptors linked to a cytoplasmic membrane complex containing the TonB-ExbB-ExbD proteins [57], are intended to facilitate iron import. But some of them have also been linked to zinc acquisition in other Gram-negative bacteria [51,58]. Therefore, B. ovis PA representative genes of the siderophore biosynthetic pathway (BOV_A0013) and genes predicted to be related to TonB systems (BOV_1306, BOV_1612, BOV_A0941 and BOV_A1093) were included in the transcriptomic analysis by RT-qPCR.

Surprisingly, a strong reduction in the transcription level of the siderophore biosynthetic gene BOV_A0013 (mean log₂ RQ values of −5,42, p ≤ 0.001) was detected in B. ovis PA cultured with TPEN (Figure 8A). The other four genes did not display important differences between both culture conditions, although a slight reduction (mean log₂ RQ values of −0.82, p ≤ 0.05) was observed for BOV_1612 (Figure 8A).

Figure 8.

Expression of genes related to zinc and histidine homeostasis, siderophore biosynthesis and/or TonB-mediated transport in B. ovis PA cultured in the presence of 35 µM TPEN (A), 3 mM ZnCl₂ (B) or 0.15 mM CdCl₂ (C). Statistically significant differences (p ≤ 0.05) when compared to B. ovis PA cultured in normal culture medium are marked with asterisks, * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.

Transcription of genes related to zinc and histidine homeostasis, as well as BOV_A0013 and BOV_A1093, was also evaluated in the parental strain incubated in the presence of 3 mM ZnCl₂ or 0.15 mM CdCl₂. This latter compound was included in the study because the ZntA zinc exporter of B. ovis PA is also a strong cadmium exporter [21].

As expected according to previous results [21], zntA showed increased transcription in B. ovis PA cultured in the presence of 3 mM ZnCl₂ (Figure 8B). Moreover, the transcription levels of zigA and znuA were reduced (p ≤ 0.05) in the same culture conditions (Figure 8B). This fact contrasts with previous results where transcription of znuA in B. ovis PA cultured in the presence of ZnCl₂ was not distinguishable from that detected in normal culture medium [21]. However, it must be noted that these previous results [21] were recorded with a ZnCl₂ concentration three times lower than that used in the present work (Figure 8B). Increased transcription of hutH, hutC and BOV_A1093 was also detected in B. ovis PA cultured in the presence of 3 mM ZnCl₂ (Figure 8B).

Contrary to the results obtained with ZnCl₂, culture with 0.15 mM CdCl₂ slightly increased transcription of znuA and reduced the transcription levels of BOV_A0013 and BOV_A1093 (Figure 8C).

3.7. Culture of B. ovis PA in the Presence of the Metal Chelator EDTA Reveals a Transcription Profile Comparable to That Obtained with TPEN

Since no growth differences were observed in TPEN-containing culture medium among the B. ovis PA mutants defective in genes related to the zinc import and export systems, EDTA was also evaluated as an additional metal chelator with B. ovis PA and the different mutants obtained in this work. To choose the optimal EDTA concentration to be used, growth curves of parental B. ovis PA were made in TSB-YE-HS supplemented with different EDTA concentrations. According to the results (Figure 9A), 0.5 mM EDTA was selected as the concentration to perform susceptibility and gene expression assays with the B. ovis parental and mutant strains. The defective growth of B. ovis PA in the presence of 0.5 mM EDTA was abolished by the addition of 0.5 mM ZnCl₂ (Figure 9A).

Figure 9.

Growth curves of parental B. ovis PA in the presence of several concentrations of the metal chelator EDTA (A) and expression of genes related to zinc and histidine homeostasis, siderophore biosynthesis and/or TonB-mediated transport in B. ovis PA cultured in the presence of 0.5 mM EDTA (B). In panel (A), incubation with 0.5 mM ZnCl₂ restores growth of B. ovis PA cultured with 0.5 mM EDTA (blue). In panel (B), statistically significant differences (p ≤ 0.05) when compared to B. ovis PA cultured in normal culture medium are marked with asterisks, * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.

Parental B. ovis PA cultured with 0.5 mM EDTA showed a transcription pattern of genes related to zinc and histidine homeostasis and to siderophore biosynthesis and TonB-related transport (Figure 9B) equivalent to that observed when cultured with 35 µM TPEN (Figure 4A and Figure 8A). Specifically, when compared to normal culture conditions, zntA (coding for the zinc exporter) was undertranscribed, while zigA, znuA and znuC (genes of the zinc import locus) were overtranscribed in B. ovis PA cultured with 0.5 mM EDTA. Reduced transcription was also observed for BOV_A0013, BOV_A1093 and BOV_1612 (genes linked to siderophore biosynthesis and TonB-related transport). Genes coding for HutH (histidine utilization pathway) and HutC (repressor of hut genes) showed higher and lower transcription levels, respectively, in response to EDTA treatment (Figure 9B).

3.8. The ΔznuACB and ΔzigAznuACB Mutants Display Growth Defects in Response to EDTA Treatment That Can Be Rescued by the Addition of ZnCl₂ and Partially Rescued with NiCl₂, MnCl₂, CoCl₂ or FeCl₂

Growth in culture medium supplemented with 0.5 mM EDTA was analyzed for all of the B. ovis strains examined in this work. All mutants behaved like the parental strain except for the ΔznuACB and ΔzigAznuACB mutants, which displayed an impaired growth (Figure 10A, Supplementary Figure S2E). Addition of 0.5 mM ZnCl₂ recovered the parental phenotype without EDTA, but a partial recovery was also observed when NiCl₂, MnCl₂, CoCl₂ or FeCl₂ were added to the culture medium, with the latter three reaching the levels of the parental strain cultured in the same conditions (Figure 10B–F, Supplementary Figure S3A–E). Additionally, 0.5 mM CuCl₂ or MgCl₂ was tested, but no differences were found when compared to the ΔznuACB and ΔzigAznuACB mutants cultured with 0.5 mM EDTA alone (Supplementary Figure S3F–I).

Figure 10.

Growth curves of B. ovis PA and the ΔznuACB or ΔzigAznuACB mutants in the presence of 0.5 mM EDTA (A) or with 0.5 mM EDTA plus 0.5 mM ZnCl₂ (B), NiCl₂ (C), MnCl₂ (D), CoCl₂ (E), or FeCl₂ (F). ICP-MS metal concentrations in TSB-YE-HS culture medium (G) and in B. ovis PA and the ΔznuACB and ΔzigAznuACB mutants after 16 h of culture with 0.5 mM EDTA (H). Growth curves with the ΔzigAznuACB mutant, equivalent to those obtained with the ΔznuACB, are shown in Supplementary Figure S3. Addition of CuCl₂ or MgCl₂ did not restore growth of parental B. ovis PA or mutant strains cultured in TSB-YE-HS 0.5 mM EDTA (Supplementary Figure S3).

In order to determine if the results described in the paragraph above could be related to metal availability in the B. ovis strains, the levels of Zn, Mn, Fe, Cu and Co present in parental B. ovis PA and the ΔznuACB and ΔzigAznuACB mutants cultured with 0.5 mM EDTA were determined by ICP-MS and compared to those present in the parental strain in normal culture conditions. Profiles for these five metals were also determined for the TSB-YE-HS culture medium, with Zn being the most abundant metal with levels of 141 ± 19 ng/mL (Figure 10G). The concentration of Fe (50 ± 2 ng/mL) was approximately three times lower than that of Zn in this medium, while Cu, Mn and Co were detected to a lesser extent (Figure 10G). Zn was also the most abundant metal detected in the B. ovis strains examined, while Co was under the detection limit (Figure 10H). However, no statistically significant differences could be demonstrated for any of the assayed metals (n = 3) in these strains when they were cultured in the presence of EDTA (Figure 10H).

3.9. The ΔznuACB and ΔzigAznuACB Mutants Exhibit Reduced Expression of the Genes Evaluated in This Work Under EDTA Treatment

Considering the growth defects observed for the ΔznuACB and ΔzigAznuACB mutants in the presence of EDTA, the expression levels of genes related to zinc and histidine homeostasis and to siderophore biosynthesis and TonB-related transport were evaluated in both mutants and in ΔzigA, ΔznuA and ΔzigAznuA mutants cultured with EDTA (Figure 11A,B, Supplementary Figure S4).

Figure 11.

Expression of genes related to zinc and histidine homeostasis, siderophore biosynthesis and TonB-mediated transport in B. ovis PA and the ΔznuACB and ΔzigAznuACB mutants cultured in the presence of 0.5 mM EDTA (A), and expression of genes related to zinc homeostasis in B. ovis PA and the ΔzigA and ΔznuA mutants (B). Target genes in blue correspond to genes that were not evaluated in the corresponding deletion mutant. Statistically significant differences (p ≤ 0.05) when compared to B. ovis PA cultured in normal culture medium are marked with black asterisks. Statistically significant differences in mutant strains cultured with 0.5 mM EDTA in comparison to B. ovis PA cultured with 0.5 mM EDTA are marked with blue asterisks, * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.

In the ΔznuACB and ΔzigAznuACB mutants cultured in the presence of EDTA, a strong reduction in transcripts for the ZntA zinc exporter was observed (Figure 11A), which suggests that the intracellular zinc levels are substantially decreased in these strains. Zur transcripts are also significantly decreased in both strains, probably due to the absence of the natural promoter region located in the znuA-znuC intergenic region (Figure 1). The same reason probably underlies the zigA subexpression detected in the ΔznuACB mutant (Figure 11A). A significant diminished transcription of hutH, hutC, BOV_A0013 and BOV_A1093 was also detected in both mutants when compared to the parental strain (Figure 11A).

Gene expression was also studied with the ΔzigA, ΔznuA and ΔzigAznuA mutants cultured in the presence of 0.5 mM EDTA. While deletion of zigA did not affect transcription of the remaining genes of the locus (Figure 11B), deletion of znuA led to a strong increased transcription of zigA, znuC and zur (znuB, located between znuC and zur, was not analyzed) and decreased zntA transcription (Figure 11B). Genes linked to histidine metabolism and siderophore-mediated metal transport displayed a similar transcription profile to that observed with B. ovis PA cultured with EDTA (Figure 11A, Supplementary Figure S4A). A similar expression pattern was also observed with the ΔzigAznuA mutant, except for the genes deleted in this strain (Supplementary Figure S4B).

3.10. The ZnuACB Import System Is Not Required for Virulence of B. ovis PA in Mice

Based on the results described in previous sections, several single and multiple mutants lacking the zinc-responsive regulator Zur or the zinc import-related proteins ZigA, ZnuA, ZnuC and ZnuB (Figure 1) were evaluated for virulence in the BALB/c mouse model of chronic infection (Figure 12).

Figure 12.

Virulence of the B. ovis mutants in mice inoculated intraperitoneally with 10⁶ CFU/mouse. Spleen colonization was evaluated at weeks 3 and 7 post-infection in five mice per strain and time point. Results are represented as means ± SD (n = 5) of the log₁₀ CFUs/spleen. Statistically significant differences (p ≤ 0.05) when compared to B. ovis PA were not detected and are not marked.

Deletion of znuA did not affect the virulence of B. ovis PA (Figure 12), which contrasts with the results previously described for the B. abortus and B. melitensis ΔznuA mutants that showed a marked attenuation in cellular models and mice [23,24,25]. In fact, the ΔzigA, ΔzigAznuA, Δzur and ΔznuACB mutants also displayed a virulence pattern that was undistinguishable from that observed with the parental strain.

4. Discussion

Zinc is an essential micronutrient that plays a crucial role in the functioning of all living organisms, including vertebrate hosts and bacterial pathogens, and is required for structural, regulatory or catalytic properties of many proteins [12]. On the contrary, the predicted high-affinity interactions of zinc with different metal-binding sites [14] results in mismetallation of non-zinc metalloproteins, causing detrimental effects on the cell, in zinc-rich environments [11,12]. Accordingly, intracellular zinc levels must be tightly balanced to preserve cell viability.

To maintain zinc homeostasis, bacterial pathogens, and particularly intracellular pathogens like Brucella, must overcome nutritional immunity that is imposed by the host through zinc availability restriction and/or zinc intoxication mechanisms [12]. Studies targeting zinc homeostasis in the genus Brucella are few, despite the relevance shown in other bacteria. The ZntA-ZntR system, made up of the zinc exporter ZntA and its transcriptional regulator ZntR, has only been studied in smooth B. abortus 2308 [20] and more recently in rough B. ovis PA [21]. Moreover, data regarding zinc uptake mediated by the ZnuABC transporter, a main zinc importer in many bacteria, has only been reported in the two smooth species B. melitensis and B. abortus where the respective ΔznuA mutants were found to be attenuated in mice [23,24,25]. Despite the high degree of DNA identity existing among the pathogenic species of the genus Brucella [59], a ΔzntR mutant of B. abortus 2308 was attenuated in mice in the chronic phase of infection [20], while a ΔzntR mutant of B. ovis PA was not [21]. Attenuation of the B. abortus ΔzntR mutant was experimentally linked with increased expression of the ZntA zinc exporter [20], but deletion of zntR in B. ovis PA had the opposite effect with a significant reduction in zntA transcription [21]. Considering the differences found between smooth Brucella strains and rough B. ovis PA regarding the ZntA-ZntR zinc export system [20,21], the main aim of the present work was to characterize the role of the ZnuABC zinc import system of B. ovis PA in zinc homeostasis and virulence.

The first finding was the identification of a gene, located at the 3′ end of znuA (Figure 1) and never studied in the genus Brucella, showing the characteristic motifs of COG0523 zinc metallochaperone GTPases whose expression is regulated by Zur (Figure 2), as ZigA of A. baumannii [44]. The B. ovis PA zigA gene was also demonstrated to be regulated by Zur (Figure 5) and to have increased expression in the presence of TPEN and EDTA chelators (Figure 4A and Figure 9B), while it showed a diminished transcription under zinc excess conditions (Figure 8B). These characteristics suggest that the B. ovis PA ZigA protein might play a relevant role in maintaining zinc homeostasis in zinc-depleted environments.

Zur and ZntR are zinc-sensitive regulators governing expression of the main bacterial zinc export–import systems under zinc excess and insufficiency conditions. They bind zinc with high affinity and work coordinately enabling the finely tuned switch-on and switch-off of zinc uptake and efflux systems [60,61]. Specifically, under excess conditions, zinc binds both ZntR (consequently activating expression of the ZntA exporter) and Zur, which represses genes involved in zinc acquisition. This balanced regulation, coupled to the reported almost absence of free zinc in the bacterial cytoplasm despite its relatively high content in the cell, raised the idea of the existence of a pool of chelated zinc that could be mobilized and transferred to essential proteins during zinc starvation [60]. Some bacterial metabolites have been associated with a labile zinc pool, among them the amino acid histidine in A. baumannii [12,13,26]. In this bacterium, the zinc metallochaperone GTPase ZigA was reported to contribute to metallation of the histidine ammonia lyase HutH that metabolizes histidine into urocanic acid, which is further converted into glutamate via the activity of other enzymes encoded by the hut operon. Given that both urocanic acid and glutamate have less affinity for zinc than histidine, the availability of zinc in the cell increases with histidine catabolism [12,13,26]. Both ZigA and HutH were shown to be required for adaptation to zinc starvation in A. baumannii, where zigA is closely located in the genome to the hut operon [26]. Although this genome distribution is not conserved in B. ovis PA (Figure 1), increased expression of hutH and decreased expression of hutC, which encodes a transcriptional repressor of the hut operon in B. abortus [47], was observed in zinc starvation conditions induced by the TPEN or EDTA chelators (Figure 4A and Figure 9B). Although these findings are exclusively based on expression studies, and additional research is required, they also suggest a possible link in B. ovis PA between histidine catabolism and zinc homeostasis under zinc scarcity conditions. This potential connection would not be mediated, at least directly, by Zur since the Δzur mutant of B. ovis PA did not show differences in expression of hutH or hutC, contrary to the highly increased transcription observed in this mutant for zigA, znuA and znuC (Figure 5).

A clear change in the transcription profile of the classical zinc import–export genes, suggestive of restricted intracellular zinc, was observed for parental B. ovis PA grown in the presence of TPEN (Figure 4A). However, unexpectedly, none of its derived mutants for the zigA-zur locus (Table 1; Figure 1), or the ΔzntA or double ΔzntAΔzur mutants (Table 1), displayed an increased growth defect in the presence of 35 µM TPEN when compared to the parental strain (Figure 3B and Figure 7A,B). Because a defined minimal medium is not available for culturing B. ovis, a zinc-restricted environment was mimicked with the TPEN chelator, which has been widely used with this purpose in many bacteria [26,29,30,48], including B. abortus [20]. However, due to its lipophilic nature, TPEN is described as a membrane-permeable chelator in studies performed with eukaryotic cells [62]. This property could be related to the lack of increased detectable defects in the presence of TPEN for the B. ovis PA mutant strains obtained and with the similar zinc content detected in the presence of TPEN for the parental strain and the ΔzigAznuACB mutant (Figure 7C). If chelated zinc enters the cell, independently of specific import systems, the metal could be sequestered by bacterial proteins in enough quantity to maintain viability. This could also explain the narrow margin of TPEN concentration that separates normal growth from the absence of replication in B. ovis PA (Figure 3A), which could be triggered when the TPEN concentration surpasses the threshold in which the cell would be unable to compete with the chelator for metal binding.

Despite the previous comments, the existence of additional mechanisms for active or passive zinc acquisition in B. ovis seems probable, particularly considering that the parental strain and all mutant strains showed a similar growth pattern also in the absence of TPEN. This possibility for additional zinc uptake mechanisms also seems plausible in the smooth Brucella species analyzed in previous reports, since the ΔznuA mutants of B. abortus strains 2308 and 544 and B. melitensis 16M also grew in rich medium [23,24,25] and the B. abortus 544 mutant showed limited growth defects in culture medium chelated with EDTA (the B. melitensis 16M ΔznuA mutant was not tested under zinc-restricted conditions). The most significant effect of znuA deletion was reported with the mutant of B. abortus 2308, which could be tested in a low-zinc minimal medium available for this species. In these zinc-restricted conditions, the ΔznuA mutant of B. abortus 2308 was unable to grow, but supplementation of the minimal medium with zinc restored growth of the mutant strain [25], which also suggests the existence of additional mechanisms for zinc acquisition. However, the ΔznuA single mutants of these three smooth Brucella strains were attenuated in a mouse model based on splenic counts [23,24,25], which demonstrates the significance of ZnuA for their infectious process and contrasts with the lack of attenuation of the ΔznuA, ΔzigAznuA and ΔznuACB mutants of rough B. ovis PA in an equivalent model of infection (Figure 12). These contrasting results do not appear to derive from interspecies differences in the zigA-zur locus at the DNA level, as the minor nucleotide substitutions detected only affected one amino acid of each of the five proteins of the locus and one nucleotide in the intergenic znuA-znuC region (Supplementary Figure S1A). No additional zinc import systems have been detected in B. ovis but a relevant differential characteristic between smooth and rough Brucella strains is the absence of O-polysaccharide chains in the bacterial surface. Consequently, membrane permeability to diverse compounds may differ between rough and smooth strains, and other OM components, such as porins, that are more exposed on the surface of B. ovis might have a role in zinc uptake. In this respect, differences among Brucella species regarding the profile of OM proteins (OMPs) and OM-related properties have been described [38,63,64] and might impact zinc homeostasis mechanisms. Additionally, a genomic island in B. ovis, absent from B. abortus and B. melitensis, encodes an ABC transporter required for virulence in mice and sheep [65,66]. The exact function of this transporter has not been elucidated but it seems to be required for expression of the type-IV secretion system [65], a recognized virulence factor in Brucella spp. including B. ovis [27]. Although no relationship with zinc uptake has been reported for this ABC transporter, this possibility cannot be discarded, also considering that in the KEGG database, the putative substrate-binding protein associated with this ABC transporter (BOV_0504) is annotated as a peptide/nickel transport system substrate-binding protein. The role in zinc uptake of other ABC transporters of B. ovis, and nonfunctional in B. abortus and B. melitensis, does not seem probable since deletions and/or pseudogenization have considerably reduced their number in B. ovis (59 versus 79 ABC systems detected in B. melitensis) [67].

Another striking finding obtained during this work was the strong subexpression of the siderophore biosynthesis gene dhbC (BOV_A0013) in B. ovis PA when exposed to TPEN (Figure 8A). This effect, although more moderate, was also replicated in the parental strain after exposure to EDTA and was more pronounced in the mutants for the znuACB locus exposed to the same chelator (Figure 9B and Figure 11A, Supplementary Figure S4). These results, which are accompanied by increased transcription of znuA and subexpression of the zntA zinc exporter gene (Figure 4A and Figure 11, Supplementary Figure S4B), suggest that zinc deprivation leads to a diminished transcription of siderophore biosynthesis genes. Although this effect could be triggered by chelation of other metals by TPEN or EDTA, and the expression of other genes related to iron uptake (that could increase intracellular iron level) has not been evaluated in this work, it must be noted that an opposed effect, with increased transcription of BOV_A0013, was observed after exposure to high zinc in the absence of chelators (Figure 8B). Moreover, exposure to cadmium in the absence of chelators also strongly influenced BOV_A0013 expression but in the opposite direction to zinc (Figure 8C). This gene, as well as BOV_A1093 that encodes a TonB-related OM receptor, exhibited a strongly diminished transcription in the presence of cadmium that was accompanied by a slight increase in znuA transcription suggestive of zinc deprivation conditions in the cell (Figure 8C). Zinc and cadmium interplay was previously described in B. ovis since the ZntA zinc exporter was shown to be a powerful cadmium exporter [21] as also reported in other bacterial species [22,68,69]. Accordingly, a link between the regulatory mechanisms controlling the intracellular levels of zinc, iron and cadmium seems to exist in B. ovis, as it has been reported in Salmonella where accumulation of intracellular cadmium also mimicked zinc starvation conditions [69].

It must also be noted that, in addition to the strong activity as a cadmium exporter discussed above, the contribution of the B. ovis ZntA zinc exporter to cobalt, copper and nickel detoxification was also previously evidenced [21]. Moreover, although expression studies with the different mutants of the zigA-zur locus suggested increased zinc deprivation when cultured with EDTA (Figure 11, Supplementary Figure S4) and the growth defects observed for the ΔznuACB and ΔzigAznuACB mutants of B. ovis PA in the presence of this chelator were restored by adding ZnCl₂ to the culture medium, a partial restoration of growth was also observed with other metals (Figure 10, Supplementary Figure S3). The limitations derived from using chelators, with different affinities for several metals, challenge the interpretation of these results, but they also suggest a complex regulatory network governing metal homeostasis in B. ovis that merits being evaluated in further studies.

In addition to this initial evidence about the interaction between metal homeostatic mechanisms in B. ovis, the identification of a new gene of the Zur regulon in the genus Brucella and a potential link between zinc homeostasis and histidine metabolism that also deserves further investigation, a notable observation of this work is the dispensability of the ZnuACB zinc import system for the virulence of B. ovis, at least in the mouse model, which contrasts with previous reports of its essential role in smooth Brucella species. Some plausible explanations for these in vivo apparent differences, concerning structural differences between B. ovis and smooth Brucella strains, have already been discussed above. However, the known slower growth of B. ovis and its particularities regarding host–pathogen interactions (including a different route of entry into phagocytes and a slower intracellular trafficking until reaching the replicative vacuole) [64,65,70] might also contribute to these apparent differences, constituting a relevant line for future comparative studies. However, considering that virulence has only been evaluated in the mouse model, a different behavior in the natural host cannot be discarded. Further characterization of the B. ovis mutants in zinc homeostasis systems in the natural host and/or ovine cell lines would provide a more detailed picture about the role of virulence in these systems.

5. Conclusions

Overall, this work provides relevant information about zinc homeostasis in the genus Brucella, an area of research largely unexplored in rough Brucella species such as B. ovis. The identification and characterization of a zigA ortholog, shown to be controlled by Zur and upregulated under zinc scarcity conditions, represents the first report in the genus Brucella about this gene and its link with zinc homeostasis. Although it is important to consider the limitations of using TPEN and EDTA chelators to induce a zinc-restricted environment, and that the evidence primarily comes from expression studies, the results obtained in this work also suggest a potential connection between histidine catabolism and zinc homeostasis as well as a crosstalk between homeostatic mechanisms of several metals that merit being further explored.

Evidence for additional zinc acquisition mechanisms, other than the ZnuABC system, has been obtained in B. ovis that also seem to exist in the smooth Brucella species previously analyzed. The full virulence in the mouse model of the B. ovis mutant defective in ZnuA, which contrasts with the attenuation previously described in the same model for smooth Brucella strains, points to a new differential characteristic between smooth zoonotic Brucella species and rough non-zoonotic B. ovis, species that are highly homologous at the DNA level, but display relevant differences in pathogenicity and host preference. Although the exact basis for these contrasting results in virulence remains to be elucidated, two plausible explanations may rely on (i) the absence of O-polysaccharide chains in B. ovis and other specific characteristic of its OM, which might lead to increased permeability to zinc, and/or (ii) the particular host–pathogen interactions reported for B. ovis. However, the presence of additional species-specific zinc uptake mechanisms in B. ovis cannot be ruled out and merits further investigation. This work pioneers the characterization of zinc uptake systems in B. ovis, opening new lines of research for future studies. The evaluation of the natural host or/and in ovine cell lines of the B. ovis mutants defective in zinc uptake and export systems would provide a more accurate view about the role of zinc homeostasis in the virulence of B. ovis.

Abbreviations

The following abbreviations are used in this manuscript:

ATPase	Adenosine triphosphatase
cDNA	Complementary DNA
CFU	Colony-forming unit(s)
COG0523	Cluster of Orthologous Groups 0523
DHBA	2,3-dihydroxybenzoic acid
DNA	Deoxyribonucleic acid
EDTA	Ethylenediaminetetraacetic acid
GTP	Guanosine triphosphate
GTPase	Guanosine triphosphatase
HS	Horse serum
iBAQ	Intensity-based absolute quantification
ICP-MS	Inductively coupled plasma mass spectrometry
LC-MS/MS	Liquid chromatography tandem mass spectrometry
LPS	Lipopolysaccharide
MS/MS	Tandem mass spectrometry
OD	Optical density
OM	Outer membrane
PCR	Polymerase chain reaction
RNA	Ribonucleic acid
RQ	Relative quantification
RT-PCR	Reverse transcription polymerase chain reaction
RT-qPCR	Reverse transcription quantitative polymerase chain reaction
SD	Standard deviation
SDS-PAGE	Sodium dodecyl sulfate polyacrylamide gel electrophoresis
TPEN	N,N,N′,N′-Tetrakis(2-pyridylmethyl)ethylenediamine
TSA	Trypticase soy agar
TSB	Trypticase soy broth
YE	Yeast extract

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/biology15191753/s1, Supplementary Table S1: Primers used in this work; Supplementary Figure S1: Comparison of nt sequences of the entire locus not shown zigA-zur of B. abortus 2308 and B. ovis PA (A), end-point RT-PCR reactions targeting zigA, znuA or their intergenic region in cDNA obtained from B. ovis PA and the Δzur and ΔznuACB mutants cultured in TSB-YE-HS for 16 h (B), and picture of the second SDS-PAGE gel used to evaluate the protein profile of selected B. ovis mutants and of parental strain cultured in the presence of 35 µM TPEN and 1 or 3 mM ZnCl₂ (C); Supplementary Figure S2: Growth curves of B. ovis PA-pBBR1MCS4, the ΔzntAΔzur-pBBRMCS4 double deletion mutant, and the mutant strain complemented with wild-type zntA or zur genes of B. ovis PA in TSB-YE-HS (A) or TSB-YE-HS supplemented with 0.5 mM ZnCl₂ (B). Growth curves of parental B. ovis PA and the ΔzigA, ΔznuA, ΔznuCB, Δzur, ΔznuACB, ΔzigAznuACB, and ΔzntAΔzur mutans in TSB-YE-HS (C) or the same medium supplemented with 5 mM ZnCl₂ (D), or 0.5 mM EDTA (E); Supplementary Figure S3: Growth curves of B. ovis PA and the ΔzigAznuACB mutant in the presence of 0.5 mM EDTA or with 0.5 mM EDTA plus 0.5 mM ZnCl₂ (A), NiCl₂ (B), MnCl₂ (C), CoCl₂ (D), or FeCl₂ (E), and growth curves of B. ovis PA and the ΔznuACB (F,G) and ΔzigAznuACB (H,I) mutants in the presence of 0.5 mM EDTA or with 0.5 mM EDTA plus 0.5 mM CuCl₂ (F,H) or MgCl₂ (G,I); Supplementary Figure S4: Expression of genes related to histidine homeostasis and TonB-mediated transport in the the Δzig and ΔznuA mutants of B. ovis PA (A) and expression of these genes, as well as genes related to zinc homeostasis, in the ΔzigAznuA mutant cultured in the presence of 0.5 mM EDTA (B); Supplementary Data S1: Data for Figures.

Author Contributions

Conceptualization, B.T.-C. and N.V.; methodology, B.T.-C., I.S.B., R.M.R.II and N.V.; validation, B.T.-C., I.S.B., R.M.R.II and N.V.; formal analysis, B.T.-C. and N.V.; investigation, B.T.-C. and N.V.; resources, B.T.-C. and N.V.; writing—original draft preparation, B.T.-C. and N.V.; writing—review and editing, B.T.-C., I.S.B., R.M.R.II and N.V.; visualization, B.T.-C., I.S.B., R.M.R.II and N.V.; supervision, N.V.; project administration, N.V.; funding acquisition, N.V. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

The animal study was reviewed and approved by the Bioethics Committee of the University of Salamanca and the competent authority of Junta de Castilla y León, Spain (protocol codes 464 and 1195 approved on 21 January 2020 and 3 June 2024).

Data Availability Statement

The original contributions presented in this study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Funding Statement

This work is part of I+D+i projects PID2019-107601RB-C33 and PID2023-146797OB-C33 funded by MCIN/AEI/10.13039/501100011033 and ERDF/EU. Financial support to BT-C was provided by projects PID2019-107601RB-C33 and PID2023-146797OB-C33 funded by MCIN/AEI/10.13039/501100011033 and ERDF/EU, and by Program Investigo from the Spanish Servicio Público de Empleo Estatal (University of Salamanca call) financed by NextGenerationEU (Recovery Plan, Transformation and Resilience). Research reported in this publication was also supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award number R01AI172822. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.