Role of the High-Affinity Zinc Uptake znuABC System in Salmonella enterica Serovar Typhimurium Virulence

Abstract

The Salmonella enterica serovar Typhimurium znuABC genes encoding a high-affinity zinc uptake system and its regulatory zur gene have been cloned. Salmonella serovar Typhimurium zur and znuC knockout mutants have been constructed by marker exchange. The 50% lethal dose of the znuC mutant increased when either orally or intraperitoneally inoculated in BALB/c mice, while virulence of the zur mutant decreased only when mice were intraperitoneally challenged.

Zinc plays an important role in living organisms in which the functioning of many enzymes and structural proteins is involved. In Escherichia coli, two systems have been described for zinc uptake: a high-affinity system, which is made up of the products of the znuABC genes, and a low-affinity system depending on the pitA gene (2, 16). The first is negatively regulated by the Zur protein (17), whereas the second seems to be constitutively expressed and is also related to cellular inorganic phosphate transport (2). The znuA, znuC, and znuB operons (encoding a periplasmic-binding protein, an ATPase, and an integral membrane protein, respectively) are divergently oriented, and their transcriptional starting points are separated by 24 bp (17). The Zur protein belongs to the Fur family of metalloregulatory proteins, and in the presence of Zn²⁺, it binds to a nearly perfect palindrome found in this region, blocking both znuA and znuCB transcription (17). If other genes under the Zur protein control exist, they remain to be elucidated.

The zur gene has also been functionally described in Bacillus subtilis, Staphylococcus aureus, and Listeria monocytogenes (5, 7, 13). Putative E. coli-like zur genes have also been identified in Pseudomonas aeruginosa (the np20 gene) and Vibrio cholerae (the iviXI gene) through in vivo expression technology (4, 23). Analysis of unfinished sequences of different organisms has revealed the presence of similar zur-znuABC systems in different bacterial families such as Enterobacteriaceae, Pseudomonadaceae, Neissereaceae, and Pasteurellaceae, among others (17). Few and sometimes contradictory data on the importance of the zinc uptake systems in bacterial virulence are available. All of these data correspond to extracellular pathogens. Thus, it has been described that Haemophilus influenzae znuA and P. aeruginosa np20 mutants are less virulent than wild-type strains (12, 23). On the other hand, virulence is not affected in an S. aureus zur mutant (13). Moreover, information about the role of zinc in intracellular pathogens does not exist. In this work, the importance of the zur and znuC genes in the virulence of the facultative intracellular pathogen Salmonella enterica serovar Typhimurium has been determined through the construction of zur and znuC knockout mutant derivatives.

Cloning of Salmonella serovar Typhimurium zur and znuC genes.

To clone the zur and znuC genes, similarity searches were performed on the incomplete Salmonella serovar Typhimurium sequence genome database (http://www.genome.wustl.edu/gsc) from the Genome Sequencing Center of Washington University by using their homologs from E. coli as a probe. In this way, sequences corresponding to the Salmonella serovar Typhimurium zur and znuABC genes were identified. Salmonella serovar Typhimurium ATCC 14028 chromosomal DNA was used as a template to amplify the zur and znuC genes, as well as their surrounding regions, with oligonucleotide primers designed from the data obtained in the sequencing search (Table 1). These data pointed out that, as in E. coli, the Salmonella serovar Typhimurium znuA and znuCB genes are separated by a very short intergenic region (25 bp) and they are divergently transcribed.

TABLE 1.

Oligonucleotide primers used in this work

Primer	Sequencea	Position
ZurAb	5′-GTGCCCCTCTCTACTGCCGC-3′	−395f
ZurBb	5′-CTCCGCTGCCATAATTCGCGG-3′	+548f
Zur1c	5′-GGATCCGTAACTCTTGCGTTGTGG-3′	+28f
Zur2c	5′-GAATTCTCGATATCAAGCTTAAGC-3′	−411f
ZnuC1de	5′-GAATTCGGCAAGTTGCGCAATGG-3′	−402g
ZnuC2d	5′-GAATTCGCTGTGCGCCATAATGC-3′	+1027g
ZnuC3e	5′-GGATCCGACACGTCAGAGAGGGAC-3′	+71g

^aWhen present, added restriction sites are shown in italics.

^bPrimer used to obtain the 950-bp fragment containing the S. enterica serovar Typhimurium zur gene.

^cPrimer used to obtain the 439-bp fragment containing the S. enterica serovar Typhimurium zur promoter.

^dPrimer used to obtain the 1,450-bp fragment containing the S. enterica serovar Typhimurium znuC gene.

^ePrimer used to obtain the 474-bp fragment containing the S. enterica serovar Typhimurium znuC promoter.

^fPosition of the 5′ end of the oligonucleotide with respect to the translational starting point of the S. enterica serovar Typhimurium zur gene.

^gPosition of the 5′ end of the oligonucleotide with respect to the translational starting point of the S. enterica serovar Typhimurium znuC gene.

PCR products of the expected size (943 bp and 1,447 bp containing the zur gene and the promoter of the znuA and znuC genes, respectively, as well as the encoding region of the former) were obtained. These fragments were cloned in pGEM-T, transformed in E. coli DH5α, and sequenced to confirm the presence of the desired gene. Similarity values existing between the Salmonella serovar Typhimurium and E. coli zur and znuAC genes were 85 and 95%, respectively.

Construction and phenotypic characterization of zur and znuC knockout mutants.

To obtain zur and znuC knockout mutants, a 3.5-kb Cm^r cassette from the pHP45ΩCm plasmid was inserted in either the NsiI or StuI sites of these genes, which afterwards were ligated to the pGP704 suicide plasmid, giving rise to plasmids pUA925 and pUA926, respectively. These two plasmids were then introduced by biparental mating into a Rif^r derivative of Salmonella serovar Typhimurium ATCC 14028 by using E. coli S17(λpir) as the mobilizing strain (4), and the presence of the mutation was tested by both PCR (Fig. 1) and Southern analyses (data not shown) among the Cm^r Amp^s transconjugants obtained. Since it has been described that Salmonella serovar Typhimurium Rif^r strains may be avirulent (3), the zur::Cm and znuC::Cm constructions were transferred by P22int7(HT)-mediated transduction to wild-type Salmonella serovar Typhimurium ATCC 14028 cells. For each mutant, the presence of either the zur::Cm or the znuC::Cm construction was tested by PCR in eight of the Cm^r transductants obtained (data not shown). Likewise, absence of the P22int7(HT) prophage in transductants was determined by streaking the mutants onto green plates in which Salmonella serovar Typhimurium P22 lysogenic colonies were dark green whereas nonlysogenic cells formed light-colored colonies (6).

FIG. 1.

Chromosomal DNAs from Salmonella serovar Typhimurium ATCC 14028 wild-type strain (lanes 2 and 4) and the zur (lane 1) and znuC (lane 3) mutants were subjected to PCR analysis with primer pairs ZurA-ZurB (lanes 1 and 2) and ZnuC1-ZnuC2 (lanes 3 and 4). Kilobase pairs are indicated on the left.

Salmonella serovar Typhimurium wild-type zur and znuC strains showed the same growth yield in Luria-Bertani (LB) broth (Table 2). This was probably due to the fact that the high concentration of zinc present in this medium (about 10 μM) enables cells to achieve the little quantity of this element required for bacterial growth (0.5 to 1 μM) through the low-affinity zinc uptake system, which is pitA dependent (14). The presence of the chelating agent EDTA dramatically decreased the growth of the znuC mutant, whereas zur and wild-type cells were practically not affected (Table 2). Addition of ZnSO₄ (1 mM) restored the wild-type growth yield of znuC cells in the presence of EDTA (Table 2), indicating that this growth defect was zinc specific. Furthermore, Salmonella serovar Typhimurium znuC cells containing plasmid pUA952 carrying the wild-type znuCB operon showed the same behavior as wild-type cells in the presence of EDTA (Table 2).

TABLE 2.

Comparative growth of the S. enterica serovar Typhimurium znuC mutant in the presence and absence of the chelating agent EDTA

Strain	Absorbance of culture grown ina:
	LB medium	LB + EDTA (1 mM)	LB + EDTA (1.5 mM)	LB + EDTA (1.5 mM) + ZnSO4 (1 mM)
Wild type
ATCC 14028	2.1	1.9	1.8	2
ATCC 14028 (pUA952 [ZnuCB+])	2	1.8	1.8	2.1
ZnuC mutant
UA1778	2	0.27	0.2	2
UA1778 (pUA952 [ZnuCB+])	2.1	1.9	1.8	2.1

^aOvernight LB cultures of each S. enterica serovar Typhimurium strain presenting the same growth yield (an A₅₅₀ of about 2) were diluted 1:50 into the indicated culture medium and incubated at 37°C, and A₅₅₀ was measured after 16 h of growth. For all cases, data presented are the means of three independent experiments (each in triplicate), and the single standard deviation of any value was never greater than 10%.

To further analyze Salmonella serovar Typhimurium zur and znuC mutants, the expression pattern of the zur and znuC genes in either the presence or absence of zinc was studied through transcriptional fusions with the promoterless lacZ gene contained in the broad-host range and low-copy-number pHRP309 plasmid to obtain the pUA927 and pUA928 plasmids. Both plasmids were afterwards introduced into the wild-type, zur::Cm, and znu::Cm strains. Figure 2 shows how the expression of znuC was higher in the zur::Cm cells than in both the wild-type and znuC::Cm cells, indicating that, as happens in E. coli, expression of the Salmonella serovar Typhimurium znuC gene is negatively controlled by the product of the zur gene. As expected, expression of the znuC gene in the wild-type strain was higher in the presence of EDTA than in its absence. Addition of ZnSO₄ to cells growing in the presence of EDTA repressed znuC transcription again (Fig. 2). It is worth noting that znuC transcription is higher in the znuC::Cm mutant than in the wild-type strain. This could be attributed to a lower level of intracellular zinc being present in the znuC mutant, which would trigger the expression of the zur network to which the znuC gene belongs. Although zur transcription is under its own control in S. aureus and L. monocytogenes (5, 13), Salmonella serovar Typhimurium zur gene transcription, as in E. coli (17), is not self regulated since expression of the zur-lacZ fusion is the same in wild-type cells as in zur::Cm cells (Fig. 2).

FIG. 2.

Expression of zur and znuC genes in wild-type (wt), Zur⁻, and ZnuC⁻ strains of Salmonella serovar Typhimurium determined as β-galactosidase activity through either zur-lacZ or znuC-lacZ fusions. Transcription of the znuC gene in wild-type cells in the presence of EDTA with or without ZnSO₄ (Zn) is also shown. Concentrations of EDTA and ZnSO₄ used were 1.5 and 1 mM, respectively. β-Galactosidase was measured in mid-logarithmic phase cultures growing in LB medium as described previously (6). Data presented are the averages of three independent experiments with duplicate samples.

Recently, it has been reported that Fur protein can bind zinc (1). For this reason, and to determine if any relationship exists between the Salmonella serovar Typhimurium fur and zur networks, the expression of these two genes in both the zur and fur mutants was studied. Results obtained indicate that expression of the fur gene was not affected by the presence of a mutation in the zur gene (data not shown). In agreement with this, no increase in siderophore production was detected in Salmonella serovar Typhimurium zur cells on specific growth plates (19) (data not shown). Likewise, zur and znuC transcription was not enhanced in a fur mutant with respect to wild-type cells (data not shown).

Animal virulence and survival of zur and znuC mutants in cell cultures.

To determine the 50% lethal dose (LD₅₀) of the zur and znuC strains, groups of four 6- to 8-week-old BALB/c female mice were inoculated either orally or intraperitoneally with seriated dilutions of the desired Salmonella serovar Typhimurium strain. Mortality was recorded up to 28 days postinfection, and the LD₅₀ value was calculated as reported (18). Table 3 indicates that, contrary to what happens when the regulatory protein of the iron uptake system (encoded by the fur gene) is mutagenized (8), zur cells showed the same level of virulence as the wild-type strain when orally inoculated. This attenuation of the orally inoculated fur mutant has been attributed to its higher sensitivity against acid pH (8), whereas the pH behavior of zur cells is like that of wild-type cells (data not shown). Another significant difference existing between fur and zur cells is that recA fur double mutants are only viable in the absence of oxygen (22) whereas recA zur strains support fully aerobic growth (data not shown).

TABLE 3.

Virulence in BALB/c mice and infecting ability of S. enterica serovar Typhimurium zur and znuC mutants

Strain	LD50a after inoculation		Invasion of epithelial cellsb	Survival in macrophage cellsc
	oral	i.p.
ATCC 14028 (wt)d	5.47 × 105	9	5.95 ± 1.79	1.00 ± 0.05
UA1776 (Zur−)	5.57 × 105	104	6.06 ± 1.22	1.12 ± 0.35
UA1778 (ZnuC−)	>3.25 × 107	1,110	5.85 ± 1.51	1.25 ± 0.59
UA1779 (Fur−)	>1.83 × 107	115	5.04 ± 0.44	1.09 ± 0.42

^aThe LD₅₀ value was determined as reported previously (18). In all cases, the single standard deviation of any value was never greater than 10%. i.p., intraperitoneal.

^bPercentage of bacteria (10⁻⁴) added that were amikacin resistant, after infecting epithelial cells for 20 min. Values represent the mean ± the standard deviation from two experiments with duplicate samples.

^cRatio between the number of viable intracellular bacteria at 24 and 4 h postinfection. Values represent the mean ± the standard deviation from at least two experiments with triplicate samples.

^dwt, wild type.

An approximately tenfold increase in the LD₅₀ of the zur mutant when mice were intraperitoneally challenged was found (Table 3). The reason for this decrease in the virulence of Salmonella serovar Typhimurium zur cells is still unknown, but it could be related to the fact that expression of the fliAZ transcriptional unit is lower in this mutant than in the wild-type strain (data not shown) and it has been recently demonstrated that the fliZ gene product is necessary for the virulence of Salmonella serovar Typhimurium cells (10).

The virulence of either orally or intraperitoneally inoculated znuC cells is decreased much more in comparison with that of both the wild-type and zur cells (Table 3). To confirm this attenuation of the znuC mutant, competitive experiments were carried out. Then, wild-type and znuC strains were grown separately and mixed in an approximate 1/1 ratio before animal injection. The total dose of mixed population inoculated was 2 × 10³ CFU per animal, and the concentration of both strains was checked by plating seriated dilutions of the bacterial suspensions onto LB medium before mixing. Samples of blood were taken by heart puncture immediately after the death of the mouse (4 to 5 days generally after inoculation), and the concentration of bacterial cells was determined by plating dilution series onto LB medium. The percentage of each strain was calculated afterwards by replica plating of cells recovered from the animal on LB plates supplemented with chloramphenicol at 34 μg/ml.

Data obtained indicate that znuC cells were almost fully excluded by the wild-type cells in these competition assays (Table 4). This exclusion did not exist when mixed cultures were carried out in LB medium (Table 4), indicating that the selective advantage of the wild type over the znuC mutant is specifically due to the environmental and growth conditions of bacterial cells during animal infection and must be attributed to the low concentration of free zinc existing in mammalian tissues (15). The poor availability of zinc could interfere in the growth of the Salmonella serovar Typhimurium znuC mutant in mice because it was necessary for either general bacterial cellular processes or for the activity of specific proteins induced during animal infection.

TABLE 4.

Competition assay of S. enterica serovar Typhimurium zur and znuC mutants with wild-type strain in mixed cultures^a

Mutant	CI
	Mouse	In vitro
UA1776 (Zur−)	0.54	0.75
UA1778 (ZnuC−)	<8.17 × 10−5	1.33
UA1779 (Fur−)	0.86	1.01

^aCompetitive indices (CIs) were calculated as described in the text and are defined as the output ratio (mutant/wild type) divided by the input ratio (mutant/wild type). Mouse CI is the average of the CIs determined from three mice. In vitro CI is the average from three independent cultures. In all cases, the single standard deviation of any value was never greater than 10%.

The behavior of the Salmonella serovar Typhimurium zur and znuC strains in cell cultures was also analyzed. The PK15 epithelial pig cell line (ATCC CCL-33) and the RAW264.7 mouse macrophage line (ATCC TIB-71) were used in invasion and intracellular survival experiments, respectively. Macrophages and epithelial cells were both grown at 37°C in 5% CO₂ in either Dulbecco's minimal Eagle's medium (Sigma) or minimal essential Eagle's medium (Sigma), respectively. Macrophage or epithelial lines (90% confluent growth) were infected (multiplicity of infection, 50:1) with Salmonella serovar Typhimurium zur or znuC cells grown in LB medium either without agitation to stationary phase or with agitation to mid-log phase, respectively. Infection of both cell cultures was allowed for 20 min. After internalization, the supernatant containing extracellular bacteria was removed. Mammalian cells were then washed with phosphate-buffered saline, and after the addition of fresh culture medium, incubation was allowed in the presence of amikacin (100 μg/ml) to kill extracellular bacteria remaining in the wells after washing (21). From this moment, samples were periodically taken and cells were lysed with phosphate-buffered saline containing 1% Triton X-100. The number of viable Salmonella serovar Typhimurium cells in the lysates was determined by plating serial dilutions on LB solid medium.

No relevant differences in intracellular survival or invasion rates were detected for the zur and znuC mutants in macrophages and epithelial cells with respect to those for the wild-type strain (Table 3). Similar behavior has been reported for Salmonella serovar Typhimurium fur cells (8). In concordance with this, it has been shown that production of the siderophore aerobactin in Shigella dysenteriae is important for extracellular multiplication but not for the intracellular stages of infection (9). Furthermore, mutants of Salmonella serovar Typhimurium in the mgtA or mgtCB genes involved in magnesium transport are not affected in their invasion efficiency or short-term survival within eukaryotic cells (20). This has been attributed to a very high extracellular magnesium concentration added to the eukaryotic cell growth medium which would enable Salmonella serovar Typhimurium cells to acquire the intracellular concentration required through a low-affinity system (20). A similar phenomenon to that described for magnesium may be responsible for the behavior of the znuC mutant in cell cultures. Likewise, it has also been suggested that invasive pathogens such as Salmonella serovar Typhimurium do not need high-affinity iron uptake systems once they are inside host cells (11).

In conclusion, our findings concerning virulence of the Salmonella serovar Typhimurium znuC mutant indicate that the znuABC system must be the principal one by which zinc uptake takes place in this bacterial species during mouse infection and give support to previous suggestions about the protective role that zinc-chelating systems of mammalian organisms may play against bacterial pathogens (16).