The Role of ZntA in Klebsiella pneumoniae Zinc Homeostasis

ABSTRACT Klebsiella pneumoniae is an opportunistic Gram-negative pathogen that is a leading cause of healthcare-associated infections, including pneumonia, urinary tract infections, and sepsis. Essential to the colonization and infection by K. pneumoniae is the acquisition of nutrients, such as the transition metal ion zinc. Zinc has crucial structural and catalytic roles in the proteome of all organisms. Nevertheless, in excess, it has the potential to mediate significant toxicity by dysregulating the homeostasis of other transition elements, disrupting enzymatic processes, and perturbing metalloprotein cofactor acquisition. Here, we sought to elucidate the zinc detoxification mechanisms of K. pneumoniae, which remain poorly defined. Using the representative K. pneumoniae AJ218 strain, we showed that the P-type ATPase, ZntA, which is upregulated in response to cellular zinc stress, was the primary zinc efflux pathway. Deletion of zntA rendered K. pneumoniae AJ218 highly susceptible to exogenous zinc stress and manifested as an impaired growth phenotype and increased cellular accumulation of the metal. Loss of zntA also increased sensitivity to cadmium stress, indicating a role for this efflux pathway in cadmium resistance. Disruption of zinc homeostasis in the K. pneumoniae AJ218 ΔzntA strain also impacted manganese and iron homeostasis and was associated with increased production of biofilm. Collectively, this work showed the critical role of ZntA in K. pneumoniae zinc tolerance and provided a foundation for further studies on zinc homeostasis and the future development of novel antimicrobials to target this pathway. IMPORTANCE Klebsiella pneumoniae is a leading cause of healthcare-associated infections, including pneumonia, urinary tract infections, and sepsis. Treatment of K. pneumoniae infections is becoming increasingly challenging due to high levels of antibiotic resistance and the rising prevalence of carbapenem-resistant, extended-spectrum β-lactamases producing strains. Zinc is essential to the colonization and infection by many bacterial pathogens but toxic in excess. This work described the first dissection of the pathways associated with resisting extracellular zinc stress in K. pneumoniae. This study revealed that the P-type ATPase ZntA was highly upregulated in response to exogenous zinc stress and played a major role in maintaining bacterial metal homeostasis. Knowledge of how this major bacterial pathogen resists zinc stress provided a foundation for antimicrobial development studies to target and abrogate their essential function.

identity, WP_002882907.1). Examination of 2,706 publicly available Klebsiella spp. genomes (18) revealed that all the putative zinc efflux genes were highly conserved in the Klebsiella population (Fig. 1A). All four genes were carried in .99.6% of genomes analyzed with high sequence conservation of present genes exceeding 99.4% pairwise identity (Table S1). Taken together, these data show that the putative zinc export pathways are highly conserved in the global Klebsiella spp. population and indicate that K. pneumoniae AJ218 is a suitable representative model for investigation of Klebsiella spp. zinc homeostasis.
The contribution of these systems to K. pneumoniae zinc tolerance was investigated using a combination of molecular and phenotypic approaches. Supplementation of Luria Bertani (LB) broth with increasing concentrations of ZnSO 4 (0 to 1000 mM) had no discernible effect on the growth phenotype of K. pneumoniae AJ218 (Fig. 1B). Wholecell accumulation of zinc in K. pneumoniae AJ218, determined by inductively coupled plasma mass spectrometry (ICP-MS), revealed no significant differences between cultures grown in 0 mM and 400 mM ZnSO 4 -supplemented media (Fig. 1C). Despite the lack of a discernible phenotypic impact, quantitative real-time PCR (qRT-PCR) revealed that zntA transcription was upregulated 5.7-fold in 400 mM ZnSO 4 -supplemented LB relative to the nonsupplemented medium. In contrast, no significant changes (,2-fold induction, P . 0.05) in the expression profiles of the other putative zinc transporter homologs, zitB, zntB, and fieF, were observed (Fig. 1D). Taken together, these data implicate ZntA as the primary K. pneumoniae AJ218 zinc efflux pathway expressed in response to zinc stress.
with the zinc/cadmium-binding P 1B -2-type ATPases Shigella sonnei ZntA (77.1% identity) and E. coli ZntA (76.9% identity) (Fig. 1E). Amino acid sequence comparisons revealed that functionally important regions, such as the CXXC and CPC zinc/copper metal-binding motifs, were conserved in K. pneumoniae ZntA (Fig. 1F). Moreover, K. pneumoniae ZntA also contains zinc ion coordinating residues within the M5 and M6 transmembrane domains, which are absent from copper-coordinating P 1B -type ATPase CopA homologs, indicating a potential capacity for cadmium binding. Notably, ZntAmediated cadmium export has been reported for E. coli and Salmonella typhimurium (20,21). Accordingly, the contribution of ZntA to the homeostasis of zinc, cadmium, and other transition metal ions was investigated.
To elucidate the contribution of ZntA to K. pneumoniae AJ218 zinc homeostasis, a zntA deletion mutant was generated, and this was complemented in trans on a lowcopy-number plasmid with its native promoter (pACYC184::zntA; pZntA). The wild-type and derivative strains were assessed for their growth phenotype and cellular accumulation of metal ions in a zinc-supplemented LB medium. All strains showed comparable growth in a nonsupplemented LB medium ( Fig. 2A). However, the DzntA strain showed perturbed growth in 400 mM ZnSO 4 relative to the wild-type and complemented strain and abrogated growth at 800 mM ZnSO 4 ( Fig. 2B and C). Whole-cell metal accumulation of the K. pneumoniae strains revealed that the loss of zntA had no significant effect on cellular zinc in nonsupplemented media (Fig. 2D). Growth in 400 mM ZnSO 4 supplemented LB resulted in a 3.4-fold increase in zinc accumulation in the DzntA mutant (P , 0.0001) relative to the wild-type and complemented strains (Fig. 2D).
The viability of DzntA in 400 mM ZnSO 4 supplemented LB medium (Fig. 2B) suggested that other putative zinc export systems might compensate for the deletion of the P-type ATPase. Accordingly, qRT-PCR analysis of the putative zinc export systems was performed on K. pneumoniae AJ218 and the mutant derived strains grown in 400 mM supplemented ZnSO 4 media. Analysis of the DzntA strain relative to the wild-type and complemented strains revealed that zitB and zntB were not significantly (P . 0.05) upregulated in the presence of exogenous zinc stress (Fig. 2E) while fieF was significantly downregulated (P , 0.05). Collectively, these data support a role for ZntA as a crucial zinc efflux pathway in K. pneumoniae AJ218. The lack of upregulation of the alternate putative zinc efflux transporters in K. pneumoniae suggested that they may not contribute in a significant manner to mitigating changes in bacterial zinc and/or their expression is uncoupled with cellular zinc abundance. Although zitB and fieF expression in response to zinc stress has been reported (14), the balance of evidence from recent studies suggests that zitB is constitutively expressed and does not respond to exogenous zinc abundance (15), while fieF expression is regulated by intracellular iron levels in a Fur-independent manner (13). Therefore, these data support a primary role for ZntA in responding to dynamic changes in K. pneumoniae zinc abundance. Although ZitB, ZntB, or FieF have the potential to contribute to Klebsiella zinc homeostasis, no definitive role(s) can be inferred based on the data arising from the experimental conditions investigated. Given the prominent contribution of ZntA to zinc homeostasis, its functional role was further investigated.
ZntA contributes to cadmium homeostasis. Whole-cell metal accumulation revealed that K. pneumoniae strains grown in the zinc-supplemented medium had decreased cadmium accumulation by comparison to untreated cells (Fig. 3A). Notably, cadmium accumulation was ;5.0-fold higher in the DzntA strain relative to the wild-type and complemented strains, which is consistent with a role for ZntA in cadmium efflux. To probe the contribution of ZntA to cadmium resistance, strains were grown in increasing CdCl 2 concentrations (1 to 200 mM). This revealed that the DzntA strain was highly susceptible to cadmium stress relative to the wild-type and complemented strains ( Fig. 3B to D). At a subinhibitory concentration of cadmium (10 mM CdCl 2 ), the qRT-PCR analysis showed that zntA transcription was upregulated relative to untreated cells in wild-type and complement strains (P , 0.001, Fig. 3E). This shows that cadmium accumulation is associated with upregulation of zntA and is suggestive of a protective role against cadmium. However, cadmium-induced upregulation of zntA may also result in a concomitant perturbation of zinc homeostasis. Therefore, the impact of cadmium stress on metal ion homeostasis was investigated in wild-type and mutant derivative K. pneumoniae strains.
Supplementation of the growth medium with 10 mM CdCl 2 increased cadmium levels in all strains (Fig. 3F). Accumulation of cadmium was higher in the DzntA strain relative to the wild-type but was not significantly different from the complemented strain. In contrast, zinc accumulation was significantly decreased in all strains upon exposure to cadmium to similar levels (Fig. 3G). The DzntA strain also showed reduced cellular zinc suggesting that upregulation of zntA in the wild-type and complemented strains ( Fig. 3E) was not the primary cause of zinc depletion. It, therefore, follows that cadmium must be disrupting zinc import in K. pneumoniae AJ218. Accordingly, we investigated the hypothesis that cadmium ions could compete with zinc for import using a transcriptional approach. Expression of znuA, the gene encoding the putative solutebinding protein associated with the conserved bacterial zinc ATP-binding cassette permease ZnuABC, was monitored. Our data show that znuA was upregulated in K. pneumoniae AJ218 upon exposure to cadmium consistent with a transcriptional response to cellular zinc limitation (Fig. 3H). Taken together, the transcriptional analyses of zntA and znuA support a model wherein zinc accumulation is impaired by cadmium, at least in part, due to competition between the two metal ions for the zinc import pathway(s) of the bacterium. Further studies to elucidate the zinc import machinery of K. pneumoniae and the relative permissiveness of these pathway(s) to interact with cadmium ions are warranted.
Loss of ZntA dysregulates the homeostasis of other transition metals. The impact on the homeostasis of other transition metal ions was also investigated in the K. pneumoniae DzntA strain. During exposure to 400 mM ZnSO 4 , iron levels were significantly decreased (2.6-fold, P , 0.0001; Fig. 4A), consistent with the observed downregulation of fieF, while manganese levels significantly increased (10.9-fold; P , 0.01; Fig. 4B) in the DzntA strain relative to the wild-type and complemented strains. This pattern of dysregulated iron and manganese homeostasis was also observed in 10 mM CdCl 2 -treated K. pneumoniae DzntA relative to the wild-type and complemented strains ( Fig. 4C and D). Notably, zinc stress did not perturb iron accumulation in wild-type AJ218 or the complemented strain (Fig. 4A), although cadmium treatment exerted a  modest, but significant, iron depletion effect (Fig. 4C). Collectively, these data indicate that loss of ZntA renders K. pneumoniae susceptible to manganese and iron dyshomeostasis via exogenous zinc or cadmium stress.
To further investigate these dyshomeostasis impacts, the transcriptional profile of a selection of genes putatively associated with iron import (fur, entB, icuB, and feoB), and manganese uptake (sitA) and export (mntP), were analyzed in the K. pneumoniae strains exposed to 400 mM ZnSO 4 (Fig. 4E). These results show that the enterobactin siderophore biosynthesis gene, entB, and ferrous iron uptake system, feoB, were significantly downregulated (P , 0.001) in the DzntA strain in the presence of extracellular zinc relative to the wild-type and complemented strains. The downregulation of these genes associated with iron uptake may be due to zinc-mismetallation of the ferric-uptake regulator Fur. However, this contrasts with E. coli, wherein increased zinc accumulation is associated with derepression of iron uptake genes and repression of iron-storage genes (8). This disparity may be explained by the observed increase in cellular manganese, also in contrast to what has been previously reported in E. coli (8). Previously it has been shown in E. coli and S. aureus that increased cellular manganese results in decreased iron levels (22,23). This has been attributed to the manganese-based blockade of Fur activity. Thus, in K. pneumoniae DzntA the increased cellular quotient of manganese may be contributing to dysregulation of the cellular iron pools either directly or concomitantly with zinc. Elucidation of the precise molecular basis warrants further investigation.
K. pneumoniae strains grown in 400 mM ZnSO 4 revealed that transcription of the manganese import (sitA) and efflux (mntP) pathways were both significantly downregulated (P , 0.0001; Fig. 4E) in the DzntA strain relative to the wild-type and complemented strains. Downregulation of both pathways may be due to zinc-mediated dysregulation of the manganese metalloregulatory control, which is comprised of MntR, Fur, and manganese-dependent riboswitch regulatory elements (24). This could arise from zinc-mediated repression of mntP, resulting in an increased level of cellular manganese. The increased abundance of manganese would then trigger repression of the sitABCD import pathway in a manganese-dependent manner. However, an alternative explanation may be that ZntA directly contributes to manganese homeostasis as an efflux pathway, which is similar to the P 1B -type ATPase CtpC (25). To test that hypothesis, K. pneumoniae AJ218 and the mutant derivative strains were grown in increasing concentrations of MnSO 4 (1 to 2000 mM) to investigate the potential for manganese toxicity (Fig. S1). No phenotypic growth impact was observed for any of the strains in response to manganese. Further analysis at the supraphysiological concentration of 400 mM MnSO 4 showed no change in growth (Fig. 4F) and a modest increase in manganese relative to untreated cells in all strains (Fig. 4G). Importantly, the DzntA strain did not have significantly more manganese than the wild-type or complemented strains, and manganese levels were less than those observed for the zinc or cadmium stressed treatments ( Fig. 4B and D). Taken together, these analyses indicate that ZntA does not serve as a manganese efflux pathway in K. pneumoniae. The increased manganese accumulation observed in the DzntA under zinc stress warrants further investigation in future studies. Accumulation of iron was unaffected by manganese stress in all strains (Fig. 4H). However, zinc levels were significantly increased in the DzntA strain relative to the wild-type and complemented strain in the presence of manganese (Fig. 4I). This may suggest that cellular manganese influences the expression of zinc uptake and/or efflux, but the extent of this cross-talk did not have an apparent physiological impact.
Biofilm formation aids in resistance to zinc stress. K. pneumoniae forms biofilms during infection of the respiratory, gastrointestinal, and urinary tracts (26). Biofilm formation is one mechanism by which some bacteria resist extracellular stress, such as heavy metal ions (27)(28)(29). Iron is established as an important trace element for Klebsiella biofilm formation (30,31), while multiple studies have shown that cellular zinc is a critical cofactor for cell surface adhesins that contribute to the formation and stability of biofilms (32)(33)(34). Despite this, high concentrations of extracellular zinc have been shown to exert mild inhibitory effects on the biofilm formation capacity of urinary tract isolates of K. pneumoniae (35). It was hypothesized that this arose due to zinc-mediated inhibition of iron uptake via mismetallation of Fur, although direct evidence remains lacking. Therefore, the influence of zinc on K. pneumoniae AJ218 biofilm formation was investigated.
The K. pneumoniae AJ218 wild-type and derivative strains were grown in M9 minimal media supplemented with subinhibitory zinc concentrations (12.5 and 25 mM ZnSO 4 ) in microtiter plates. The surface-adhered biomass was measured by crystal violet staining and standardized to the colony forming unit (CFU) within the biofilm to account for potential zinc stress-related impacts to cell density. At subinhibitory concentrations of zinc, the DzntA strain produced more biofilm, while the wild-type and complemented strains did not significantly change (Fig. 5A). These data indicate that the inability to efflux zinc efficiently leads to increased biofilm formation when excess zinc is present in the external milieu.
To ascertain whether biofilm formation was enhancing stress resistance, the relative susceptibility of the planktonic and biofilm K. pneumoniae cells to zinc stress was investigated. This was addressed by growing the cells in M9 minimal media, isolating biofilm and planktonic cells, and then challenging them with 800 mM ZnSO 4 and measuring survival over 2 h of the challenge. This revealed that all strains had comparable survival of biofilm cells over the first 1 h of challenge with only the DzntA strain showing a lower survival by 2 h, by comparison to the wild-type and complemented strains. The planktonic cells of all strains showed reduced survival to zinc challenge relative to the biofilm-forming cells with less than 10% survival by 1 h (Fig. 5B and C). Notably, the planktonic DzntA strain succumbed to zinc stress more rapidly than the wild-type or complemented strains (Fig. 5B). These data suggest that K. pneumoniae AJ218 biofilm formation contributes to resistance against zinc stress. This could arise from the biofilm matrix components chelating or slowing permeation of zinc ion diffusion and thereby protecting K. pneumoniae. In particular, extracellular DNA (eDNA), which is a major constituent of biofilm matrices, has known cation-chelating properties (36).
Collectively, this work elucidates the primacy of ZntA as the major efflux pathway for zinc and cadmium detoxification in K. pneumoniae. Our data highlights that strict regulation of zinc homeostasis is critical to prevent dyshomeostasis of iron and manganese, suggesting that zinc intoxication has the potential to disrupt other cellular processes. Resistance to zinc intoxication can be enhanced by the formation of biofilm, which suggests that components of the biofilm matrix may act to buffer or sequester metal ions. These insights advance our understanding of the molecular mechanisms of K. pneumoniae zinc homeostasis, which may serve as potential targets in future antimicrobial development approaches.

MATERIALS AND METHODS
Bioinformatic analyses. Putative K. pneumoniae AJ218 zinc resistance mechanisms were identified by BLASTP alignment to E. coli K-12 MG1655 candidates. Maximum likelihood phylogenetic analyses of the P 1B -type ATPases were performed through the web server NGPhylogeny workflow using MAFFT, BMGE, and PhyML software following the LG substitution model (37). Branch statistics were calculated using 1000 bootstrap replications.
Comparative genomics. A database of 2,706 publicly available Klebsiella genomes and associated maximum likelihood phylogenetic tree (18) was screened for zinc homeostasis genes by the BLASTN screening tool, Screen Assembly (v1.2.7) (38), applying cutoffs of 80% identity and 80% reference length. To reduce the false-negative rate (e.g., due to contig breaks), gene absence was further validated by screening 300 bp segments of each target gene (3 segments for zntB, zitB, fieF; 5 for zntA; Table S1), and hits were mapped onto the phylogenetic tree using Interactive Tree Of Life (iTOL version 6.1) (39). Translated full-length protein sequences were used for variation analysis by MUSCLE Alignment (version 3.8.425) (40).
Bacterial strains, chemicals, media, and growth. Bacterial strains and plasmids used in this study are listed in Table 1. All strains were routinely cultured with Luria Bertani broth (LB). For growth assays, overnight cultures of K. pneumoniae were standardized (optical density at 600 nm (OD 600 ) = 0.05) and incubated at 37°C with shaking in a FLUOStar Omega spectrophotometer. Media were supplemented, where appropriate, with kanamycin (Km) 50 mg/mL; chloramphenicol (Chl), 30 mg/mL for E. coli, and 80 mg/mL for K. pneumoniae; and ZnSO 4 , MnSO 4, or CdCl 2 as stated in the text.
Construction of K. pneumoniae gene deletion mutants. The zntA deletion mutant was generated by adapting the gene gorging method (41) as follows: using the NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs, NEB) to flank a kanamycin resistance cassette from pKD4 with ;500 bp upstream and downstream of the zntA gene. Two hundred nanograms of the linear construct were electroporated into electrocompetent K. pneumoniae AJ218 cells containing the plasmid pACBSR, encoding  (45) an L-arabinose inducible lambda Red recombinase gene. Complementation vectors were made by amplification of AJ218 zntA and 300 bp upstream, including its native promoter and insertion into the tet gene of pACYC184, using NEBuilder. All primers are listed in Table S2. Cellular metal ion content analysis. Overnight K. pneumoniae cultures were standardized to an OD 600 of 0.05 and grown in biological triplicate to mid-log-phase (OD 600 = 0.8 to 1) in untreated and metal supplemented media. Cultures were washed twice with phosphate-buffered saline (PBS) containing 5 mM ethylenediaminetetraacetic acid, and twice with PBS. Bacterial pellets were desiccated at 95°C overnight then digested in 250 mL 65% (vol/vol) HNO 3 at 95°C for 20 min. Samples were centrifuged at 18,000 Â g for 25 min and soluble material was diluted to a final concentration of 3.25% HNO 3 using MilliQ-H 2 O. Elemental content was quantitatively analyzed in technical triplicate using an Agilent 8900 ICP-QQQ-MS (Agilent Technologies).
RNA isolation and qRT-PCR. K. pneumoniae strains were grown to mid-log-phase (OD 600 = 0.8 to 1) and harvested by centrifugation (18,000 Â g, 5 min) and resuspended in RNA Protect Bacteria Reagent (Qiagen). Total RNA was extracted using the RNeasy minikit, digested twice with on-the-column DNase I digestion, as per the manufacturer's guidelines. Samples were quantified using the FLUOStar Omega Spectrophotometer.
Quantitative real-time PCR was performed on the QuantStudio 7 real-time PCR system (Thermo Fisher Scientific) using the SuperScript III Platinum SYBR Green one-step qPCR Mix according to the manufacturer's instructions. Oligonucleotides used for qRT-PCR are listed in Table S2. The gene rpoD was used to normalize gene expression. The data represent at least three biological replicates.
Biofilm culture and quantification. Biofilm cultures were grown in duplicate in 24-well microtiter plates. Bacterial cultures were normalized (OD 600 =0.05) in M9 minimal media (42) and statically incubated for 24 h at 37°C with ZnSO 4 supplementation. After 24 h growth, planktonic cells were discarded, and biofilm cells were washed with PBS. For analysis, one replicate was used to determine CFU by a detachment of biofilm cells by pipetting into 1 mL PBS followed by serial dilution and viable count on LB agar. The other replicate was used to quantify biofilm, adapting the crystal violet stain method (43) with wash by aspiration. Biofilm mass was quantified by measuring absorbance (595 nm) relative to CFU.
Zinc susceptibility assay. Biofilm cultures were grown for 24 h in untreated media. From each well, 1 ml of planktonic culture was centrifuged at 18,000 Â g for 10 min and the remaining biofilm was washed once with saline solution (0.85% NaCl). A saline solution containing 800 mM ZnSO 4 was added directly onto the biofilm cells and used to resuspend the planktonic culture. The cells were incubated statically at 37°C and CFU/mL was determined by viable count. Relative survival was determined by comparison to an untreated saline condition.
Statistical analysis. Unless otherwise stated, the data represent the mean of biological triplicates (6standard error of the mean [SEM]). Statistical analyses were performed using a two-tailed Student's t test when comparing two data sets or a one-way ANOVA test for .2 data sets. Comparisons to the wildtype or untreated condition were realized by the Tukey or Sidak posttest as specified in figure legends; ns = not significant; *, P , 0.05; **, P , 0.01; ***, P , 0.001; and ****, P , 0.0001.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, XLSX file, 0.2 MB. SUPPLEMENTAL FILE 2, PDF file, 0.3 MB.