Salmonella Utilizes Zinc To Subvert Antimicrobial Host Defense of Macrophages via Modulation of NF-κB Signaling

We utilized the RAW264.7 macrophage cell line and the intracellular bacterium Salmonella Typhimurium to study the impact of cellular zinc availability on host-pathogen interactions.

Infection of RAW264.7 macrophages with S. Typhimurium resulted in a significant increase in the level of labile intracellular zinc, as measured by Fluozin fluorescence (Fig. 1A). In parallel, the mRNA expression of two important Zn-binding metallothioneins (MTs), MT1 and MT2 (33), was induced, which could be considered a host response mechanism to limit labile zinc levels in cells (Fig. 1B). Of interest, we detected only a minor increase in the total (i.e., free and protein-bound) cellular zinc content following infection using atomic absorption spectrometry (Fig. 1C). This suggests that primarily a shift of zinc from protein-bound sources to the free intracellular zinc pool, rather than the acquisition of the metal from the extracellular space, may occur in macrophages upon Salmonella infection.

To study the biological importance of altered zinc availability for the course of Salmonella infection in macrophages, we modulated zinc availability by adding either ZnCl₂ or N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN) (34), which specifically chelates intracellular zinc, to macrophages infected with green fluorescent protein (GFP)-expressing Salmonella bacteria (see Fig. S1A in the supplemental material). Neither of the treatments diminished cell viability (Fig. S1B). Interestingly, zinc supplementation significantly elevated the percentage of Salmonella-infected macrophages and total numbers of bacteria (Fig. 2A and Fig. S2A). In turn, zinc chelation had an opposite effect on both readouts of infection (Fig. 2A). Of note, the same effects of zinc modulation on the infection rate were observed in mouse primary bone marrow-derived macrophages (Fig. S3). This observation could not be explained by an alteration of the phagocytic uptake of particles such as bacteria or beads, as zinc supplementation decreased rather than increased the phagocytic capacity of macrophages (Fig. S2B and S2C). This led us to hypothesize that the infection-promoting action of increased cellular zinc levels occurs after the entry of the bacteria into the cell. Zinc addition had no effects on bacterial survival and proliferation in in vitro cultures, hence excluding the possibility that additional zinc favors extra- or intracellular bacterial growth, resulting in a higher macrophage infection rate (Fig. S4).

To test whether free zinc directly alters intracellular bacterial proliferation, we determined the number of viable bacteria per cell following ZnCl₂ or TPEN treatment. Relative numbers of bacteria per cell remained unaffected upon this treatment (Fig. S5A), suggesting that not intracellular pathogen proliferation but rather the susceptibility of macrophages to bacterial invasion is increased upon higher zinc availability.

We then sought to test whether zinc can inhibit the immediate killing of Salmonella by macrophages. We thus quantified the percentages of macrophages containing viable and dead bacteria upon zinc modulation and infection with the red fluorescent protein (RFP)-expressing Salmonella reporter strain. Intracellular bacteria were additionally stained with an anti-CSA-1 antibody, which recognizes both live and dead Salmonella bacteria. Since only metabolically active bacteria produce RFP, we defined viable Salmonella bacteria as being CSA-1 positive (CSA-1⁺) RFP⁺ and dead ones as being CSA-1⁺ RFP negative (RFP⁻) (Fig. S5B). As shown in Fig. 2B, surplus zinc significantly reduced the number of macrophages containing dead bacteria, whereas TPEN treatment increased it. These data led us to hypothesize that an increase in the level of free zinc in macrophages may be employed by Salmonella to resist the antimicrobial elimination pathways of the host.

The generation of ROS and RNS by NOX and iNOS in macrophages serves as an efficacious mechanism to eliminate intracellular pathogens (2, 22). These pathways are also of importance in our in vitro infection system, since treatment of macrophages with a ROS scavenger, N-acetyl-cysteine (NAC), or an iNOS inhibitor, N⁶-(1-iminoethyl)-l-lysine (l-NIL), increased the numbers of intracellular bacteria (details not shown).

Since elevated free zinc levels hindered the effective clearance of Salmonella, we investigated whether the modulation of zinc concentrations affected ROS and RNS production upon bacterial infection. We detected a drop in nitrite production (the stable end product of the NO pathway) after zinc supplementation (Fig. 3A). The formation of ROS and RNS is tightly controlled by the transcriptional, posttranscriptional, and posttranslational regulation of ROS/RNS-generating enzymes. Upon zinc supplementation, the mRNA levels of iNOS and the key NOX subunit p47phox were significantly reduced (Fig. 3B). Accordingly, a significant reduction of the iNOS protein level was observed as well (Fig. 3C). Of note, treatment with TPEN did not increase p47phox mRNA and iNOS transcript and protein amounts above the control levels, suggesting that the TPEN-mediated improvement of the elimination of bacteria may also involve some other radical-independent antimicrobial pathways.

To test for the possibility that zinc stimulation can protect bacteria against the detrimental reactive oxygen and nitrogen species in a macrophage-independent manner, we monitored the growth of wild-type (WT) Salmonella bacteria in the presence of ROS (paraquat) and RNS (S-nitroso-N-acetyl-dl-penicillamine [SNAP]) generators with zinc and TPEN supplementation. As shown in Fig. S6 in the supplemental material, neither the addition of zinc nor its chelation impacted the survival and proliferation of the bacteria under conditions of oxidative and nitrosative stress.

Nuclear factor kappa B (NF-κB) is a master regulator of iNOS and p47phox mRNA expression (2, 35). Having observed that surplus zinc downregulates both of these transcripts, we asked whether zinc can affect NF-κB transcriptional activity. The phosphorylation and nuclear accumulation of the p65 subunit are surrogates for NF-κB activation. The addition of zinc reduced p65 phosphorylation in infected macrophages, whereas zinc chelation strongly enhanced it (Fig. 4A). Similarly, we observed impaired p65 nuclear trans-localization after zinc addition and the opposite phenomenon after TPEN treatment (Fig. 4B; see also Fig. S7 in the supplemental material). To further corroborate the impact of zinc on the transcriptional activity of p65 NF-κB, we performed a chromatin immunoprecipitation assay with a phospho-p65 antibody. As shown in Fig. S5A in the supplemental material, zinc almost completely abrogated p65 binding to the proximal site in the iNos promoter after Salmonella infection as well as to promoters of other classical p65 target genes, Icam and Cxcl2. Zinc chelation in turn increased the affinity of p65 for both binding sites in the iNos promoter region (Fig. S8A). Our finding that zinc supplementation impairs p65 transcriptional activity in general is further supported by the effects of surplus zinc on mRNA levels of the NF-κB-regulated cytokine genes Il1b, Il6, Il10, and Tnf (Fig. S8B). Notably, these effects were specific for the NF-κB pathway, since the activity of extracellular signal-regulated kinase (ERK), STAT1, and STAT3 signaling remained unchanged (Fig. S9).

Taken together, these data show that the expansion of cellular zinc levels impairs the killing capacity of macrophages by inhibiting NF-κB activation and the downstream transcriptional activation of iNOS, phox47, and several proinflammatory cytokines.

As an intracellular pathogen, Salmonella developed a plethora of strategies to bypass antimicrobial effector pathways of macrophages. We hypothesized that the cellular accumulation of free cytoplasmic zinc causing the inhibition of ROS/RNS formation may pose such a bacterium-driven protective mechanism.

To test this, we infected macrophages with RFP reporter-expressing Salmonella bacteria and sorted bacterium-hosting (RFP⁺) and bacterium-cleared/uninfected (RFP⁻) macrophages from these cultures. Macrophages containing viable Salmonella bacteria accumulated significantly more free zinc (Fig. 5A) and expressed more zinc-sensitive Mt1 and Mt2 transcripts (Fig. 5B) than did Salmonella-naive or Salmonella-clearing macrophages. As with zinc supplementation, we also observed significantly reduced ROS production (Fig. 5C) and iNOS mRNA (Fig. 5D) and iNOS protein (Fig. 5E) expression levels in Salmonella-hosting macrophages compared to bacterium-negative cells. Importantly, this was paralleled by a profound downregulation of phosphorylated p65 and an inhibition of the downstream p38 pathway in RFP⁺ macrophages compared to Salmonella-clearing ones (Fig. 5F). No such consistent regulatory pattern was seen among other signaling pathways involved in antibacterial activity, such as STAT1, STAT3, and ERK (see Fig. S10 in the supplemental material). These data support the hypothesis that either increased zinc levels in macrophage are a predisposing factor for Salmonella infection or Salmonella invasion of macrophages may promote cytoplasmic zinc accumulation, which impairs the immediate NF-κB-mediated antimicrobial host defense.

The zinc-binding metallothioneins MT1 and MT2 are vital regulators of cellular free zinc levels. Their expression is induced upon Salmonella infection and specifically increased in Salmonella-hosting macrophages. We hypothesized that the upregulation of MT1 and MT2 may pose a host response mechanism to limit zinc availability for intracellular bacteria. To study their specific role in our infection model, we disrupted the Mt1 and Mt2 genes in macrophages using the clustered regularly interspaced short palindromic repeat (CRISPR)-Cas9 technique (see Fig. S11A in the supplemental material). This event results in a boost of cellular free zinc levels, even without zinc supplementation, in resting and infected cells (Fig. S11B and S11C). Comparably to zinc supplementation, Mt knockout cells displayed increased Salmonella colonization (Fig. 6A and Fig. S11D), reduced ROS levels (Fig. 6B), and impaired RNS production as assessed by nitrite and iNOS protein levels (Fig. 6C) compared with wild-type macrophages.

These results further corroborate the key impact of labile intracellular zinc on intracellular bacterial survival by modulating antimicrobial ROS and RNS formation in Salmonella-infected macrophages. In addition, we identify infection-driven MT1 and MT2 expression with the subsequent reduction of intracellular zinc levels as an important antibacterial host response mechanism.

Wild-type Salmonella enterica serovar Typhimurium bacteria were obtained from the ATCC (ATCC 14028). Reporter GFP-expressing (ST-GFP) and RFP-expressing (ST-RFP) Salmonella strains were described previously (49, 50).

The murine macrophage cell line RAW264.7 was kept in Dulbecco's modified Eagle's medium (DMEM) with 2 mM l-glutamine and 10% fetal calf serum (FCS) (Biochrom). Cells were routinely collected by scraping. Cells were seeded at a density of 5 × 10⁴ cells/cm² a day before infection. Infections with WT, ST-GFP, and ST-RFP Salmonella bacteria were performed at a multiplicity of infection (MOI) of 1:10, as described previously (23), 1 h after bacterial challenge medium was changed to a gentamicin-containing one (25 μg/ml). A total of 100 μM ZnCl₂ or 5 μM TPEN (Sigma) was added to the cultures concomitantly with the bacteria and was present for the whole infection period.

Flow cytometry measurements were performed with a Beckman-Coulter Gallios device and analyzed with FlowJo software (FlowJo LLC). ΔMFI is defined as the difference in the signal intensity between an unstained control and a stained sample. For sorting, a FACSAria I device (Becton Dickinson) was used. Microscopy images were acquired with an Axioskop fluorescence microscope (Zeiss) and analyzed with ImageJ software.

Macrophage cultures were incubated with 5 μM Fluozin 3-AM (Life Technologies) for 1 h at 37°C. Subsequently, medium was removed, and cells were incubated in phosphate-buffered saline (PBS) for 30 min at 37°C. The green fluorescence of the probe was detected by fluorescence microscopy and flow cytometry. Total zinc concentrations were measured by atomic absorption spectrometry as described previously (51).

RNA was isolated and reverse transcribed as described previously (52). Relative gene expression was calculated with the ΔΔC_T method, normalizing the results to the value for the Hprt gene. The list of primer oligonucleotide sequences can be found in Table S1 in the supplemental material. Protein extraction and Western blotting were performed as described previously (52), using primary rabbit antibodies against NF-κB p65 (catalog no. 4764), pSer536–NF-κB p65 (catalog no. 3033), p38 (catalog no. 9212), phosphorylated p38 (p-p38) (catalog no. 4511), ERK (catalog no. 4695), p-ERK (catalog no. 4370), STAT1 (catalog no. 9172), p-STAT1 (catalog no. 9167), STAT3 (catalog no. 9132), p-STAT3 (catalog no. 9134), and actin (catalog no. A2066) (all from Cell Signaling Technology, except for actin, which was purchased from Sigma-Aldrich). The dilution of primary antibodies was 1:1,000. As a secondary antibody, a goat anti-rabbit antibody conjugated to horseradish peroxidase (HRP) (catalog number P0448; Dako) was used (1:2,000).

Salmonella-infected cells were fixed with 4% formalin in PBS for 15 min at room temperature, washed twice with PBS, and blocked in blocking buffer (5% horse serum and 0.3% Triton X-100 in PBS) for 1 h. Cells were stained with a 1:200 dilution of a primary antibody against NF-κB p65 (catalog no. 4764; Cell Signaling Technologies), diluted in antibody dilution buffer (1% bovine serum albumin [BSA] and 0.3% Triton X-100 in PBS) at 4°C overnight. Next, staining with a 1:200 dilution of Dylight488 donkey anti-rabbit IgG (BioLegend) for 1 h was performed, followed by 3 PBS wash steps and mounting with 4′,6-diamidino-2-phenylindole (DAPI)-containing fluorescence mounting medium (Dako). Cytoplasmic and nuclear NF-κB mean fluorescence intensities were measured by using ImageJ software (see Fig. S4 in the supplemental material).

Macrophages infected with RFP-expressing Salmonella bacteria were harvested and fixed with 4% formalin in PBS for 15 min at room temperature. Fixed cells were permeabilized with permeabilization buffer (0.05% Triton X-100 in PBS) at room temperature for 15 min. For intracellular iNOS staining, cells were incubated for 1 h with a 1:200 dilution of allophycocyanin (APC)–anti-iNOS antibody (eBioscience). For the killing assay, cells were incubated for 2 h with a 1:200 dilution of biotinylated antibody to Salmonella CSA-1 (catalog no. 16-91-99; KPL) in permeabilization buffer, followed by secondary staining with a 1:200 dilution of Alexa Fluor 647-streptavidin (BioLegend) for an additional 2 h.

For the detection of ROS, 5 μM CellROX Deep Red reagent (Life Technologies) was incubated with macrophage cultures for 30 min at 37°C. Subsequently, cells were washed three times with PBS and analyzed by flow cytometry. The nitrite concentration was determined with the Griess assay (Sigma) according to the manufacturer's instructions.

The metallothionein genes Mt1 (Entrez identification no. 17748) and Mt2 (Entrez identification no.17750) were disrupted in RAW264.7 macrophages by using the CRISPR-Cas9 technique (53, 54). In brief, two guide RNAs (gRNAs) targeting the first exon of the Mt1 and Mt2 genes, TAGTCGTTGGACGAGTCC (Mt#3) and TCCGAGATCTGGTGAAGC (Mt#4), along with the appropriate restrictase adapters and complementary oligonucleotides were designed. The annealed gRNA duplexes were cloned into the lentiCRISPRv2 vector (a gift from Feng Zhang; Addgene plasmid 52961) (55) by using the BsmBI restriction site. The gRNA-expressing vectors were transfected into HEK293T cells, and cell culture supernatants containing viral particles were harvested and used for infection of RAW264.7 cells. RAW264.7 cells were subsequently subjected to puromycin selection for 3 weeks. Selected clones were tested for Mt1 and Mt2 mRNA expression by reverse transcription-PCR (RT-PCR). A representative clone from each of the two CRISPR constructs was used in the experiments described here.

Each of the experiments described here was performed in at least two independent biological replicates. A replicate utilizing macrophages and bacteria of different passages or batches is defined as a biological replicate. n in the figure legends refers to the number of biological replicates.

Statistical analysis was performed by using GraphPad Prism and R software. All results are presented as means ± standard errors of the means (SEM). Statistical tests included unpaired 2-tailed Student's t test and one-way or two-way analysis of variance (ANOVA), as appropriate, with Bonferroni post hoc tests. A Kolmogorov-Smirnov test was used to compare distributions. P values of <0.05 were considered to be statistically significant. On the graphs, P values of the main ANOVA effects as well as significant P values from post hoc tests are presented.