Frontline Science: LPS‐inducible SLC30A1 drives human macrophage‐mediated zinc toxicity against intracellular Escherichia coli

Abstract TLR‐inducible zinc toxicity is an antimicrobial mechanism utilized by macrophages, however knowledge of molecular mechanisms mediating this response is limited. Here, we show that E. coli exposed to zinc stress within primary human macrophages reside in membrane‐bound vesicular compartments. Since SLC30A zinc exporters can deliver zinc into the lumen of vesicles, we examined LPS‐regulated mRNA expression of Slc30a/SLC30A family members in primary mouse and human macrophages. A number of these transporters were dynamically regulated in both cell populations. In human monocyte‐derived macrophages, LPS strongly up‐regulated SLC30A1 mRNA and protein expression. In contrast, SLC30A1 was not LPS‐inducible in macrophage‐like PMA‐differentiated THP‐1 cells. We therefore ectopically expressed SLC30A1 in these cells, finding that this was sufficient to promote zinc‐containing vesicle formation. The response was similar to that observed following LPS stimulation. Ectopically expressed SLC30A1 localized to both the plasma membrane and intracellular zinc‐containing vesicles within LPS‐stimulated THP‐1 cells. Inducible overexpression of SLC30A1 in THP‐1 cells infected with the Escherichia coli K‐12 strain MG1655 augmented the zinc stress response of intracellular bacteria and promoted clearance. Furthermore, in THP‐1 cells infected with an MG1655 zinc stress reporter strain, all bacteria contained within SLC30A1‐positive compartments were subjected to zinc stress. Thus, SLC30A1 marks zinc‐containing compartments associated with TLR‐inducible zinc toxicity in human macrophages, and its ectopic over‐expression is sufficient to initiate this antimicrobial pathway in these cells. Finally, SLC30A1 silencing did not compromise E. coli clearance by primary human macrophages, suggesting that other zinc exporters may also contribute to the zinc toxicity response.


INTRODUCTION
macrophages and other innate immune cells utilize an array of antimicrobial mechanisms to defend against pathogenic bacteria. 1 One such strategy, nutritional immunity, refers to the competition between host cells and invading microorganisms for essential nutrients and metal ions. 2 For example, neutrophils secrete the zinc and manganese-binding protein calprotectin to restrict availability of these critical trace elements to bacteria, thus limiting their growth and survival. 3 Several pathogens counteract these and other innate immune responses by employing sophisticated mechanisms to acquire essential metal ions. 4 For example, Salmonella uses the ZnuABC zinc uptake system for host colonization 5 and to defend against innate immune-mediated nitrosative stress. 6 In recent years, it has become clear that specific trace elements such as zinc can also be harnessed by innate immune cells as antimicrobial agents to combat bacterial infection. 7 In Mycobacteria-infected macrophages exposed to inflammatory cytokines, phagosomal zinc concentrations were reported to reach millimolar concentrations. 8 Mycobacterium tuberculosis upregulates heavy metal efflux P-type ATPases in the intramacrophage environment, with an accumulation of zinc within bacteria-containing phagosomes also being observed. 9 Consistent with this, TLR-mediated macrophage activation triggers the mobilization of zinc into vesicular-like structures that co-localize with engulfed E. coli, with intracellular bacteria producing a transcriptional response consistent with zinc poisoning. 10 The zinc toxicity response is also deployed by the soil amoeba Dictyostelium discoideum, suggesting this is an ancient host defense pathway. 11 In keeping with a central role of zinc poisoning in innate immune antimicrobial responses, the bacterial pathogens Salmonella 10 and uropathogenic E.
coli 12 are able to both resist and evade this response. Mechanisms by which mobilized zinc exerts antimicrobial effects within innate immune cells are unknown. However, studies on bacteria alone have implicated induced copper deficiency, 13 manganese deficiency resulting in increased sensitivity to oxidative stress, 14,15 the replacement of other cations in essential enzymes, 7 and disruption of iron-sulfur (4Fe-S) biogenesis with consequential arrest of key metabolic pathways 16,17 as factors that contribute to the antimicrobial effect of zinc. Interestingly, zinc can synergize with reactive oxygen species to limit E. coli growth. 12 Thus, zinc toxicity could also act in a combinatorial fashion with other innate immune antimicrobial pathways.

Mammalian cell culture
To generate human monocyte-derived macrophages (HMDM), CD14 + monocytes were isolated from human buffy coats (Australian Red Cross Blood Service) and cultured for 7 days with 150 ng/ml recombinant human CSF-1 (University of Queensland Protein Expression Facility), as previously described. 12 To generate mouse bone marrowderived macrophages (BMM), bone marrow cells from the femurs and tibias of C57BL/6 mice were cultured for 7 days as previously described. 12 Human monocytic THP-1 cells were obtained from the American Tissue Culture Collection. Non-adherent THP-1 cells were differentiated into macrophages in presence of 30 ng/ml PMA (Sigma-Aldrich) for 48 h.

Confocal microscopy
Immunofluoresence imaging of fixed cells was performed using a Zeiss

Statistical analyses
Individual experiments were typically performed in experimental duplicate or triplicate, with mean values taken from each experiment for combining data from at least 3 experiments for statistical analyses.
Prism 5 and/or 7 software was used to perform the specific statistical tests that are indicated in the individual figure legends.

Zinc-stressed E. coli reside within membrane-bound compartments in human macrophages
E. coli zntA encodes a zinc efflux system that confers zinc resistance, as originally demonstrated through transposon mutagenesis, targeted gene deletion and gene complementation studies. 27, 28 We previously showed that the E. coli K-12 strain MG1655 is exposed to zinc stress during macrophage infection, as evidenced by intra-macrophage expression of zinc-inducible zntA and reduced intramacrophage survival of a zntA mutant strain. 10,12 To track this antimicrobial response, we also developed a zinc stress reporter strain (MG1655 pGcCzntAp), which constitutively expresses GFP and inducibly expresses mCherry when bacteria are subjected to zinc stress. 12 To further investigate this phenomenon, we utilized CLEM to visualize MG1655 pGcCzntAp within HMDM (

Zinc transporter SLC30A1 is constitutively expressed and further upregulated by LPS in human macrophages
To gain mechanistic insights into the delivery of antimicrobial zinc, we next profiled the LPS-regulated expression of the family of Green arrows indicate GFP +ve only bacteria, while red arrows indicate GFP +ve , mCherry +ve bacteria. In both cases, bacteria indicated in (A) correlate with putative bacteria indicated in (B) and are shown in a green or red inset, respectively. Black arrows indicate visible membrane surrounding bacteria. Images are of a single cell and are representative of 3 cells from 1 experiment macrophages. 19 We note that the 75 kDa monomer was the most abundantly expressed form of SLC30A1 in HMDM (Fig. 2C,

Ectopic expression of SLC30A1 in THP-1 cells drives zinc-containing vesicle formation
In contrast to observations in HMDM, LPS did not up-regulate SLC30A1 mRNA expression in PMA-differentiated macrophage-like THP-1 cells ( Supplementary Fig. S2A), despite these cells being LPSresponsive as assessed by TNF expression (Supplementary Fig. S2B).
Thus, we used a doxycycline-inducible system to ectopically express SLC30A1 in these cells for investigation of its contributions to the macrophage zinc toxicity response. Doxycycline induced SLC30A1 mRNA expression in THP-1 cells transduced with a lentiviral construct for SLC30A1, but not empty vector (EV), as expected ( Supplementary   Fig. S2C). Doxycycline-inducible epitope-tagged (V5) SLC30A1 expression in these cells was confirmed by both immunoblotting (Fig. 3A) and flow cytometry (Fig. 3B). In these cells, overexpressed SLC30A1_V5 was detectable at ∼75 kDa (Fig. 3A). However, a weakly expressed higher order complex (∼150 kDa, high exposure blot) was again detected, similar to our observation in HMDM (Fig. 2C).
We next examined the localization of SLC30A1 in macrophages, particularly in relation to LPS-inducible zinc-containing vesicles.
SLC30A1 has been reported to localize exclusively to the plasma membrane of baby hamster kidney cells, 31,32 rat astroglial cells, 33 (Fig. 3C). Notably, the staining pattern of these LPS-inducible zinc vesicles was not as pronounced as we have previously observed in HMDM, 10,12 being much more sparse and heterogeneous on a population and individual cell level.
Upon induction of V5-tagged SLC30A1 in THP-1 cells, we observed that SLC30A1 localized to both the plasma membrane and intracellular vesicular structures (Fig. 3D). Interestingly, treatment with doxycycline to induce SLC30A1 expression increased the percentage of FluoZin-3-positive THP-1 cells, with the effect being similar to that observed with LPS stimulation (Fig. 3E). Importantly, doxycycline alone did not have this effect in pLenti_EV-transduced control cells (Fig. 3F).
This suggests that the overexpression of SLC30A1 alone is sufficient to mobilize zinc in THP-1 cells in a similar fashion to the response trig-gered by LPS. This hypothesis is further supported by the observation that FluoZin-3-stained zinc was often detected within SLC30A1_V5positive vesicles (Fig. 3D). We note, however, that this co-localization was not uniform within a population of macrophages, likely because of the dynamic nature of zinc mobilization.
Although SLC30A1 can prevent cellular zinc toxicity via plasma membrane efflux, 30,31 examination across a broader range of cell types also supports a contribution to intracellular zinc redistribution. In a human keratinocyte cell line, SLC30A1 localized to the endoplasmic reticulum, nuclear membrane, and Golgi, and was found to regulate the intracellular distribution of zinc, but not the overall cellular concentration. 35 SLC30A1 was also reported to be expressed at both the plasma membrane and in intracellular compartments in rat

F I G U R E 3 Ectopically-expressed SLC30A1_V5 in THP-1 cells localizes to the plasma membrane and zinc-containing intracellular compartments.
PMA-differentiated THP-1 cells stably transduced with lentivirus expressing either empty vector (EV) or SLC30A1_V5 were left unstimulated (-, Con) or were stimulated with (A and B) 100 ng/mL or (C-F) 500 ng/mL doxycycline (Dox) for 24 h. (A) Cells were simultaneously stimulated ± 100 ng/ml LPS for 24 h, after which samples were lysed, processed and analyzed by western blot. Blots were probed with Abs against V5 or GAPDH as a loading control. The

SLC30A1 subjects engulfed E. coli to a zinc stress antimicrobial response in PMA-differentiated THP-1 cells
We next investigated whether the induction of SLC30A1 increased zntA mRNA expression within macrophages. In pLenti_SLC30A1_V5-  Table 3). This confirms that zinc export contributes to the survival of MG1655 in THP-1 cells, as was previously observed in E. coliinfected HMDM, where zntA was both deleted and complemented. 9 We note, however, that the effect of zntA deletion on survival within THP-1 cells was less pronounced than previously observed for primary ( Supplementary Fig. S2D).

F I G U R E 4 SLC30A1_V5 promotes a zinc stress response and bacterial killing in THP-1 cells, and localizes to both the plasma membrane and compartments containing zinc-stressed
We next utilized the MG1655 zinc stress reporter strain 12 (Fig. 1

SLC30A1 is not essential for bacterial killing in primary human macrophages
Somewhat surprisingly, silencing of SLC30A1 in HMDM ( SLC30A3, SLC30A4, and SLC30A7. 38 As a homodimer, SLC30A1 localized to the plasma membrane, 38 however subsequent studies revealed that SLC30A1 also heterodimerized with SLC30A2 and SLC30A4, and that this altered its sub-cellular localization. 39 Thus, SLC30A1 may also function in conjunction with other zinc transporters to achieve vesicular delivery of zinc to engulfed bacteria. In summary, the LPS-inducible zinc exporter SLC30A1 permits mobilization of zinc toward intracellular E. coli and the generation of a zinc-mediated antimicrobial response in macrophages. On the basis of studies on the E. coli homologue, YiiP, it is likely that SLC30A1 requires a proton gradient to deliver zinc in a pH-driven, sodium-independent, and calcium-sensitive manner (1:1, Zn 2+ /H + ). 44