The Extracellular Electron Transport Pathway Reduces Copper for Sensing by the CopRS Two-Component System under Anaerobic Conditions in Listeria monocytogenes

ABSTRACT

The renowned antimicrobial activity of copper stems in part from its ability to undergo redox cycling between Cu^1+/2+ oxidation states. Bacteria counter copper toxicity with a network of sensors that often include two-component signaling systems to direct transcriptional responses. As in typical two-component systems, ligand binding by the extracellular domain of the membrane bound copper sensor component leads to phosphorylation and activation of the cognate response regulator transcription factor. In Listeria monocytogenes, the plasmid-borne CopRS two-component system upregulates both copper resistance and lipoprotein remodeling genes upon copper challenge, but the oxidation state of copper bound by CopS is unknown. Herein, we show CopS utilizes a triad of key residues (His-His-Phe) that are predicted to be at the dimerization interface and that are analogous with the Escherichia coli CusS copper sensor to specifically bind Cu¹⁺/Ag¹⁺ and activate CopR transcription. We demonstrate Cu²⁺ only induces CopRS if first reduced by electron transport systems, as strains lacking menaquinone carriers were unable to respond to Cu²⁺. The flavin-dependent extracellular electron transport system (EET) was the main mechanism for metal reduction, capable of either generating inducing ligand (Cu²⁺ to Cu¹⁺) or removing it by precipitation (Ag¹⁺ to Ag⁰). We show that EET flux is directly proportional to the rate of Cu²⁺ reduction and that since EET activity is low under oxygenated conditions when a competing respiratory chain is operating, CopRS signaling in turn is activated only under anaerobic conditions. EET metal reduction thus sensitizes cells to copper while providing resistance to silver under anaerobic growth.

IMPORTANCE Two-component extracellular copper sensing from the periplasm of Gram-negative bacteria has been well studied, but copper detection at the cell surface of the Gram-positive L. monocytogenes is less understood. Collectively, our results show that EET is most active under anaerobic conditions and reduces Cu²⁺ and Ag¹⁺ to, respectively, generate or remove the monovalent ligands that directly bind to CopS and lead to the induction of lipoprotein remodeling genes. This reducing activity regulates CopRS signaling and links the upregulation of copper resistance genes with increasing EET flux. Our studies provide insight into how a two-component copper sensing system is integrated into a model monoderm Firmicute to take cues from the electron transport chain activity.

INTRODUCTION

Bacterial two-component systems figure prominently in the recognition and response to extracellular copper in many different bacteria (1, 2). As with typical two-component systems (3–5), one component is a transmembrane signaling protein with an extracellular sensor domain that surveils for extracellular copper. Ligand coordination drives the formation and stabilization of subunit dimers, transmitting protein conformational changes to the cytoplasmic histidine kinase (HK) domain to promote autophosphorylation. The HK domain transfers the phosphate onto the second component, the cognate response regulator (RR) transcription factors. RR phosphorylation alters DNA binding affinity to direct transcription of key genes needed to respond to the environmental input signal. In the case of copper challenge, transcriptional responses include upregulation of P-type ATPase heavy metal efflux pumps (6–8), multicopper oxidases to detoxify copper by oxidation (9), and various metallochaperone proteins involved in copper trafficking and metal sequestration (10–12). These copper resistance mechanisms work in concert to alleviate toxicity and maintain negligible amounts of free cytosolic copper (1), which could otherwise impair iron-sulfur cluster assembly (13), nonspecifically inhibit enzymes, and generate reactive oxygen species by redox cycling (14).

We recently identified plasmids circulating in Listeria monocytogenes strains harboring a copper-inducible operon encoding canonical copper-resistance-associated genes in tandem with lipoprotein remodeling genes (15). In the companion article, we showed the RR from the CopRS two-component system regulates expression from three distinct promoter elements with a conserved DNA binding motif (16). Genes under the control of CopRS that are known to be involved in copper resistance include orthologs of the multicopper oxidase cueO, while lipoprotein-related genes include copies of the prolipoprotein diacylglyceryl transferase (lgt2) (17) and lipoprotein intramolecular acyltransferase (lit2) (15, 18). Lit2 expression converts the lipoprotein chemotype from diacylglycerol to lyso-form, where single acyl chains are located on the α-amino group of the N-terminal cysteine residue and on the glycerol moiety. How lipoproteins are involved in copper resistance is not entirely understood, but the coregulation with copper resistance genes under direct control of the same two-component system suggests a functional link. We and others have proposed excess copper may directly bind certain lipoprotein chemotypes and interfere with lipoprotein biogenesis (15, 19, 20).

We are thus interested in understanding under which conditions lipoproteins get remodeled and, in particular, the oxidation states of copper that are directly detected by the two-component CopRS system in L. monocytogenes. We reasoned that determining the oxidation state of copper detected by the heavy metal binding domain of the sensor CopS might provide insight into when lyso-form lipoprotein synthesis is most beneficial. The Escherichia coli two-component system copper sensor CusS binds Cu¹⁺/Ag¹⁺ (21–23), CopS from the cyanobacterium Synechocystis sp. PCC 6803 binds Cu²⁺ (24), and in cases where both oxidation states were explicitly measured in vitro, CopS from Pseudomonas aeruginosa bound both Cu¹⁺ and Cu²⁺ with near equal femtomolar affinities (25). There is little known about how two-component system sensors recognize copper ligands outside a Gram-negative periplasmic compartment, and even within Gram-negative organisms, there appears to be diversity in preferences for copper oxidation state. There is also comparatively less known about copper trafficking and the redox state at the cell envelope in a typical Firmicute, which, being a monoderm, presents a very different extracellular environment than Gram-negative bacteria. Within Firmicutes, there is no outer membrane boundary to define the periplasm and retain soluble copper metallochaperones important for extracellular heavy metal sensing systems (26, 27). Herein, we show that CopRS from L. monocytogenes specifically recognize the Cu¹⁺ oxidation state and that extracellular electron transport (EET) activity in reducing Cu²⁺ to Cu¹⁺ is central to activating CopRS signaling and, in turn, lit2 induction.

RESULTS

CopS utilizes a His-His-Phe triad of residues characteristic of monovalent Cu¹⁺ binding to induce lit2 expression.

To address which form(s) of copper ion and the potential of other metal cations, if any, to bind CopS and induce lit2 expression in L. monocytogenes, computational structural modeling was used to predict the metal binding site of CopS (Fig. S1 and S2 in the supplemental material). The HK copper sensor from the two-component CusRS system in E. coli has been well characterized and is typical in having an N-terminal extracellular heavy metal binding sensor domain connected to the cytoplasmic C-terminal HK domain by a transmembrane helix (21–23, 28). CopS from L. monocytogenes shares these features with CusS, although the amino acid sequence similarity is noticeably lower in the extracellular metal binding domain (Fig. S1). Biochemical binding studies and crystal structure data have identified two pairs of distinct Ag¹⁺ metal binding sites in CusS (21, 23). Ag¹⁺ and Cu¹⁺ have similar ionic radii and valence electron shell occupancies and generally bind to proteins using common residues with analogous coordination geometries (29). Ag¹⁺ is thus a redox stable mimetic for Cu¹⁺ binding. One set of Ag¹⁺/Cu¹⁺ binding sites in CusS is located at the dimer interface and coordinates two ions using residues from both subunits, while periphery intrasubunit sites on each subunit bind two more metal ions (Fig. S2A and S2B). The predicted tertiary CopS structure (using AlphaFold [30]) is missing periphery sites (Fig. S2C and S2D), which largely accounts for the lower primary sequence conservation specific to the heavy metal binding domain. However, the three key residues constituting the metal binding site appear to be conserved (His42, Phe43, His176 in CusS corresponding to His36, Phe37, and His159 in CopS). AlphaFold predicts these residues to be in close proximity on the dimerization interface, analogous to CusS. Replacement of any of these residues with Ala decreased CopRS signaling, as evidenced by the loss of lit2 expression under copper-inducing conditions. Lack of induction for His36Ala or His159Ala mutants was equivalent to mutating the cytosolic side His255 required for CopR phosphorylation but more modest in the Phe37Ala construct (Fig. 1). As sequence conservation flanking His159 is particularly low, we also mutated residues that could possibly coordinate cations (E158 and D160) or satisfy copper’s thiophillicity (Met167 and Met172). None of these mutants impacted CopRS copper sensing. Assuming equivalent protein expression and/or stability among the CopS mutant panel, the data suggest CopS may use a triad of residues to chelate monovalent ligand (Cu¹⁺ and Ag¹⁺) in a mechanism akin to CusS.

FIG 1.

CopS active site residues (H36A, H159A, and H255A) support a Cu⁺ binding site. (A) Wt L. monocytogenes, ΔcopS, and back complemented strains cells carrying CopS point mutant alleles were grown in TSB with or without 1 mM CuSO₄ under standard conditions. (B) Total RNA was extracted and probed for lit2 expression by Northern blotting. Ethidium bromide staining of total RNA was utilized as an internal loading control.

To further support monovalent metal binding specificity, recombinant CopS protein encompassing the soluble extracellular metal binding domain between the transmembrane helices (Fig. S1) was tested for interactions with metals. Migration in native PAGE gels was consistent with monomeric forms for all preparations (Fig. 2A). Addition of Ag¹⁺, but not Cu²⁺, in 4-fold molar excess over protein quantitatively shifted protein to a presumably dimeric form for both the wild-type CopS and CusS constructs. Converting His36 and Phe37 to Ala residues (HF->AA) prevented any Ag¹⁺ induced shifting, consistent with a role for these residues in the ligands of coordinating monovalent metals. Moreover, while Cu²⁺ does appear to interact with CopS, the interaction is nonquantitative even at ~10-fold higher concentrations and produces multiple bands migrating faster than the Ag¹⁺-CopS complex (Fig. 2B). All of this suggests weak and nonspecific Cu²⁺-CopS interactions that are unlikely to be physiologically relevant.

FIG 2.

Native PAGE analysis of the extracellular metal binding domain of recombinant CopS. (A) Recombinant proteins (5 μM) were incubated for 30 min at room temperature either in water only (Con) or with 20 μM EDTA, Cu²⁺, or Ag¹⁺ before separation using a nondenaturing 15% native PAGE gel. Gels were stained with Coomassie blue for imaging. (B) A 2-fold serial dilution of either Cu²⁺ or Ag¹⁺ solution was added to protein solution (5 μM final) and separated as above.

Cellular reducing activity requiring menaquinones imparts (Cu²⁺) or prevents (Ag¹⁺) lit2 induction by heavy metals.

The predicted monovalent coordination chemistry in CopS was contrary to expectations. Previous work from our laboratory found that Ag¹⁺ does not efficiently induce lit2 expression in L. monocytogenes, while divalent Cu²⁺ does under standard culture conditions (Fig. 1 and reference 15). We hypothesized that if Ag¹⁺ does bind to CopS, then extracellular Ag¹⁺ must be made unavailable for binding under the conditions previously used for measuring lit2 expression. Tryptic soy broth (TSB) media have high levels of chloride anion, and AgCl has low μM-level solubility in aqueous media. As such, we switched to a low-chloride media, supplemented ISP1 (5 g/L tryptone, 3 g/L yeast extract, 5 mM glucose, and 30 mM HEPES [pH 7.4]), and again looked at the expression of lit2 (Fig. 3). However, supplemented ISP1 media did not promote lit2 expression by Ag¹⁺. We thus reasoned extracellular reduction of Ag¹⁺ to insoluble Ag⁰ could be removing CopS ligand from the solution. Such redox activity with Ag¹⁺ as electron acceptor would explain the lack of CopS activation as well as generating extracellular Cu¹⁺-inducing ligand from Cu²⁺. MenE is an O-succinylbenzoic acid CoA ligase essential in the menaquinone biosynthesis pathway, and mutants have deficient electron transport chains (ETC) (31). Blocking electron transport in L. monocytogenes ΔmenE inverted CopRS induction under standard culture conditions, with Ag¹⁺ now a strongly inducing ligand and Cu²⁺ having no effect. These experiments demonstrate menaquinone electron carriers are required for Cu²⁺ to be a ligand and to prevent induction by Ag¹⁺.

FIG 3.

Menaquinones are required for induction by Cu²⁺ but prevent induction by Ag⁺. Wt L. monocytogenes and the menaquinone-deficient ΔmenE strain were grown in supplemented ISP1 media containing low chloride to maintain high silver ion solubility. Cells were incubated with the indicated concentrations of CuSO₄ or AgNO₃ under standard culture conditions. (A) Total RNA was extracted and probed for lit2 expression by Northern blotting as in Fig. 1. (B) Ethidium bromide staining of total RNA was utilized as an internal loading control.

Cu¹⁺ is produced from extracellular Cu²⁺ utilizing reducing equivalents provided by the EET pathway.

In L. monocytogenes, menaquinones operate in both the aerobic respiratory ETC and in the recently discovered extracellular EET (32). Like aerobic respiration, EET uses NADH as the initial source of reducing equivalents but can reduce Fe³⁺ and other alternate extracellular terminal electron acceptors using flavin redox mediators (32–36). EET-related genes encode orthologs for a type II NADH dehydrogenase (Ndh2), demethylmenaquinone-synthesizing enzymes (DmkAB), uncharacterized membrane proteins (EetAB), and a series of proteins required for the secretion of flavin cofactor used by terminal reductases (FmnAB and PplA). Since MenE is required for a common menaquinone precursor used by both electron transfer pathways, we were interested in determining the relative contribution of each pathway to Cu²⁺ and Ag¹⁺ reduction. To address this question, we constructed L. monocytogenes strains with deletions specific to the respiratory ETC (Ndh1, MenA, and the cytochrome bd-type QoxAB and cytochrome aa 3-type menaquinol CydAB terminal oxidases), EET (Ndh2, DmkA, EetAB, FmnA, FmnB, and PplA), or in combination. Strains with defects in respiratory ETC components retained robust Cu²⁺ reducing activity, while strains with EET pathway gene deletions had markedly lower activity in comparison to wild-type (Wt) or respiratory ETC mutants (Fig. 4A, top). Deletion of respiratory ETC genes in tandem with EET mutants had minimal additive effect (Fig. 4B), indicating that the majority of the observed Cu¹⁺ formed is due to EET flux. We repeated the assay using Fe³⁺, a well-established EET substrate in L. monocytogenes (32), as the terminal electron acceptor. Across the entire mutant strain panel, Fe³⁺ reducing activity closely mirrored those for Cu²⁺ (Fig. 4A, bottom). It therefore appears electrons traverse a common route from NADH donor to Cu²⁺ and Fe³⁺ metal acceptors using EET.

FIG 4.

EET activity reduces both cupric and ferric metal ions in L. monocytogenes. (A) The BCS assay (top) was used to monitor copper reduction by Wt and strains with deletions in aerobic respiration, EET, or in both pathways. Cells were incubated in supplemented ISP1 media containing 1 mM BCS and 0.1% agarose to slow diffusion. Reduction reactions were initiated by adding CuSO₄ to a final concentration of 0.25 mM. The production of a dark orange color indicates metal reduction and BCS-Cu complex formation. Tubes were imaged after 30 min. Iron reduction assays were likewise conducted (bottom), with detection of ferrous iron using ferrozine as chelator [1 mM ferrozine + 0.25 mM ferric citrate]. Image is from a representative experiment. (B) Copper reduction was quantified by measuring the absorption at A_480nm after 30 min of reaction. The measurements were normalized against a no-cell control. The data are shown as the means ± standard deviations of the results of three independent experiments. Statistical significance compared to the Wt were obtained using an unpaired two-sided t test: *, P ≤ 0.05; ***, P ≤ 0.001; ****, P ≤ 0.0001.

Cu²⁺ reduction by EET is active under anaerobic conditions.

Stark variations in the kinetics of copper and iron reduction were observed among the aerobic respiration ETC deletion strains (Fig. 4A). The Δndh1 and ΔqoxAB strains behaved similarly to the Wt strain, showing gradual Cu¹⁺-bathocuproine sulfonate (Cu¹⁺-BCS) chromophore formation starting at the bottom of the tube and progressively building upward toward the air-liquid interface. Distinct biphasic boundaries of high and low Cu¹⁺-BCS were clearly discernible. The remaining strains (ΔmenA, ΔcydAB, and ΔΔcydAB/qoxAB) immediately produced Cu¹⁺-BCS upon reaction initiation, indicating instant metal ion reduction. While Ndh1 and QoxAB are respiratory ETC components, neither is integral for oxygen reduction. There is functional redundancy between Ndh2 and Ndh1 with respect to supplying electrons to aerobic ETC (35). Once more, QoxAB has a minor role in L. monocytogenes where CydAB is the main terminal oxidase under aerobic conditions (37). This observation suggests that respiratory-proficient strains must consume oxygen before reducing equivalents are funneled to EET for Cu²⁺ reduction. The pattern of Cu¹⁺-BCS color change is then reflective of achieving an anaerobic state. For the respiration-deficient strains, we reasoned EET flux is immediate since the machinery is missing to utilize oxygen as a terminal electron acceptor at the onset of the assay.

We monitored oxygen consumption in tandem with the rate of Cu²⁺ reduction to test the relationship between oxygen, electron pathway flux, and the preferred terminal acceptor. For Wt cells, there was an extended lag in Cu¹⁺-BCS formation that lasted until oxygen was mostly consumed (Fig. 5A). The EET deletion strain ΔfmnB showed a similar rate in oxygen consumption as Wt but with a much lower steady rate of copper reduction (Fig. 5B). Oxygen consumption in the aerobic respiratory ETC deletion strain ΔΔcydAB/qoxAB was expectedly much slower than the Wt (Fig. 5C). Reduction of Cu²⁺ in this strain initiated without any initial lag, at a rate that was similar to Wt but decoupled from oxygen levels. The ΔΔndh1/2 strain, lacking flux through both respiratory ETC and EET pathways, neither consumed oxygen nor reduced copper beyond basal rates observed in ΔfmnB (Fig. 5D). Collectively, these data provide evidence that electron flux through EET to Cu²⁺ is secondary to the respiratory ETC, assuming the pathway is intact and oxygen is available.

FIG 5.

EET-mediated copper reduction increases with oxygen depletion. Oxygen levels (black trace) were measured using an optical oxygen meter in tandem with copper reduction (orange trace) using the BCS assay. Cells were grown in supplemented ISP1 media were incubated with 1 mM BCS and copper reduction was initiated with 0.25 mM CuSO₄ in sealed cuvettes with continuous stirring. The absorbance at A_480nm was measured every minute over 1 h and corrected for background reduction using a no-cell control. (A to D) Four representative strains with either functioning aerobic respiration and EET electron transport (WT; A), only aerobic respiration (ΔfmnB; B), only EET (ΔcydAB ΔqoxAB; C), or with both electron transport chains blocked (Δndh1/2; D) were tested. Data are representative of at least three independent experimental replicates.

EET reduces Ag¹⁺ under anaerobic conditions.

Supplemental Ag¹⁺ did not lead to robust induction of lit2 transcripts through CopS in Wt L. monocytogenes under standard culture conditions (Fig. 3). To determine whether reduction of Ag¹⁺ by EET flux was causative, we used a silver-loaded filter disk diffusion assay to visually score silver reduction after incubation under aerobic or strict anaerobic incubation (Fig. 6). A dark black/brown precipitate at the interphase between bacterial growth and the diffused Ag¹⁺ concentration gradient silver due to insoluble Ag⁰ was observed for Wt and the ΔΔcydAB/qoxAB deletion strains under anaerobic conditions. In contrast, EET-deficient strains (ΔfmnB and ΔΔndh1/2) exhibited much lower reduction levels. Under aerobic conditions, competition with oxygen for reducing equivalents should limit EET activity. No strain reduced Ag¹⁺ except ΔΔcydAB/qoxAB, which has a broken respiratory ETC. EET thus reduces Ag¹⁺ as well as Cu²⁺ in oxygen-limited environments.

FIG 6.

EET reduces Ag¹⁺ to form an insoluble Ag⁰ precipitate. A modified disc-diffusion assay was used to detect silver reduction. Filter paper disks soaked with a 0.5 mM AgNO₃ were placed in the center of ISP1 agarose plate from which the indicated bacterial strains were radially struck. Plates were incubated for 24 h under either strict anaerobic (left) or aerobic (right) conditions. The formation of a dark black-brown precipitate (Ag⁰) at the interphase between bacterial growth and the diffused gradient of silver that permitted growth qualitatively indicates silver reduction rates.

Expression of lit2 depends on oxygen levels.

Since EET generates extracellular Cu¹⁺ (Fig. 4) and EET flux is ramped up in the absence of oxygen (Fig. 5), CopRS signaling should in turn be subject to oxygen levels. We again measured the lit2 expression under the standard culture conditions (as in Fig. 3) but under continually oxygenated or strictly anaerobic culture conditions (Fig. 7). Under aerobic conditions, Cu²⁺ no longer induced lit2 in all strains tested except ΔΔcydAB/qoxAB, which was very weakly upregulated. Even with the competition for reducing equivalents from terminal oxidases removed, the rate of Cu¹⁺ formation by EET was still apparently eclipsed by the rate of reoxidation by molecular oxygen to keep levels below the CopS detection limit. The opposite was true with the more oxygen-stable Ag¹⁺ ion, with all strains now upregulated lit2 transcripts under high aeration conditions. Once again, the metal-reducing rate in ΔΔcydAB/qoxAB was not robust, as some Ag¹⁺ ligands remained available for CopS.

FIG 7.

Oxygen levels modulate copper and silver ion induction of CopRS. (A) Cells were grown in supplemented ISP1 media were incubated for 1 h with 0.25 mM CuSO₄ or 0.125 mM AgNO₃ under highly aerated (top) or anaerobic (bottom) conditions. Total RNA was extracted and probed for lit2 expression by Northern blotting as in Fig. 1. Ethidium bromide staining of total RNA was utilized as an internal loading control. (B) ISP1-grown cells were incubated anaerobically with 0.25 mM CuCl. Metabolism was quenched at the indicated time points (0, 5, 10, 20, and 40 min) using TRIzol, followed by total RNA extraction and Northern blotting to probe for lit2 expression as in Fig. 1.

Under anaerobic culture conditions, only the Cu¹⁺ disproportionation rate needs to be considered and not reoxidation by molecular oxygen. Accumulation of extracellular Cu¹⁺ from Cu²⁺ is thus largely determined by EET flux, in tandem with the rates of Cu²⁺ uptake-cytosolic reduction-efflux activity and the possible direct reduction by menaquinones (38, 39). Both Wt and ΔΔcydAB/qoxAB, having an intact EET, showed strong lit2 expression under anaerobic incubation with Cu²⁺ in comparison to EET-impaired strains (Fig. 7A). Some induction still occurs without EET in ΔfmnB and ΔeetAB strains, however, likely reflecting secondary Cu¹⁺ sources. The complete lack of lit2 expression in ΔΔndh1/2, where reduced menaquinones are in short supply but intracellular reducing equivalents are readily available, points toward the direct reduction by menaquinones as the secondary Cu¹⁺ source in EET-impaired strains. First impairing the EET pathway (ΔΔndh1/2, ΔfmnB, and ΔeetAB) was required for upregulation of lit2 in response to Ag¹⁺, consistent with the high anaerobic EET flux rate efficiently removing Ag¹⁺ ligand from solution through reduction. To further prove that EET generates Cu¹⁺ ligand for CopS under anaerobic conditions, we compared the rates of lit2 upregulation using freshly prepared Cu¹⁺ solutions as inducers. Unlike Cu²⁺, which weakly induced the EET-deficient ΔfmnB strain in comparison to wild type (Fig. 7A), direct addition of Cu¹⁺ ligand induced lit2 transcripts at comparable rates in both strains (Fig. 7B).

EET modulates heavy metal resistance.

Both silver and copper are toxic to bacterial cells. Silver cations can damage bacterial membrane, disrupt metabolic pathways, and induce DNA damage (40). Copper cations share many of the same antibacterial mechanisms, in addition to disrupting iron-sulfur cluster cofactor assembly and directly generating reactive oxygen species during redox cycling between Cu²⁺ and the more toxic Cu¹⁺ oxidation state (41). Considering how EET removes Ag¹⁺ from solution (Fig. 6) and increases the concentration of Cu¹⁺ (Fig. 4), we were interested in determining whether EET alters intrinsic resistance under anaerobic conditions. The Wt and ΔfmnB strains retained a similar degree of viability in the absence of metals, indicating EET provided no inherent growth advantage under these conditions (Fig. 8). A modest yet reproducible 10-fold viability advantage for ΔfmnB over the Wt strain was observed with 2 mM Cu²⁺ (Fig. 8B). Conversely, the plating efficiency of ΔfmnB dropped by 100-fold in comparison to Wt after exposure to 0.125 mM Ag¹⁺ ions. When preincubated samples were subcultured into rich TSB media to monitor growth recovery kinetics, Wt cells were slightly disadvantaged after Cu²⁺ exposure but enjoyed a marked advantage after Ag¹⁺ treatment (Fig. 8C). This is consistent with EET-reducing activity converting Cu²⁺ to the more toxic Cu¹⁺ state, while detoxifying Ag¹⁺ to Ag⁰ through precipitation.

FIG 8.

EET activity sensitizes cells to copper while detoxifying silver by precipitation. (A) The Wt and the EET deletion strain ΔfmnB (~10⁸ cells) were incubated for ~20 h at room temperature in supplemented ISP1 media with or without the indicated concentrations of CuSO₄or AgNO₃ under anaerobic conditions. A 10-fold serial dilution series was then spot platted on TSA plates for each sample to count viable cells. (B) The number of viable cells after metal exposure was extrapolated from the colony count in panel A and plotted for Wt (black bars) and ΔfmnB (gray bars). Data represents colony counts obtained from three independent experiments. Statistical significances were obtained using an unpaired two-sided t test, *, P ≤ 0.05; **, P ≤ 0.01. (C) Aliquots from cells that been previously incubated metal at the indicated concentrations were diluted 100-folds into rich TSB media lacking any inhibitory metals. The OD_600nm was measured over 24 h to monitor the growth recovery kinetics for Wt (black trace) and ΔfmnB (gray trace) cells that had been exposed to ISP1 media only (left), with 2 mM copper (middle), or with 0.125 mM silver (right).

DISCUSSION

Unlike Gram-negative bacteria, which have a defined periplasmic compartment for copper sensing among a network of dedicated copper chaperones (1, 2), extracellular copper detection and trafficking at the monoderm cell envelope surface in Firmicutes are less well understood. Initial studies suggested Cu²⁺, but not the redox stable Cu¹⁺ isostere Ag¹⁺, binds to CopS and induces lit2 (Fig. 1 and reference 15). This might be expected in an exposed and presumably oxidizing, noncytosolic environment where Cu¹⁺ would have a short half-life. Cytosolic detection occurs in a generally reducing compartment due to oxidative catabolism of organic substrates, and all known cytosolic one-component copper sensors interact with Cu¹⁺ (1). Extracellular copper detection by metal binding domains in two-component sensors appears to be more complex, likely reflecting the less certain redox environment. We have now shown that extracellular metal detection by CopS from L. monocytogenes is absolutely required for lit2 transcription (Fig. 1) and that the metal binding sensor domain specifically detects monovalent Cu¹⁺/Ag¹⁺ ligands using three key residues at the dimer interface in a fashion that is likely analogous to E. coli CusS (Fig. 1 and 2; Fig. S2).

A fundamental difference between copper sensing in E. coli with CusS and CopS in L. monocytogenes, however, may concern how copper is presented to the membrane-bound two-component sensor. The concentration of free copper in any oxidation state in the Gram-negative periplasm is considered negligible (1), with effectively all copper being bound to high-affinity chaperones during metal trafficking. Chaperones can bind Cu²⁺, Cu¹⁺, or both at distinct coordination sites, as seen with the widely distributed CopC periplasmic chaperone family (12). In Bacillus subtilis and other Firmicutes, including L. monocytogenes, the copper importing transmembrane protein YcnJ does have an N-terminal CopC-like extracellular domain but is unlikely to have a general role in copper trafficking (42). Outside of low affinity adventitious binding sites to negatively charged cell envelope components, such as phospholipids and teichoic acids, the absence of chaperones in a defined compartment suggests potential for the existence of meaningful pools of free extracellular copper. Free or weakly bound metals at the cell envelope surface in L. monocytogenes provides opportunities for any number of redox reactions. Indeed, reduced menaquinones have previously been postulated to directly interact and reduce extracellular copper in a nonenzymatic process in L. lactis IL-1403 (38, 39). Here, we show EET also has a major role in reducing both Cu²⁺ and Ag¹⁺, using an electron transfer route that overlaps with the Fe³⁺ pathway (32–36). The EET pathway uses a series of proteins to transfer reducing equivalents generated from NADH oxidation by Ndh2 to protein-bound flavins outside the cell. Flavinylated reductases then complete the electron transfer pathway using their own respective terminal electron acceptor substrates, including free flavins. In L. monocytogenes, multiple reductases within the cell envelope have been identified (34). Since PplA reductase required for Fe³⁺ reduction also factored into Cu²⁺ reduction (Fig. 4), a dedicated Cu²⁺/Ag¹⁺-binding reductase is unlikely. Rather, a shared chemical reduction between all three metals and PplA bound flavin or soluble flavin mediator seems to be the most probable mechanism. The standard reduction potentials of Fe³⁺/Cu²⁺/Ag¹⁺ are within expectation for thermodynamically favorable redox couples with flavins (43). An EET-flavin-metal electron transfer pathway operates in the Gram-negative Shewanella sp. as well, where secreted flavins are reduced by the c-type cytochrome-based Mtr EET pathway (44–46). Among the metals subject to reduction by Mtr EET are Fe³⁺ and Cu²⁺ (47).

EET flux indirectly exerts significant control over CopRS two-component signaling in Wt L. monocytogenes (Fig. 7). Metal reduction by EET inverts which metals induce lit2 transcripts by converting a nonligand (Cu²⁺) into a ligand (Cu¹⁺) and a ligand (Ag¹⁺) into nonligand (Ag⁰). In the latter case, ligand is removed from solution by precipitation (Fig. 6), while for the former we could not find any evidence for direct activation of CopS by Cu²⁺ without prior in vivo reduction. EET-mediated metal reducing activity was inversely related to oxygen levels (Fig. 5). This could in part be attributed to simple thermodynamic hierarchy (48), where oxygen being the more energetically favorable redox partner must be consumed by the respiratory ETC before any electrons flow through EET to Cu²⁺. Rapid downregulation of EET flux in favor of aerobic respiration has been demonstrated in other Firmicutes, including L. lactis (38, 39), L. cremoris (49), and E. faecalis (33, 50). This bifurcation in flux is made particularly sharp considering that Ndh2-derived reducing equivalents can flow through the aerobic respiratory chain when oxygen is present (35). Altering the isoprenoid repeat number and the demethylmenaquinone content within the menaquinone pool has also been proposed as a mechanism for segregating electron flux between the two pathways (32, 39, 49). In strains with blocked respiratory ETCs (ΔmenA, ΔcydAB, and ΔΔcydAB/qoxAB), electrons immediately flowed through EET and Cu²⁺ was reduced without an initial lag for oxygen consumption (Fig. 4 and 5). These results have important consequences for Cu²⁺-induced lipoprotein remodeling through CopRS signaling since, in effect, lit2 is only induced under low oxygen conditions when EET flux is robust. Under highly oxygenated conditions, the short Cu¹⁺ half-life of an unchaperoned state further prevents accumulation of the CopS-inducing ligand. In the terminal oxidase mutant ΔΔcydAB/qoxA where only EET is intact, there was seemingly strong Cu²⁺-reducing activity without prior consumption of oxygen (Fig. 4 and reference 4). However, the reducing activity did not translate to lit2 induction (Fig. 7). One explanation is that while any Cu¹⁺ produced can be captured in the oxygen-stable BCS₂-Cu¹⁺ complex, without such chelation the reoxidation rates prevent the accumulation of Cu¹⁺ to levels needed to induce CopS. It should also be noted that with the more oxygen-stable isostere Ag¹⁺, EET-mediated reduction was weaker under aerobic conditions (Fig. 6). This suggests that in the terminal oxidase mutant, electron transport flux is not diverted from aerobic respiratory ETC to EET in its entirety, and alternative factors may be at play in this mutant, which has a highly altered metabolism. Anaerobic conditions, by contrast, not only lead to more EET flux but also allow secondary Cu²⁺ reduction mechanisms to contribute as Cu¹⁺ produced is more stable. Secondary mechanisms may include direct reduction by reduced menaquinones at the membrane or cytosolic reduction followed by efflux as observed in L. lactis (38, 39), thus providing the Cu¹⁺ ligand for the weak induction observed in the EET-deficient ΔeetAB and ΔfmnB strains (Fig. 7). The net result is a complex, multilayered regulation by the CopRS two-component systems that integrates multiple environmental cues, including oxygen and EET flux in addition to copper levels (Fig. 9).

FIG 9.

Model for activation of the CopRS two-component system by the EET pathway in L. monocytogenes. Reducing equivalents extracted from organic substrates are shuttled into either the respiratory ETC (teal) or the EET pathway (purple) by the respective type II NADH dehydrogenases (Ndh1 and Ndh2). Dedicated menaquinone pools for each pathway have been proposed (32), with electron carriers being discriminated by either the methylation state (MK in respiratory ETC verse DMK in EET) and/or the isoprenoid (IPP) repeat number (X IPP in respiratory ETC versus Y IPP in EET) (left inset). Under high oxygenation, respiratory ETC flux using Ndh1, MK carrier, and the cytochrome bd-type terminal oxidase (CydAB) predominates, while under microaerobic conditions the proton pumping cytochrome aa 3-type menaquinol (QoxAB) terminal oxidase also contributes (37). Any Cu²⁺ reduction that occurs likely arises from the direct reduction of menaquinones as has been proposed in L. lactis (38, 39). The apparent rate is too low to eclipse the reoxidation rate by molecular oxygen (dotted curved arrow) however, keeping Cu¹⁺ levels below the CopS detection threshold. Under these conditions, any reducing equivalents extracted by Ndh2 may also be used by the branched respiratory terminal oxidase pathway as the aerobic growth of the Δndh1 mutants is similar to wild-type (35). It is unclear whether this cross talk is mediated through Ndh2 catalyzed reduction of MK electron carriers, utilization of reduced DMK carriers from EET by respiratory ETC components, or simply reflects a thermodynamic hierarchy that is funneling a common pool of reduced menaquinone carriers to the respiratory ETC (dotted lines). When oxygen is absent, electron flux proceeds through EET to the reduced terminal reductase PplA which has previously been flavinylated (right inset). Reduced flavin redox mediators (including a free flavin pool) then generate Cu¹⁺ from Cu²⁺ by reduction, providing ligand for CopS binding (blue). CopS signaling leads to the phosphorylation of the RR regulator CopR, which upregulates copper resistance and lipoprotein (lit2) remodeling genes (16). In this model, Ag¹⁺ ligand will only remain available for CopS binding under high oxygenation conditions when EET flux is low (not shown). Enzymes that were deleted and characterized for Cu^2+/Ag¹⁺ reduction in this study are underlined.

Anerobic conditions, and by extension EET activity, decrease the Cu²⁺ tolerance of L. monocytogenes Wt in comparison to ΔfmnB (Fig. 8). Copper reduction potentiates toxicity by generating the more redox active Cu¹⁺ oxidation state (51), which may also have an increased cellular uptake rate due to more permeability or active transport. Conversely, EET enhanced Ag¹⁺ resistance. EET activity is clearly detrimental to copper tolerance in L. monocytogenes, as also observed in Lactobacillus sp. (38, 39, 49). By using the CopRS two-component system sensing system, L. monocytogenes can supplement defenses by increasing copper resistance and lipoprotein remodeling gene expression in parallel with EET flux to protect itself against increased Cu¹⁺production. While EET still remains a liability with respect to copper tolerance, the cost is almost certainly outweighed by the many benefits (52). EET activity helps maintain the NADH/NAD⁺ redox balance, provides respiratory ETCs options that are not dependent on oxygen, and enhances bioavailability of extracellular ferric iron through reduction. Further understanding of extracellular copper trafficking in Firmicutes, and in turn the additional mechanisms of partitioning electron flux between EET and respiratory ETCs, will be needed to fully address what environmental conditions trigger lipoprotein remodeling through copper sensing two-component signaling.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

L. monocytogenes CFSAN023459 contains two plasmids, CFSAN023459_01 (12,949 bp; GenBank accession no. NZ_CP014253.1) and CFSAN023459_02 (52,687 bp; GenBank accession no. NZ_CP014254.1); the latter contains the lit2-copper resistance operon. Listeria strains were grown at 37°C with agitation in TSB or in supplemented ISP1. Standard culture conditions were defined as 2-mL aliquots of culture in 14-mL Falcon snap cap culture tubes grown with shaking (250 rpm). Where noted, cultures were grown under microaerobic, high aeration, or anaerobically without shaking in degassed media using an anerobic chamber (COY lab). Metal stock solutions (1 M for CuSO₄ or AgNO₃ in water and 0.5 M CuCl in degassed 6 N HCl) were made daily before each experiment. Plasmids and strains used in this study are listed in Table 1, and primers are listed in Table S1 in the supplemental material.

TABLE 1.

Bacterial strains and plasmids used in this study

Plasmid or strain	Relevant genotype/phenotypea	Source
Plasmid
pKFC-oriT	E. coli-Gram-positive shuttle vector with temp sensitive replicon and oriT origin of transfer for RP4 mediated conjugation; carbR chlR	Lab stock
pTXM1332	pPL2-SacI-Ppen β-lactamase strep tag-AscI-TT; chlR	Lab stock
pTXM1724	pPL2-PcueOLmCFSANcopS; chlR	This study
pTXM1162	pET22b(+)-N-terminal 6×His tag-TEV cut site-extracellular domain of CopS (KO07_15370, Ser31 to Ser170) from L. monocytogenes	This study
pTXM1405	pTXM1162 His36Phe37 to Ala36Ala37 mutant	This study
pTXM1555	pGEX6P-1-N-terminal 6×His tag-GST fusion tag-TEV cut site-periplasmic domain of CusS (Ser38 to Leu187) from E. coli	This study
Escherichia coli
TXM1152	E. coli HST04 F-, endA1, supE44, thi-1, recA1, relA1, gyrA96, gyrA96, phoA, Φ80d lacZΔ M15, Δ(lacZYA-argF) U169, Δ(mrr-hsdRMS-mcrBC), ΔmcrA, λ-; pRK2013 expressing RP4; kanR	Lab stock
Listeria monocytogenes
Wild-type (Wt)	L. monocytogenes CFSAN023459 with plasmid pCFSAN023459_02	Dwayne Roberson, FDA
TXM1641	Wt ΔcopS	This study
TXM1371	Wt ΔmenE	This study
AAR1846	Wt Δndh1	This study
AAR1848	Wt Δndh2	This study
AAR1857	AAR1846 Δndh2 (Δndh1 Δndh2)	This study
TXM1994	Wt ΔqoxAB	This study
TXM1996	AAR1846 ΔmenE (Δndh1 ΔmenE)	This study
TXM2015	Wt ΔdmkA	This study
TXM2016	Wt ΔcydAB	This study
TXM2017	AAR1848 ΔmenE (Δndh2 ΔmenE)	This study
TXM2018	TXM2046 ΔmenE (ΔeetAB ΔmenE)	This study
TXM2020	Wt ΔmenA	This study
TXM2021	Wt ΔfmnB	This study
TXM2029	TXM1994 ΔcydAB (ΔqoxAB ΔcydAB)	This study
TXM2045	Wt ΔfmnA	This study
TXM2046	Wt ΔeetAB	This study
TXM2065	AAR1846 ΔdmkA (Δndh1 ΔdmkA)	This study
TXM2066	AAR1846 ΔeetAB (Δndh1 ΔeetAB)	This study
TXM2067	Wt ΔpplA	This study

^achl, Chloramphenicol; carb, carbenicillin; kan, kanamycin; TT, transcriptional terminator.

Construction of in-frame deletion and point mutation strains.

In-frame internal gene deletions were constructed using the temperature-sensitive shuttle vector pKFC (53) or an RP4-conjugation competent version (pKFC-oriT). Plasmids were assembled from two 1-kb DNA long PCR amplicons flanking the targeted gene using the primer pairs P1-P2 and P3-P4 (Table S1) which were designed to maintain ~10 coding triplets from both ends of the targeted gene. Electroporation/conjugation, integration, and outcrossing was performed as described previously (16). All deletion alleles were confirmed by PCR using up- and downcheck primers annealing outside the targeted locus (Table S1).

To create point mutations in CopS under the control of the native P_cueO promoter (16), PCR cassettes encoding wild-type copS (TM2232-2233) and P_cueO (TM2311-2317) were introduced into SacI/AscI digested plasmid pTXM1332 by a three-piece assembly (InFusion; TaKaRa Biosciences). The plasmid pTXM1332 is a modified version of the E. coli-L. monocytogenes integration vector pPL2 (54) that had been fitted with a strong transcriptional terminator (AscI-blaZ-TT). The resulting plasmid, pTXM1724, was then used as the template to generate copS variant alleles by inverse PCR (TM2234-2244, TM2318-2319). Sequence verified constructs were integrated into L. monocytogenes ΔcopS (TXM1641) by conjugation to test in trans functional complementation.

Recombinant protein expression and purification.

The extracellular metal binding domain of L. monocytogenes CopS (Ser31 to Ser170) was amplified by PCR (TM1806-1807) and cloned into a pET22b(+) expression vector modified with an N-terminal 6×His tag followed by a TEV protease cut site using NheI/HindIII restriction sites to make pTXM1162. The His36Phe37 to Ala36Ala37 (HF->AA) double mutant expression construct pTXM1405 was constructed by inverse PCR (TM2058-2059) using pTXM1162 as the template. E. coli BL21(DE3) cells carrying either plasmid were grown in 1 L of LB at 37°C until optical density at 600 nm (OD_{600 nm}) reached ~0.5. After cooling to 16°C, the cultures were induced with 1 mM isopropyl b-d-1-thiogalactopyranoside and grown at 16°C overnight with shaking (250 rpm). Cells were pelleted by centrifugation (4,000 × g for 10 min) and frozen at −20°C. Cells were resuspended in column buffer (20 mM Tris [pH 8], 300 mM NaCl, 5 mM imidazole, and 1 mM phenylmethylsulfonyl fluoride protease inhibitor), lysed with a French pressure cell (3 passes at 14,000 lb/in²), and clarified by centrifugation (20,000 × g for 30 min). Recombinant 6×His proteins were then purified from the supernatant using HisPur cobalt resin (ThermoFisher Scientific) according to the manufacturer’s instructions. The 6×His-tag was cleaved by incubation with TurboTEV protease (1:200 6×His-CopS; BioVision) at 4°C overnight in dialysis buffer (5 mM DTT, 10% glycerol, 300 mM NaCl, and 50 mM sodium phosphate [pH 7.4]). Recombinant proteins were then passed through HisPur cobalt resin to remove the cleaved His tag fragment. CopS preparations were treated with 10 mM EDTA and further purified on a Superdex 200 gel filtration column (GE Healthcare) preequilibrated with running buffer containing 20 mM sodium phosphate (pH 7.4), 150 mM NaCl, and 10% glycerol. Fractions containing purified proteins were pooled, concentrated, and supplemented with glycerol to a 20% final concentration before being flash frozen in liquid nitrogen for storage at −80°C. The periplasmic metal binding domain of CusS from E. coli (Ser38 to Leu187) was amplified (TM2130-2148) and cloned into a modified pGEX-6P-1 expression vector with an N-terminal 6×His tag-GST fusion-TEV protease cut site using NheI/XhoI restriction sites to make pTXM1555. Recombinant CusS was expressed as above, and isolated using Pierce Glutathione Superflow agarose according to the manufacturer’s instructions. The equilibration and wash buffer contained 125 mM Tris (pH 8), 150 mM NaCl, and 10% glycerol, with 10 mM reduced glutathione added to elution buffer. After cleaving the GST tag with TurboTEV protease overnight, the protein was passed through the glutathione resin again to remove the cleaved tag. After treatment with 10 mM EDTA, purified protein was obtained using a Superdex200 column run in buffer containing 20 mM Tris (pH 7.4), 150 mM NaCl, and 10% glycerol. Protein aliquots were supplemented to 20% glycerol and flash frozen in liquid nitrogen for storage at −80°C.

Native PAGE analysis.

Recombinant CopS from L. monocytogenes, the HF-> AA mutant, and E. coli CusS stock solutions were diluted in incubation buffer (20% glycerol, 40 mM HEPES [pH 7.4], and 0.02% bromophenol blue) and mixed with an equal volume of water, EDTA, and Cu²⁺ or Ag¹⁺ stock solutions. Final concentrations of proteins were 5 μM, and metals/EDTA ranged from 200 to 0.09 μM. Solutions were incubated for 30 min at room temperature before loading 10 μL into a 15% acrylamide minigel prepared as described (55), but with the following modifications. The separating gel was made by mixing 4.5 mL acrylamide-bisacrylamide (AB)-3 mix (49.5% T, 3% C), 5 mL gel buffer 3× (150 mM Bis-Tris, pH adjusted to 7.4 using HCl), 3 g 100% glycerol, 2.4 mL water, 75 μL 10% APS, and 7.5 μL TEMED. The 3.5% acrylamide sample gel was composed of 0.4 mL AB-3 mix, 2 mL gel buffer 3×, 3.4 mL water, 50 μL 10% APS, and 5 μL TEMED. Gels were run for ~3 total hours at 4°C using a 15 mM Bis-Tris, 50 mM tricine (pH ~7) running buffer with no other additives since proteins are soluble and acidic (calculated pI ~ 4.5 to 5.5). A voltage of 30 V was applied until the samples entered the separating gel, and then the voltage was increased gradually until a maximum of 400 V to keep the current kept below 7 mA. The NativeMark Unstained Protein Standard ladder (Invitrogen) was used to estimate size. After the bromphenol blue front exited the gels, the gels were stained with Coomassie blue and imaged using Bio-Rad Image Lab 5.0 software.

Culture conditions for lit2 induction by metals.

To test CopS-mediated induction of lit2 by copper, overnight cultures were diluted 1:200 vol/vol in TSB media and grown at 37°C in TSB supplemented with 1 mM CuSO₄ under standard culture conditions. When the OD₆₀₀ reached 1.0, cells from 1-mL aliquots (~10⁸ cells) were harvested by centrifugation (3,000 × g for 5 min) and metabolism was immediately quenched by resuspension in 900 μL of ice-cold Tri reagent (Sigma). Samples were stored at −80°C until ready for RNA extraction.

To test lit2 induction in ETC, aerobic respiration, and EET deletion strains, cells were grown in supplemented ISP1 media. This medium was formulated to be pH buffered, to prevent silver ion salt precipitation by limiting chloride ions, and to have a small amount of adventitious metal chelating nutrients. Overnight cultures grown in supplemented ISP1 were diluted 1:200 vol/vol into fresh media and grown to an OD₆₀₀ of ~1 at 37°C under standard culture conditions. Cells from 1 mL aliquots were collected by centrifugation (3,000 × g for 5 min), the spent media was removed by pipette, and the resulting cell pellets were treated as follows. For microaerobic conditions, pellets were washed once with supplemented ISP1 before resuspension in 1 mL of supplemented ISP1 containing various concentrations of CuSO₄ or AgNO₃. Cultures were incubated in closed tubes with minimal headspace. For aerobic conditions, washed cells were likewise resuspended but transferred to uncovered 12-well microplates that were continually aerated with high-speed shaking (250 rpm). For anaerobic conditions, cell pellets were transferred to an anaerobic chamber and washed with degassed supplemented ISP1 before resuspension in media with added metal (CuSO₄, CuCl, or AgNO₃). Cultures were incubated for the indicated times before being harvested by centrifugation, quenched with Tri reagent, and stored at −80°C.

Total RNA isolation.

RNA extraction and analysis by Northern blotting for lit2 expression were according to published protocols (15, 16). Briefly, RNA was extracted using a using a PureLink RNA minikit (Invitrogen) with the following modifications. Thawed cells in Tri Reagent were combined with an equal volume of 0.1-mm zirconium beads and disrupted using a mini-bead-beater (Biospec) at 3,800 rpm for four 30-s cycles with a 3-min rest on ice in between. Beads and unbroken cells were removed by centrifugation (3,000 g for 2 min), and 600 μL of the supernatants was collected. After chloroform (100 μL) induced phase separation and centrifugation, 400 μL of the upper aqueous phase was collected. An equal amount of 70% ethanol was added, and the entire volume was transferred to a silica spin cartridge for RNA isolation according to the manufacturer’s instructions. Total RNA was eluted with 40 μL of RNase-free water and quantified by absorbance at 260 nm.

Northern blotting.

Northern blots were performed using a NorthernMax kit (Ambion) according to published protocols (15) and the manufacturer’s instructions. Briefly, total RNA was separated on a MOPS-formaldehyde-agarose gel and transferred to a BrightStar-Plus positively charged nylon membrane (Invitrogen) using a Whatman Nytran SuPerCharge TurboBlotter kit (GE Healthcare Life Sciences) with a 3-h transfer time. Nucleic acids were fixed by baking at 80°C for 20 min. Biotin-labeled lit2 RNA probes were synthesized from DNA using a Maxiscript T7 transcription kit (Thermo Fisher), including the optional DNase digestion and cleanup with NucAway spin columns (Invitrogen), along with gene-T7-specific primer sets (15). Probes were added to 10 ng/mL in Ultrahyb ultrasensitive hybridization buffer (Invitrogen) and incubated at 72°C for 20 h. The membranes were washed as directed in the NorthernMax kit, with the two high-stringency washes performed at 68°C. RNA was visualized with a chemiluminescent nucleic acid detection module kit (Thermo Fisher) according to the manufacturer’s instructions.

BCS and ferrozine assays for Cu²⁺ or Fe³⁺ reduction.

L. monocytogenes cells grown in supplemented ISP1 media to mid-log phase under standard culture conditions were washed with fresh media, and the cell density normalized to an OD₆₀₀ of ~1 by resuspension in fresh medium aliquots containing 2 mM the Cu¹⁺ chelator bathocuproine sulfonate (BCS; Sigma). BCS is a Cu¹⁺chelator that forms oxygen stable complexes and has low intrinsic cell permeability (56). Experiments were initiated by adding 100 μL of cell suspensions to an equal volume of supplemented ISP1 with 0.5 mM CuSO₄ and 0.2% agarose in uncapped 200 μL PCR tubes. Colorimetric change was visually monitored for 1 h at room temperature without agitation. For quantification of copper reduction, reactions were conducted in triplicates in 96-well microplates setup as described above but lacking agarose. End-point absorbances at 480 nm (SpectraMax Plus 384) were recorded to measure the amount of BCS₂-Cu¹⁺ complex formed. The average value of a no-cell background control was subtracted from each value to account for noncell-catalyzed copper reduction. Detection of Fe³⁺ reduction was performed as described above for copper, except using ferrozine (1 mM final) and iron (III) citrate (0.25 mM final) to monitor formation of the dark blue ferrozine-Fe²⁺ complex. Ferrozine stably binds Fe²⁺ to create a navy-blue color in a complex containing other amino acid ligands (57).

Oxygen consumption continuous assay.

Reactions were conducted using sealed quartz cuvettes with a silicone septum to block gas exchange at the entry point of the oxygen-sensing probe. Cells were pregrown under standard culture conditions in supplemented ISP1, washed, and adjusted to an OD_600nm of ~0.5 (5 × 10⁷ CFU/mL) by resuspension in fresh medium containing 2 mM BCS. Aliquots (2 mL) were added to the cuvette, and reactions initiated by syringe injection of 2 mL of 0.5 mM CuSO₄. A_480nm measurements (Cary 100 UV-Vis; Agilent) were made every minute over the 1-h incubation period at 25°C with continuous stirring to determine the rate of copper reduction. A paired no-cell control measurement was subtracted from the tested sample to account for background. The oxygen sensing probe (Pyroscience, Fixed Fiber Oxygen Minisensor OXF1100) was used to simultaneously measure oxygen levels at 1-s intervals over the incubation period. The probe measures oxygen concentrations through the quenching of the luminescence of phosphorescent porphyrins embedded within the probe tip.

Filter disk-diffusion assay for whole-cell silver reductase activity.

A modified version of the disk-diffusion assay was used to detect silver reduction whereby 10 μL aliquot of a 0.5 mM AgNO₃ solution was added to a filter paper disk (3 mm) and placed in the center of a supplemented ISP1 agarose (1.5%) plate. Exponentially dividing cells grown under standard culture conditions in supplemented ISP1 media until an OD_600nm of ~1 were washed, diluted, and then radially struck on the ISP1 plate. Plates were incubated overnight aerobically or anaerobically at 37°C. The formation of a black/brown dark precipitate at the interphase between bacterial growth and the diffused silver indicates reduction to insoluble Ag⁰.

MIC measurements and growth assay.

Strains were grown under standard culture conditions in supplemented ISP1 to an OD_600nm of ~1, transferred to an anaerobic chamber and washed with degassed media. Around 10⁸ cells were then incubated overnight at room temperature under anaerobic conditions in supplemented ISP1 media with added CuSO₄ or AgNO₃. After 20 h of metal challenge, a 10-fold serial dilution series in TSB was made and aliquots were spotted on TSA plates for colony counting. Growth recovery kinetics were monitored by measuring the OD_600nm of the diluted cultures over 24 h at 37°C, with measurements recorded every 20 min.