Methionine-rich domains emerge as facilitators of copper recruitment in detoxification systems

Significance

The Copper stress represents a virulence factor for bacteria, including pathogenic ones, requiring resistance mechanisms to cope with elevated copper levels. Copper efflux oxidase (CueO), one of the richest methionine-containing proteins in Escherichia coli proteome, is part of the copper detoxification system. It oxidizes Cu+ into less toxic Cu2+ in the periplasm limiting reactive oxygen species production. A prominent structural feature of CueO is a partially disordered methionine-rich domain believed to participate in copper detoxification. Being often present in homologous proteins, the exact role of this domain and related copper-binding sites in the enzymatic mechanism has remained elusive. Using an original combination of in vivo and in vitro methods, we address this question toward improved understanding of bacterial copper homeostasis.

Abstract

Copper homeostasis mechanisms are critical for bacterial resistance to copper-induced stress. The Escherichia coli multicopper oxidase copper efflux oxidase (CueO) is part of the copper detoxification system in aerobic conditions. CueO contains a methionine-rich (Met-rich) domain believed to interact with copper, but its exact function and the importance of related copper-binding sites remain unclear. This study investigates these open questions by employing a multimodal and multiscale approach. Through the design of various E. coli CueO (EcCueO) variants with altered copper-coordinating residues and domain deletions, we employ biological, biochemical, and physico-chemical approaches to unravel in vitro CueO catalytic properties and in vivo copper resistance. Strong correlation between the different methods enables evaluation of EcCueO variants’ activity as a function of Cu+ availability. Our findings demonstrate the Met-rich domain is not essential for cuprous oxidation, but it facilitates Cu+ recruitment from strongly chelated forms, acting as transient copper binding domain thanks to multiple methionines. They also indicate that the Cu6/7 copper-binding sites previously observed within the Met-rich domain play a negligible role. Meanwhile, Cu5, located at the interface with the Met-rich domain, emerges as the primary and sole substrate-binding active site for cuprous oxidation. The Cu5 coordination sphere strongly affects the enzyme activity and the in vivo copper resistance. This study provides insights into the nuanced role of CueO Met-rich domain, enabling the functions of copper-binding sites and the entire domain itself to be decoupled. This paves the way for a deeper understanding of Met-rich domains in the context of bacterial copper homeostasis.

Copper homeostasis is a vital aspect of bacterial physiology, as it allows microorganisms to cope with toxic concentration of copper in the environment. This ability represents an important virulence factor used in many host–pathogen interactions (1–5). The resistance of bacteria to Cu-mediated lysis by macrophages mainly lies in their capability to export Cu⁺ from the bacterial cell (6, 7). In gram-negative bacteria, such as Escherichia coli, three main systems control the copper homeostasis: CusCFBA, CopA, and copper efflux oxidase (CueO). CusCFBA is responsible for the excretion of copper into the extracellular environment (8, 9). CopA is an inner membrane ATPase that pumps cuprous ions (Cu⁺) from the cytoplasm to the periplasm (10). The CueO then couples the oxidation of Cu⁺ to less toxic cupric (Cu²⁺) ions with the concomitant oxygen reduction reaction (ORR) to water, limiting the production of reactive oxygen species (11, 12). CueO is a member of the multicopper oxidase (MCO) family, alongside the well-described bilirubin oxidases (BODs) and laccases (LACs). MCOs share a common coordination of four copper atoms that are spatially and functionally organized into two centers: a mononuclear type-I copper center (Cu-T1) and a trinuclear copper center (TNC), composed of one type-II (Cu-T2) and two type-III (Cu-T3) copper atoms (Fig. 1B). MCOs catalyze the oxidation of an extensive range of aromatic substrates, including polyphenols such as lignin (13). Substrate oxidation occurs at the Cu-T1 center that also plays the role of electron relay to the TNC, which then catalyzes the ORR. In contrast to BODs and LACs, CueO-like enzymes exhibit an additional protein domain rich in methionine residues (Met-rich) covering the Cu-T1 center (Fig. 1). CueO proteins are extensively distributed among pathogenic and nonpathogenic bacteria across diverse environments (14), and the structural organization and number of methionine residues in the Met-rich domain differ depending on the microorganism (15). E. coli CueO (EcCueO) has been especially, and almost exclusively, studied. Its crystallographic structures have revealed three additional copper-binding sites located in the Met-rich insert (Fig. 1B) (11). Cu5 is coordinated by two Met and two Asp residues, while Cu6 and Cu7 are coordinated by Met residues only.

Fig. 1.

(A) Ribbon diagram of superimposed EcCueO X-ray structures (PDB entries: 3OD3 and 3NT0). The Met-rich domain is highlighted in lime green. Copper atoms are represented as colored spheres, Cu-T2, and Cu-T3, respectively, colored in brown and dark red. (B) Copper-binding sites found in the X-ray structures of EcCueO: Cu5, Cu6, and Cu7 coordination from 3NT0, Cu-T1, TNC, and ribbon diagram from 3OD3. The structure 3NT0 is of ΔCu-T1 mutant that allows the occupied Cu6 and Cu7 to be observed, yet the whole Met-rich domain was not resolved in this structure (11). The WT-structure 3OD3 ribbon diagram was therefore used to construct the complete picture. Cu6 is coordinated by M358, M362, and a water molecule (small red sphere) [electron density for an unknown molecule has been modeled as a water molecule that is more likely to be a histidine or a methionine as suggested by authors (11)]. Cu7 is coordinated by M364, M368, and M376. Cu5 is coordinated by M355, M441, D439, and D360. The latter being a residue of the methionine-rich helix is lacking in the EcCueO ΔMet^358-407 variant.

The function of the Met-rich domain and additional copper-binding sites in relation to the role of CueO in copper homeostasis remains poorly understood. Typically, the catalytic activity of CueOs is assessed in vitro using two categories of substrates: phenolic compounds, representing typical substrates for MCOs (referred to as phenol oxidase activity), and Cu⁺ complexes, the oxidation of which is specific to CueO subfamily (referred to as cuprous oxidase activity). The phenol oxidase activity of CueOs is typically low because the Met-rich insert functions as a physical barrier to phenol substrate access. It was shown that Cu²⁺ ion addition in solution can enhance this activity (16–19), thanks to Cu²⁺ coordination in the Cu5 site acting as an electron relay (20). Accordingly, in EcCueO, it has been demonstrated that the removal of the Met-rich domain (ΔMet) induces a 33-fold increase in phenol oxidase activity and mutation of one of the EcCueO Cu5 ligands induced a decrease in the Cu²⁺-stimulated phenol oxidase activity (17). However, in contrast with the above findings, no increase in phenol oxidase activity was observed for Thermus thermophilus CueO (TtCueO) after deletion of a Met-rich hairpin also present near Cu-T1. The Cu²⁺-dependent activation by ΔMet protein was also similar to that observed for the WT enzyme, thus exhibiting a marked different behavior than EcCueO (18). Regarding cuprous oxidase activity, it has been demonstrated that the removal of the Met-rich domain in EcCueO results in a dramatic reduction of Cu(I) oxidation rate in solution by approximately 90% compared to the WT enzyme (16). On the contrary, the cuprous oxidase activity of ΔMet TtCueO quantified by electrochemistry thanks to electrogeneration of Cu⁺ in close proximity to the enzyme failed to demonstrate a role of the Met-rich domain (18). The specific role of Cu6 and Cu7 in EcCueO seems controversial as well. It has been proposed that Cu6 and Cu7 act as active sites, their absence inducing cuprous and phenol oxidase inactivity of EcCueO (21). In another report, Cu6 and Cu7 removal did not abolish the cuprous oxidase activity but reduced it by four times compared to the WT enzyme (11).

The present study provides a perspective of the role of Met-rich domains in CueO by examining the oxidative activities of EcCueO using a multimodal strategy. Notably, in vivo copper resistance experiments are shown to corroborate the mechanisms drawn from in vitro analysis. Several EcCueO variants with different copper-coordinating residues and entire domain deletions were designed. Complementary information was gained from each of the approaches used, namely biochemistry, spectroscopy, electrochemistry, and genotype–phenotype relationship, showing strong correlation and allowing us to propose a specific role of the Met-rich domain and each copper-binding site. Our multimodal strategy, employing copper from weakly complexed to strongly chelated forms, firmly demonstrates the link between cuprous oxidase activity in vitro, copper resistance in vivo and Cu⁺ availability. In this context, the multiple methionine residues of the Met-rich domain help in Cu⁺ recruitment. Single or double mutants of Cu6 and Cu7 further demonstrate that both copper sites are unnecessary for in vitro and in vivo catalysis, which is also assessed by the retention of catalytic activity of the enzyme with a deleted Met-rich domain. In-depth study of the different variants properties and catalytic activities provides robust evidence that Cu5 is the sole substrate-binding active site for cuprous oxidation. We also show that removal of at least two Cu5 ligands is required to abolish cuprous oxidase activity. In addition, the nature of the Cu5 coordination sphere strongly affects the affinity for Cu⁺, hence affecting the electron transfer to Cu-T1 and the in vivo resistance to copper stress.

Results

Biochemical and Spectroscopic Properties of EcCueO Enzymes Recombinant Expressed in E. coli.

In order to better understand the specific function of the Met-rich domain and each of the additional coppers, namely Cu5, Cu6, and Cu7, a series of mutations were carried out (SI Appendix, Table S1, Fig. S1). First, two distinct deletions of the Met-rich domain were designed. Deletion from P₃₅₈ to H₄₀₇ (ΔMet^358-407) leads to the elimination of the entire domain along with the D₃₆₀ Cu5 ligand. On the other hand, deletion from M₃₆₁ to M₃₉₆ (ΔMet^361-396) leads to the removal of the majority of the Met-rich domain but retains D₃₆₀ as a fourth Cu5 ligand (SI Appendix, Fig. S2). Second, mutants that lost the coordination of either Cu6 or Cu7 were obtained by mutating the involved methionine residues to glutamine residues (ΔCu6: M₃₅₈,₃₆₂Q and ΔCu7: M₃₆₄,₃₆₈,₃₇₆Q). A mutant that lost the coordination of both Cu6 and Cu7 (ΔCu6/7: M₃₅₈,₃₆₁,₃₆₂,_364,366,₃₆₈Q) was obtained by replacing two Cu6- and Cu7-methionine residues and additionally two other neighboring methionine residues not involved in their coordination but accessible to solvents (M₃₆₁ and M₃₆₆) (SI Appendix, Fig. S1). Finally, two mutants of a Cu5 site, a single residue ΔCu5^M441Q mutant, and one in which two residues were mutated (ΔCu5^D439A,M441Q) were also constructed to specifically target the influence of Cu5 coordination. A catalytically inactive ΔCu-T1 (C₅₀₀S) enzyme was used as a negative control. Indeed, the loss of Cu-T1 coordination disrupts the electron transfer to the TNC required to catalyze the ORR.

EcCueO WT and mutants were produced in DH5α E. coli cells. One-step purification yielded pure and homogeneous enzymes, as demonstrated by analytical SDS-PAGE and size exclusion chromatography (SI Appendix, Fig. S3). Circular dichroism spectra analysis confirms that residue substitution or deletion of the Met-rich insert does not affect nor induce major changes in the secondary structures of EcCueO mutants (SI Appendix, Fig. S4). All EcCueO preparations show the typical blue coloration due to the presence of the Cu-T1 center, apart from the ΔCu-T1 mutant as expected. The presence or absence of copper centers is also supported by the UV-visible absorption spectra (SI Appendix, Fig. S5), which show a typical shoulder at 330 nm and a peak at 610 nm due to Cu-T3 and Cu-T1 absorption, respectively. Determination of copper content by ICP-OES (SI Appendix, Table S3) confirms the presence of about 4 copper atoms for EcCueO. Attempts to assay by ICP-OES the possibility of binding Cu⁺ to the additional sites present in the Met-rich domain were not successful suggesting only transient binding of Cu⁺ to these sites (SI Appendix, for more details). EPR spectroscopy confirms the integrity of the Cu clusters. All the isolated EcCueOs show Cu-T1 and Cu-T2 signatures (SI Appendix, Fig. S6). The Cu-T1 presents a near-axial signal with g_// = 2.241, g_⊥ = 2.04 and A_//(Cu) = 62.10⁻⁴ cm⁻¹; whereas the Cu-T2 site is characterized by g_// = 2.278 and A_//(Cu) = 146.10⁻⁴ cm⁻¹. These values are in agreement with those previously published for EcCueO (22–26).

In Vitro Cu⁺ Oxidase Activity of EcCueOs in Solution.

Measurement of cuprous oxidase activity requires overcoming the instability of Cu⁺ ions in aerobic aqueous solutions. The use of [Cu^I(BCA)₂]³⁻ as a copper complex allows the combination of sufficient stability and a priori suitable affinity (β₂ = 10^17.2 M⁻²) for the transfer of Cu⁺ to the enzymatic active site (12, 21, 27). Complexes with either higher ([Cu^I(BCS)₂]³⁻, β₂ = 10^20.8 M⁻²) or lower ([Cu^I(Ferene)₂]³⁻, β₂ = 10^13.7 M⁻²) affinity were also studied for comparison, resulting in lower and higher V_max, respectively (SI Appendix, Fig. S7 and Tables S4–S6). In the presence of active EcCueO, [Cu^I(BCA)₂]³⁻ is expected to be oxidized releasing free BCA and Cu²⁺ as reaction products (SI Appendix, Fig. S7A). Control experiments were performed with Bacillus pumilus BOD and Acidithiobacillus ferrooxidans AcoP (SI Appendix, Fig. S7 B and C). The former is a MCO that naturally lacks the Met-rich domain (28, 29), whereas the latter is a cupredoxin protein with no enzymatic function (29, 30). No cuprous oxidase activities were detected for either protein, attesting that the oxidation of [Cu^I(BCA)₂]³⁻ is inherent to EcCueOs. Furthermore, in the presence of an inactive EcCueO, Cu⁺ transfer to the enzyme was observed without catalytic oxidation (SI Appendix, Fig. S8). Additionally, no shift in the visible spectrum of [Cu^I(BCA)₂]³⁻ was observed that would have indicated the formation of a stable Cu⁺-mediated ternary complex between the EcCueO and [Cu^I(BCA)₂]³⁻ (31).

As shown in Fig. 2A, the oxidation of [Cu^I(BCA)₂]³⁻ followed standard Michaelis–Menten behavior. Derived kinetic parameters are given in Fig. 2D and SI Appendix, Table S6. As expected, the loss of Cu-T1 coordination (ΔCu-T1) completely abolished cuprous oxidase activity, highlighting the requirement of the electron transfer chain from Cu-T1 to the TNC. The abolishment of the activity is also observed upon loss of Cu5 coordination (ΔCu5^D439A,M441Q). In contrast, mutation of a single Cu5 ligand (ΔCu5^M441Q) results in an active enzyme with approximately 20% residual catalysis. The loss of Cu6 or Cu7 coordination does not affect the activity. Indeed, the V_max values for the single mutants (ΔCu6 V_max = 1.05 U mg⁻¹, ΔCu7 V_max = 1.00 U mg⁻¹) are comparable to that of the WT enzyme (WT V_max = 1.06 U mg⁻¹). The double mutant shows about half of the V_max (ΔCu6/7 V_max = 0.57 U mg⁻¹), which will be explained below considering in addition phenol oxidase activity. As a first conclusion, these results demonstrate the functional role of Cu5 and strongly suggest that Cu6 and Cu7 binding sites should not be required for cuprous oxidase activity, at least with [Cu^I(BCA)₂]³⁻ as substrate. ΔMet preparations on the other hand showed 10 to 20% of remaining catalysis (ΔMet^358-407 V_max = 0.11 U mg⁻¹, ΔMet^361-396 V_max = 0.22 U mg⁻¹). The marked different activity of ΔMet compared to ΔCu6 or ΔCu7 implies that the Met-rich domain participates in cuprous oxidation. Its exact role will be investigated further below using alternative substrates.

Fig. 2.

(A) Cuprous oxidase activity of EcCueOs preparations: WT (), ΔCu7 (), ΔCu6 (), ΔCu6/7 (), ΔCu5^M441Q (), ΔCu5^D439A,M441Q (), ΔMet^361-396 (), ΔMet^358-407 (), and ΔCu-T1 (), the latter points overlay with ΔCu5^D439A,M441Q. Dashed lines represent Michaelis–Menten fit for [Cu^I(BCA)₂]³⁻ oxidation. (B) ABTS oxidation activity for EcCueOs preparations: WT (a), ΔCu7 (b), ΔCu6 (c), ΔCu6/7 (d), ΔCu5^M441Q (e), ΔCu5^D439A,M441Q (f), ΔMet^361-396 (g), and ΔMet^358-407 (h). Dashed lines represent Michaelis–Menten fit for ABTS oxidation: as prepared (empty squares) and in presence of 500 µM CuSO₄ (filled squares). EcCueO ΔCu-T1 () does not show any activity and it is reported in all graph as negative control. (C) V_max correlation between the cuprous and ABTS oxidase activities. V_max for [Cu^I(BCA)₂]³⁻ oxidation in U mg⁻¹, V_max for ABTS oxidation enhancement in presence of 500 µM CuSO₄ in %. (D) Resulting V_max and K_M from Michaelis–Menten fit of cuprous oxidase activity of EcCueOs. Columns coloration referred to (A). (E) Resulting V_max and K_M from Michaelis–Menten fit of ABTS oxidase activity of EcCueOs. Columns coloration referred to (A), empty and filled columns for without and with 500 µM CuSO₄ respectively.

In Vitro ABTS Oxidase Activities of EcCueOs in Solution.

Phenol oxidase activity is usually studied using ABTS as a model substrate for many MCOs. As with phenols, CueO-like enzymes typically demonstrated low activity with this substrate, presumably due to steric hindrance induced by the Met-rich domain. Nevertheless, ABTS-oxidation activity was shown to be enhanced by the addition of Cu²⁺, reportedly due to the Cu²⁺-binding and acting as an electron relay (17). In the current study, this approach complements the [Cu^I(BCA)₂]³⁻ one in the sense that it allows an indirect probe of cuprous oxidase activity vice versa, starting from Cu²⁺-binding, while avoiding the strongly chelating BCA agent. The steady-state ABTS oxidase activities of EcCueOs preparations were measured in the absence and presence of 500 µM CuSO₄ (Fig. 2B) (16, 18). The activities obtained are in accordance with the Michaelis–Menten equation, thereby enabling the retrieval of kinetic parameters (SI Appendix, Table S6 and Fig. 2E).

In the absence of Cu²⁺, the catalytic efficiency (V_max/K_M) of the WT enzyme is significantly reduced for ABTS oxidase activity (148 U mg⁻¹ M⁻¹), being 70 times lower than for cuprous oxidase activity (10,303 U mg⁻¹ M⁻¹). While ΔCu6 and ΔCu7-mutants show similar ABTS activities to WT, the deletion of the bulky Met-rich domain favors the approach of ABTS to Cu-T1, increasing the inherent ABTS oxidation activity by an order of magnitude in comparison to the WT (Fig. 2E and SI Appendix, Table S6) (16, 32). Interestingly, the activity of ΔCu5^D439A,M441Q was equivalent to that of WT activity, indicating that ABTS transfers electrons directly to Cu-T1 in the absence of the copper atom bound to Cu5 binding site, albeit at a slow rate.

Upon addition of Cu²⁺, all mutants, except ΔCu5^D439A,M441Q, displayed an enhancement of the ABTS oxidation activity (Fig. 2E). A positive correlation between [Cu^I(BCA)₂]³⁻ oxidation activity and enhancement of ABTS oxidation activity upon Cu²⁺ addition is obtained (Fig. 2C), which suggests that the same residues are involved in both activities. The sole mutant not activated by the addition of Cu²⁺ was ΔCu5^D439A,M441Q, in line with the pivotal role of Cu5-residues in copper binding. In addition, this result suggests that electrons are transferred from ABTS through Cu5 in the presence of Cu²⁺. In cases where a single Cu5 ligand was mutated or deleted (ΔCu5^M441Q and ΔMet^358-407), some activation upon Cu²⁺ addition was preserved, in agreement with the [Cu^I(BCA)₂]³⁻ oxidation results. WT, ΔCu6, ΔCu7, and ΔCu6/7 exhibit comparable enhancements in activity, ranging from 500 to 650% (Fig. 2C), which confirms our initial findings on the noninvolvement of Cu6 and Cu7 in the cuprous oxidase activity.

Activation upon Cu²⁺ addition was still observed for both ΔMet-mutants, yet to a lesser extent than in WT enzyme. These findings, along with our analysis of [Cu^I(BCA)₂]³⁻ activities, confirm that the deletion of the Met-rich domain alone is not sufficient to abolish cuprous oxidase activity. The activation of ΔMet^358-407 which lacks D₃₆₀ residue, is more pronounced than that of ΔMet^361-396. This was a priori unexpected as Cu5 in ΔMet^361-396 presents a complete coordination sphere. Nevertheless, the activation of ABTS oxidation is a complex process that is subject to a number of factors, including the binding of copper to the Cu5-site, as well as the binding and accessibility of ABTS. It is possible that the remaining portion of the helix in ΔMet^361-396 may impede access to ABTS. Molecular dynamics simulations conducted in the absence and presence of copper bound to Cu5-site indicate that the Cu-T1 is less accessible to the solvent, thereby limiting the availability of ABTS, upon copper fixation (SI Appendix, Fig. S9). This could explain the lower extent of the activation of this mutant compared with ΔMet^358-407 which does not possess the remaining helix.

The case of the ΔCu6/7 mutant deserves to be discussed separately. The ABTS oxidation activity is 42% lower than for the WT enzyme both in the absence and presence of Cu²⁺ (Fig. 2E and SI Appendix, Table S6). Considering the lower [Cu^I(BCA)₂]³⁻ oxidation activity of this mutant compared to WT, such findings suggest that, more than a specific function of these residues in the cuprous oxidation, the intrinsic enzyme activity has been modified due to replacement of multiple amino acids. Indeed, for ΔCu6/7 mutant, six hydrophobic methionine residues have been substituted to more hydrophilic glutamine residues that could impact the functional folding of the Met-rich domain.

Electrochemical Cuprous Oxidase Activity of EcCueOs.

[Cu^I(BCA)₂]³⁻ is a suitable substrate for the measurement of cuprous oxidase activity in solution; however, it is not representative of physiological conditions. Seeking a more accurate method capable of mimicking physiological Cu⁺ transfer, we recently reported the possibility of detecting the cuprous oxidase activity of CueO-like proteins by electrochemistry (18). The principle involves immobilizing enzymes onto an electrode in an orientation that, unlike classical protein film electrochemistry, impedes most direct electron transfer between the electrode surface and the first electron acceptor (Cu-T1). Upon addition of Cu²⁺ into the oxygenated electrochemical cell and application of a sufficient reductive potential, Cu⁺ can be electrogenerated from the reduction of Cu²⁺ at electrode surface, i.e., in close proximity to the immobilized enzyme. If the enzyme catalyzes cuprous oxidation, it can oxidize Cu⁺ and regenerate Cu²⁺ which can be electroreduced at the electrode again, resulting in a catalytic recycling and an absolute increase in reductive current. We propose that this electrocatalytic signal can be employed as a measurement of cuprous oxidase activity. One benefit of the electrochemical approach is that the amount of Cu⁺ produced is controlled by the applied potential and its delivery to the enzyme is rapid, obviating the requirement for strongly complexing agents.

We previously demonstrated that a negatively charged electrode surface induces the unfavorable orientation we were looking for toward the direct electrocatalysis of O₂ by immobilized EcCueO and TtCueO (18, 33). This was attributed to the extensive negatively charged region near Cu-T1, which places it too far away from the electrode surface for efficient electron transfer (SI Appendix, Fig. S8). Similarly, in the current study, EcCueO preparations were immobilized on the negatively charged surface of electrodes based on carboxylic-functionalized carbon nanotubes (COOH-CNT). All EcCueOs exhibit weak reductive catalytic currents toward the direct electroreduction of O₂ (Fig. 3A, red curves), 5 to 15 times lower than when the enzyme is immobilized on positively charged surfaces (NH₂-CNT) (SI Appendix, Fig. S10 and Table S7). Only ΔCu-T1 bioelectrode does not achieve any electrocatalytic activity, showing that similarly to other MCOs, direct ORR by EcCueO occurs exclusively through the Cu-T1 center (SI Appendix, Table S7 and Fig. 3A). The variations in the steady-state current densities (j) may be linked to several factors, such as the loading of enzymes and the distance between the Cu-T1 and the electrode surface, which can be adjusted by minimal conformational changes related to the mutations.

Fig. 3.

(A) Electrochemical activity of EcCueO preparations: WT (a), ΔCu7 (b), ΔCu6 (c), ΔCu6/7 (d), ΔMet^358-407 (e), ΔMet^361-396 (f), ΔCu5^M441Q (g), ΔCu5^D439A,M441Q (h) and ΔCu-T1 (i). Cyclic voltammograms obtained with CueO functionalized COOH-CNT-based electrodes under N₂ (black dotted line), O₂ (red line), and O₂ in presence of increasing CuSO₄ concentrations (blue lines). CuSO₄ concentrations: 0.1, 0.2, 0.5, 1, 2, 5 mM for a–e and g–i, and 0.02, 0.05, 0.1, 0.2, 0.5, 1 for f. Experimental conditions: 100 mM Acetate buffer, pH 5.0 at 30 °C, ω = 3,000 rpm, v = 5 mV s⁻¹. (B) Chronoamperometry at 0.3 V vs. NHE, under O₂ and increasing concentrations of CuSO₄: 0.1, 0.2, 0.5, 1, 2, 5, and 10 mM for WT (), ΔCu7 (), ΔCu6 (), ΔCu6/7 (), ΔCu5^M441Q (), ΔCu5 ^D439A,M441Q () and ΔMet^358-407 (). CuSO₄ concentrations: 0.02, 0.05, 0.1, 0.2, 0.5, 1, and 2 mM for ΔMet^361-396 (). (C) Apparent Michaelis–Menten fit of the Δj at 0.3 V vs. NHE in presence of electrogenerated Cu⁺ species. (D) Resulting Δj_max and _appK_M from apparent Michaelis–Menten fit of the Δj at 0.3 V vs. NHE. Columns coloration referred to (B). (E) Correlation between Δj_max and V_max for ABTS oxidation activity. (F) Correlation between Δj_max and V_max for cuprous oxidase activity.

The addition of CuSO₄ to the electrolyte leads to an increase in catalytic currents for EcCueO enzymes (Fig. 3A, blue curves). Control experiments carried out using COOH-CNT electrodes without immobilized EcCueO (SI Appendix, Fig. S12 A and B) confirm that these reductive currents are closely related to the enzymatic activity of cuprous oxidase. The only two cases where there is no increase in the catalytic current are with ΔCu-T1 and ΔCu5^D439A,M441Q, the mutant that lacks two Cu5 ligands. These results are in agreement with the activity kinetic data for ABTS and [Cu^I(BCA)₂]³⁻ in solution. Hence, although a very different method of analysis than previous ones in solution, electrochemistry validates the role of Cu-T1 to transfer electrons from Cu⁺-oxidation site to the TNC, and the pivotal function of Cu5, which acts as a binding site for electroreduced Cu⁺ ions.

The electrochemical cuprous oxidase activity of EcCueOs can be quantitatively analyzed through chronoamperometry (Fig. 3B). Applying a potential of 0.3 V vs. NHE and subsequently injecting CuSO₄ in the solution, induce an increase in the catalytic current until a limiting value is reached, corresponding to maximum activation. Based on the known thermodynamic constants, the equilibrium concentration of Cu⁺ at 0.3 V vs. NHE can be calculated (SI Appendix, Fig. S12 C and D). By determining the difference between the current density at each concentration of CuSO₄ and in the absence of CuSO₄ (Δj), a Michaelis–Menten-like behavior is obtained (Fig. 3C). The apparent kinetics parameters of cuprous oxidase activity are provided in SI Appendix, Table S8 and Fig. 3D.

Interestingly, the resulting affinities for the electrogenerated Cu⁺ ions are within the nanomolar range, hence in accordance with prior literature findings and the availability of Cu⁺ in vivo (12, 34, 35). The Δj_max obtained from the chronoamperometry data can be regarded as a substitute for V_max and it correlates reasonably well with the outcomes of [Cu^I(BCA)₂]³⁻ and ABTS oxidation in the presence of copper (Fig. 3 E and F). A weaker correlation is obtained with the two ΔMet variants, which show cuprous oxidase activity through electrochemistry comparable to the WT contrary to the solution-based assays. The possible reasons for this disparity will be considered in the discussion section, but it emphasizes the significance of the method used for activity measurement and especially the form in which Cu⁺ is provided.

In Vivo Cuprous Oxidase Activity of EcCueOs.

Since the CueO’s primary function is to protect the cells from copper stress, we ultimately opted to validate the role of the Met-rich domain and copper-binding sites under physiological conditions in vivo, and to confront the conclusions with the in vitro methods. For that purpose, we carried out an in vivo trans-complementation test based on the copper sensitivity of the ΔcusB ΔcueO strain. The cusB gene encodes a subunit of the copper efflux pump CusCFBA (Fig. 4B), and its deletion increases the strain’s copper sensitivity, rendering this trans-complementation test more sensitive. It is noteworthy that the ΔcusB ΔcueO strain carrying an empty plasmid is highly sensitive to copper. The trans-complementation successfully restored the phenotype of the ΔcusB ΔcueO strain when the gene expressing the EcCueO WT was introduced (Fig. 4A). This emphasizes the crucial role of CueO in aerobic copper detoxification in the absence of a functional CusCFBA system. The expression of the ΔCu-T1 variant did not allow complementation, leading to an enzyme completely inactive in vivo and a copper-stressed phenotype similar to that of cells transformed with the empty plasmid (Fig. 4A). We found that the expression of mutated variants on Cu6, Cu7, or both copper centers complemented the strain with the same efficiency as WT EcCueO, as indicated in Fig. 4A. Similarly, high complementation efficiency was observed with the mutants ΔMet^361-396 and ΔMet^358-407. These results indicate that, under the tested conditions, Cu6 and Cu7 binding sites and even whole Met-rich domain are not required for the EcCueO to help the cells to resist the copper-induced stress in vivo. It is also notable that in vivo data correlate the best to cuprous oxidase activity measured by electrochemistry.

Fig. 4.

(A) Effect of copper stress on E. coli cell viability. Cells have been deleted of native cueO and cusB genes, in order to increase the strain’s copper sensitivity, and trans-complemented with the EcCueO plasmids. The presence of exogenous copper stress reflects the differences in efficiency of cuprous oxidase activity of the different EcCueO enzymes. (B) Schematic view of copper homeostasis systems in E. coli (36). CopA translocates Cu⁺ ions from the cytoplasm into the periplasm. CueO oxidizes Cu⁺ to less toxic Cu²⁺. CusCBA efflux system pumps out copper to the extracellular environment. CusF is a periplasmic metallochaperone which supplies copper to the CuCBA pump. (C) Proposed cuprous oxidase (a) and ABTS oxidation (b) reaction mechanisms for EcCueO. The sole active site for cuprous oxidase activity is Cu5. The mutation or loss of only one Cu5-residue is not enough to completely abolish cuprous oxidation (ΔMet^358-407and ΔCu5^M441Q). Conversely, at least two residues must be mutated to fully abolish the catalysis in vitro and in vivo (ΔCu5^D439A,M441Q). The mutation of methionine residues involved in Cu6 and Cu7 coordination virtually does not affect the cuprous oxidation (ΔMet^361-396, ΔCu6, ΔCu7, and ΔCu6/7). ABTS oxidase activity in absence of CuSO₄ (dashed arrow) proceeds via Cu-T1, similarly to MCOs. On the other hand, Cu²⁺ bound to Cu5 might allow the oxidation of ABTS (full arrow), such as in WT, ΔCu7, ΔCu6, ΔCu6/7, ΔCu5^M441Q, ΔMet^358-407, and ΔMet^361-396 preparations. The removal of two Cu5 ligands (ΔCu5^D439A,M441Q) results in an enzyme insensitive to CuSO₄ activation but still active for ABTS oxidation. The loss of Cu-T1 coordination totally abolishes the catalysis, confirming Cu-T1 role in mediating the electron transfer to the buried TNC.

The importance of the Cu5 binding site is clearly visible in copper resistance tests. Indeed, particularly visible at 100 µM CuSO₄, the cells become, to some extent, more sensitive to copper whenever one of the Cu5-binding residues (ΔMet^358-407, ΔCu5^M441Q, or ΔCu5^D439A-M441Q) is affected. As predicted from the enzymatic assays and electrochemistry, the ΔCu5^D439A-M441Q mutant failed to complement, and the cells exhibited a phenotype similar to the empty plasmid or ΔCu-T1 mutant. The ΔCu5^M441Q showed relatively good complementation but not as effective as the WT or Cu6/Cu7-mutants. In particular, in the medium with the addition of 70 µM CuSO₄, ΔCu5^M441Q cells exhibit a phenotype similar to that of WT cells but lower viability at 100 µM CuSO₄ (Fig. 4A), in agreement with most in vitro results.

Discussion

We have employed a multimodal and multiscale approach, combining different and complementary in vitro and in vivo methods aiming at deciphering the cuprous oxidase mechanism of EcCueO. An overall consensus on data analysis allows grouping all the mutants into three categories regarding the effect of the mutations on cuprous oxidase activity: a) the mutants whose cuprous oxidase activity was not or almost not changed compared to the WT, namely ΔCu6, ΔCu7, ΔCu6/7; b) the mutants whose cuprous oxidase activity was significantly reduced compared to WT, but not completely abolished, namely ΔMet^358-407, ΔMet^361-396, ΔCu5^M441Q; c) the mutants which lost completely the cuprous oxidase activity, namely ΔCu-T1 and ΔCu5^D439A,M441Q.

It is remarkable that the mutations of either Cu6 or Cu7 do not alter the cuprous oxidase activity of EcCueO compared to WT in both solution and electrochemical assays, under both in vitro and in vivo conditions (Figs. 2, 3, and 4). It must be reminded that the sole experimental evidence of Cu6 and Cu7 coordination has been acquired in an inactive enzyme protein crystal (ΔCu-T1) that was soaked in Cu⁺ ions (11). Our work suggests that there is only transient copper binding to these enzyme coordination centers, as we were unable to detect additional Cu⁺ binding through ICP-OES after removal of excess unbound Cu⁺. From the functional point of view, our results differ from a previous report in the literature (21), where the loss of Cu6 or Cu7 led to a significant reduction of cuprous oxidase activity. In ref. 21 however, phenol oxidation in the absence of Cu²⁺ was also significantly affected in Cu6 and Cu7-mutants compared to the WT enzyme. Hence, the low cuprous oxidase activity could have been due to intrinsically less active enzymes (21), as we concluded above in our work for the ΔCu6/7 mutant.

In addition to the noninvolvement of Cu6 and Cu7, our Cu5-mutations prove that Cu5 is the main and only substrate-binding active site for cuprous oxidase activity (Fig. 4C). Several roles were proposed before for this site: a regulatory site (rCu) (17), a substrate-binding site (sCu) (11, 12) or electron transfer site (21). Conserved cuprous oxidase activity of ΔMet^361-396, which lost all Met-domain but Cu5-residues, clearly confirms that this site is responsible for Cu⁺-binding and oxidation (Fig. 4C). Remarkably, the residual activities of ΔCu5^M441Q and ΔMet^358-407 variants suggest that the mutation of only one Cu5-residue is not enough to prevent copper binding. Most probably the Cu5 coordination in the ΔMet^358-407 preparation can be completed by water molecules. Indeed, in the crystallographic structures of EcCueO WT and ΔCu6/7 (1N68 and 3NSY) only the O_δ1 atom of D₃₆₀ participates in Cu5 coordination, the other O-donor being a water molecule (11, 17). Therefore, at least two residues must be mutated (ΔCu5^D439A,M441Q) to abolish cuprous oxidase activity completely in vitro and in vivo.

Our conclusion asserting Cu5 as the principal substrate-binding site, while excluding the involvement of Cu6 and Cu7 in cuprous oxidase activity, raises an intriguing question about the true role of the Met-rich domain. It should be noted that the ΔMet^358-407 and ΔMet^361-396 mutants are those correlating the least among different methods of cuprous oxidase activity measurement (Fig. 3 E and F). For example, both ΔMet-mutants demonstrate only 10 to 20% of the WT activity with [Cu^I(BCA)₂]³⁻ as a substrate but essentially the same activity in the electrochemical assays (Figs. 2 and 3). Noteworthy, the in vivo tests support the electrochemical assays, since ΔMet-mutants show a similar copper resistance as WT. One difference between [Cu^I(BCA)₂]³⁻ assays and electrochemistry is that copper is provided in a strongly chelated form in the former one and as copper ions in the latter one. We can therefore speculate that the hydrophobic pool of several methionine residues presents in the Met-rich domain, and not only those involved in the Cu6 and Cu7 coordination observed in the crystallographic structure (Fig. 1 B and C and SI Appendix, Fig. S1), allow the formation of a transient complex that helps to destabilize the strongly chelated Cu⁺ bound to BCA and to transfer it to Cu5. In electrochemistry, the copper is provided in an unchelated form, rendering the Met-pool unnecessary. Consequently, in this assay the two ΔMet variants exhibit an activity profile comparable to that of the WT. This is supported by the fact that this domain sequence and position are relatively weakly conserved among homologous proteins (15, 37). Endorsing this hypothesis, when Cu⁺ is supplied in a weaker chelated form, as in the case of [Cu^I(Ferene)₂]³⁻, the two ΔMet enzymes showed higher relative activity than with [Cu^I(BCA)₂]³⁻, 75 and 57% V_max compared to WT, with or without the complete Cu5 coordination sphere, respectively (SI Appendix, Fig. S7D and Table S5). As can be seen, a more robust correlation between electrochemistry and cuprous oxidase activity with [Cu^I(Ferene)₂]³⁻ was observed for all EcCueOs, including ΔMet variants (SI Appendix, Fig. S7E). The almost completely conserved resistance to copper stress of the cells complemented with ΔMet-mutants in vivo suggests that Cu⁺ is not strongly chelated in the periplasm, at least in our experimental conditions.

In addition to the role we propose for the Met-rich domain to recruit strongly chelated Cu⁺, other roles, not intrinsically linked to the catalytic mechanism, cannot be excluded. One alternative role could be an interaction with a Cu⁺-binding partner. However, our complementation results are not in favor of this role since in this case Met-rich domain deletion would have influenced more importantly copper resistance. In other stress conditions, such interaction cannot be ruled out. Another role could be in enzyme maturation. CueO is a substrate of the twin-arginine translocation (TAT) pathway, a machinery known to be involved in the translocation of folded proteins. There are differing opinions on whether copper centers are inserted into CueO during the folding in the cytoplasm (38) or after the translocation in the periplasm (39). In any scenario, this domain, acting as a facilitator in Cu⁺ recruitment, could allow apo-CueO to become active in the conditions of low copper availability and of competition with other biological chelators. If this process takes place in the cytoplasm, CueO with bound coppers would additionally help to export it from the cytoplasm to the periplasm. If it takes place in the periplasm, CueO would be able to mature faster and start oxidizing remaining Cu⁺ to avoid reactive oxygen species production.

In summary, our multimodal and multiscale approach provides clear insights into the mechanism of CueO and functional role of the Met-rich domain and its copper-binding sites. Our results demonstrate that the well-defined Cu6 and Cu7 binding sites observed in the crystallographic structure do not contribute directly to the cuprous oxidase activity. On the other hand, we confirm the pivotal role of the Cu5 allowing to assert it as a sole substrate-binding active site for cuprous oxidase activity of CueO. A double mutation of this site is necessary to prevent copper coordination and to abolish cuprous oxidase activity completely. We also demonstrate that the whole Met-rich domain is not required for cuprous oxidase activity in vitro but it acts as a facilitator of Cu⁺ recruitment from the strongly copper chelated forms. Accordingly, the methionines of the Met-rich domain collectively appear to play a role in the cuprous oxidation when Cu⁺ is strongly complexed, presumably by forming a transient complex. Although the functional implication of this facilitator role for the bacteria must still be considered, a first answer provided by our in vivo test shows that CueO lacking the Met-rich domain can still successfully detoxify the cell from copper. We speculate that the high affinity of the Met-rich domain for Cu⁺ might also have another function, such as in CueO maturation under conditions of low copper availability. This role may explain the presence and structural diversity of Met-rich domains notably present in many copper homeostasis actors.

Materials and Methods

Plasmid Construction.

The plasmids used in this study are listed in SI Appendix, Table S1. The TAT signal sequence vector was constructed by amplifying the TAT sequence of cueO gene from the chromosome (MG1655) using primers AV179 and AV180. The PCR product was cloned into PJF119EH using EcoRI and SacI restriction sites, generating plasmid pAV94 (pJF119EH-TAT). The CueO, Isopropyl β-D-thiogalactopyranoside (IPTG) induced, expression vector was constructed by amplifying the cueO gene (without its signal sequence) from the chromosome (MG1655) using primers AV184 and AV185, which resulted in the fusion of a Strep-tag II coding sequence at the 3′ end. The PCR product was cloned into pAV94 using NotI and SacI restriction sites, generating plasmid pAV97 (pJF119EH-TAT-CueO E. coli).

cueO Directed Mutagenesis.

First, 50 μL PCRs were performed using Q5 Hot start High-Fidelity DNA polymerase (New England Biolabs), pAV97 (pJF119EH-TAT-CueO E. coli) as the template. Then, suitable primers were used to introduce mutations (SI Appendix, Table S2). ΔCu5^M441Q and ΔCu5^D439A,M441Q variants also present the M₄₄₀Q mutation, which is omitted in the name for simplicity, to avoid possible replacement of the mutated methionine by the adjacent one in the Cu5 coordination sphere. The resulting PCR products were digested using DpnI, purified using the GeneJET PCR purification kit (Thermo Fisher) and transformed into E. coli DH5α. Three colonies were randomly selected from each transformation, and the plasmids were isolated using the GeneJET Plasmid Miniprep kit (Thermo Fisher). DNA sequencing was carried out to assess the fidelity of the mutagenesis reaction.

Expression and Purification of EcCueOs.

E. coli cells containing the desired plasmid were grown in LB Broth (Miller’s) with 100 µg mL⁻¹ ampicillin at 37 °C up to OD_{600 nm} 0.6 and 0.1 mM IPTG was then added to induce gene expression. Cultures were supplemented with 0.5 mM CuSO₄ and then grown for 24 h at 25 °C up to OD_{600 nm} 5-6. Cells were harvested by centrifugation at 5,000 rpm, washed with 50 mM phosphate-buffered saline (PBS) at pH 7.4. Bacteria pellets were resuspended at 4 °C in 50 mM PBS at pH 7.4 containing 1 mM phenylmethylsulfonyl fluoride (PMSF), 10 µg mL⁻¹ DNase I, and 2 mM MgSO₄. Cell lysis was performed using French Press (Thermo Scientific) at 4 °C, 1,000 psi two times. The cell debris was removed by ultracentrifugation at 40 krpm for 45 min at 4 °C. The cleared lysate was passed through a filter, cutoff 0.45 µm, prior to be loaded onto Strep-Tactin®XT 4Flow® (IBA Lifesciences) 5 mL column, nonadsorbed proteins were washed with 50 mM PBS at pH 7.4 and 50 mM PBS, NaCl 150 mM at pH 7.4 and EcCueO eluted with 50 mM PBS, biotin 5 mM at pH 7.4. Eluted CueO was concentrated up to 100 µM and desalted with 40 mM MOPS buffer at pH 7.0, using a Sephadex® G-25 (Cytiva) 5 mL column. An in vitro copper incorporation step was carried out for all EcCueO enzymatic preparations by dialyzing the protein overnight at 4 °C in 40 mM MOPS buffer at pH 7.0 with addition of 5 equivalents of CuSO₄. The excess of copper was removed by desalting with 40 mM MOPS buffer at pH 7.0, using a Sephadex® G-25 (Cytiva) 5 mL column. Pure EcCueOs preparations were aliquoted, frozen in liquid nitrogen, and stored at −80 °C.

Cuprous Oxidase Activity of EcCueOs in Solution.

Air-stable substrate [CuI(BCA)₂]³⁻ was used for measuring the cuprous oxidase activity of EcCueOs. 800 µM [Cu^I(BCA)₂]³⁻ stock solution was prepared according to ref. 12, by dissolving [Cu^I(CH₃CN)₄]PF₆ in 50 mM BisTris buffer, 2.4 mM bicinchoninic acid disodium salt hydrate (BCA) at pH 7.0 under anaerobic conditions. The final ratio was 1/2.5 Cu⁺/BCA to avoid competition of free BCA for Cu⁺ coordination. The BCA complexation of Cu⁺, under these experimental conditions takes about 30 min to reach the final concentration, calculated from ε_562nm = 7,900 M⁻¹ cm⁻¹ (12). [Cu^I(BCA)₂]³⁻ concentration was checked prior to measuring EcCueO activity. Enzymatic assays were performed under the air at 30 °C by measuring the absorption decrease at 562 nm using a spectrophotometer microplate reader (Spark 10 M, Tecan, Swiss). To avoid any product inhibition during kinetic measurements, the cuprous oxidase activity was measured in 50 mM BisTris buffer at pH 7.0, for its ability to coordinate the excess of Cu²⁺ (12). Values reported in SI Appendix, Table S6 come from at least 6 measures for each concentration of substrate, normalized for the substrate oxidation in absence of enzyme. EcCueO concentration in well was 0.1 ± 0.025 µM for all tested mutants. The reaction buffer was composed of 50 mM BisTris at pH 7.0 with increasing concentrations of [Cu^I(BCA)₂]³⁻ from 0 to 400 µM. [Cu^I(BCS)₂]³⁻ preparation and assays referred to SI Appendix, Fig. S7 and Table S4 were performed as above, except the wavelength of absorption (483 nm) and the EcCueO WT concentration in well was 1 ± 0.01 µM. [Cu^I(Ferene)₂]³⁻ preparation and assays referred to SI Appendix, Fig. S7 and Table S5 were performed as above, except the wavelength of absorption (484 nm), the Cu⁺/Ferene ratio of 1/5 and the EcCueOs concentration in well was 10 ± 5 nM.

[Cu^I(ligand)₂]³⁻ kinetic parameters of EcCueOs were calculated by OriginPro 2016 (OriginLab Corporation, Massachusetts, USA), using the first 250, 300, and 120 s to make a linear fit and estimate the reaction rate, respectively for BCA, BCS, and Ferene as ligand. The resulting reaction rates were fitted with the Michaelis–Menten equation with one site saturation: V_i = (V_max x [S]₀)/(K_M + [S]₀). Where V_i = initial speed, V_max = maximal initial speed, K_M = Michaelis constant, and [S]₀ = [Cu^I(ligand)₂]^3- concentration.

ABTS Oxidase Activity of EcCueOs in Solution.

2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonique) (ABTS, ε_{420 nm} = 36,000 M⁻¹ cm⁻¹) was used as an electron donor for measuring the activity of EcCueOs. Enzymatic assays were performed as above, except the wavelength of absorption (420 nm) and the reaction buffer was 100 mM Acetate buffer at pH 5.0 with increasing concentrations of ABTS from 0 to 60 mM. ABTS kinetic parameters of EcCueOs were calculated by OriginPro 2016 (OriginLab Corporation, Massachusetts, USA), using the first 250 s to make a linear fit and estimate the reaction rate. The resulting reaction rates were fitted with the Michaelis–Menten equation with one site saturation: V_i = (V_max × [S]₀)/(K_M + [S]₀). Where V_i = initial speed, V_max = maximal initial speed, K_M = Michaelis constant and [S]₀ = ABTS concentration.

Preparation of the EcCueO Functionalized COOH-CNT-Based Working Electrodes.

EcCueO bioelectrodes were prepared as previously reported (18). Planar glassy carbon electrodes (GCs; diameter: 3 mm) were polished with an alumina slurry, sonicated for 10 min first in acetone/ethanol 1:1 and then in MilliQ water, and finally rinsed with MilliQ water. Then, COOH-CNT slurry dispersed in water was drop cast onto the surface of GCs and dried at 60 °C. A 20 μL aliquot of 15 μM EcCueO preparation was drop cast on the CNT-based electrode, and incubated under the air for 30 min at 4 °C. The enzyme-modified electrode was washed with 40 mM MOPS buffer at pH 7.0 and was stored in the same buffer under air at 4 °C when not used.

Electrochemical Measurements.

Electrochemical experiments (cyclic voltammetry and chronoamperometry) were carried out in a three-electrode electrochemical cell using Autolab PGSTAT30 potentiostat, controlled by Nova software, and a rotating electrode instrument (Metrohm Autolab, Switzerland). The rotation speed (ω) of the working electrode was set to 3,000 rpm. The CNT-based bioelectrodes were used as working electrodes. Pt wire was used as the counterelectrode and the Hg|Hg₂SO₄|sat. K₂SO₄ served as reference electrode. Potentials are referred to the normal hydrogen electrode according to E_NHE = E_MSE + 0.640 V. All current densities are normalized toward the geometrical surface of the glassy carbon electrode (0.071 cm²). The experiments were conducted at 30 °C and under controlled oxygen- or nitrogen-saturated atmosphere, by continuously bubbling either O₂ or N₂ gas in the electrolyte solution. The supporting electrolyte was 100 mM Acetate buffer at pH 5.0 for all experiments. Additions of CuSO₄ to the electrolyte solution were performed every 100 s for chronoamperometry experiments and at an applied potential E = 0.74 V vs. NHE for cyclic voltammetry experiments. For each EcCueO preparation at least four bioelectrodes were used for each electrochemical measurement.

Copper Survival Assays.

The ΔcusB ΔcueO MG1655 cells carrying CueO expressing plasmids were grown aerobically at 37 °C under agitation in 5 mL of M9 minimal medium with Ampicillin (50 µg/mL) and IPTG (100 µM). When cultures reached OD_600nm ≈ 0.1, cells were harvested and diluted in phosphate-buffered saline (PBS): 5 µL of 10-time serial dilutions were spotted onto M9 minimal medium-agar plates with Ampicillin (50 µg/mL) supplemented or not with CuSO₄ (70 and 100 µM). Plates were incubated at 37 °C for 3 d.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.