The copper supply pathway to a Salmonella Cu,Zn-superoxide dismutase (SodCII) involves P1B-type ATPase copper efflux and periplasmic CueP


For correspondence. E-mail; Tel. (+44) 161 275 1543; Fax (+44) 161 275 5082.


Periplasmic Cu,Zn-superoxide dismutases (Cu,Zn-SODs) are implicated in bacterial virulence. It has been proposed that some bacterial P1B-type ATPases supply copper to periplasmic cupro-proteins and such transporters have also been implicated in virulence. Here we show that either of two P1B-type ATPases, CopA or GolT, is needed to activate a periplasmic Cu,Zn-SOD (SodCII) in Salmonella enterica serovar Typhimurium. A ΔcopAgolT mutant accumulates inactive Zn-SodCII which can be activated by copper-supplementation in vitro. In contrast, either single ATPase mutant accumulates fully active Cu,Zn-SodCII. A contribution of GolT to copper handling is consistent with its copper-responsive transcription mediated by DNA-binding metal-responsive activator GolS. The requirement for duplicate transcriptional activators CueR and GolS remains unclear since both have similar tight KCu. Mutants lacking periplasmic cupro-protein CueP also accumulate inactive Zn-SodCII and while CopA and GolT show functional redundancy, both require CueP to activate SodCII in vivo. Zn-SodCII is also activated in vitro by incubation with Cu-CueP and this coincides with copper transfer as monitored by electron paramagnetic resonance spectroscopy. These experiments establish a role for CueP within the copper supply pathway for Salmonella Cu,Zn-SodCII. Copper binding by CueP in this pathogen may confer protection of the periplasm from copper-mediated damage while sustaining vital cupro-enzyme activity.


To cope with both the essentiality and toxicity of copper, organisms have evolved elaborate copper homeostatic mechanisms that act to tightly control cellular copper pools (O'Halloran and Culotta, 2000; Huffman and O'Halloran, 2001; Robinson and Winge, 2010). These include copper-responsive transcriptional regulators, membrane transporters and copper chaperones. The latter assist copper in reaching vital destinations in cells while inhibiting the deleterious side reactions of copper en route (Tottey et al., 2012). So far, three major eukaryotic copper-supply routes have been described: via the copper chaperone Atx1 to a P1B-type ATPase for copper transport into the trans-Golgi, via CCS to copper,zinc-superoxide dismutase (Cu,Zn-SOD), or via Cox17, Sco1 and Cox11 to cytochrome c oxidase in mitochondria (Robinson and Winge, 2010). Copper chaperones have also been identified in some bacteria that are components of copper-resistance determinants (Cobine et al., 1999; Radford et al., 2003; Bagai et al., 2008; Gonzalez-Guerrero and Arguello, 2008; Tottey et al., 2012) and/or linked to cupro-protein assembly; including for plastocyanin inside cyanobacterial thylakoids (Tottey et al., 2012) or for cytochrome c oxidase assembly (Winge, 2003; Thompson et al., 2012). Here we present evidence for a bacterial copper-supply pathway for periplasmic Cu,Zn-SOD activity that involves a novel periplasmic copper-binding protein.

Salmonella enterica serovar Typhimurium (S. Typhimurium) is an important human pathogen as a major contributor to the 93.8 million annual cases of food-associated Salmonella gastrointestinal disease worldwide (Coburn et al., 2007; Majowicz et al., 2010). Its virulence during systemic disease is associated with the ability to survive and replicate in macrophage phagosomes (Fields et al., 1986). Elevated copper is a feature of this bactericidal compartment and the ability of S. Typhimurium to export copper via either one of two related P1B-type ATPases, CopA and GolT, aids survival in this environment (Osman et al., 2010). Mutants of S. Typhimurium lacking CopA and GolT are extremely copper-sensitive and hyper-accumulate copper, relative to wild-type or either single mutant demonstrating functional redundancy with respect to copper export and resistance (Osman et al., 2010). The S. Typhimurium Cue system consists of the copper-sensing MerR-family transcriptional regulator CueR that mediates copper-responsive expression of CopA, the multi-copper oxidase CueO, and the periplasmic copper-binding protein CueP (Espariz et al., 2007; Pontel and Soncini, 2009; Achard et al., 2010; Osman et al., 2010). The Gol system consists of a second CueR-like sensor GolS which has primarily been associated with gold-sensing (Checa et al., 2007; Pontel et al., 2007; Perez Audero et al., 2010), GolT, and an Atx1/CopZ copper chaperone-like protein GolB (Checa et al., 2007; Pontel et al., 2007; Osman et al., 2010; Perez Audero et al., 2010).

There is growing evidence that a subgroup of P1B-type ATPases assist the assembly of bacterial cupro-proteins (Tottey et al., 2001; Gonzalez-Guerrero et al., 2010; Waldron et al., 2010; Raimunda et al., 2011). These include the FixI/CopA2-like ATPases that are required for aa3-type cytochrome c oxidase activity and that have also been associated with bacterial virulence (Gonzalez-Guerrero et al., 2010; Raimunda et al., 2011). The cyanobacterial ATPases PacS and CtaA are additionally required for plastocyanin-mediated electron transport and copper loading of periplasmic CucA (Tottey et al., 2001; Waldron et al., 2010). While these copper targets are absent from S. Typhimurium (McClelland et al., 2001), virulent strains possess two periplasmic Cu,Zn-SODs, SodCI and SodCII, encoded by prophage (Gifsy-2) and chromosomal genes respectively (Fang et al., 1999). Both SodCI and SodCII are expressed during infection, although a contribution to virulence by detoxifying phagocytic superoxide is only detectable for SodCI under normal conditions (De Groote et al., 1997; Krishnakumar et al., 2007; Kim et al., 2010; Rushing and Slauch, 2011). Both proteins are exported into the periplasm via the general secretory pathway (Sec) for unfolded proteins, and hence anticipated to acquire their metal cofactors within this compartment.

In this study we have further examined the roles of CopA and GolT in S. Typhimurium with respect to cupro-protein assembly and established that copper transport to the periplasm by either protein is needed for SodCII activity. This is the first evidence of a copper-trafficking pathway for the activation of a periplasmic Cu,Zn-SOD. We confirm that GolT, in addition to CopA, is expressed in response to copper. GolS binds copper with a similar affinity to CueR consistent with a role in sensing cytosolic copper to trigger copper efflux via GolT. We demonstrate that activation of SodCII in the S. Typhimurium periplasm is dependent upon the periplasmic copper-binding protein CueP. There is copper transfer between CueP and SodCII in vitro implicating CueP in the delivery of copper.


Either CopA or GolT is required for active SodCII

A role for some bacterial copper-transporting P1B-type ATPases in periplasmic cupro-enzyme assembly has been reported (Tottey et al., 2001; Gonzalez-Guerrero et al., 2010; Waldron et al., 2010; Raimunda et al., 2011). S. Typhimurium strain SL1344 possesses two periplasmic SodC proteins, SodCI and SodCII (Fang et al., 1999), known to require copper for activity (Ammendola et al., 2008), and thus representing potential targets for activation by P1B-type ATPase-mediated copper export. To check for SodCI and SodCII associated activity in the S. Typhimurium periplasm, SOD activity was monitored in periplasmic extracts from wild-type S. Typhimurium and mutants lacking sodCI or sodCII. Periplasm contents were liberated by cold osmotic shock and resolved by anion exchange chromatography followed by isoelectric focusing gel electrophoresis prior to SOD activity assays based upon inhibition of nitro blue tetrazolium reduction. A band of SOD activity was detected in periplasmic extracts from wild-type cells, which was absent from ΔsodCII but not ΔsodCI (Fig. 1A), and thus attributed to active SodCII. No band corresponding to SodCI activity was detected, consistent with SodCI not being released by osmotic shock due to it being tethered to the membrane (Krishnakumar et al., 2004; 2007).

Figure 1.

Mutants lacking CopA and GolT have decreased SodCII activity.

A. Isoelectric focusing gels of periplasmic proteins resolved by anion exchange chromatography from 2 l of cultures (eluted in 100 mM NaCl) of wild-type S. Typhimurium SL1344, ΔsodCII or ΔsodCI stained for SOD activity.

B and C. Isoelectric focusing gels of periplasmic proteins resolved by anion exchange chromatography from 2 l of cultures (eluted in 100 mM NaCl) of SL1344 (B, lane 1) and ΔcopAgolT (B, lane 2), or from 0.5 l of cultures (eluted in 200 mM NaCl) of strain SodCIIUP (C, lane 1) and SodCIIUP ΔcopAgolT (C, lane 2), stained for protein (upper panel) and SOD activity (lower panel). Arrow indicates overexpressed SodCII.

D and E. SOD activity (D) in fractions 22–30 (0.5 ml) obtained following size-exclusion chromatography of proteins eluted by 200 mM NaCl during anion exchange chromatography of periplasmic proteins from SodCIIUP (open circles), SodCIIUP ΔcopA (open squares), SodCIIUP ΔgolT (open triangles) and SodCIIUP ΔcopAgolT (open diamonds), and SodCII abundance visualized by SDS-PAGE gels stained with Instant Blue (E).

Having confirmed SodCII enzymatic activity in the S. Typhimurium periplasm, the role of the copper-transporting P1B-type ATPases CopA and GolT in SodCII activation was examined. Comparison of SodCII activity in periplasmic extracts from wild-type S. Typhimurium versus a copper-export mutant, lacking both copA and golT, indicated decreased activity in the latter (Fig. 1B). However, due to the low concentrations of SodCII in these cells, activity levels could not be accurately quantified. Hence, to further examine the roles of CopA and GolT in SodCII activation, SodCII was overexpressed from a plasmid in strain SodCIIUP (Fig. 1C). As anticipated, SodCII activity levels were substantially increased in periplasmic extracts from SodCIIUP compared with wild-type S. Typhimurium (compare lane 1 in Fig. 1B and C which show activity in periplasmic extracts from 2 l and 0.5 l of cultures respectively). Notably, SodCII activity levels remained low in periplasmic extracts from SodCIIUP cells lacking copA and golT, despite increased SodCII protein (Fig. 1C). These data are therefore consistent with a requirement for copper export to the periplasm for SodCII activation implying that copper is trafficked to apo-SodCII. To further quantify SodCII activity levels in these cells, liberated periplasmic proteins were resolved by native two-dimensional liquid chromatography (anion exchange followed by size exclusion) and eluant fractions analysed for SOD activity via inhibition of pyrogallol autooxidation. A peak of SOD activity was detected for strain SodCIIUP, coinciding with SodCII elution, and which was absent from a control strain lacking SodCII expression (Fig. S1). Crucially, similar SodCII activity levels were detected in periplasmic fractions from SodCIIUP cells lacking either copA or golT but again (as in Fig. 1C) there was substantially decreased activity in equivalent fractions from a mutant deficient in both copA and golT (Fig. 1D), despite similar protein abundance (Fig. 1E). These data therefore demonstrate a requirement for copper transport to the periplasm by either CopA or GolT for periplasmic SodCII activity.

SodCII is deficient in copper but not zinc in cells lacking CopA and GolT

To determine whether or not the loss of SodCII activity in the copper transport mutants was due to the inability of SodCII to acquire copper in these cells, the metal contents of SodCII containing periplasmic fractions from strain SodCIIUP and the ΔcopAgolT derivative were determined by inductively coupled plasma mass spectrometry (ICP-MS). Fractions from strain SodCIIUP contained equimolar protein-bound copper and zinc (Fig. 2A), consistent with the presence of active Cu,Zn-SodCII (Fig. 1D). It is noteworthy that the periplasmic copper-binding protein CueP co-elutes in these fractions (Osman et al., 2010) but its contribution to the copper content of these SodCII enriched fractions (Fig. 1), due to SodCII overexpression, must be negligible. In contrast, while a similar concentration of protein-bound zinc was present in the corresponding fractions from SodCIIUP ΔcopAgolT, there was a lack of copper (Fig. 2B). ΔcopAgolT cells have increased CueP because hyper-accumulated copper activates CueR (Pontel and Soncini, 2009; Osman et al., 2010), but again in the SodCIIUP background this appears to make a negligible contribution to the metal content of these fractions relative to Zn-SodCII. Copper was also retained in SodCII containing fractions from SodCIIUP ΔcopA and SodCIIUP ΔgolT (∼ 3 μM and ∼ 5 μM copper, respectively, compared with 0.3 μM for SodCIIUP ΔcopAgolT). Supplementation of pooled SodCII containing fractions from SodCIIUP ΔcopAgolT with approximately twofold excess of copper resulted in restoration of SodCII activity to levels observed for the same amount of SodCII isolated from strain SodCIIUP (Figs 2C and S2). Addition of copper to the latter did not increase activity levels consistent with the isolation of fully metallated enzyme from SodCIIUP cells. These data confirm that the low level of SodCII activity in the copper transport mutant corresponds to accumulation of inactive Zn-SodCII resulting from its inability to acquire copper in these cells.

Figure 2.

The low level of SodCII activity in SodCIIUP ΔcopAgolT corresponds with decreased bound copper.

A and B. [Copper] (filled circles) and [zinc] (filled triangles) in fractions 22–30 (0.5 ml) following size-exclusion chromatography of proteins eluted by 200 mM NaCl during anion exchange chromatography of periplasmic proteins from strain SodCIIUP (A) and SodCIIUP ΔcopAgolT (B).

C. SodCII activity in combined SodCII-containing fractions from SodCIIUP (black bars) and SodCIIUP ΔcopAgolT (open bars) adjusted to 2 μM protein following incubation with buffer or CuSO4 at 20°C for 1 h and three rounds of dialysis at 4°C to remove unbound metal.

GolS mediates copper-responsive expression of GolT

The identification of a role for GolT in copper loading of SodCII is somewhat unexpected since this transporter has largely been associated with gold detoxification (Checa et al., 2007; Pontel et al., 2007; Perez Audero et al., 2010). CueR and GolS are proposed to sense copper and gold, respectively, in S. Typhimurium (Checa et al., 2007; Perez Audero et al., 2010) with the selective recognition of their target promoters resulting in the copper and gold responsiveness of CopA and GolT respectively (Perez Audero et al., 2010). Hence, we have re-examined expression from the golTS promoter in response to a range of copper concentrations, up to inhibitory levels. Elevated β-galactosidase activity was detected in response to increasing concentrations of copper (Fig. 3A). This copper responsiveness was completely abolished in a ΔcueRgolS double mutant but retained in ΔcueR and ΔgolS single mutants (Fig. 3B), and could be restored to ΔcueRgolS by providing cueR or golS in trans (Fig. S3). Both CueR and GolS can therefore activate expression from the golTS promoter in response to copper (at least in the absence of each other). In wild-type bacteria, golTS regulation by CueR is, however, thought to be prevented due to the preferential binding of GolS to the golTS promoter (Perez Audero et al., 2010). Nonetheless, the retention of golTS copper responsiveness in ΔcueR demonstrates that GolS is able to detect copper within the S. Typhimurium cytosol.

Figure 3.

GolS-mediated expression of PgolTS can be induced by copper.

A. β-Galactosidase activity measured in S. Typhimurium containing PgolTS fused to lacZ following growth (150 min) in LB medium with no metal supplement and up to inhibitory concentrations of CuSO4.

B. Expression from PgolTS in S. Typhimurium (WT), ΔcueR, ΔgolS or ΔcueRgolS following growth in LB medium with no metal supplement or maximum permissive concentrations of CuSO4.

GolS and CueR have similar tight KCu

Why does S. Typhimurium possess duplicate CueR-like copper sensors? To explore the ability of GolS and CueR to sense copper in S. Typhimurium further, their affinities for copper were examined in vitro. Competition for copper between protein ligands and bathocuproine disulphonate (BCS) can be used to measure protein copper binding affinities (Xiao et al., 2004). Apo-CueR or apo-GolS were added to a Cu+-BCS mix (using equimolar CueR or GolS and BCS) and the formation of Cu+-CueR and Cu+-GolS measured by loss of absorbance. This established that at equilibrium Cu+ fully partitions from BCS to CueR and GolS (Fig. 4A and B), placing the Cu+ affinities of both sensors at least an order of magnitude greater than BCS. CueR was able to withhold copper from a 10-fold excess of BCS, whereas at 50-fold and 100-fold excess of BCS Cu+ partitioned between CueR and chelator (Fig. S4). Equal partitioning between molecules at 100-fold excess of BCS implies KCu of CueR ∼ 100-fold tighter than BCS (β2 = 1019.8 M−2). In addition, titration of Cu+ into a 10:1 mix of BCS and CueR caused no increase in absorbance until ∼ 0.5 Cu+ per CueR monomer was added (Fig. S5). This suggests that CueR can bind 1 Cu+ per homodimer with very high affinity (at least two orders of magnitude greater than BCS). Some precipitation of GolS was observed at ≥ 10-fold excess BCS preventing an estimation of an equilibration concentration for Cu+-GolS.

Figure 4.

CueR and GolS have similar tight affinities for copper.

A and B. Absorbance spectra of apo-subtracted Cu-BCS (10 μM BCS; 4 μM Cu+) before (solid lines) and after (broken lines) incubation to equilibrium with purified CueR (10 μM, A) or GolS (10 μM, B).

C. GolS (2.5 μM) equilibrated with Cu+ (1 μM) (left panel) and after subsequent addition of apo-CueR (2.5 μM, right panel).

D. CueR (2.5 μM) equilibrated with Cu+ (1 μM) (left panel) and after subsequent addition of apo-GolS (2.5 μM, right panel).

Proteins were separated by cation exchange chromatography and fractions (1 ml) assayed for copper by ICP-MS and protein by SDS-PAGE. CueR was present in the flow-through (fractions 5–14) whereas GolS eluted in 1 M NaCl (fractions 17–18).

Hence, in order to establish which protein has a higher affinity for Cu+, competitive Cu+ binding between CueR and GolS was examined. Apo-CueR was added to Cu+-bound GolS (Fig. 4C) and in a reciprocal experiment apo-GolS was added to Cu+ bound CueR (Fig. 4D), using 0.8 Cu+ per homodimer. Metal transfer was monitored by ICP-MS after separation of the proteins using a cation exchange column. CueR does not bind to the column and is eluted in the flow-through, whereas GolS binds the column and is eluted in 1 M NaCl. In both reactions, Cu+ partitioned between CueR and GolS (Fig. 4C and D) establishing that equilibrium had been reached and that both proteins have similar affinities for Cu+. This argues against a model in which one sensor has a weaker affinity and solely triggers export of surplus copper while the other is adapted to activate export for cofactoring proteins in the periplasm. However, our results (Fig. S5) suggest that CueR binds only one atom of copper per homodimer with very tight affinity (far tighter than BCS) and hence this would be consistent with the two metal binding sites within the homodimer possessing step-wise affinities. This is the case in other metal-sensors, such as the Staphylococcus aureus zinc-sensor CzrA which binds two equivalents of Zn2+ per homodimer with negative cooperativity (Pennella et al., 2006). It is not known whether Cu+ occupancy of one or both metal binding sites is needed for CueR or GolS activation, and it is possible that this differs between the proteins. Thus, it remains formally possible that there are dissimilarities in the set points for the expression of the cue and gol systems in elevated copper. Differing interactions of the sensors with a cytosolic copper donor (such as GolB) and differences in the intracellular concentrations of the two sensors may also be factors.

Periplasmic CueP is also required for active copper-loaded SodCII

We previously identified Cu-CueP as a major periplasmic metal complex in S. Typhimurium (Osman et al., 2010). CueP is under the control of CueR and can also contribute to copper resistance (Pontel and Soncini, 2009). To establish whether or not CueP could play a role in the copper loading of SodCII we examined the SOD activity of SodCII containing periplasmic protein fractions from a cueP-deficient mutant (SodCIIUP ΔcueP). The absence of cueP led to accumulation of inactive SodCII (Fig. 5A and B), and this loss of activity was unaffected by the absence of copA or golT individually (Fig. 5C and D). This demonstrates that while the copper exporters CopA and GolT show functional redundancy with respect to supplying copper for SodCII, both require CueP to activate SodCII in vivo. Measurement of the metal content of the SodCII containing fractions from the cueP mutant revealed the presence of zinc but a lack of copper (Fig. 5E) and supplementation of pooled fractions from strain SodCIIUP ΔcueP with approximately twofold excess of copper restored SOD activity to levels observed for the same amount of SodCII isolated from strain SodCIIUP (Figs 5F and S2). CueP is therefore needed for SodCII to acquire copper within the S. Typhimurium periplasm.

Figure 5.

CueP is required for copper loading of SodCII.

A–D. SOD activity (A and C) in fractions 22–30 (0.5 ml) following size-exclusion chromatography of proteins eluted by 200 mM NaCl during anion exchange chromatography of periplasmic proteins from strain SodCIIUP (open circles), SodCIIUP ΔcueP (open diamonds), SodCIIUP ΔcuePcopA (open triangles), SodCIIUP ΔcuePgolT (open squares), with the abundance of SodCII in the fractions visualized on SDS-PAGE gels stained with Instant Blue (B and D).

E. [Copper] (filled circles) and [zinc] (filled triangles) in fractions 22–30 (0.5 ml) following size-exclusion chromatography of proteins eluted by 200 mM NaCl during anion exchange of periplasmic proteins from SodCIIUP ΔcueP.

F. SOD activity in combined SodCII-containing fractions from SodCIIUP ΔcueP adjusted to 2 μM protein following incubation with buffer or CuSO4 at 20°C for 1 h and three rounds of dialysis at 4°C to remove unbound metal.

G. SOD activity in combined SodCII-containing fractions (2 μM protein) from SodCIIUP, SodCIIUP ΔcopAgolT and SodCIIUP ΔcueP following incubation with equimolar Cu-CueP at 25°C for 30 min in the absence (black bars) or presence (grey bars) of 50 μM reduced glutathione. Final bar; SOD activity of Cu-CueP.

Cu-CueP activates Zn-SodCII in vitro

To establish whether or not CueP could play a direct role in SodCII activation by supplying copper, we next tested its ability to activate SodCII in vitro. Cu-CueP was isolated, following its overexpression in Escherichia coli grown in copper-supplemented media, and confirmed to contain ∼ 1 molar equivalent of copper by ICP-MS (10.31 μM copper and 0.97 μM zinc per 10 μM CueP). Cu-CueP lacks SOD activity and its addition to active Cu,Zn-SodCII, isolated from strain SodCIIUP, did not affect measured SOD activity (Fig. 5G). However, addition of Cu-CueP to inactive Zn-SodCII from SodCIIUP ΔcopAgolT or SodCIIUP ΔcueP caused a substantial increase in SOD activity (Fig. 5G), demonstrating that Cu-CueP can activate SodCII in vitro. To determine whether copper competitors can interfere with the activation of SodCII by Cu-CueP we also examined the ability of Cu-CueP to activate Zn-SodCII in the presence of reduced glutathione. Glutathione plays a role in both cytosolic and periplasmic redox homeostasis (Pittman et al., 2005), and has been used previously as a low-affinity copper competitor in assays for direct copper transfer between the copper chaperone Atx1 and partner protein Ccc2 (Huffman and O'Halloran, 2000). The addition of a 25-fold excess of reduced glutathione did not affect the ability of Cu-CueP to activate Zn-SodCII (Fig. 5G).

Cu-CueP can transfer copper to SodCII in vitro

If CueP supplies copper for SodCII activity in vivo it is proposed that there could be detectable transfer of copper from CueP to SodCII in vitro. The co-elution of SodCII and CueP during anion exchange and size-exclusion chromatography prevented their separation, following co-incubation, to allow determination of their metal status. To observe copper transfer between CueP and SodCII we therefore employed electron paramagnetic resonance (EPR) spectroscopy to monitor the Cu2+ environment in these proteins. For these assays, S. Typhimurium SodCII was purified following overexpression in E. coli (in zinc-supplemented media). The resulting protein (Zn-SodCII) exhibited low SOD activity levels which could be substantially increased following incubation (30 min) with equimolar copper or Cu-CueP (Fig. 6A). Comparison of the EPR spectra (at 20 K) for apo-CueP, Cu-CueP, Zn-SodCII and the copper activated Zn-SodCII showed that only the latter contains appreciable quantities of Cu2+ (Fig. 6B and S6), suggesting that Cu-CueP primarily binds and stabilizes Cu+. Incubation of inactive Zn-SodCII with Cu-CueP for 30 min caused the appearance of a Cu2+-spectrum with A|| = 140 G (Fig. 6C), and characteristic of a specific Cu2+ species present in Cu,Zn-SodCII pre-activated by incubation with copper (Fig. 6B). Spectra with very similar characteristics have been reported for other Cu,Zn-SODs isolated from prokaryotes (Stroppolo et al., 1998; Pesce et al., 2000). This demonstrates transfer of copper, presumably as Cu+, to Zn-SodCII where it is oxidized to give a SodCII associated Cu2+ EPR spectrum. Prolonged incubation of SodCII with CueP resulted in a more intense EPR spectrum, suggesting time-dependent Cu2+ occupancy of the SodCII active site. The lack of a signal for Cu(H2O)62+ species, A|| = 160 G (Fig. S6, ∼ 1 μM Cu2+ detection limit), implies that the metal is transferred directly between CueP and SodCII rather than via a dissociative process. The spectra arising from freshly frozen samples of Cu-CueP with Zn-SodCII incubated over long time periods show the appearance of a second Cu2+ species, A|| = 150 G (Fig. S7). The latter possesses similar properties to a low intensity Cu2+ spectrum detected for samples containing Cu-CueP (Figs S6 and S7) and hence may relate to equilibration of Cu2+ between CueP and SodCII. However, the presence of this (or a closely related) Cu2+-species in samples containing Cu,Zn-SodCII pre-activated by incubation with copper (Fig. 6B), rather than Cu-CueP, shows that it can also form in the absence of CueP. It is possible that (at least for the latter) the A|| = 150 G species therefore represents a Cu2+ intermediate formed during Cu2+ incorporation, rather than CueP-mediated Cu+ incorporation, into Zn-SodCII. Crucially, these experiments demonstrate facile transfer of copper to Zn-SodCII from Cu-CueP.

Figure 6.

CueP transfers copper to SodCII in vitro.

A. SOD activity in samples containing 2 μM Zn-SodCII isolated from E. coli following incubation with or without equimolar CuSO4 or Cu-CueP (at 25°C for 30 min), with the final bar showing SOD activity of Cu-CueP.

B. X-band EPR spectra as indicated. The Cu,Zn-SodCII spectrum shows a minor second species with a narrower line width and A|| of 150 G as indicated by the asterisk. These spectra are the sums of two scans.

C. X-band EPR spectra following incubation of Cu-CueP with Zn-SodCII for the shown time periods. Each spectrum is the sum of eight scans. Protein concentrations in all spectra are 50 μM in 10 mM HEPES, 100 mM NaCl pH 7.8. Experimental conditions: microwave power 0.5 mW, modulation frequency 100 KHz, modulation amplitude 5 G, temperature 20 K.


We have demonstrated that the two copper-transporting P1B-type ATPases, CopA and GolT, in S. Typhimurium have a role in activating a periplasmic Cu,Zn-SOD. The absence of both of these transporters results in the accumulation of inactive SodCII, whereas SodCII activity is retained in mutants possessing either one of the transporters (Fig. 1). The loss of SodCII activity in a copA and golT double mutant is due to its inability to acquire copper, but not zinc, in these cells (Fig. 2B), and activity can be fully restored by addition of copper to isolated SodCII in vitro (Fig. 2C). Consistent with golT, in addition to copA, having a role in copper homeostasis, we show that golT expression is activated by copper (Fig. 3A) and this can be mediated by GolS (Fig. 3B). The affinities of GolS and CueR for copper are closely matched (Fig. 4), which is also in harmony with the overlapping functions of their target genes. We demonstrate a role for the periplasmic copper-binding protein CueP in the delivery of copper to SodCII. CueP is required for SodCII activity in vivo (Fig. 5A) and inactive Zn-SodCII accumulates in its absence (Fig. 5E). We show that CueP can activate Zn-SodCII in vitro (Figs 5G and 6A) and this is due to copper transfer between the two proteins and EPR data support transfer via an associative process (Fig. 6C). Both the CopA and GolT copper supply pathways in S. Typhimurium require CueP for SodCII activation (Fig. 5C).

CopA and GolT play dual roles within S. Typhimurium. Via copper export to the periplasm, both of these proteins confer copper resistance (Osman et al., 2010) as well as supplying copper for periplasmic SOD activity. This differs from the related CopA1 and CopA2 proteins from Pseudomonas aeruginosa that have roles in copper resistance and aa3-type cytochrome c oxidase activation, respectively, and neither protein can functionally substitute for the other (Gonzalez-Guerrero et al., 2010). In this case, the differences in function correlate with differences in copper-transport rates, with the proteins involved in copper resistance having copper-transport rates 10 times higher than those in cupro-protein assembly (Raimunda et al., 2011). The copper-transport rates of CopA and GolT have yet to be determined and it remains possible that one or other performs an additional, yet to be identified, cellular function and/or they are active under different copper-stress conditions. It is notable that expression levels from the golTS promoter (Fig. 3) are substantially lower than those from the copA promoter (Osman et al., 2010), in both the absence and the presence of copper. Furthermore, golT and copA show different expression patterns in response to other environmental factors, including acid stress and anaerobiosis (Fig. S8); conditions encountered during infection. Nonetheless, it is clear that CopA and GolT have substantial functional redundancy with respect to their roles in copper resistance and SodCII activation.

Copper transport by CopA and GolT aids S. Typhimurium survival in macrophage phagosomes (Osman et al., 2010). This requirement for copper transport coincides with an increased bacterial copper load within the phagosome (Osman et al., 2010), and hence one can envisage that this correlates with a need for copper resistance. However, by providing resistance to reactive oxygen and nitrogen species, the S. Typhimurium SodC proteins also confer resistance to phagocyte killing (De Groote et al., 1997). Hence, the role of CopA and GolT in SodCII activation may also be a contributing factor to the decreased survival of a copper transport mutant in macrophage phagosomes. On the contrary, the dismutation of superoxide into oxygen and hydrogen peroxide, due to SOD activity, may lead to increased H2O2 production, which in the presence of copper will drive Fenton chemistry and the production of highly toxic hydroxyl radicals. As such, the copper-transport activity of CopA and GolT during infection may well be double edged. Ongoing experiments are attempting to tease out how the activities of the S. Typhimurium copper-homeostatic proteins influence survival during infection.

The overlapping functions of CopA and GolT are in complete concurrence with their regulation by the two related transcriptional regulators CueR and GolS, respectively (Perez Audero et al., 2010), with similar tight affinities for copper (Fig. 4). S. Typhimurium CueR and GolS share 91% and 42% identity, respectively, with E. coli CueR. The latter has been structurally characterized and shown to possess a buried metal binding site (with a Cu+ affinity of 1021 M−1) providing a high level of selectivity for monovalent metal ions (Changela et al., 2003). The homodimer interface forms the site of metal binding and Cu+ is co-ordinated in a linear two co-ordinate geometry involving two cysteines (Cys-112 and Cys-120) from the same monomer. A conserved serine (Ser-77) also contributes to the metal co-ordination environment by promoting an interaction between the dimerization domain of one monomer and the metal binding loop of the second monomer (Changela et al., 2003). These features are conserved in S. Typhimurium CueR and GolS implying similar metal co-ordination environments, with two identical sites per homodimer. Some variance does, however, exist between the metal-binding loops of CueR and GolS which has been proposed to affect their metal responsiveness (Checa et al., 2007). It might be of interest to examine the redox poise and pKa of the thiols in GolS and CueR in view of the possibility that one may be better adapted to more oxidizing or acidic conditions.

SodCII activation in S. Typhimurium requires CueP and this is irrespective of the presence of CopA and/or GolT. We therefore propose a model in which copper is transported into the periplasm by either of the two P1B-type ATPases where it becomes bound by CueP for transfer to SodCII. It is currently not known how CueP acquires copper within the periplasm and it remains possible that CueP could interact with the P1B-type ATPases to accept the released Cu+. The periplasmically exposed residues of CopA and GolT (Fig. S9), may therefore allow for specific interactions with CueP, and by analogy to known interactions between copper chaperones and their partner proteins (Bagai et al., 2008; Robinson and Winge, 2010), it is likely that any interactions are metal-dependent. It is notable that residues corresponding to the putative P1B-type ATPase Cu+ exit site (Gourdon et al., 2011) are conserved in both CopA and GolT and hence are likely candidates for having a role in metal transfer (Fig. S9). Our data (Fig. 6B) are consistent with CueP primarily binding and stabilizing Cu+. As such, this may provide some selectivity for copper, against divalent metals, within the predominantly oxidizing environment of the periplasm.

The need for CueP to activate SodCII in S. Typhimurium differs to the situation in E. coli which possesses a SodCII orthologue (Fang et al., 1999) but lacks cueP, raising the question of how the E. coli enzyme acquires its copper. It is possible that E. coli SodC is able to acquire copper directly from its copper-exporting P1B-type ATPase CopA without the requirement for an intermediate. Other notable differences between these two closely related organisms include the lack of a Gol system in E. coli and the lack of a CusCFBA system for copper export across the outer membrane (Outten et al., 2001; Franke et al., 2003) in S. Typhimurium. The latter has led to the suggestion that CueP functionally replaces the CusCFBA system in S. Typhimurium for periplasmic copper resistance (Pontel and Soncini, 2009). Indeed, the ability of CueP to bind Cu+ within the periplasm may limit its toxicity in this compartment. Binding of Cu+ in the periplasm has also been reported for E. coli CusF (Loftin et al., 2007). CusF acts as a periplasmic copper chaperone and transfers Cu+ to the membrane fusion protein CusB for efflux (Bagai et al., 2008). It will be of interest to determine whether or not CusF also has a wider role in periplasmic cupro-protein assembly.

It is tempting to speculate that the specialized copper homeostatic systems in S. Typhimurium allows for the adaptation of this organism to the different challenges associated with copper handling in the host. The Cue and Gol systems may provide a mechanism by which copper is supplied for crucial cupro-enzyme activity, even if all copper ions are tightly bound and buffered in the periplasm (Raimunda et al., 2011), but at the same time preventing deleterious side reactions of copper: Inside host cells such side reactions might otherwise include the conversion of SOD products to deadly hydroxyl radicals via the Fenton reaction.

Experimental procedures

Bacterial strains and DNA manipulations

Salmonella enterica serovar Typhimurium strain SL1344 was used as wild-type and strain LB5010a was used as a restriction-deficient modification-proficient host for DNA manipulations, both were obtained from the Salmonella Genetic Stock Centre. E. coli strains JM109 and DH5α were used for routine cloning and BL21(DE3) for recombinant protein expression. Bacteria were cultured with shaking at 37°C in Luria–Bertani (LB) medium supplemented with ampicillin (100 μg ml−1), kanamycin (50 μg ml−1) and/or chloramphenicol (10 μg ml−1), where appropriate. Cells were transformed to antibiotic resistance as described (Datsenko and Wanner, 2000; Sambrook and Russell, 2001). All generated plasmid constructs were checked by sequence analysis.

Generation of S. Typhimurium deletion mutants

Deletion derivatives of S. Typhimurium SL1344 either were generated previously (Osman et al., 2010) or were obtained using the λ Red method (Datsenko and Wanner, 2000) using primers: 5′-GCTTTTATTAATGGTATTTACGATACAACCAAAAAACGAGGTAACTAATGGTGTAGGCTGGAGCTGCTT-3′ and 5′-GAAATTATGACGATATGGCTATGTTGCTGTTATTTCTCAATGACACATATGAATATCCTCCTTA-3′ for sodCI; and 5′-CGCCAGTGGTTTACACTTAACAGGCGACCACATGTAACGGAGGTTTTATGGTGTAGGCTGGAGCTGCTT-3′ and 5′-GCTCCAGCGCAGGGAACACTGGCTCCGGGTTATTTAATGACGCCGCAGGCCATATGAATATCCTCCTTA-3′ for sodCII. Mutagenesis was performed using strain LB5010a and selection of mutants achieved using LB medium supplemented with chloramphenicol. Mutations were subsequently moved into SL1344 or derivatives using P22 phage transduction. Antibiotic-resistance cassettes were removed using the helper plasmid pCP20 carrying the FLP recombinase. Constructs for complementation of mutant strains have been described previously (Osman et al., 2010).

Expression of SodCII in S. Typhimurium, isolation and SOD activity assays

The entire sodCII coding region was amplified from S. Typhimurium SL1344 genomic DNA using primers: 5′-GTTTGAGCTCGTGGTTTACACTTAACAGGCG-3′ and 5′-CACTGTCTAGAGGTTATTTAATGACGCCGC-3′, digested with SacI and XbaI and ligated into the SacI/XbaI site of pBAD28 (Guzman et al., 1995). The resulting construct was introduced into S. Typhimurium strain LB5010a prior to strain SL1344 lacking sodCI and sodCII to generate strain SodCIIUP, or into derivatives to generate strains SodCIIUP ΔcopA, SodCIIUP ΔgolT, SodCIIUP ΔcopAgolT and SodCIIUP ΔcueP. For expression of SodCII, overnight cultures of SodCIIUP and derivatives were diluted 1:100 in 500 ml of fresh LB medium and incubated at 37°C with shaking until an OD600 of 0.4–0.6 was reached, at which point expression of sodCII was induced by addition of 0.1% arabinose and incubation for a further 16 h. Periplasmic proteins were then isolated as described previously (Osman et al., 2010) but with the following modifications: periplasmic contents were liberated using 100 ml of ice-cold H2O, adjusted to 50 mM Tris pH 8.8 and once bound to a 1 ml Hi-Trap Q HP anion exchange column (GE Healthcare), proteins were eluted in 1 ml of the same buffer with addition of 200 mM NaCl and 0.2 ml loaded onto a Superdex 75 10/300 column (GE Healthcare) for size-exclusion chromatography (collecting 0.5 ml fractions). SodCII eluted with an apparent molecular mass of 16 kDa as reported previously (Krishnakumar et al., 2007). Periplasmic proteins were also isolated from strains SL1344 and ΔcopAgolT, with natively expressed SodCII, as described previously (Osman et al., 2010) using 2 l of culture.

To assay for SOD activity, anion exchange fractions containing SodCII were resolved by isoelectric focusing gel electrophoresis (Life Technologies) and stained for protein using Instant Blue (Expedeon) according to manufacturer's instructions, or in gel SOD activity (Beauchamp and Fridovich, 1971). Following size-exclusion chromatography, fractions were assayed for SOD activity by the pyrogallol method (Marklund and Marklund, 1974), where 1 unit of activity is defined as the amount of SOD required for 50% inhibition of pyrogallol autooxidation, and protein content analysed by SDS-PAGE (15% polyacrylamide) and gels scanned using a Li-Cor Odyssey infrared imager.

Activation of SodCII by copper or Cu-CueP in vitro

Size-exclusion chromatography fractions containing SodCII were combined and adjusted to 40 μg ml−1 protein (corresponding to ∼ 2.5 μM SodCII) and incubated with 5 μM CuSO4 for 1 h at 20°C with continuous gentle mixing before being dialysed into chelex-treated 50 mM NaCl, 5 mM HEPES pH 7.8 to remove unbound copper using an EDTA-washed 3.5 K MWCO Slide-A-Lyzer (Thermo-Scientific) rinsed with dialysis buffer according to manufacturer's instructions. SodCII recovered from dialysis was adjusted to 2 μM and assayed for SOD activity. For SodCII activation by Cu-CueP, SodCII was incubated for 30 min at 25°C in 80 mM NaCl, 10 mM HEPES pH 7.8 with continuous gentle mixing in the absence or presence of equimolar CueP (co-purified with copper), before being assayed for SOD activity. Reduced glutathione was included as a competitor when indicated.

Expression and purification of CueR, GolS, CueP and SodCII from E. coli

The cueR, golS, cueP and sodCII coding regions were amplified from S. Typhimurium genomic DNA using primers: 5′-GCATATGAATATTAGCGATGTGGCG-3′ and 5′-GGATCCTCAACGTGGCTTTTGCGC-3′ for cueR; 5′-GCATATGAACATCGGTAAAGCAGC-3′ and 5′-GGATCCGACTTACAGACGCTTTGC-3′ for golS; 5′-CATATGTCGAAATCATCATGG-3′ and 5′-GGATCCGGGCATTTTTTTAACG-3′ for cueP; and 5′-ACGGAGGTCATATGAAGCGATTAAGTTTAGCG-3′ and 5′-GGAACACTGGATCCGGGTTATTTAATGACGCCGC-3′ for sodCII, ligated into pGEM-T, prior to subcloning into the NdeI/BamHI site of pET29a (Novagen). Proteins were expressed in E. coli BL21(DE3) for 3 h at 37°C using 0.1 mM 1-thio-β-d-galactopyranoside (IPTG) for CueR and GolS or for 3 h at 30°C using 1 mM IPTG for SodCII and CueP. The medium was supplemented with 250 μM CuSO4 or 40 μM ZnSO4 for CueP and SodCII respectively (to ensure protein metallation).

Soluble cell lysates (post sonication) for CueR in buffer A (10 mM HEPES, pH 7.8, 1 mM EDTA, 1 mM DTT, 100 mM NaCl) were applied to a HiTrap Heparin affinity column (GE Healthcare) pre-equilibrated with buffer A, washed with 25 column volumes of buffer A, and bound protein eluted in buffer B (300 mM NaCl, 1 mM EDTA, 5 mM DTT, 10 mM HEPES, pH 7.8). Pooled fractions containing CueR were then applied to a HiLoad Superdex 75 26/60 size-exclusion column (GE Healthcare) running in buffer A prior to concentration using a 1 ml Heparin-affinity column and protein eluted in buffer C (500 mM NaCl, 1 mM EDTA, 5 mM DTT, 10 mM HEPES, pH 7.8).

Crude cell lysates containing GolS were resuspended in buffer C, applied to a HiTrap SP HP cation exchange column (GE Healthcare) pre-equilibrated with buffer C, washed with 25 column volumes of buffer C, and bound protein eluted in 800 mM NaCl, 1 mM EDTA, 5 mM DTT, 10 mM HEPES, pH 7.8. Pooled fractions containing GolS were then separated by size-exclusion chromatography as for CueR, except using buffer C. GolS containing fractions were then pooled, diluted in buffer C containing 10% glycerol and concentrated using a 1 ml HiTrap SP HP column and protein eluted in 1 M NaCl, 1 mM EDTA, 5 mM DTT, 10% glycerol, 10 mM HEPES pH 7.8.

For CueP and SodCII, periplasmic extracts were isolated as described for SodCIIUP strains except proteins were liberated using 20 ml of ice-cold H2O, and following centrifugation, adjusted to 10 mM NaCl, 10 mM HEPES pH 7.8 (buffer D) before being subject to anion exchange chromatography at room temperature. Bound proteins were washed in buffer D eluted with 100 mM NaCl, 10 mM HEPES pH 7.8 and those enriched for the desired protein were concentrated using a 5000 MWCO centrifugal concentrator and subject to size-exclusion chromatography in 50 mM NaCl, 5 mM HEPES pH 7.8. Fractions containing SodCII were then incubated with a 50-fold excess of BCS followed by equimolar ZnSO4 at 25°C for 1 h to generate Zn-SodCII. Apo-CueP was generated by incubation with a 50-fold excess of BCS at 25°C for 1 h. Proteins were subsequently concentrated, following dilution in buffer D, using a Hi-Trap Q HP anion exchange column and proteins eluted in 200 mM NaCl, 10 mM HEPES pH 7.8. All proteins were confirmed to be > 95% pure by SDS-PAGE. CueR, CueP and SodCII concentrations were determined by Bradford assay calibrated against BSA and with a 0.735 conversion factor for CueR estimated from sequence composition and validated by quantitative amino acid analysis (Alta Bioscience). GolS concentration was estimated by measurement of A280 via the theoretical extinction coefficient of 11 460 M−1 cm−1.

Analyses of copper binding

Anaerobic protein stocks were prepared by moving a heparin or cation exchange column loaded with approximately 100 μM CueR or GolS, respectively, into an anaerobic chamber. Washing and elution was performed under anaerobic conditions using Chelex-treated, N2-purged buffers, with final buffer compositions containing 10% glycerol, 10 mM HEPES pH 7.8 with 500 mM NaCl or 1 M NaCl for CueR and GolS, respectively, and protein concentrations re-estimated. Cys residues were > 95% reduced as determined with 5,5′-dithiobis-(2-nitrobenzoic acid) and proteins verified to be > 95% metal free by ICP-MS. Cu+ stocks were prepared as described previously (Dainty et al., 2010) and verified by ICP-MS and titration against BCS.

Cu+ affinities of CueR and GolS were examined by competition with BCS in Chelex-treated, N2-purged 1 M NaCl, 10% glycerol, 10 mM HEPES pH 7.8 under anaerobic conditions. BCS (10–1000 μM) was incubated with 4 μM Cu+ for 10 min before addition of 10 μM protein. Absorbance was monitored (Perkin Elmer λ35 spectrophotometer) to equilibrium. In addition, A483 was monitored following titration of BCS (100 μM) with Cu+ in the absence or presence of 10 μM CueR.

For interprotein copper exchange experiments, CueR (2.5 μM) was pre-incubated with Cu+ (0.4 molar equivalents) for 10 min before addition of 2.5 μM GolS (or vice versa) or buffer and incubation at room temperature for 16 h (under anaerobic conditions). Proteins were then diluted to 500 mM NaCl, 10% glycerol, 10 mM HEPES pH 7.8, loaded onto a pre-equilibrated 1 ml cation exchange column, washed with three column volumes of buffer and bound protein eluted with three column volumes 1 M NaCl (1 ml fractions collected throughout). Copper was assayed by ICP-MS and protein by SDS-PAGE.

β-Galactosidase assays

A construct containing PgolTS fused to lacZ in pRS415 (Osman et al., 2010) was introduced into S. Typhimurium strain LB5010a prior to SL1344 or derivatives. β-Galactosidase assays were performed as described (Osman et al., 2010) following growth of cultures in medium supplemented with various concentrations of CuSO4.

EPR spectroscopy

Zn-SodCII and Cu-CueP were diluted to 50 μM in 100 mM NaCl, 10 mM HEPES pH 7.8. For the time-course experiments samples were either frozen after the incubation time, then thawed and refrozen for EPR experiments, or frozen directly in the EPR tube for EPR analysis. EPR spectra were obtained using a Bruker ELEXSYS E500 EPR spectrometer, operating at X-band (approximately 9.4 GHz) equipped with a Bruker ER 4122-SHQE cylindrical mode (TE102) cavity and an Oxford Instruments ESR900 helium flow cryostat. Temperature control was effected using an Oxford Instruments ITC503 temperature controller. Spectra were recorded at 20 K, further acquisition conditions are provided in the figure legends.


This work was supported by Grant BB/G010765/1 from the Biotechnology and Biological Science Research Council (BBSRC).