Copper Chaperone CupA and Zinc Control CopY Regulation of the Pneumococcal cop Operon

As mechanisms of copper toxicity are emerging, bacterial processing of intracellular copper, specifically inside Streptococcus pneumoniae, remains unclear. In this study, we investigated two proteins encoded by the copper export operon: the repressor, CopY, and the copper chaperone, CupA. Zinc suppressed transcription of the copper export operon by increasing the affinity of CopY for DNA. Furthermore, CupA was able to chelate copper from CopY not bound to DNA and reduce it from Cu2+ to Cu1+. This reduced copper state is essential for bacterial copper export via CopA. In view of the fact that innate immune cells use copper to kill pathogenic bacteria, understanding the mechanisms of copper export could expose new small-molecule therapeutic targets that could work synergistically with copper against pathogenic bacteria.

copper into an oxidized phagolysosome to kill pathogens (8,10). This antimicrobial effect has been exploited since antiquity and is currently used in hospitals to reduce nosocomial infections (11,12). Several different types of bacterial copper export systems are used to mitigate toxic effects. In general, these systems consist of an operon repressor that releases DNA upon copper binding, a copper chaperone, and the exporter. Further understanding of how copper export systems are regulated in different bacterial pathogens is an important step in combating these infections.
Streptococcus pneumoniae (the pneumococcus) is a Gram-positive pathogen and a causative agent of pneumonia, otitis media, and meningitis. Often residing in the upper respiratory tract, the pneumococcus encounters innate immune responses that use copper to combat infection (3). The pneumococcus contains a copper export system, the cop operon, that encodes a repressor, CopY, the copper chaperone, CupA, and the copper exporter, CopA. CopA, which exports Cu 1ϩ from the bacterium, is important for pneumococcal virulence and resistance to copper toxicity (3,(13)(14)(15).
S. pneumoniae CopY is highly conserved (Ͼ90% identity) in many streptococcal species and is related to other copper operon repressors, such as COPR in Lactococcus lactis (34% identity, 57% similarity) (16). S. pneumoniae CopY protein structure predictions reveal similarity to the L. lactis COPR winged helix-turn-helix DNA-binding domain (see Fig. S1A and C in the supplemental material) (17,18). The L. lactis COPR nuclear magnetic resonance structure (PDB code 2K4B) contains only the sequence of the DNA-binding domain and therefore does not include the metal-binding domain. Thus, how CopY interacts with copper or other metals remains unknown. The known protein structure most homologous to the generated pneumococcal CopY model is that of BlaI (PDB code 1SD4), the penicillinase repressor in Staphylococcus aureus ( Fig. S1B and D) (19). There are several structures of copper operon repressors/sensors containing copper, such as in CueR from Escherichia coli and CsoR from Mycobacterium tuberculosis (16,20,21). However, these proteins share no homology with pneumococcal CopY.
A further feature distinguishing pneumococcal CopY from other homologous proteins is the lack of a C-terminal CXC Cu 1ϩ -binding motif compared with COPR from L. lactis, while the beta-lactamase repressor from S. aureus lacks both CXC motifs (Fig. S1A). The number of CXC motifs could affect the coordination and number of copper or other metal atoms in CopY. Interestingly, Zn 2ϩ can be displaced by two Cu 1ϩ atoms in Enterococcus hirae CopY (22). Additionally, zinc was shown to suppress copper-induced pneumococcal cop operon expression while bacterial zinc export is upregulated upon copper stress (8,13). The mechanism of how zinc, copper, and other metals interact with COPR/CopY repressors in context with DNA binding remains unclear.
In contrast to CopY, the S. pneumoniae CupA sequence is unique and is present in only a few streptococcal species. Although the CupA copper chaperone is not needed for virulence in mice, it is required for copper resistance in culture (3,13,23). Copper chaperones in different organisms have been shown to traffic Cu 1ϩ both to the membrane copper exporter CopA as CupA in S. pneumoniae and to the operon repressor COPR as CopZ in E. hirae (22,23). Notably, E. hirae CopZ and pneumococcal CupA share no homology. S. pneumoniae CupA differs from other copper chaperones in that it has a cupredoxin fold and, as a result, may play a role in reducing or facilitating the reduction of copper inside the bacteria for efficient export by CopA. Such activity would also facilitate the need for the copper chaperone to also coordinate Cu 2ϩ .
In this study, we examined the influence of copper and other metals on CopY, CupA's effect on copper, and the interplay between the two proteins. We tested how S. pneumoniae CopY interacts with copper to release it from the cop operon and how this process is affected by zinc, if CupA could facilitate the copper reduction, and/or if CupA could affect CopY's repressor function. We found that zinc caused tetramerization of CopY and thus increased its affinity for DNA by approximately an order of magnitude, CupA facilitated the reduction of Cu 2ϩ to Cu 1ϩ , and CupA chelated copper from CopY.
changes. For example, copper repressors have previously been shown to bind zinc, which results in multimerization (22,30). Additionally, zinc export is upregulated under copper stress in the pneumococcus (8). Thus, we determined whether zinc interacts with CopY's ability to bind DNA and form multimeric structures. In the EMSA, we saw an upward shift occur in a zinc-dependent manner, implying that pneumococcal CopY was multimerizing in the presence of zinc (Fig. 2D). To further investigate the multimerization of CopY in the presence of zinc, we performed size exclusion chromatography with multiangle light scattering (SEC-MALS) to ascertain the molecular mass in both the apo and zinc-bound states. Because of the construct used, the CopY monomer should be 18.1 kDa. However, the apo form of CopY appeared at a molecular mass of 40.9 kDa, implying that without a metal bound, CopY exists as a dimer (Fig. S3A, B, and E). Interestingly, the molecular mass of the zinc-bound form of CopY was 79.4 kDa, suggesting that it exists as a tetramer in solution (Fig. S3C, D, and E). Taken together, the results show that zinc caused the CopY dimer to dimerize into a tetramer, a form consistent with other copper operon repressors (30)(31)(32).
Next, we determined the binding affinity of CopY for pSP_0727 and what effect zinc multimerization has on that binding. Using fluorescence polarization assay, we observed a 4.24 Ϯ 0.78 M binding affinity for apo CopY to pSP_0727 and nonspecific binding to the negative control, pSP_1541 ( Fig. 3A; Fig. S4). Binding increased by almost an order of magnitude with the addition of Zn 2ϩ to 0.52 Ϯ 0.13 M (Fig. 3A). Additionally, starting with a protein concentration that led to 40% CopY binding to pSP_0727, we observed an increase in signal in the presence of Zn 2ϩ with a 50%  (Fig. 3B). Taken together, the results show that CopY had a higher affinity for the cop operon promoter in the presence of zinc binding, which could be due to the tetrameric state of CopY.
Combinatory effects of copper and zinc in pneumococcus killing. To examine if the increased affinity of CopY for DNA with zinc present has a detrimental effect on bacterial survival under copper stress, we made a ΔczcD (SP_1857 zinc exporter) mutant of wild-type S. pneumoniae (TIGR4) to increase the internal concentration of zinc (Table S1) (3,8). As expected, in an overnight growth assay, the ΔczcD mutant was significantly more susceptible to zinc stress than the wild-type and ΔcopA mutant strains, and the ΔcopA mutant strain was more susceptible to zinc stress than wild-type S. pneumoniae (Fig. 4A). Additionally, there was a significant growth lag of the ΔczcD mutant under copper stress compared to S. pneumoniae TIGR4 (Fig. 4B). When combined with a sublethal dose of zinc (75 M), copper was more toxic to the ΔczcD mutant than to S. pneumoniae TIGR4 (Fig. 4C). Taken together, the results show that excess zinc is detrimental to the ability of the pneumococcus to combat copper stress and sublethal amounts of zinc with copper have additive effects that inhibit bacterial growth.
CupA facilitates copper reduction. In previous toxicity assays, copper was added to the pneumococcus as Cu 2ϩ under nonreducing conditions (8,13). Furthermore, it has been shown in E. coli that copper can enter the bacterium as Cu 2ϩ via the iron siderophore Ybt (33,34). However, for copper to leave the bacterium, it has to be in the Cu 1ϩ redox state and therefore must be reduced for functional CopA export (14,25). In S. pneumoniae, CupA has a predicted cupredoxin fold and a membrane leader sequence, which it uses to shuttle bound Cu 1ϩ to the membrane-bound copper exporter CopA (23). Using the Cu 1ϩ colorimetric reporter bathocuproinedisulfonic acid (BCS) and the crystal structure of CupA to design amino acid exchanges, we determined whether CupA could facilitate the reduction of Cu 2ϩ to Cu 1ϩ and which amino acid residues are involved in that reaction (Table S1) (23). Because the leader sequence made CupA insoluble, we used CupA amino acids 23 to 123 for biochemical assays. In these assays, there is incomplete accessibility of Cu 1ϩ to BCS when using the CupA protein because the signal depends upon CupA's ability to reduce and release Cu 1ϩ . Given these caveats, we still observed a reduction of Cu 2ϩ to Cu 1ϩ that was not detected with buffer, Cu 2ϩ , and BCS alone ( Fig. 5A to D). Thus, recombinant CupA is able to reduce Cu 2ϩ to Cu 1ϩ in a manner detectable by a Cu 1ϩ chelator.
Next, we wanted to examine the residues involved in this reduction. The CupA structure contains two Cu 1ϩ atoms: one coordinated by C74 and C111 (the high-affinity site) and one coordinated by C74, M113, and M115 (the low-affinity site) (23). Therefore, we mutated each of these residues to alanine and also made a double cysteine mutant form to represent a completely null mutant form (Table S1). The C74A C111A double variant was unable to facilitate the reduction of Cu 2ϩ to Cu 1ϩ in a concentrationdependent manner, suggesting that both of these residues are used to catalyze copper reduction. However, each single cysteine variant was independently able to catalyze the reaction, albeit significantly more slowly than wild-type CupA ( Fig. 5A to D). This result implies that C74 and C111 are at least sterically involved in the coordination of Cu 2ϩ before reduction takes place. Additionally, a time course at a fixed concentration confirmed that the double amino acid exchange variant was unable to facilitate the reduction of copper ( Fig. 5C and D). The M113A and M115A mutants yielded higher rates and V max values than did wild-type CupA ( Fig. 5A and B). The increases in the M113A and M115A mutants' V max values are consistent with the assay in that BCS needs to detect the CupA reduced and released Cu 1ϩ and these exchanges were made to the protein's low-affinity Cu 1ϩ -binding domain (23).
CupA's role in alleviation of copper toxicity. To examine the effect of these mutations on resistance to copper toxicity, we put full-length wild-type and full-length  (Table S1). In both endpoint growth curve and zone-of-inhibition assays, each Cu 1ϩ -coordinating residue was necessary to restore wild-type cupA-complemented copper resistance in the bacteria ( Fig. 5E and F). Taken together, these results show that although the single amino acid exchanges still allowed copper reduction, each copper-binding residue was necessary to restore wild-type resistance to copper toxicity.
CupA's effect on CopY's DNA binding. Metallochaperones can provide copper to sensors and exporters (4,22,23). CupA has already been shown to take copper to the CopA exporter (23). Therefore, we examined interactions between pneumococcal CopY and CupA. We determined whether CupA could chelate copper from CopY and if this chelation could restore CopY's ability to bind to DNA in the presence of copper. We found that, in vitro, CupA can restore binding of CopY to DNA in the presence of Cu 2ϩ by chelating copper from CopY ( Fig. 6A and B). These results are consistent with CupA having a membrane leader sequence and thus not trafficking copper to the repressor for activation the operon but instead for the copper to be chelated from copper-bound CopY so that it can return to repress the cop operon (23).

DISCUSSION
Here we have shown that, in vitro, zinc can lead to homotetramerization of CopY, which binds upstream of the pneumococcal cop operon, in addition to the protein existing as a dimer in the apo form ( Fig. 2D; Fig. S3). Through this tetramerization event, the DNA-binding domain structure is presumably altered, or zinc increases the affinity of CopY for DNA upstream from the cop operon by roughly an order of magnitude ( Fig. 1B and 3). These data support the finding that zinc limitation leads to cop operon upregulation, as under this zinc limitation, CopY would bind copper more easily, leading to release of the cop operon (26).
When Cu 2ϩ enters the cell, it might displace bound zinc from CopY to upregulate copY operon expression (Fig. 7). This displacement of zinc could also explain the upregulation of the zinc exporter under copper stress (8). However, it is also possible that copper indirectly induces the upregulation of this zinc export system through binding of the MerR family transcriptional regulator soxR (SP_1856) directly downstream of zinc exporter czcD (SP_1857) on the negative strand. Like czcD, soxR is also upregulated under copper stress (8). This copper-specific regulation does not occur through SczA (SP_1858), a protein known to regulate czcD expression, or CopY, as the promoter sequence is not present (35).
Accordingly, zinc stress, caused either by its addition with copper or by elimination of its efflux system, was detrimental to the pneumococcus under copper stress (Fig. 4). During infection, host zinc and copper levels increase and host deficiencies in these metals lead to an immunocompromised state (36,37). Silencing of ATP7A expression in macrophages decreased their ability to kill E. coli, and while limited data exist on the  CupA traffics to the membrane and, either en route or once bound to the membrane, chelates copper from CopY or the cytosol. The copper export protein CopA and the zinc export protein CzcD are also made to export Cu 1ϩ and Zn, respectively. CupA, if necessary, facilitates the reduction of Cu 2ϩ to Cu 1ϩ and transfers Cu 1ϩ to membrane-bound CopA for export. Once free copper is chelated or exported, apo dimeric or zinc-bound tetrameric CopY can return to repress the cop operon.
role of phagolysosomal zinc importers in the SLC30 and SLC39 family members, SLC30A1 expression in macrophages is increased during M. tuberculosis infection (10,38). Zinc's ability to further repress the cop operon gives a glimpse of how the "brass dagger" of copper and zinc toxicity inside the macrophage can be so detrimental to bacteria (Fig. 3) (39, 40). Understanding how zinc and copper interact with CopY to modulate both its structure and its ability to bind DNA will be the subject of future studies.
In the assays used in this study and some other assays studying copper toxicity in the pneumococcus, the copper ion used is Cu 2ϩ (3,8,13). This form has previously been shown to be toxic in S. pneumoniae by mismetallating ribonucleotide reductase, NrdF, an essential protein in aerobic nucleotide synthesis (8). However, the form of copper exported by the bacteria is Cu 1ϩ , thus creating a constant need for electrons to maintain the reducing environment under copper stress (25). This study showed that CupA can facilitate the reduction of Cu 2ϩ to Cu 1ϩ , thus offering another function for membrane-bound CupA, which acts to chelate copper as it enters the bacteria, and then pass copper to the metal-binding domain of CopA (Fig. 5, 7, and 8) (23). This is supported by previous data that show that the ΔcopY mutant, in which cupA expression is no longer repressed, has significantly higher levels of protein oxidation under increasing copper stress than does wild-type S. pneumoniae TIGR4, while the ΔcupA mutant has significantly lower levels than wild-type S. pneumoniae TIGR4 (8). One hypothesis that is consistent with our findings is that CupA must acquire electrons to reduce the copper that it has taken up as Cu 2ϩ for export. Therefore, CupA oxidizes other proteins in the cell to acquire sufficient reducing equivalents, which might explain why the measured copper oxidation level under copper stress is dependent on internal CupA levels. One protein source of these electrons could be thioredoxin (SP_1776), which is also upregulated under copper stress (8). However, the identity of the external source of electrons, and any additional protein(s) responsible for reducing Cu 2ϩ to Cu 1ϩ , remains to be established.
CupA has a conserved cupredoxin fold, and there are multiple sources showing that Cu 2ϩ can be reduced depending on its binding coordination at sulfur-containing sites, i.e., cysteine-and methionine-rich sequences (41)(42)(43). Although there is no established copper import system, there are data supporting the idea that Cu 2ϩ enters bacteria through iron siderophores (33,34). Therefore, given nonreducing conditions where Cu 2ϩ is significantly more soluble than Cu 1ϩ , reduction of Cu 2ϩ and efflux of Cu 1ϩ could serve to decrease the extracellular pool of Cu 2ϩ . The result would be relatively insoluble Cu 1ϩ , which would have more difficulty reentering S. pneumoniae than Cu 2ϩ . This soluble copper accessibility could be one of the reasons why the oxidizing Our copper reduction assay was dependent on the ability of CupA to reduce and release Cu 1ϩ for BCS chelation. In our proposed copper reduction model of CupA, Cu 2ϩ is chelated by CupA and reduced to Cu 1ϩ (Fig. 8). This process does not negate CupA's ability to bind Cu 1ϩ . The reduction process is facilitated by C111 and C74, as shown by the inability of the C74A C111A protein variant to either facilitate reduction of Cu 2ϩ or restore CopY binding to DNA in the presence of Cu 2ϩ (Fig. 5A to D). CupA's crystal structure denotes C74 and C111 as high-affinity site 1 for the binding of Cu 1ϩ and C74, M113, and M115 as low-affinity site 2 for the binding of Cu 1ϩ (23). Our data suggest that the copper-reducing event occurs in conjunction with site 1 and then the resultant Cu 1ϩ atom is passed to site 2 to facilitate export via CopA (Fig. 8).
Within the confines of our assay, mutation of M113 at site 2 increased the V max and decreased the K m of CupA copper reduction relative to those of wild-type CupA. This result was in support of previous studies showing that site 2 is the recipient of the reduced copper, which hands it off to CopA (23). Furthermore, our data suggested, via M113A's shorter T 50 (the time it takes for 50% of the copper reduction and release to take place at a given copper concentration) and lower K m than the wild type, that copper reduction by site 1 is enhanced by the ability of site 2 to flux copper. Mutations of site 2 resulted in a reduction in growth in the presence of copper, as previously reported (44). Thus, even though the M113A variant facilitated copper reduction and release faster than the wild-type protein in vitro, in vivo, M113A CupA likely lost bound copper before it was able to hand copper off to the exporter or was unable to hand off copper to the exporter, leading to decreased bacterial copper resistance.
Mutations of site 1 in our growth and zone-of-inhibition assays also resulted in copper-stunted growth. This difference from previously published results could be due to the concentration (1 M) used in the zone-of-inhibition assay, the medium used, or the mutated residue. Furthermore, the copper-reducing function of CupA explains why, in a ΔcupA mutant background of S. pneumoniae expressing the Bacillus subtilis CopZ metallochaperone, growth cannot be restored in the presence of copper stress (44).
Previous models reported that copper chaperones, such as CopZ in E. hirae, carry copper to the repressor for activation of the operon (4,22). We observed that, at least in vitro, the pneumococcal copper chaperone CupA can chelate copper from CopY ( Fig. 6A and B). Unlike E. hirae CopZ, however, CupA of S. pneumoniae has a membrane leader sequence. Full deletions of the CupA membrane leader sequence increase pneumococcal susceptibility to copper stress (23). The membrane leader sequence might facilitate CupA's interaction with CopA, although this is speculative at this stage. Nevertheless, our current model suggests that in S. pneumoniae when CopY binds copper, it does not bind DNA and instead acts as a source of copper for CupA (Fig. 7). CupA reduces Cu 2ϩ to Cu 1ϩ and transfers copper to CopA for export from the cell as Cu 1ϩ . After releasing copper to CupA, CopY either binds to the DNA or picks up excess cytosolic Cu 1ϩ or Cu 2ϩ (Fig. 7). Finally, after the bacteria have sufficiently detoxified copper to a tolerable level, zinc interacts with CopY, allowing it to repress cop operon expression (Fig. 7) (36, 45).

MATERIALS AND METHODS
Protein homology and domain prediction. Protein sequences were entered into the NIH Protein Basic Local Alignment Search Tool (BLAST) (16). The BLAST output predicted protein domains and aligned them with similar sequences. Models were generated by using I-TASSER, and models and empirically determined structures were aligned by using PyMOL (18,46,47).
Protein production. DNA encoding full-length CopY and residues 23 to 123 of CupA was amplified from S. pneumoniae TIGR4 DNA and inserted into a pMCSG7 ligation-independent cloning vector (Table S1). The first 22 amino acids are a leader sequence that makes the protein insoluble. Site-directed mutagenesis was performed to produce C74A, C111A, M113A, M115A, and C74A C111A mutations in the gene encoding CupA. Vectors for CopY and CupA production were used to transform E. coli BL21 Gold DE3 cells (Stratagene) with selection on ampicillin plates (100 mg/ml). A single colony was used to inoculate a 100-ml flask of Luria broth (LB) containing ampicillin (100 mg/ml), and the cells in the culture were grown overnight. Cell were collected by centrifugation at 3,000 ϫ g, and the supernatant was discarded. The resultant pellet of cells was used to inoculate a 2.8-liter shaker flask of LB containing ampicillin (100 mg/ml). Cells were grown in Terrific broth at 37°C until the optical density at 600nm (OD 600 ) of the culture was 0.4. The temperature was lowered to 17°C for overnight growth, and protein production was induced by adding 0.3 mM isopropyl-␤-D-thiogalactopyranoside. After overnight growth, cells were harvested by centrifugation at 6,000 ϫ g for 20 min at 4°C. The supernatant was discarded, and cell pellets were stored at Ϫ80°C.
Cell pellets were thawed by using a buffer consisting of phosphate-buffered saline (PBS) containing 10 mM imidazole, 5% glycerol, 200 U of Benzonase nuclease (Novagen), 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor tablets (Roche). Cells were passed twice through a microfluidizer and centrifuged at 16,000 ϫ g. The soluble fraction was filtered through a 0.22-m filter and nickel purified on gravity HisPur Ni-nitrilotriacetic acid (NTA) resin (Thermo Fisher) by using PBS containing 250 mM imidazole and 5% glycerol. Elution fractions were concentrated to 10 ml and dialyzed into either 1 liter of PBS containing 5% glycerol or 1 liter of 50 mM Tris (pH 7.5), 100 mM NaCl, and 5% glycerol each at least three times. The results of an SDS-PAGE assay confirmed that the protein was 95% pure, and gel filtration assay results confirmed that the protein was not a soluble aggregate. Purified proteins were concentrated to~500 to 1,000 M for CupA and the corresponding variants and~150 M for CopY and stored at Ϫ80°C.
As needed, eluted His-tagged CupA was incubated with tobacco etch virus (TEV) protease at 1 mg/ml to cleave the 6ϫHis tag and dialyzed in 1 liter of 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, and 1 mM dithiothreitol (DTT) for 18 h at 4°C. The sample was then dialyzed four times in 50 mM Tris-HCl (pH 7.5) and 100 mM NaCl to eliminate the DTT. Samples were then filtered and passed through a nickel column preequilibrated with 50 mM Tris-HCl (pH 7.5) and 100 mM NaCl. The flowthrough as CupA was collected and concentrated to~500 M and frozen at Ϫ80°C. When necessary, samples and the corresponding buffers were treated with a gravity flow column consisting of Chelex 100 resin.
Induction of copA. S. pneumoniae in THYB was incubated with copper, silver, and zinc at 100 M for 30 min. Bacteria were collected, and RNA extraction and quantitative real-time PCR were performed as previously described (3). SuperScript III First-Strand Synthesis SuperMix (Invitrogen) was used to synthesize cDNA from the isolated bacterial RNA (50 ng/pl). SYBR green (Invitrogen) was used to monitor DNA amplification on an ABI Prism 7300 real-time PCR machine (Applied Biosystems). After samples were normalized relative to gyrA expression, fold change was obtained by the ΔΔC T method.
CD. CopY in PBS with 5% glycerol was chelated and diluted to 10 M with or without 50 M CuSO 4 , ZnSO 4 , and AgCl 2 . A wavelength scan from 200 to 250 nm was performed with a 62 DS CD spectrometer (Aviv) at 23°C with a 15-s averaging time. The equation used to calculate mean residue ellipticity was ϫ MW/10 ϫ c ϫ l ϫ r, where is the experimental wavelength, MW is the molecular weight (18,000), c is the concentration (10 M or 0.18 mg/ml), l is the path length (1 mm), and r is the number of residues (133 from CopY and 24 from the His tag and TRV protease cleavage site).
EMSA. DNA leading strands pSP_0727 (TCTATAATTGACAAATGTAGATTTTAAGAGTATACTGATGAGTG TAATTGACAAATG/3Bio/) and pSP_1541 (GCGAACCTCACTTACCCCTTGCAAAGTCTTGGGGTCATTAGA/3Bio) were purchased from IDT and annealed to the lagging strand partner CATTTGTCAATTACACTCATCAGT ATACTCTTAAAATCTACATTTGTCAATTATAGA (SP0727) or TCTAATGACCCCAAGACTTTGCAAGGGGTAAGTG AGGTTCGC (SP_1541) by heating the reaction mixture containing the strands to 95°C and then reducing the temperature by 1°C/min to 22°C. DNA was then diluted to the desired concentration. Next, 0.5ϫ Tris-borate-EDTA (TBE) electrophoresis buffer with no EDTA (so as not to chelate the metals used in the assay) was prepared and used as the buffer for both the 6% TBE DNA retardation gel (Invitrogen) and transfer to Zeta Probe blotting membrane (Bio-Rad). Unless indicated otherwise in the figure legend, 50 M CopY was added with 50 ng of promoter DNA and brought to the desired volume of 20 l with Tris-buffered saline (TBS), taking into account future changes in volume by metals added (from a 1 mM stock) and/or CupA (100 M), if applicable. Samples were incubated at 4°C for 5 min at each step, loaded onto a TBE polyacrylamide gel, and then electrophoresed at 40 V for 120 min (100 V for 60 min for zinc). The polyacrylamide gel was incubated in distilled H 2 O (dH 2 O) with 0.02% ethidium bromide (AMRESCO) for 10 min and washed twice for 5 min each time in dH 2 O, and the gel was visualized via UV on a Gel Doc XRϩ system (Bio-Rad).
SEC-MALS. SEC-MALS experiments were performed with a WTC-010S5 (molecular mass range, 100 to 100,000 Da) size exclusion column (Wyatt Technologies, Santa Barbara, CA, USA) with three detectors connected in series, i.e., an Agilent 1200 UV detector (Agilent Technologies, Santa Clara, CA), a Wyatt DAWN-HELEOS MALS detector, and a Wyatt Optilab rEX differential refractive index detector (Wyatt Technologies, Santa Barbara, CA, USA). The DAWN-HELEOS MALS detector uses a laser wavelength of 658 nm and was calibrated using the manufacturer's specifications. The columns were equilibrated with 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, and 5% glycerol with or without 100 M zinc, and the experiment was conducted at 25°C. A 100-l sample was placed on the column with the autosampler, and a flow rate of 0.40 ml/min was maintained throughout the experiments. Protein in the eluent was detected and measured via light scattering, absorbance at 280 nm, and refractive index determination, and the data were recorded and analyzed by using Wyatt Astra software (version 6.0.5.3). The refractive index increment, dn/dc, was taken as 0.185 ml/g to measure the concentration of the protein sample. EASI Graphs (Astra software) were exported and plotted as a molar mass distribution superimposed on (i) a chromatogram of absorbance at 280 nm versus the elution volume and (ii) the same graphs with the addition of the light scattering and refractive index signal distributions (48).
Fluorescence polarization assay. DNA was annealed as previously described for EMSAs but with a 6-carboxyfluorescein (FAM) tag on the leading strand. Samples were diluted in Chelex 100-treated buffer containing 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, and 5% glycerol to obtain the desired CopY concentration (25 to 50 or 2 M). Experiments were run in quadruplicate 20-l aliquots pipetted into a  (Table S1). Fragments approximately 1 kb upstream and downstream of the target gene were amplified and spliced to an erythromycin resistance cassette. SOE PCR products were subsequently used to transform S. pneumoniae TIGR4. Briefly, for transformation, bacteria were grown to an OD 600 of 0.1 and induced with competence-stimulating peptide 2 (CSP2) for 14 min. The SOE PCR product was then added to bacteria for a 2-h recovery period at 37°C before plating of bacteria on tryptic soy agar (TSA) blood agar plates with 20 g/ml neomycin and 1 g/ml erythromycin. Single colonies were verified by PCR to confirm correct insertion of the SOE PCR product and deletion of the target gene (49).
Growth curves. Using THYB (30 g of Todd-Hewitt broth [Sigma], 2 g of yeast extract [Sigma], and 1 liter of dH 2 O, pH 6.5), S. pneumoniae was grown to an OD 620 of 0.1 and diluted 1:50 into THYB with or without metal stress. The OD 600 was measured after 18 h or every 30 min for 10 h of growth in a 96-well plate on a BioTek Cytation 5 and normalized to maximal growth without metal stress. Data are representative of four individual experiments.
Cupredoxin assay. For V max and K m studies, BCS was added at 1 mM with 50 M CupA (or variants thereof) and various amounts of CuSO 4 (5 to 625 M) in TBS to a final volume of 1 ml to a cuvette and the absorbance at 490 nm was read in a spectrophotometer with TBS, CuSO 4 , and BCS used as the blank. For maximal binding (B max ) and T 50 in the timed measurements, TBS, 1 mM BCS, 50 CupA (or variants thereof), and 50 M CuSO 4 (in that order) were added to a cuvette in a final volume of 1 ml and monitored over time, with TBS with CuSO 4 and BCS used as the blank. Data were plotted as the average of three individual experiments.
CupA mutants. The ΔcupA mutants in the S. pneumoniae TIGR4 background were made as described in reference 3 (Table S1). Full-length cupA was cloned from S. pneumoniae TIGR4 and ligated into pABG5 by using EcoRI. Site-directed mutagenesis was performed as previously described to obtain mutants. After incubation with CSP2 as described above, the pABG5 vector containing sequence-confirmed mutant forms was added to competent cells of the S. pneumoniae ΔcupA mutant grown to an OD of 0.1. After a 2-h recovery, bacteria were plated on TSA blood agar plates containing kanamycin at 400 g/ml.
Zone-of-inhibition assay. Bacteria were grown in THYB at pH 6.5 to an OD 620 of 0.3. A 100-l volume of culture was then spread onto a blood agar plate. A disc of filter paper with 10 l of 1 M CuSO 4 was placed on the plate. After overnight growth, the distance between the outer edge of the disc and the bacterial growth was reported as the zone of inhibition.
Inductively coupled plasma mass spectrometry. First, 10 nmol of chelated CopY with the 6ϫHis tag was incubated with or without 50 nmol of CuSO 4 for 10 min at 4°C, and then the CopY-Cu mixture or apo chelated CopY was incubated with 100 l of Ni-NTA Superflow resin for 10 min. Samples were gently centrifuged at 500 ϫ g for 1 min and washed three times with a 10ϫ resin volume in 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, and 5% glycerol. Chelated, TRV protease-cleaved CupA (no 6ϫHis tag) was incubated with both the CopY-Cu mixture or apo chelated CopY bound to Ni-NTA resin for 10 min at 4°C. Samples were gently centrifuged at 500 ϫ g for 1 min, and CupA was collected. CupA samples and eluted CopY samples were then mixed in 5 ml of 1% nitric acid and filtered for spectrometry (820 ICPMS System; Varian, Inc.). The fold difference between the amounts of copper bound to CupA in the non-copper-treated and copper-treated CopY samples was determined. Data shown represent the average of three individual experiments.