The Escherichia coli Copper-responsivecopA Promoter Is Activated by Gold

Abstract

The copA gene of Escherichia coli encodes a copper transporter and its promoter is normally regulated by Cu(I) ions and CueR, a MerR-like transcriptional activator. We show that CueR can also be activated by gold salts and that Cys112 and Cys120 are involved in recognition of gold, silver, and copper salts. Gold activation is unaffected by copper chelating agents but is affected by general metal chelators. This is the first example of specific regulation of transcription by gold, and we briefly speculate that the biological effects of gold antiarthritic drugs may be through their effects on copper management in eukaryotic systems. The copA gene of Escherichia coli encodes a copper transporter and its promoter is normally regulated by Cu(I) ions and CueR, a MerR-like transcriptional activator. We show that CueR can also be activated by gold salts and that Cys112 and Cys120 are involved in recognition of gold, silver, and copper salts. Gold activation is unaffected by copper chelating agents but is affected by general metal chelators. This is the first example of specific regulation of transcription by gold, and we briefly speculate that the biological effects of gold antiarthritic drugs may be through their effects on copper management in eukaryotic systems. There is no known natural biological function of gold or its salts, but it has biological effects. Solutions of gold chloride are toxic to bacteria (1Alper J. ASM News. 1992; 58: 650-651Google Scholar), disimilatory Fe(III)-reducing bacteria and archaea are capable of reducing Au(III) to Au(0), precipitating the metal (2Kashefi K. Tor J.M. Nevin K.P. Loveley D.R. Appl. Environ. Microbiol. 2001; 67: 3275-3279Crossref PubMed Scopus (166) Google Scholar), and may be responsible in part for the formation of some gold deposits (1Alper J. ASM News. 1992; 58: 650-651Google Scholar). Medical uses of gold have been limited largely to prosthetics and antiarthritic Au(I)-thiolate drugs, although the basis of antiarthritic action has been unclear (3Kendrick M.J. May M.T. Plishka M.J. Robinson K.D. Burgess J. Metals in biological Systems. Ellis Horwood, London1992Google Scholar). In contrast, the biological requirements for copper are quite well understood. Mammalian systems require copper for many enzymes (4Linder M.C. Biochemistry of Copper.in: Frieden E. Biochemistry of the Elements. Platinum Press, New York1991Crossref Google Scholar). Recently, new elements of copper homeostasis in bacteria have been identified, including transporters, multicopper oxidases, and two-component and MerR-like regulators (5Rensing C. Fan B. Sharma R. Mitra B. Rosen B.P. Proc. Natl. Acad. Sci. U. S. A. 2000; 94: 652-656Crossref Scopus (408) Google Scholar, 6Munson G.P. Lam D.L. Outen F.W. O'Halloran T.V. J. Bacteriol. 2000; 182: 5864-5871Crossref PubMed Scopus (252) Google Scholar, 7Outten F.W. Outten C.E. Hale J. O'Halloran T.V. J. Biol. Chem. 2000; 275: 31024-31029Abstract Full Text Full Text PDF PubMed Scopus (251) Google Scholar, 8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar, 9Franke S. Grass G. Nies D.H. Microbiology. 2001; 147: 965-972Crossref PubMed Scopus (166) Google Scholar). The Escherichia coli CopA protein is a Cu(I)/Ag(I)-translocating P-type ATPase involved in copper export and resistance (5Rensing C. Fan B. Sharma R. Mitra B. Rosen B.P. Proc. Natl. Acad. Sci. U. S. A. 2000; 94: 652-656Crossref Scopus (408) Google Scholar). It shows similarity to copper pumps from several prokaryotic and eukaryotic sources, including 31 and 29% identity to the human Wilson's and Menkes' proteins, respectively. The transcription of copA is regulated solely by CueR (GeneBankTMAF318185), a copper- and silver-responsive (7Outten F.W. Outten C.E. Hale J. O'Halloran T.V. J. Biol. Chem. 2000; 275: 31024-31029Abstract Full Text Full Text PDF PubMed Scopus (251) Google Scholar,8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar, 10Petersen C. Moller L.B. Gene ( Amst. ). 2000; 261: 289-298Crossref PubMed Scopus (90) Google Scholar) member of the MerR family of transcriptional activators (Fig. 1). MerR acts as a dimer on the promoter of bacterial mercury resistance (mer) genes recruiting RNA polymerase to the promoter in the process (11Summers A.O. J. Bacteriol. 1992; 174: 3097-3101Crossref PubMed Scopus (184) Google Scholar), and it is activated by Hg(II) binding to 3 cysteine residues in a trigonal planar fashion (12Utschig L.M. Bryson J.W. O'Halloran T.V. Science. 1995; 268: 380-385Crossref PubMed Scopus (147) Google Scholar). CueR activates expression of the copA promoter (PcopA) upon addition of copper or silver, but not nickel, lead, cadmium, mercury, or zinc salts (8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar). Here we show that CueR activates the PcopA promoter in the presence of gold salts and that cysteine residues, equivalent to those in MerR, are required for activation by copper, silver, and gold. All experiments, other than the construction of mutants and some β-galactosidase assays, were carried out in E. coli strain TG1 (K-12, lac-pro supE thi hsdD5 [F′ traD36proA + B + lacIq lacZ M15]). E. coli WK6 Δ(lac-proAB)galEstrA [F′ lacIq lacZΔM15 proAB+] and its mutS215:Tn10variant were used for gapped duplex mutagenesis (13Stanssens P. Opsomer C. McKeown Y.M. Kramer W. Zabeau M. Fritz H.J. Nucleic Acids Res. 1989; 17: 4441-4454Crossref PubMed Scopus (259) Google Scholar). E. coli CSH26Δ(lac-proAB)thi ara ΔrecA (14Miller J.H. Experiments in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1972Google Scholar) was used in some β-galactosidase assays to show that the observed effects were not strain specific. The following plasmids have been described elsewhere: pMU2385 (15Praszkier J. Wilson I.W. Pittard A.J. J. Bacteriol. 1992; 174: 2376-2383Crossref PubMed Google Scholar), pSU18 (16Bartolome B. Jubete Y. Martinez E. de la Cruz F. Gene ( Amst. ). 1991; 102: 75-78Crossref PubMed Scopus (428) Google Scholar), pMUPcopA (8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar), pSUcueR (8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar), pMUPmerT (17Parkhill J. Lawley B. Hobman J.L. Brown N.L. Microbiology. 1998; 144: 2855-2864Crossref PubMed Scopus (13) Google Scholar), and pSUmerR (17Parkhill J. Lawley B. Hobman J.L. Brown N.L. Microbiology. 1998; 144: 2855-2864Crossref PubMed Scopus (13) Google Scholar). Bacterial cells were grown at 37 °C in Luria-Bertani (LB) medium or M9 medium supplemented with casamino acids and thiamine (18Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Press, New York1989Google Scholar). For metal induction assays, sodium chloride was omitted to avoid AgCl precipitation (called SLB and SM9 media). Antibiotics were used as described earlier (17Parkhill J. Lawley B. Hobman J.L. Brown N.L. Microbiology. 1998; 144: 2855-2864Crossref PubMed Scopus (13) Google Scholar). Standard DNA manipulations were carried out as described elsewhere (18Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Press, New York1989Google Scholar). Oligonucleotides were synthesized by Alta Bioscience Ltd., University of Birmingham, by Invitrogen, or by MWG Biotech. Cloned PCR products were sequenced using an PE Applied Biosystems Big DyeTM sequencing kit according to the manufacturers' instructions and analyzed on an ABI 3700 sequencing apparatus. E. coli cells transformed with plasmid pMUPcopA (8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar) or pMUPmerT(17Parkhill J. Lawley B. Hobman J.L. Brown N.L. Microbiology. 1998; 144: 2855-2864Crossref PubMed Scopus (13) Google Scholar) were assayed for β-galactosidase activity (14Miller J.H. Experiments in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1972Google Scholar) as described previously (17Parkhill J. Lawley B. Hobman J.L. Brown N.L. Microbiology. 1998; 144: 2855-2864Crossref PubMed Scopus (13) Google Scholar). E. coli TG1 [cueR::kan] cells containing pMUPcopA(8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar) were transformed with pSUcueR or other pSU18 vectors containing mutant cueR, and induced overnight with 1 mmisopropyl-β-d-thiogalactopyranoside. Cultures for metal induction experiments were grown overnight in the required medium with antibiotic selection to maintain resident plasmids. The overnight culture was diluted 100-fold into fresh medium without antibiotics and grown to mid-exponential phase. Metal induction was performed for 2 h at 37 °C. Assays were performed in triplicate and repeated several times. The direct effect of gold salts on β-galactosidase was tested by a standard enzyme assay using 96 units of β-galactosidase in 1 ml of Z-buffer (14Miller J.H. Experiments in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1972Google Scholar) in the presence of 0, 50, or 150 μmAuCl3. Reactions were started with 0.1 ml 4 mg ml−1 o-nitrophenyl-β-d-galactoside and stopped with 0.5 ml 1 m Na2CO3. Z-buffer is reducing, containing 50 mm 2-mercaptoethanol. No inhibition by gold was seen. The cueR mutants C112S; C120S; C129S,C130S; and H131N,H132N were generated by the gapped duplex method (13Stanssens P. Opsomer C. McKeown Y.M. Kramer W. Zabeau M. Fritz H.J. Nucleic Acids Res. 1989; 17: 4441-4454Crossref PubMed Scopus (259) Google Scholar) using 24- to 30-mer oligonucleotides containing the 1 or 2 mismatched bases. Site-directed mutagenesis by PCR overlap extension (19Horton R. Pease L. McPherson M.J. Directed Mutagenesis: A Practical Approach. Oxford University Press, New York1991: 217-246Google Scholar) was used in later stages for creating H76N,H94N using primer pairs containing complementary mutating nucleotides, and flanking forward and reverse primers containing EcoRI andBamHI sites, on pSU18cueR as template. TruncatedcueR mutants were generated by PCR using a downstream primer containing a stop codon and a BamHI site, with a flanking upstream primer containing an EcoRI site. PCR products were digested with EcoRI and BamHI and cloned intoEcoRI-BamHI-cut pSU18, creating pSUcueR-mutant. All mutated DNA fragments were fully sequenced to confirm the required mutation and screen against undesired mutations. Solutions of cuprous ions, Cu(I), were made freshly on a daily basis as described elsewhere (20Byrd J. Berger R.M. McMillin D.R. Wright C.F. Hamer D.H. Winge D. J. Biol. Chem. 1988; 263: 6688-6694Abstract Full Text PDF PubMed Google Scholar). The dynamic response of the copA promoter was tested at a range of copper, silver, and gold ion concentrations in minimal medium (Fig. 2 A), and maximal responses were observed with additions of 100 μmCuSO4, 2.5 μm AgNO3, or 25 μm AuCl3 for mid-exponential phase cells. Gold toxicity caused a reduction in cell growth at higher metal concentrations. Similar relative concentrations for maximal induction were observed at different growth stages, although the absolute concentrations differed, presumably due to the production of metal-chelating extracellular material. AuCl3 does not inhibit β-galactosidase in vitro (data not shown), so reduction in induction at high gold concentrations is not due to direct effects on the reporter enzyme activity. The values of half-maximal induction, calculated from the induction profiles, were: for copper, ∼21 μm; for gold, ∼15.2 μm; and for silver, ∼1.9 μm; these values include differences in access of the metal to the regulator (sequestration, uptake, and reduction) as well as the induction per se. The values for silver and copper agree with those determined previously (8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar). The increases in induction with increasing metal ion concentrations are different for the three metals. Induction by gold of PcopA was abolished in a disruptant strain, E. coliTG1[cueR::kan R] (8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar). As with previous results for copper and silver (8Stoyanov J.V. Hobman J.L. Brown N.L. Mol. Microbiol. 2001; 39: 502-511Crossref PubMed Scopus (193) Google Scholar), metal induction of PcopA could be recovered by providing the cueRgene in trans (Fig. 2 B). Therefore, CueR activates transcription of PcopA in response to gold ions. To check that the response of CueR to gold was specific, a second member of this class of metal regulators, MerR and its promoter PmerTPAD, was tested in both E. coli TG1 and inE. coli CSH26ΔrecA (17Parkhill J. Lawley B. Hobman J.L. Brown N.L. Microbiology. 1998; 144: 2855-2864Crossref PubMed Scopus (13) Google Scholar). Induction of the mercury resistance promoter, PmerTPAD, was observed only when the merR gene was provided in trans and induced with 1 μm HgCl2, but was not observed after induction with copper, silver, and gold or when merRwas not present, even though cueR was present on the chromosome (data not shown). The growth response of E. coli W3110 ΔcopA (5Rensing C. Fan B. Sharma R. Mitra B. Rosen B.P. Proc. Natl. Acad. Sci. U. S. A. 2000; 94: 652-656Crossref Scopus (408) Google Scholar) to increasing concentrations of AuCl3 in LB medium was indistinguishable from that of the parental W3110 strain (data not shown), indicating that the CopA transporter does not confer gold resistance. The CueR protein contains several cysteine and histidine residues which are candidates for recognition and specific binding of activating metals. Cysteine residues at positions 112, 120, 129, and 130 were mutated to serine and histidines at positions 76, 94, 131, and 132 were altered to asparagine. These variants of the cueR gene were expressed in pSU18 vector (16Bartolome B. Jubete Y. Martinez E. de la Cruz F. Gene ( Amst. ). 1991; 102: 75-78Crossref PubMed Scopus (428) Google Scholar) in E. coliTG1[cueR::kan R](pMUPcopA). The response of the complemented E. coliTG1[cueR::kan R] strain to metals was more variable between experiments than when the wild-typeE. coli TG1 was used, possibly due to the variable copy number of pSU18::cueR plasmids. However, qualitatively clear results were obtained, showing that mutations C112S or C120S eliminated the response to copper, silver, and gold salts (Fig. 3) as did all double and truncated mutants containing one of these mutations (data not shown). Furthermore, none of these mutants responded to lead, mercury, nickel, cadmium, or cobalt. A deletion from position 120 failed to respond to metals, while a deletion from position 128 had an attenuated response. The cysteines at positions C112 and C120 are analogous to C117 and C126 of Tn501 and Tn21 MerR, which are important for binding Hg(II) and activation of the transcription of themer operon (11Summers A.O. J. Bacteriol. 1992; 174: 3097-3101Crossref PubMed Scopus (184) Google Scholar). Alanine A78 was replaced by cysteine to mimic the equivalent essential cysteine C82 in MerR (Fig. 1), but the mutant protein did not activate transcription in response to Hg(II)in vivo (data not shown). Double mutations H131N,H132N and, separately, C129S,C130S, were generated to investigate the importance of these distal, potentially metal-binding, amino acids and were tested in the induction assay (Fig. 3). The mutations H131N,H132N gave no significant change in the response to copper but gave an altered response to silver (decreased) and gold (increased) compared with the wild type. There was a slight increase of the background response without metal. The mutant C129S,C130S also showed relatively small changes in the response to copper and silver but there was a slight decrease in the response to gold. For neither mutant was the specificity of reaction with other metals altered. The retention of a response to copper, silver, and gold salts, albeit altered in magnitude, by these double mutations and by the seven amino acid C-terminal deletion from Gly128indicates that the full CCHH motif at positions 129–132 is not responsible for metal recognition by CueR, as we had originally hypothesized. A decrease of 30–40% but not abolition of the response to copper was found with the separate H76N and H94N mutations (data not shown), indicating that these two histidines are not essential to metal recognition. The effect of different chelating agents on the induction of PcopA was tested (Fig. 4). Bathocuproine disulfonate (21Chen S.H. Liu S.H. Liang Y.C. Lin J.K. Lin-Shiau S.Y. Glia. 2000; 31: 249-261Crossref PubMed Scopus (54) Google Scholar), cuprizone (22Pfeiffer S. Schrammel A. Schmidt K. Mayer B. Anal. Biochem. 1998; 258: 68-73Crossref PubMed Scopus (42) Google Scholar), and neocuproine (21Chen S.H. Liu S.H. Liang Y.C. Lin J.K. Lin-Shiau S.Y. Glia. 2000; 31: 249-261Crossref PubMed Scopus (54) Google Scholar, 22Pfeiffer S. Schrammel A. Schmidt K. Mayer B. Anal. Biochem. 1998; 258: 68-73Crossref PubMed Scopus (42) Google Scholar) were used as copper chelators; EDTA was used as a nonspecific metal chelator. When PcopAwas induced in the presence of 100 μm bathocuproine, activation by Cu(I) was reduced by and the activation by by activation by silver was reduced by and activation by gold was by of the mercury resistance promoter, PmerTPAD, in the presence and 1 μm Hg(II) showed no significant in the presence and of 100 μm activation by Cu(I) was reduced and by that cuprizone may not be specific for as by silver by 29% and activation by gold by both of which may be due to the of relatively more copper ions by the chelator. No was observed for the activation of by of PcopA was observed in the presence of the copper neocuproine even in the of in increase (Fig. 4). The induction by different metals was calculated by from the background in the with but without metal and with the equivalent without the no direct of the The values were as increase in activation by Cu(I), increase in activation by no change in activation by silver or with gold. there was no change in the response of and These results that neocuproine copper into the The background may have due of copper present in SM9 Gold extracellular is not by and activates the copA promoter in their presence to the as in the presence or of these chelators. salts are to in EDTA the activation of the copA promoter to background by all metals (Fig. 4). of the copper had effect on the induction of themer promoter by MerR and mercury, which that all of the effects observed for copper, silver, and gold are and not caused by the chelating is a metal ion with a of and and of to those of silver, but different to those of copper, and J. The Elements. Oxford University Press, Scholar). The the due to the of the a due to the of the for gold, copper, and silver is at No other has the The activation of PcopA by CueR and gold salts was from a biological of The of gold ions that are Cu(I) and in the metal binding of CueR, to activation of is a for this and may be in the reducing of the or in the by a as Biochem. PubMed Scopus Google Scholar, L. Biochem. PubMed Scopus Google Scholar). is in which has from vivo with in vitro experiments, the production of salts is the for the of Cu(I) by gold salts, the the three The of metal required for half-maximal induction of the promoter are extracellular values and and reduction may for the differences between the metal not be involved in and has the for half-maximal for in vitro of the effect of the metals on CueR to be performed to confirm the of gold required for activation of The of effect of a ΔcopA mutation on the gold of growth of E. coli cells may be due to gold salts not by the CopA indicating that there are different for between copper and gold salts, or may be due to change to the of gold as it is to the of the or may not confer resistance to the metal. an effect has been for the copper which Cu(I) and but resistance only to Cu(I) and not to M. A. J. Biol. Chem. 1995; Full Text Full Text PDF PubMed Scopus Google Scholar). The of recognition of Cu(I) and by the CueR regulator to be fully The that cysteine residues and 120 are required for activation by all three metals that the metals to the on the and the of these amino acids to and in MerR (Fig. that the metal binding is in the relative position on both it is not that with cysteine to mimic the equivalent essential cysteine in MerR (Fig. has no effect on CueR response to as the between Cys112 and Cys120 in CueR is different to that between and in MerR, and more and effects be in metal recognition. The described from M.J. Mol. Microbiol. PubMed Scopus Google Scholar) contains cysteines equivalent to Cys112 and but a equivalent to the CCHH motif at positions providing of The of metal binding require on the CueR The of MerR family regulators are for to other than metals J. Biol. Chem. 2001; Full Text Full Text PDF PubMed Scopus Google Scholar, 2001; PubMed Scopus Google Scholar), and cysteines and 120 the of similarity between CueR and these The for copper in to be The toxicity of copper is due in part to its ions can be by to ions and which can and in biological and enzymes S. G. J. Biol. PubMed Scopus Google Scholar). This of is in with copper due to a in the copper protein Mol. 2001; Full Text Full Text PDF PubMed Scopus Google Scholar). is a by of and other into and and of these by activated from J. M. 1998; PubMed Scopus Google Scholar). copper were found in the of M.C. S. S. PubMed Scopus Google Scholar), copper toxicity and were in a with Wilson's P. Google Scholar), and a was to in R. M. F. G. L. Res. 2000; PubMed Scopus Google Scholar). is that by copper, gold may as an of the expression of copper and the copper homeostasis the by copper from it The Medical provided in We are to J. L. T. and R. for to R. for and to P. A. C. and C. for of and plasmids.