1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Transcriptional response of Escherichia coli upon exposure to external copper was studied using DNA microarray and in vivo and in vitro transcription assays. Transcription of three hitherto-identified copper-responsive genes, copA (copper efflux transporter), cueO (multicopper oxidase) and cusC (tripartite copper pump component) became maximum at 5 min after addition of copper sulphate, and thereafter decreased to the preshift levels within 30 min. Microarray analysis at 5 min after addition of copper indicated that a total of at least 29 genes including these three known genes were markedly and specifically affected (28 upregulated and one downregulated). Transcription of the divergent operons, cusCFB and cusRS, was found to be activated by CusR, which bound to a CusR box between the cusC and cusR promoters. Except for this site, the CusR box was not identified in the entire E. coli genome. On the other hand, transcription of copA and cueO was found to be activated by another copper-responsive factor CueR, which bound to a conserved inverted repeat sequence, CueR box. A total of 197 CueR boxes were identified on the E. coli genome, including the CueR box associated with the moa operon for molybdenum cofactor synthesis. At least 10 copper-induced genes were found to be under the control of CpxAR two-component system, indicating that copper is one of the signals for activation of the CpxAR system. In addition, transcription of yedWV, a putative two-component system, was activated by copper in CusR-dependent manner. Taken together we conclude that the copper-responsive genes are organized into a hierarchy of the regulation network, forming at least four regulons, i.e. CueR, CusR, CpxR and YedW regulons. These copper-responsive regulons appear to sense and respond to different concentrations of external copper.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Copper is an essential trace metal element for both eukaryotes and prokaryotes, and is associated with various metal enzymes, which are involved in electron transport, reduction of nitrite and nitrous oxides, and electron carriers. Intracellular free copper is, however, toxic even at low concentrations, and thus both prokaryotes and eukaryotes carry cellular systems to maintain the copper homeostasis (for instance Rensing and Grass, 2003).

The intracellular level of copper in Escherichia coli is controlled by the export of excess copper, but the entire systems of copper uptake and intracellular copper delivery are not fully understood. Two regulatory systems, the CueR and CusR systems, have been identified to be involved in transcription regulation of the genes for copper homeostasis (Rensing et al., 2000; Rensing and Grass, 2003). CueR, a MerR-family transcription factor, stimulates copper-induced transcription of both copA encoding Cu(I)-translocating P-type ATPase pump (exporter), that is the central component for maintenance of the copper homeostasis, and cueO encoding a periplasmic multicopper oxidase for detoxification (Outten et al., 2000; Petersen and Moller, 2000). In the presence of excess copper, CopA plays a key role for export of free copper from cytoplasm to periplasm. Under aerobic conditions, the sensitivity to external copper increases in E. coli mutants lacking CopA and/or CueO, indicating that the CueR-regulated system (or the CueR regulon) plays a key role of aerobic copper tolerance (Grass and Rensing, 2001; Outten et al., 2001).

The cusCFBA operon encoding components of the tripartite transporter (exporter) (Franke et al., 2001, 2003) is organized on the E. coli genome next to the genes for CusSR two-component system, which regulates cusCFBA transcription in copper-dependent manner (Munson et al., 2000). The cusSR genes are, however, transcribed in opposite direction from cusCFBA. The cusR null mutant show a high level of copper sensitivity only in the absence of oxygen. The induction level of cusC at low concentrations of copper is higher under anaerobic conditions than aerobic conditions. These observations indicate that the CusSR system is involved, in addition to the CopA copper efflux, to establish full copper tolerance under anaerobic conditions, where copper toxicity increases (Outten et al., 2001).

Even though two regulatory systems, the CueR and CusR regulons, have been identified in E. coli for controlled transport of excess copper and for maintenance of the copper homeostasis, but the entire genetic systems for response to environmental copper is not clear yet. In order to get insights into the genome-wide regulation of E. coli upon exposure to external copper, we have performed in this study the DNA microarray to search for the whole set of copper-responsive genes. Detailed in vivo and in vitro analyses of transcription regulation was then carried for the copper-responsive genes thus identified. Results will show that besides the known CusR and CueR regulons, two additional cellular systems, CpxR and YedW regulons, are involved for defence against increased levels of external copper. Physiological meanings of the four copper-response regulons are discussed.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Transcriptional response of E. coli upon exposure to copper-shock

Up to the present time, only three genes, copA (copper efflux transporter), cueO (multicopper oxidase) and cusC (tripartite copper pump component), have been identified, which are induced in E. coli upon exposure to external copper ion. The present study was conducted to reveal the entire genetic system that operates in response to the copper ion. For this purpose, we first analysed kinetics of transcriptional response of the known copper-responsive genes upon sudden exposure to external copper. When CuSO4 was added at various concentrations into E. coli W3110 type-A culture, cell growth was significantly retarded above 1.0 mM (data not shown). Transcription pattern was then analysed at various times after addition of 0.5 mM CuSO4. The S1 assay indicated that in the absence of copper addition, transcription was hardly detected for cueO, cusC and cusR (Fig. 1A, none) [cusR is located next to the cusC operon but is transcribed toward opposite direction from the cusC operon (Fig. 1C)], but copA transcription was always detected at a low level (Fig. 1A, none). After the addition of CuSO4, transcription of all four genes was stimulated, reaching the maximum level at 5 min after copper addition, and then gradually decreased (Fig. 1A, +CuSO4). After 30 min, the level of cusC and cusR transcripts decreased to half of the maximum level, whereas the transcripts of copA and cueO decreased to the preshift levels.


Figure 1. Transcription regulation of the cusS and cusR genes by CusR. A. Escherichia coli W3110 type-A was grown at 37°C in LB medium with shaking. In the middle of exponential growth phase, the culture was divided into two equal aliquots, and one aliquot was treated with CuSO4. Total RNAs were prepared at 0 (lanes 2 and 7), 5 (lanes 3 and 8), 10 (lanes 4 and 9) and 30 min (lanes 5 and 10), and analysed by S1 nuclease assay. The radio-labelled gene-specific probes for cusRp (a), cusCp (b), copAp (c) and cueOp (d) were prepared by PCR using primer pairs as described in Experimental procedures (for sequences see Table 3). Total RNA (100 µg) was subjected to S1 nuclease mapping using 104 cpm of the 32P-end-labelled probe as described in Experimental procedures. AG lanes (lanes 1 and 6) show the A + G ladder of Maxam–Gilbert sequencing reaction. B. CusR (4 µM) was incubated with 32P-labelled cusS-cusR gene-specific probes (1.6 µM) for 10 min at 37°C. The mixtures were digested with DNase I for 30 s at 25°C, and then the digested products were extracted with phenol, precipitated with ethanol, and analysed by 6% PAGE containing 8 M urea. AG lanes (lanes 1 and 6) show the A + G ladder of Maxam–Gilbert sequencing reaction. Bars on right indicate the protected regions, whereas arrows indicate the hypersensitive sites. The numbers represent the distance from the transcription start site of cusC. C. The sequence represents the intergenic region between transcription initiation sites of cusC and cusR. The CusR-binding regions, which were determined as described in Fig. 1B, are indicated by dotted lines. These CusR-binding sequences include the inverted repeat CusR box. The −10 and −35 regions of the cusCp and cusRp promoters are shown by grey boxes. The DNase-hypersensitive sites are indicated by vertical arrows.

Download figure to PowerPoint

The S1 assay also indicated that the initiation site of copA transcription was the same as that reported (Munson et al., 2000; Outten et al., 2000), whereas the initiation site of cusC transcription was two nucleotides upstream from the site previously identified (Munson et al., 2000). Transcription start sites were determined for the first time for both cusR and cueO.

Genome-wide transcription profile upon exposure to copper-shock

To search for the whole set of copper-responsive genes, we performed the DNA microarray analysis of E. coli cells after copper-shock. Among a total of 4000 genes on the DNA chip, transcription of a total of 37 genes exhibited more than threefold increase, while more than threefold decrease in mRNA level was detected only for the aphA gene encoding acid phosphatase. Among a total of 38 genes that showed marked increase or decrease in response to copper-shock, nine genes [cysK (cysteine synthase A), cysN (sulphate adenylyltransferase), cysP (thiosulphate-binding protein), hyfG (formate hydrogenlyase), htrA/degP (periplasmic serine protease), ybiK, yeeD, yeeE and zntA (zinc efflux P-type ATPase)] were also induced by zinc ion (K. Yamamoto, T. Miki and A. Ishihama, in preparation), indicating that a total of 29 genes [28 upregulated and one downregulated genes] were specifically affected by copper-shock (Table 1). The hitherto-identified copper-responsive genes, i.e. the cusCFBA operon and the copA and the cueO genes, are all included in this group of 28 highly induced genes. However, cusR transcript for the response regulator of cusRS two-component system was not included in this group, but transcript of the cusS gene, immediate downstream of cusR, increased more than twofold by copper-shock (data not shown).

Table 1. Genes markedly affected by copper-shock.
  • *

    This list shows the genes, which are more than threefold upregulated by copper, including hitherto identified copper-response genes ().

  • **

    One exception is adhA (**), which showed more than threefold downregulation.

Regulon gene
 cusB*copper/silver transporter
 cusC*copper/silver transporter
 cusF*periplamic copper-binding protein
 yedVputative sensor kinase
 yedWputative response regulator
 copA*copper exporter
 cueO*periplasmic multicopper oxidase
 moaBmolybdenum cofactor synthesis protein B
 moaCmolybdenum cofactor synthesis protein C
 moaDmolybdopterin [MPT] converting factor subunit 1
 moaEmolybdopterin [MPT] converting factor subunit 2
 cpxPperiplasmic protein; extracytplasmic stress response
 ebrsmall multidrug exporter; ethidium bromide resistance
 spyperiplasmic protein related to spheroplast formation
 yccAmembrane-associated putative carrier/transporter
 yecIferritin-like protein
Unclassified genes
 aldHaldehyde dehydrogenase
 aphA**acid phosphatase
 mutLmethyl-directed mismatch repair enzyme
 tesAacyl-CoA thioesterase I
 htpXheat shock-induced membrane protease
 soxStranscription factor for super-oxide response regulon
 yciWputative oxidoreductase
 yeeFputative amino acid/amine transporter

The changes of transcription pattern observed above were mainly attributable to the addition of external copper, because no significant changes were observed in pH and oxygen level in culture media after copper addition (data not shown).

Regulation of the copper-responsive genes by CusS-CusR two-component system

Two transcription factors, CusR and CueR, have been identified to be involved in copper-response transcription regulation. In an E. coli mutant lacking CusR, transcription of not only cusS but also cusR was not detected as analysed by S1 mapping, but transcription of both copA and cueO was detected (data not shown). To get insight into the regulatory mode of CusR action, we determined the binding site of CusR on the cusC and cusR promoters by DNase-I footprinting. Results, shown in Fig. 1B, indicate that CusR binds at a site between the cusC and cusR genes. The protected region by CusR were −42 to −67 on top strand and −44 to −70 on bottom strand [the numbers represent the distance from the initiation site of cusC transcription] (for sequence see Fig. 1C). This CusR binding region, −42 to −67 on the cusC promoter, corresponds to −41 to −67 on the cusR promoter. Two DNase hyper-sensitive sites appeared in the presence of CusR (Fig. 1B and C). Taken together, we conclude that the same CusR molecule activates transcription towards both directions. This CusR-binding sequence between the cusC and cusR promoters (Fig. 1C) includes the inverted repeat, AAAAT GACAANNTTGTCATTTT, which has been proposed to constitute the CusR box (Munson et al., 2000). Except for the cusC/cusR-associated CusR box herewith identified, however, we could not detect the CusR box in the entire E. coli genome even after allowing 2-nucleotide mismatch.

To search for the genes under the control of CusR, we next performed the DNA microarray analysis using a CusR-defective mutant JD20810 [KP7600 cusR::Km]. Copper shock-induced transcripts were compared between the cusR mutant and its parent KP7600. Among a total of 28 copper shock-induced genes detected in the wild-type strain, transcription induction was not detected in the mutant for only three genes, cusC, cusF (ylcC) and cusB (Table 1), which are all organized in the same cusCFBA operon. The other 25 genes exhibited essentially the same level induction in both wild-type and cusR mutant strains. The results of microarray assay agree well with the finding that the CusR box is located only in the intergenic region between the cusRS and cusCFB operons.

Regulation of the copper-responsive genes by CueR

Another copper-responsive transcription factor CueR activates transcription of copA and cueO in the presence of copper ion (Outten et al., 2000). To get insight into the mode of CueR action, we carried out DNase I footprinting for both copA and cueO promoters. On the copA promoter, CueR protected −34 to −10 on top strand and −36 to −12 on bottom strand (Fig. 2A), while on the cueO promoter, it protected −41 to −17 on top strand and −42 to −19 on bottom strand (Fig. 2B). The CueR binding site within the copA promoter agrees well with that previously identified by other groups (Outten et al., 2000; Stoyanov et al., 2001). Here we identified the second CueR binding sequence on the cueO promoter and thus allowed the prediction of CueR binding consensus sequence, which includes the inverted repeat, CCTTCCNNNNNNGGA AGG, designated hereafter as the CueR box (Fig. 2C). This CueR box sequence agrees with the dyad symmetrical sequence proposed by Stoyanov et al. (2001). DNase-I footprinting showed two hypersensitive sites within the middle of CueR box on all four DNA probes (Fig. 2A and B).


Figure 2. Transcription regulation of the copA and cueO genes by CueR. A. DNase-I footprinting was performed, as described in Fig. 1, using copA-specific probes. 32P-labelled copAp-specific probes (1.6 µM) were incubated with CueR (4 µM) and then was digested by DNase I. Undigested products were purified and analysed by 6% PAGE containing 8 M urea. Bars and arrows, on right of each panel, indicate the protected regions and the hypersensitive sites respectively. The numbers, added to the protected region bars, represent the distance from the transcription start site of copA. AG lanes (lanes 1 and 4) show the A + G ladder of Maxam–Gilbert sequencing reaction. B. DNase-I footprinting of CueR on the cueOp promoter was carried out using cueO-specific probes. C. The sequences show the copAp and cueOp promoter regions. The CueR binding regions, which were determined in Fig. 1C, are indicated by dotted lines, which include the inverted repeat CueR box. The −10 and −35 regions on each promoter are boxed. The DNase-hypersensitive sites are indicated by vertical arrows.

Download figure to PowerPoint

We then searched for the CueR box within the entire E. coli genome. By allowing two nucleotides mismatch, a total of 129 CueR box-like sequences were identified, of which 74 are associated with the known genes but 55 are with the unidentified genes. Among the known genes with the CueR box, copper-response transcription activation was observed for the moaABCDE operon, besides the copA and cueO genes (see Table 1). The moaABCDE operon encodes the enzymes required for molybdopterin biosynthesis. Possible influence of copper on transcription of the mao operon was then analysed in details. In the presence of copper addition, two transcription start sites, moaAp1 and moaAp2, were detected [Fig. 3A(a)], which are the same as those reported (Anderson et al., 2000). The upstream promoter, moaAp1, was repressed by copper addition while moaAp2, the downstream promoter, was not affected by copper-shock. In addition to the maoAp1 and maoAp2 promoters, we detected the copper-induced transcription start site, moaBp, at 78-bp upstream of moaB start codon [Fig. 3A(b), major band]. Besides this major maoB RNA, a low level of second maoB transcript was detected, of which 5′-terminus was located at around −40 bp from the maoB initiation codon [Fig. 3A(b), minor band]. However, we can not exclude the possibility that the second RNA represents a degradation product of full-sized maoB transcript. In any event, both maoB transcripts significantly increased after the addition of copper. Therefore, we concluded that the moaABCDE operon carries at least three promoters, moaAp1, moaAp2 and moaBp, in which only the moaBp promoter is induced by copper-shock. The CueR box organized in the moa operon is located upstream of the first gene, maoA, the CueR box centre being at 16-bp upstream of its translation initiation site (Fig. 3B). Because neither CueR box nor CusR box exists within the moaB promoter region, the mechanism of copper-induced transcription activation from maoBp awaits further studies.


Figure 3. Transcription of the moaABCDE operon. A. Escherichia coli KP7600 (W3110 derivative) and JD20810 (cusR::Kmr) were grown in LB medium until the middle of exponential phase, and then divided into two portions. For one aliquot, CuSO4 was added (lanes 3 and 5). Total RNAs were subjected to S1 mapping using moaA-specific probes (Table 4). Probe sequences are shown in Table 3. AG lanes (lanes 1 and 6) show A + G ladder of Maxam–Gilbert sequencing reaction. Arrows show the undigested probe DNAs by S1 nuclease. The numbers, on left of each panel, represent the distance from the respective initiation codon. B. Transcription initiation sites are shown along the moa operon. The CueR box is located upstream of the initiation codon of moaA but downstream of its transcription sites.

Download figure to PowerPoint

Intracellular levels of CusR and CueR transcription factors

Possible influence of the increase in external copper on the intracellular level of two copper-responsive transcription factors, CusR and CueR, was examined by quantitative immunoblot assay. When the culture of wild-type E. coli W3110 was exposed to 0.5 mM CuSO4 for 5 min in the middle of exponential growth, no significant change was observed in the level of CusR (Fig. 4A, short exposure). The CusR level also stayed constant at 1–2% the level of RNA polymerase α subunit even after prolonged exposure to high-level copper (Fig. 4A, long exposure). The level of RNA polymerase α subunit stays constant at a level of 4000–5000 molecules per genome-equivalent of DNA (Ishihama, 2000), and thus the level of CusR might be 50–100 molecules per genome.


Figure 4. Intracellular levels of copper-response transcription factors. A and B (short exposure). Escherichia coli W3110 was grown in LB medium until either the middle of exponential phase or the stationary phase, and then CuSO4 was added at the final concentration of 0.5 mM. After 5 min, cells were harvested and the whole cell lysates were subjected to quantitative immuno-blot analysis of CusR (A, short exposure) and CueR (B, short exposure) proteins as described in Experimental procedures. A and B (long exposure). Escherichia coli W3110 was also grown in LB medium containing 0.5 mM CuSO4 until either the middle of exponential growth or stationary phase. The amounts of CusR (A, long exposure) and CueR (B, long exposure) were determined as described in Experimental procedures.

Download figure to PowerPoint

On the other hand, the intracellular level of CueR in exponential-phase E. coli was about 1.5% the level of RNA polymerase α subunit, and after short exposure to copper, it increased about twofold to reach 3% the level of α subunit. After prolonged exposure to copper, the CueR level increased to more than threefold to 5.5% the level of α subunit (about 250 molecules per genome) (Fig. 4B). In the stationary phase of cell growth, the level of CueR increased even in the absence of external copper, and thus further increase of CueR was not significant after short-term exposure to copper. After prolong exposure to copper, however, detectable increase was observed for the CueR level, reaching to the maximum level of abut 8% (320–400 molecules per genome) the level of α subunit (Fig. 4B). We then conclude that the induction of CueR regulon is, at least in part, resulting from the increase in CueR level.

Regulation of the copper-responsive genes by the CpxA-CpxR two-component system

Our microarray assays indicated that the addition of copper induces transcription of cpxP and spy (see Table 1), which are both under the control of CpxA-CpxR two-component system (Raivio and Silhavy, 1999). Among the set of highly upregulated genes by copper (see Table 1), we also identified additional 10 genes, ebr, JW1832, tesA, ybaJ, yccA, ycfS, ydeH, yebE, yecI and yqjA, which are all upregulated upon over-expression of CpxR (H. Aiba, pers. comm.). To confirm whether CpxR is directly involved in the regulation of these 12 copper-induced genes, S1 nuclease assay was performed for RNAs from both CpxR-defective mutant BW27559 [BW25113 ΔcpxRA] and its wild-type parental strain BW25113. Total RNA was isolated from each culture at 5 min after addition of CuSO4, and subjected to S1 nuclease assay using gene-specific probes. Except for tesA, transcripts of all 11 genes (cpxP, ebr, spy, JW1832, ybaJ, yccA, ycfS, ydeH, yebE, yecI and yqjA) were found to be induced by copper in wild-type BW25113 but not in the cpxAR null mutant (Fig. 5). Transcription start site of ebr was not identified with probes used in this study (data not shown). Transcript of tesA was detected but not induced by copper-shock in both wild-type and cpxAR null-mutant cells (data not shown). The promoters from these 10 genes carry sequences, which are similar to the complementary sequence of the proposed CpxR recognition site, GTAAANNNNNGTAAA (Pogliano et al., 1997; Wulf et al., 2002).


Figure 5. Copper-inducible transcription regulation of the Cpx regulon genes. Total RNA was isolated from each of exponential-phase cultures, BW25113 (lanes 2 and 3) and BW27559 (lanes 4–5) with (lanes 3 and 5) and without (lanes 2 and 4) treatment with CuSO4 for 5 min, and subjected to S1 nuclease assay as described in Fig. 1. Gene-specific probes used are: cpxP probe I (a); spy probe L (b); JW1832 probe K (c); ybaJ probe N (d); yccA probe O (e); ycfS probe P (f); ydeH probe Q (g); yebE probe R (h); yecI probe S (i) and yqjA probe T (j) (for probes see Table 4). Lane AG indicates the Maxam–Gilbert sequencing reaction ladder.

Download figure to PowerPoint

Copper-induced transcription activation of yedWV two-component system

Our DNA microarray analysis also showed that transcription of putative two-component system, yedW and yedV, each encoding a response regulator and a sensor kinase, respectively, was significantly increased by copper-shock in wild-type E. coli but not in cusR-null mutant JD20810 (Fig. 6A). This finding indicates that this putative two-component system is under the control of CusR, and regulates another set of genes, tentatively designated as YedW regulon. To confirm this observation of copper-induced transcription of yedW and yedV, S1 nuclease assay was carried out for total RNAs from wild-type KP7600 and cusR-defective mutant JD20810. Transcript of yedV (the down-stream gene) was detected after copper-shock in wild-type but not in cusR-null mutant (Fig. 6B) even though there is no CusR box-like sequence in the yedVp promoter region. On the other hands, the yedW (the upstream gene) transcript was not detected with and without copper-shock in both KP7600 and JD20810 (Fig. 6B). Transcription of yedW transcript was, however, markedly enhanced when YedW (the response regulator) was over-produced using an expression plasmid (Fig. 6B). A direct repeat sequence, TGCTNCCG, was identified in the yedWp promoter, suggesting that YedW binds to this repeat sequence and activates yedW transcription in autogenous manner.


Figure 6. Organization of the yedW-yedV operon. A. Organization of the yedVW operons. Initiation sites and directions of transcription are shown by arrows. B. Escherichia coli KP7600 and JD20810 (cusR::Kmr) were grown in LB medium and at exponential growth phase, treated with (lanes 3 and 5) and without (lanes 2 and 4) CuSO4. As controls, E. coli W3110 type-A transformants containing pCA24N (lane 7) or pCA24NΔGFP (lane 8) and the yedW, expression vectors were grown in LB medium, containing ampicillin. In the middle of exponential growth phase, total RNAs were prepared and analysed by the S1 nuclease assay as described in Experimental procedures. Probes used were: yedW probe U (left panel) and yedV probe V (right panel) (for probes see Table 4). After hybridization and digested with S1 nuclease (Takara), the protected probe DNAs were analysed by 6% polyacrylamide gel electrophoresis containing 8 M urea. AG lanes (lanes 1 and 6) show the A + G ladder of Maxam–Gilbert sequencing reaction.

Download figure to PowerPoint


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The copper stimulon

Copper homeostasis in E. coli is maintained mainly by controlling the export of excess copper out of the cells. At present, our knowledge of the genes that respond to variation in external copper level is limited. The copA gene codes for Cu(I)-translocating P-type ATPase pump (Outten et al., 2000; Petersen and Moller, 2000), while the cueO gene encodes the enzyme that is considered to convert periplasmic Cu(I) into the less-toxic oxidative form Cu(II). Under certain conditions such as anaerobic circumstances, the cusCBA genes coding for tripartite copper-transporter is considered to play the copper resistance (Franke et al., 2001, 2003). To reveal the entire system of genome-wide response to external copper, we first carried out in this study a systematic search for copper-responsive genes by using DNA microarray technique.

In exponential-phase E. coli, the response to sudden exposure to copper took place within a few minutes, leading to rapid increase in mRNA level for a set of the copper-responsive genes (or the copper stimulon). The copper stimulon mRNAs, however, decreased to preshift levels within next 30 min as monitored by measuring transcripts of the hitherto-identified copper-responsive gene (see Fig. 1). We have notified such a quick transcriptional response when E. coli is exposed to chemical compounds and other metals (Yamamoto et al. in preparation).

Our microarray assays indicated that the copper-stimulon includes at least 28 induced genes and one repressed gene. Among 21 induced genes (except for seven genes of unknown function), the majority (more than half) appears to be involved in cellular processes for detoxification of copper, including seven genes for transporters (cusB, cusC, cusF, copA, ebr, yccA and yeeF), three genes for periplasmic proteins (cpxP, cueO and spy) and four genes for molybdenum cofactor biosynthesis (moaB, moaC, moaD and moaE). Six copper-induced transporter genes constitute five different systems: (i) CusABC and CopA, the copper exporters involved in transport of copper ion out of the cells; (ii) Ebr (alternatively QacE), a small multidrug resistance (SMR) family pump; (iii) YccA, a putative membrane-associated transporter and (iv) YeeE, a putative membrane-associated amino acid/amine transporter. Ebr and YccA form transport systems, which must be involved in export of multicompounds including copper ion, because the expression of these genes are under the control of CpxAR that senses denatured periplasmic proteins caused by several stress (see below for detailed discussion). All these membrane proteins might play roles in sensing and detoxification of copper.

The moa operon encodes enzymes required for molybdopterin biosynthesis. Transcription of moa is controlled at two sigma 70 (RpoD)-dependent promoters. The distal promoter is the site of the anaerobic enhancement, which is Fnr-dependent, while the proximal promoter is activated by transcription factor ModE (Anderson et al., 2000). Under conditions of high molybdate availability, transcription from this downstream promoter is repressed through yet unidentified mechanism. Tungstate brings about the loss of this molybdenum cofactor repression. Here we found that the upstream promoter of moaA operon is repressed by copper (see Fig. 3). In addition, we detected the copper-induced transcription start site at 78-bp upstream of the start codon of the second gene moaB in this operon. These observations altogether indicate that the moa operon is controlled by the availability of various metals in environments.

Besides these membrane proteins, copper induces a set of the genes for cysteine biosynthesis as in the case of zinc, leading to the increase in the intracellular level of cysteine (Yamamoto et al. in preparation). One possibility is that high levels of intracellular cysteine chelate excess metal ions such as zinc and copper. This prediction is being examined.

Two copper-response transcription factors, CueR and CusR

Regulation involving the copper-response factor CueR is the main system under aerobic conditions (Outten et al., 2001). Under the aerobic experimental conditions herewith employed, CueR senses lower concentrations of copper than the CusSR system, and the induction levels of CueR-dependent copA and cueO after exposure to copper are higher than those of cusC and cusR, which are controlled by CusSR (see Fig. 1).

CueR, a MerR-family transcription factor, binds to the copA promoter region, including the extended spacer between −10 and −35 signals of 19-bp in length. As in the case of MerR, the copper-bound CueR may induce a conformational change of the copA promoter in the CueR-copA promoter-RNA polymerase ternary complex so as to reduce the actual distance between −10 and −35 signals (Outten et al., 2000; Stoyanov et al., 2001).

On the other hand, the induction levels of cusC and cusR, which are regulated by copper-response CusSR two-component system, was not so high under the current experimental conditions. Activation of the CusSR-regulated genes depends on the phosphorylation of CusR by phosphorylated CusS (Yamamoto et al., 2004). The low-level induction of the CusSR-regulated genes, cusC and cusR, suggests that the experimental conditions are not suitable for self-phosphorylation of the CusS sensor kinase and/or trans-phosphorylation of CusR by phosphorylated CusS. The CusSR system may play an important role in copper tolerance under the anaerobic conditions because self-phosphorylation of CusS is activated under the anaerobic conditions (K. Yamamoto and A. Ishihama, unpubl.). Under the aerobic conditions, the CusSR system is activated only at higher concentration of copper (> 0.5 mM). Possible influence of pH changes by the addition of CuSO4 may also result in marked changes in the genome-wide transcription pattern. Under the experimental conditions herein employed, however, the pH change was negligible (data not shown).

Involvement of CpxAR and YedVW two-component systems in copper response

One interesting finding of our transcriptome analysis is that copper induces a total of 14 genes, including dsbA, htrA/degP and ppiA, which are all under the control of CpxAR two-component system (see Table 1 and Fig. 5). The CpxA sensor kinase is activated after recognition of the denaturation of membrane proteins (Raivio et al., 1999), leading to activation of the CpxP protease for degradation of denatured proteins. The induction of CpxR regulon may be a reflection of the denaturation of some membrane proteins upon exposure to increased levels of external copper.

The finding of involvement of the CpxAR system in copper response agrees well with the previous report that the gain-of-function mutant, termed as cpxA*, which retains the kinase activity but lack the ability to dephosphorylate the phosphorylated CpxR, showed the increase in sensitivity to external copper (Raivio and Silhavy, 1997; Wulf and Lin, 2000). Moreover, mutations in the nlpE gene, whose overproduction activates the CpxA sensor kinase, lead to decrease in the tolerance against copper (Gupta et al., 1995). These observations altogether suggest that the CpxAR system is activated after sensing copper ion via NlpE outer membrane protein, and stimulates the transcription initiation of various genes organized in the CpxR regulon (Fig. 7).


Figure 7. Hierarchy within the copper stimulon. Multiple regulons are involved for maintenance of copper homeostasis. For details see text. X and Y, unidentified effectors.

Download figure to PowerPoint

The yedV and yedW genes are considered to code for a sensor kinase and a response regulator, respectively, and are organized into a single operon (see Fig. 6A). We identified two promoters within this operon, one upstream of the first gene yedW, and another upstream of the second gene yedV (or between the yedW and yedV genes). Under the steady state of exponential growth, both yedWp and yedVp promoters are silent, but only the yedVp promoter is activated after copper-shock, suggesting that only YedV sensor kinase is activated upon sudden exposure to copper. Interestingly, the yedVp was not induced by copper in cusR null mutant, even though it is not under the direct control of CusR (see Fig. 6). Recently, we found that the YedV kinase transfers phosphate not only to YedW but also CusR (Yamamoto et al., 2004). Thus, the activation of YedV by copper-shock may play a role in the activation of CusR. Along this line, it could be possible that YedV responds to yet unidentified external signal(s), leading to activation of not only YedW but also CusR via signal cross-talk.

Taken together we propose that the copper stimulon in E. coli consists of at least four regulons, each sensing various environmental signals including copper and being connected each other through signal cross-talks (Fig. 7).

Cross-talks within the copper stimulon

The cellular systems that operate in response to changes in external metal have been identified for magnesium, iron, zinc, copper, cadmium, silver, cobalt. In some cases, the metal signal is monitored by the sensor, which directly regulates transcription of the metal responsive genes, but in other cases, the regulation system is composed of two components, the sensor and the response regulator. Little is known, however, about the metal specificity of each metal response regulation system.

Cross-talk between different metal response systems has been indicated. In the case of copper-response systems, for instance, not only copper but also silver and gold are known to bind to the C-terminus of CueR (the sensor/regulator protein of the copper-response one-component system) and stimulate transcription in vivo and in vitro of the CueR-dependent copA gene (Stoyanov et al., 2001; Stoyanov and Brown, 2003; Changela et al., 2003). Franke et al. (2001) reported that the metal sensitivity of CusS, the sensor of copper-response two-component system, is rather higher against silver than copper. We have performed a systematic analysis of cross-talks in signal transduction between two component systems (Yamamoto et al., 2004).

Cross-talk also takes place at the step of metal transport. For instance, CopA ATPase, the copper exporter, is used for export of not only copper but also silver (Stoyanov et al., 2003). The CusABC pump is also suggested to play a role in silver export (Franke et al., 2001). These findings altogether indicate that the cross-talk between metal response systems takes place at various steps, including (i) the metal sensing (in the case of one-component systems) or both the metal sensing and the signal transduction via protein trans-phosphorylation (in the case of two-component systems) and (ii) the import or export of metals through the membrane transporters. To get the whole view of cross-talk between metal response systems and to understand the physiological meanings of cross-talks, systematic studies are needed at each step of these processes.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains, plasmids and growth conditions

Escherichia coli strains and plasmids used in this study are shown in Table 2. Escherichia coli cells were grown at 37°C under aerobic condition in Luria-broth (LB) medium. Cell growth was monitored by measuring the turbidity with a Klett-Summerson photometer. The culture conditions used in this study were fixed as follows: a single colony was isolated from overnight culture on LB agar plate and inoculated into 5 ml of fresh LB medium. The overnight liquid culture was diluted 100-fold into 200 ml of fresh LB medium, and then the incubation was continued at 37°C with reciprocal shaking (120 min−1).

Table 2. E. coli strains and plasmids used in this study.
Bacterial strains
 W3110 type-AWiild-type with complete σ set Jishage and Ishihama (1997)
 KP7600W3110 type-A lacIqlacZΔM15 galK2 galT22 Shimada et al. (2004)
 JD20810KP7600 cusR::KmrT. Miki (unpubl.)
 BW25113FlacIqrrnB3 lacZ4787 hsdR514 (araBAD)567 Δ(rhaBAD)568 rph-1 Haldimann and Wanner (2001)
 BW27559BW25113, ΔcpxRA Oshima et al. (2002)
 pET21a(+)plasmid vecto for overproductionNovagen
 pCueRpET21a(+), cueRThis study
 pCusRpET21a(+), cusRThis study
 pCA24NΔGFPcloning vector Mori et al. (2000)
 19–4DΔGFPpCA24NΔGFP yedWH. Aiba (unpubl.)

Preparation of 32P-end-labelled DNA probes

Primers (10 pmol, for sequences see Table 3), were phosphorylated with 10 mCi [γ-32P]ATP (5000 Ci mmol−1) by T4 polynucleotide kinase (Toyobo). The promoter fragments were amplified by polymerase chain reaction (PCR) using W3110 genomic DNA as template and Ex Taq DNA polymerase (Takara). The probes prepared using the pair of primers, i.e. 32P-labelled primer and unlabelled reverse primer, were listed in Table 4. The labelled fragment was purified by PAGE. The radioactivity of a labelled fragment was measured with liquid scintillation counter.

Table 3. Oligonucleotides used in this study.
NameSequence (5′–> 3′)
For construction of protein expression plasmids
For S1 mapping
Table 4. Probes used in this study.
NamePromoter 32P-labelled oligoCold oligoLength (bp)
Probe A cusC-cusR CUSCR-1CUSCF-1200
Probe B cusC-cusR CUSCF-1CUSCR-1200
Probe C copA COPAR-1COPAF-1290
Probe D copA COPAF-1COPAR-1290
Probe E cueO YACKR-1YACKF-1240
Probe F cueO YACKF-1YACKR-1240
Probe J ebr EBRPREBRPF540
Probe K JW1832 JW1832PRJW1832PF600
Probe L spy SPYPRSPYPF420
Probe O yccA YCCAPRYCCAPF-1330
Probe T yqjA YQJAPR-1YQJAPF270

Preparation of total RNA from copper-treated E. coli culture

Escherichia coli was grown in LB medium at 37°C until early log-phase (20 units measured with Klett-Summerson photometer) and separated two fractions. Copper sulphate was added to one culture fraction and not to another. After 5 min shaking at 37°C, cells were harvested and total RNAs were prepared with hot phenol method. In brief, total RNA was extracted with acid-saturated phenol and precipitated with ethanol. After digestion with RNase-free DNase I (Takara), total RNA was extracted with phenol and precipitated with ethanol, dissolved with RNase-free water. The concentration of total RNA was determined by measuring the absorbance at 260 nm. The purity of total RNA was checked by agarose gel electrophoresis.

S1 nuclease protection assay

The S1 nuclease protection assay was carried out as described previously (Yamamoto et al., 2002). Mixtures of the 32P-end-labelled probe (104 cpm) and total RNA (100 µg) were incubated for 10 min at 75°C for denaturation, and then incubated at 37°C overnight for hybridization. After digestion with S1 nuclease (Takara) at 37°C for 10 min, undigested products were extracted with phenol, precipitated with ethanol, and analysed by electrophoresis on polyacrylamide gels containing 8 M urea. The intensity of undigested probe bands on gels was measured with BAS1000 (Fuji).

DNA microarray analysis

To prepare fluorescence labelled cDNA, total RNA (20 µg) was mixed in 40 µl of 1× reaction buffer (for AMV Reverse transcriptase XL, Life Science) containing 5.3 nmol of random primer (Takara), 0.5 mM each of dATP, dGTP and dGTP, 0.2 mM dTTP and 4 nmol of Cy3-dUTP or Cy5-dUTP (Amersham Pharmacia Biotech). For all experiments, two independent sets were carried out with pairs of the fluorescence dye. In one experiment, control sample was labelled with Cy3 and test sample was labelled with Cy5, and in another experiment, a reciprocal labelling was carried out, i.e. Cy5 label for control sample and Cy3 label for test sample. The mixture was heated at 65°C for 5 min and cooled down to room temperature. After addition of 50 units of AMV reverse transcriptase XL (Life science), cDNA synthesis was carried out at 42°C for 1 h and, after addition of another 50 units of the reverse transcriptase, was continued for additional 1 h. Fluorescence labelled cDNA was purified by Centri-Sep (Princeton Separations), followed by phenol-chloroform extraction and ethanol-precipitation. After drying, the cDNA was dissolved in 8 µl of water.

Escherichia coli CHIP version 1.0 (Takara) was used. Before hybridization, DNA chip was incubated at 65°C for 1 h in 20 µl of prehybridization buffer (6× SSC, 0.2% SDS, 5× Denhardt's solution and 130 µg ml−1 of denatured salmon sperm DNA) and washed with 2× SSC at 65°C. Both Cy3- and Cy5-labelled cDNA were added to 22 µl of hybridization buffer (6× SSC, 0.2% SDS, 5× Denhardt's solution and 13 µg ml−1 of denatured salmon sperm DNA) and heated at 98°C for 2 min The denatured cDNA mixture was applied to DNA chip under slide glass, and the hybridization was carried out at 65°C for 16 h. The DNA chip was washed at 65°C with 2× SSC for 5 min, then at 65°C with 2× SSC containing 0.1% SDS for 5 min, and finally at room temperature with 0.2× SSC. The DNA chip was scanned with an Affimetrix laser scanner, and the intensity of hybridized Cy3 and Cy5 was quantified by Image-Quant (Molecular Dynamics). Background was achieved by measuring the fluorescence intensity outside each DNA spot and used to correct the intensity of each spot. In addition, the mean values of each intensity was collected from spots, whose the intensity was higher than the averaged signal with +1 of standard deviation (SD) of the 24 negative control spots of human beta actin. The ratio (Cy3/Cy5 and Cy5/Cy3) was estimated and normalized by defining the mean of ratios of all spots as 1.0. Among spots that showed significant fluorescent intensity of either Cy3 or Cy5, we selected the gene spots, which showed the high-level intensity (>500), as up- or downregulated genes.

Purification of CusR and CueR

To construct pCueR and pCusR for overproduction of CueR and CusR, respectively, DNA fragments containing the CueR- or CusR-coding regions were amplified by PCR with E. coli W3110 genome DNA as a template and a pair of primers, in which the NdeI and NotI sites were included (see sequences in Table 3). After digestion with NdeI and NotI, the PCR-amplified fragments were cloned into pET21a(+) (Novegen) between NdeI and NotI sites. The plasmid constructs were confirmed by DNA sequencing. The CueR- and CusR-expression plasmids were transformed into E. coli BL21(DE3). Each transformant was grown in 200 ml of LB broth and at the cell density of 0.6 OD600 nm, IPTG was added at the final concentration of 1 mM. After 3 h, cells were harvested by centrifugation, washed with lysis buffer (50 mM Tris-HCl,  pH 8.0  at  4°C,  100 mM  NaCl),  and  then  stored  at −80°C until use. Frozen cells were suspended in 3 ml of lysis buffer containing 100 mM PMSF. After addition of 80 µl of lysozyme (10 mg ml−1), the cell suspension was stored on ice for 30 min, and then lysed by sonication. After centrifugation at 15 000 r.p.m. for 20 min at 4°C, the resulting supernatant was mixed with 2 ml of 50% Ni-NTA agarose solution (Qiagen) and loaded onto a column. After washing with 10 ml of lysis buffer, the column was washed with 10 ml of washing buffer (50 mM Tris-HCl, pH 8.0 at 4°C, 100 mM NaCl). Proteins were then eluted with 2 ml each of the lysis buffer containing 0.1 M, 0.2 M, or 0.5 M imidazole. The recovery and purity of protein in each eluate was checked by SDS-PAGE. The purified protein fractions were pooled and dialysed against storage buffer (50 mM Tris-HCl, pH 7.6 at 4°C, 200 mM KCl, 10 mM MgCl2, 0.1 mM EDTA, 1 mM DTT and 50% glycerol). The protein concentration was determined by the Protein assay kit (Bio-Rad) and the purity was checked by SDS-PAGE.

DNase I footprinting

DNase I footprinting was carried out as described previously (Yamamoto et al., 2002). In brief, the test protein was incubated with a 32P-labelled DNA fragment in transcription buffer [50 mM Tris-HCl (pH 7.8 at 37°C), 50 mM NaCl, 3 mM magnesium acetate, 0.1 mM EDTA, 0.1 mM DTT and 25 mg ml−1 BSA] at 37°C for 10 min DNA digestion was initiated by the addition of 5 ng DNase I (Takara). After incubation for 30 s at 25°C, the reaction was terminated by adding 45 ml of the stop solution. DNA was precipitated with ethanol, dissolved in formamide-dye solution, and analysed by 6% PAGE in the presence of 8 M urea.

Quantitative Western blot

For the measurement of intracellular levels of transcription factors, a quantitative Western blot analysis was employed as described previously (Jishage and Ishihama, 1997) using specific rabbit antibodies raised against purified proteins. In brief, cell lysates were separated, in parallel with various amounts of purified transcription factors, by electrophoresis on 7.5 or 10% polyacrylamide gels. Proteins in the gels were directly electroblotted onto polyvinylidene difluoride membranes. Blots were blocked with BSA, probed with antibodies, and then incubated with goat antirabbit IgG conjugated with hydroxyperoxidase. The blots were developed with 3,3′-diaminobenzidine, tetrahydrochloride. Staining intensity was measured with BAS image analyzer (Fuji, Tokyo). To enhance the detection intensity, the ECL reagent system (Amersham) was used in some experiments.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Takeyoshi Miki (Fukuoka Dental College) for E. coli strain JD20810, Etsuko Koshio, Nobuyuki Fujita and Olga N. Ozoline (National Institute of Genetics) for technical support and discussion. We also thank Taku Oshima (Nara Institute of Science and Technology), Hirofumi Aiba (Nagoya University) for E. coli strain BW27559 and discussion. This work was supported by Grants-Aid from the Ministry of Education, Culture, Sports, Science and Technology, and the CREST fund from the Japan Science Corporation.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References