Copper-specific Transcriptional Repression of Yeast Genes Encoding Critical Components in the Copper Transport Pathway*

Copper is an essential micronutrient that is toxic in excess. To maintain an adequate yet non-toxic concentration of copper, cells possess several modes of control. One involves copper uptake mediated by genes encoding proteins that play key roles in high affinity copper transport. These include the FRE1-encoded Cu2+/Fe3+ reductase and the CTR1and CTR3-encoded membrane-associated copper transport proteins. Each of these genes is transcriptionally regulated as a function of copper availability: repressed when cells are grown in the presence of copper and highly activated during copper starvation. Our data demonstrate that repression of CTR3 transcription is exquisitely copper-sensitive and specific. Although copper repressesCTR3 gene expression at picomolar metal concentrations, cadmium and mercury down-regulate CTR3 expression only at concentrations 3 orders magnitude greater. Furthermore, copper-starvation rapidly and potently induces CTR3 gene expression. We demonstrate that the CTR1, CTR3, and FRE1 genes involved in high affinity copper uptake share a common promoter element, TTTGCTC, which is necessary for both copper repression and copper-starvation activation of gene expression. Furthermore, the Mac1p is essential for down- or up-regulation of the copper-transport genes. In vivo footprinting studies reveal that the cis-acting element, termed CuRE (copper-response element), is occupied under copper-starvation and accessible to DNA modifying agents in response to copper repression, and that this regulated occupancy requires a functional MAC1 gene. Therefore, yeast cells coordinately express genes involved in high affinity copper transport through the action of a common signaling pathway.

Copper is an essential trace element that is required for a number of cellular enzymes including cytochrome c oxidase, Cu/Zn-superoxide dismutase, lysyl oxidase, and dopamine-␤monooxygenase (1). Copper also plays a critical role in the assimilation of iron both in microbial and mammalian systems (2,3). However, when allowed to accumulate in excess, copper is highly toxic due to its proclivity to engage in redox reactions which result in the formation of hydroxyl radical, a reactive species that causes extensive damage to nucleic acids, proteins, and lipids (4,5). Furthermore, copper may also be toxic through inappropriate incorporation into proteins, such as the estrogen receptor, which normally bind other metal ligands (6). Therefore, all organisms must be able to sense both toxic and nutritional levels of copper to allow sufficient copper to accumulate to drive biochemical reactions, yet prevent the accumulation to toxic levels. Indeed, the failure to appropriately establish and maintain copper homeostasis results in at least two human genetic disorders, Wilson's disease and Menkes disease (1,7,8).
A number of cellular regulatory responses that result from fluctuations in environmental copper levels have been reported including transcriptional activation or repression, changes in protein stability, and the modulation of protein trafficking (9 -13).
Yeast cells have provided an excellent model system for studies of copper transport, distribution, and detoxification (5,8,14). In response to high concentrations of copper, yeast cells activate the transcription of the CUP1 (15)(16)(17) and CRS5 (18) genes, which encode copper-sequestering proteins called metallothioneins, as well as the SOD1 gene, encoding Cu/Zn-superoxide dismutase (19). This transcriptional activation involves the copper metalloregulatory transcription factors (MRTFs) 1 Ace1p (20) and Amt1p (21) from the bakers' yeast Saccharomyces cerevisiae and the opportunistic pathogenic yeast Candida glabrata, as well as cis-acting promoter regulatory sequences with the consensus sequence HTHXXGCTG (H ϭ A, C or T; X ϭ any residue). The binding of Cu(I) to Ace1p or Amt1p to form a tetra-copper cluster activates their DNAbinding domains via a conformational change, thereby providing a direct link between the toxic copper sensor and the activation of detoxification genes (22,23).
S. cerevisiae cells acquire copper as Cu(I) under high affinity conditions through the action of a plasma membrane-associated Cu(II)-Fe(III) reductase activity encoded by the FRE1 gene (24,25) and two high affinity copper transport proteins encoded by the CTR1 (26) and CTR3 (27) genes. Indeed, cells that are defective in high affinity copper transport exhibit a number of phenotypes which can be corrected by exogenous copper that include respiratory deficiency, sensitivity to superoxide generating agents due to a defect in Cu/Zn-superoxide dismutase activity, and severely diminished iron accumulation due to a defective copper-dependent ferroxidase (Fet3) required for high affinity iron transport (28,29). As would be expected, the genes encoding proteins involved in high affinity copper or iron transport are repressed, at the level of steady-state mRNA, by low concentrations of their respective metals and induced by copper or iron starvation, respectively (10,24,27). Furthermore, consistent with the involvement of the FRE1encoded Cu(II)-Fe(III) reductase in both high affinity copper and iron transport, the expression of this gene is repressed by both copper and iron (24,25). Insight into the mechanism for iron-dependent transcriptional repression in yeast has been gained by the recent identification and analysis of the ironresponsive DNA-binding protein Aft1 (30), which binds to the consensus sequence PyPuCACCCPu found in the promoters of the FET3, FTR1, FTH1, CCC2, FRE1, and FRE2 genes (31).
Currently, it is unclear how yeast cells sense very low concentrations of copper and respond by repressing the expression of the FRE1, CTR1, and CTR3 genes. One protein implicated in this signaling pathway, Mac1, was discovered based on significant homology to the amino-terminal 40 amino acid residues of the copper-activated DNA-binding domains of the Ace1 and Amt1 proteins (32). Interestingly, mac1⌬ cells exhibit phenotypes consistent with defective copper transport including respiratory deficiency, defective Cu(II) and Fe(III) reductase activity, impaired regulation of FRE1 transcription, hypersensitivity to a variety of stresses including metal and oxidative stress, and poor growth on rich medium (25,32). Furthermore, a dominant mutation in a duplicated region rich in cysteine exhibits high levels of Cu(II) reductase activity and FRE1 mRNA that are virtually not repressed by copper (25). These characteristics of Mac1p, as well as the nuclear localization of a Mac1-␤-galactosidase fusion protein (32), suggests that Mac1 plays a key role in copper homeostasis.
In this report, we demonstrate that the regulation of FRE1, CTR1, and CTR3 transcription is exquisitely copper-sensitive and specific. We demonstrate that these three genes share a common promoter element, TTTGCTC, which is necessary for both copper repression and copper-starvation activation of expression. Furthermore, the Mac1p is essential for down-or up-regulation of the copper transport genes. In vivo footprinting studies reveal that the cis-acting element, termed CuRE (copper-response element), is occupied under copper-starvation and accessible to DNA modifying agents in response to copper repression, and that this copper-responsive occupation of the CuREs requires a functional MAC1 gene. Taken together these results show that yeast cells coordinately express genes involved in high affinity copper transport through the action of a common signaling pathway to acquire this micronutrient from the environment.  (34). The mac1::URA3 allele was constructed using the disruption plasmid pmac1::URA3 (generously provided by Daniel Kosman). This disruption rendered the MAC1 gene nonfunctional as ascertained by the lack of growth on YPE media. Furthermore, the allele status of each of these loci in the strains generated in this study was verified using diagnostic polymerase chain reaction (35).

Chemicals
Yeast cells were grown in a modified minimal medium (SD), which was depleted for copper as described previously (10). This depleted medium contains 16 nM copper as determined by atomic absorption spectroscopy while standard synthetic complete media contains 150 nM copper. Copper administration or copper starvation of yeast strains were carried out by adding the indicated amount of CuSO 4 or bathocuproinedisulfonate (BCS) to cells grown to mid-logarithmic phase (optical density at 600 nm ϭ 1.1 to 1.3) in this modified SD medium. Under nonselective conditions, yeast cells were grown in YPD (1% yeast extract, 2% bactopeptone, and 2% dextrose).
Plasmids and Gene Analyses Methods-Plasmid YEp357RTRP1 was constructed by insertion of a blunt-ended 1.7-kilobase TRP1 fragment isolated from p330 (gift of Robert Fuller) into the StuI and Klenow filled-in NcoI sites of YEp357R, disrupting the URA3 genetic marker. The oligonucleotides CTR3-A (5Ј-CTCGCGGATCCAGTCATAGCAT-GAACAATTC-3Ј) and CTR3-B (5Ј-GTCCGGAATTCGAAGCAGTGCT-GCTACTGCCTCC-3Ј) were used to polymerase chain reaction amplify the CTR3 promoter (1116 bp of the 5Ј-noncoding region) and the first 10 codons of CTR3 gene from genomic DNA of strain DTY1, which expresses both the Ctr1p and Ctr3p copper transporters. The polymerase chain reaction product was sequenced in its entirety and fused in-frame to the Escherichia coli lacZ gene using the BamHI and EcoRI sites of YEp357RTRP1 to generate YEpCTR3-lacZ. A low-copy number plasmid pRSCTR3-lacZ was also constructed. To accomplish this construct, a 4.8-kilobase DNA fragment from plasmid YEpCTR3-lacZ containing the CTR3 sequence from Ϫ1116 to ϩ10 fused to the E. coli lacZ gene was inserted into the BamHI and SmaI sites of pRS314. To perform the RNase protection analyses as described previously (36), two plasmids for making antisense RNA probes were made. pKSlacZ was constructed by the insertion of the 233-bp EcoRV-BclI fragment from the E. coli lacZ gene into the EcoRV and BamHI sites of pBluescript II KS. pKSACT1 was made by the insertion of the 132-bp HindIII-EcoRI fragment from the S. cerevisiae ACT1 gene into the HindIII-EcoRI sites of pBluescript II KS. A series of plasmids containing sequential deletions from the 5Ј end of the CTR3 promoter ( Fig. 3B) were created from plasmid YEpCTR3-lacZ using standard protocols (Exo III/Mung Bean Nuclease Deletion Kit, Stratagene). To assess the capability of the CuREs to mediate copper repression and copper starvation induction, a series of purified oligonucleotides (Figs. 4C and 5C) were annealed pairwise to form double-stranded DNA and then ligated into the BglII and XhoI sites of a CYC1-lacZ fusion plasmid pCM64 (generously provided by Charles Moehle).
RNA was extracted by the hot phenol method as described previously (37). Northern blot analyses were carried out by standard protocols (38). The radioactive bands corresponding to specific transcripts were quantitated using a PhosphorImager SP and ImageQuant 3.3 software (Molecular Dynamics). The data derived from the PhosphorImage quantitation were plotted and analyzed using Kaleidagraph software 3.02 (Synergy Software, Reading, PA). DNA isolation and polymerase chain reaction were performed using standard protocols (38,35). DNA sequencing was carried out using Sequenase according to the manufacturer's protocol (U. S. Biochemical Corp.).
In Vivo Dimethyl Sulfate Footprinting-Cultures of the isogenic strains DTY1 (MAC1), DTY205 (MAC1 up1 ), or SLY2 (mac1⌬) containing YEpCTR3-lacZ were grown to early log-phase in modified SD media. Untreated and copper-treated (10 nM) cultures were incubated for 1 h at 30°C, 375 rpm. In vivo footprinting was carried out as described previously (39). Isolated DNA samples were digested by BstYI prior to G ϩ A-specific DNA cleavage of the dimethyl sulfate-treated DNA. The purified oligonucleotide used as primer in the extension reactions was CTR3E (5Ј-GCCTCCCATATTCATCTTTGTATAGCCC-3Ј), which hybridizes to CTR3 gene positions ϩ15 to Ϫ14 with respect to A of the translational start codon.

Metal Ion Specificity and Sensitivity of CTR3 mRNA Downregulation-Previous
investigations have demonstrated that all three yeast genes known to be involved in high affinity copper transport, CTR1, CTR3, and FRE1, are repressed by copper at the level of steady-state mRNA (10,27). This regulation is independent of the S. cerevisiae copper MRTF Ace1p, which activates the CUP1, CRS5, and SOD1 genes in response to micromolar copper concentrations (10). 2 To characterize in detail the regulation of CTR3 by metals, the low-copy number plasmid pRSCTR3-lacZ was transformed into DTY1, a strain that expresses both high affinity copper transporters Ctr1p and Ctr3p, and the level of lacZ steady-state mRNA was assayed by RNase protection experiments. Among the metal ions tested at many concentrations, Cu(II), Fe(II), Fe(III), Ag(I), Cd(II), Zn(II), Hg(II), Pb(II), Co(II), Mn(II), Ni(II), Mg(II), Ca(II), Au(I), and Au(III), only four were capable of repressing expression of lacZ mRNA under the control of the CTR3 promoter: copper, silver, cadmium, and mercury ( Fig. 1). Moreover by using a range of concentrations for each of these four metal ions, CuSO 4 and AgNO 3 , were most effective for the repression of CTR3-lacZ transcription. Only 2.0 ϫ 10 Ϫ11 M and 1.4 ϫ 10 Ϫ11 M CuSO 4 and AgNO 3 , respectively, was required for half-maximal repression (repression index 50% (RI 50 )) of expression from the CTR3 promoter. In contrast, CdCl 2 and HgCl 2 repress transcription from the CTR3 promoter, but at concentrations 3 orders of magnitude greater than for CuSO 4 and AgNO 3 . The RI 50 for CdCl 2 and HgCl 2 were 4.0 ϫ 10 Ϫ8 M and 1.7 ϫ 10 Ϫ8 M, respectively. Although we cannot eliminate the possibility that the CdCl 2 and HgCl 2 solutions contain trace levels of copper, atomic absorption spectroscopy failed to detect any copper in the CdCl 2 and HgCl 2 stock solutions. Therefore, the repression of CTR3 gene expression exhibits a high degree of selectivity for copper ions. Moreover, the observation that Ag(I), a metal which is electronically similar to the reduced form of cupric Cu(II), represses CTR3 gene expression with an efficacy similar to CuSO 4 , suggests that cuprous Cu(I) might be the active form of copper in the signaling process resulting in repressing CTR3 gene expression. Time Course of Repression and Activation of CTR3 Gene Expression-Because yeast cells alter the expression of the copper transport machinery as a function of changing environmental copper levels, we analyzed CTR3-lacZ steady-state mRNA levels over time in response to either copper repletion or starvation. Using the low-copy number plasmid pRSCTR3-lacZ in strain DTY1, we followed the time course of down-and up-regulation of CTR3 gene expression in the presence of copper (1 M) or BCS (100 M), respectively. The derepression of CTR3-lacZ gene expression is rapid with 91% of the maximal level of transcript detected 10 min after treatment with BCS (Fig. 2). On the other hand, 79% of the CTR3-lacZ transcript levels remained detectable after a 10-min exposure to copper. The time course data observed using the low-copy number plasmid pRSCTR3-lacZ were virtually identical to those observed for the endogenous CTR3 gene. 3 These data demonstrate that yeast cells respond to changes in environmental copper levels by rapidly altering steady-state levels of the CTR3 copper transport mRNA. The temporal differences observed between derepression and repression of the CTR3-lacZ reporter gene expression may be due to changes in the stability of the mRNA or may reflect unidentified differences in the repression versus the derepression signaling pathways.

Identification of cis-Acting Elements Necessary for Copper
Repression and Copper-starvation Activation of CTR3, CTR1, and FRE1 Gene Expression-To identify cis-acting elements necessary for the copper repression and copper-starvation-mediated induction of CTR3 expression, we fused 1116 bp of the 5Ј-noncoding region and the first 10 codons of CTR3 in-frame with the E. coli lacZ gene on both low-and high-copy number plasmids pRSCTR3-lacZ and YEpCTR3-lacZ, respectively. CTR3-lacZ expression from both reporter plasmids was downregulated in the presence of copper (approximately 5-fold), and up-regulated in the presence of BCS (approximately 8-fold) ( Fig. 3 and data not shown). To identify the copper-responsive CTR3 promoter elements, a series of nested 5Ј deletions of promoter sequences beginning at position Ϫ1116 were created in the high-copy number plasmid YEpCTR3-lacZ (Fig. 3). Removal of the CTR3 upstream region between Ϫ1116 and Ϫ248 had little effect on the copper and BCS-regulated expression of the CTR3-lacZ fusion, although the magnitude of the response decreased progressively. When the CTR3 promoter was further deleted to position Ϫ214, the overall magnitude of the response decreased by approximately 75% as compared with the parental plasmid, but the reporter gene was still regulated in response to copper levels. Further deletion to position Ϫ160 was found to completely abolish both the down-and up-regulation of the reporter gene (Fig. 3). Due to the observation that the integrity of the region between positions Ϫ248 and Ϫ160 was essential for driving copper repression and copper-starvation activation of the CTR3-lacZ fusion, we examined whether this sequence could regulate a heterologous reporter gene in a copperdependent fashion. A double-stranded DNA oligomer derived from the CTR3 promoter (positions Ϫ237 to Ϫ173) was inserted in its natural orientation into the minimal promoter of the iso-1-cytochrome c (CYC1) gene fused to lacZ (40). This fusion promoter was able to down-regulate (approximately 5-fold) and up-regulate (approximately 9-fold) lacZ mRNA expression in the presence of copper or BCS, respectively (Fig. 4). Within this 64-bp CTR3 promoter DNA fragment, we noted the presence of two copies of a repeated sequence, TTTGCTC, that are similar to the binding sites for the Ace1 and Amt1 copper MRTFs. To ascertain if this element plays a role in CTR3 regulation by copper, all seven of these residues were mutationally altered in either or both repeats and the cells carrying CTR3-CYC1-lacZ fusion plasmids were assayed for copper-regulated expression of lacZ mRNA (Fig. 4C). Multiple point mutations in either or both elements abolished copper responsiveness of the CTR3-CYC1-lacZ fusions (Fig. 4). Moreover, a CTR3-CYC1-lacZ promoter fusion plasmid in which the last C in each TTTGCTC element was changed to G, failed to respond to the presence of a wide range of copper concentrations or copper-starvation to repress or activate gene expression ( Fig. 4 and data not shown). Our data do not allow us to establish the reason why a single CuRE cannot regulate the CTR3-CYC1-lacZ fusion derivative, but does regulate with a low magnitude a 5Ј-truncated CTR3 promoter-lacZ derivative which has one element (Fig. 3).
Interestingly, the TTTGCTC element was also observed to be perfectly conserved and repeated in the CTR1 and FRE1 promoters, both of which drive the expression of proteins involved in high affinity copper transport. As shown in Fig. 5, fusion of the regions from the CTR1 or FRE1 promoters encompassing these TTTGCTC repeats mediated copper repression and copper-starvation activation of lacZ mRNA expression. For both the CTR1-CYC1-lacZ and FRE1-CYC1-lacZ promoter fusions, the integrity of the TTTGCTC repeats was essential for copper- responsive gene expression (Fig. 5). Therefore, based upon these studies the TTTGCTC element, denoted CuRE, plays a critical role in copper-regulated gene expression for all three yeast genes encoding components of the high affinity copper transport machinery. Furthermore, the CuREs in CTR1, CTR3, and FRE1 function in both copper-repression and copper-starvation-mediated gene expression.
Yeast High Affinity Copper Transport Genes Require MAC1 for Expression and Regulation by Copper-The repression of CTR3 gene expression is highly specific for copper and occurs at exquisitely low extracellular copper concentrations. Since CTR1 and CTR3 gene repression by copper is independent of the Ace1p copper MRTF (Ref. 10, and data not shown), we sought to identify components of the copper-signaling pathway that play a key role in the copper-regulated expression of the high affinity copper transport genes. The S. cerevisiae MAC1 gene encodes a nuclear protein with a high level of homology to the first 40 amino acids of the Ace1 and Amt1 copper MRTFs. A hallmark of mac1⌬ mutants is that virtually all of the pleiotropic phenotypes associated with these cells can be overcome by exogenous copper, thereby clearly implicating Mac1p in copper metabolism (32). Using isogenic strains harboring a wild type MAC1 gene, an insertionally inactivated mac1 allele (mac1⌬), and a dominant gain-of-function allele (MAC1 up1 ), we ascertained whether MAC1 plays an essential role in CTR1, CTR3, and FRE1 gene regulation as a function of cellular copper status by Northern blot analysis (Fig. 6). In the wild type strain (DTY1), basal levels of CTR1, CTR3, and FRE1 mRNA are clearly visible. However, in the presence of 1 or 10 M copper the steady-state levels of CTR1, CTR3, and FRE1 mRNA were strongly repressed. Conversely, in the presence of 100 M BCS, CTR1, CTR3, and FRE1 mRNA levels were induced 3-, 14-, and 6-fold over basal levels, respectively. No mRNA was detected for CTR1 and CTR3 in the mac1⌬ mutant strain (SLY2) even under conditions of copper starvation, although a low level of FRE1 mRNA was still observed, perhaps due to the action of Aft1p (30). In the MAC1 up1 strain (DTY205), all three genes were highly expressed (8-fold for CTR1, 37-fold for CTR3, and 8-fold for FRE1 with respect to the basal level detected in DTY1), and were virtually unregulated by copper or copper-starvation. Taken together, these data demonstrate that Mac1p is an essential trans-acting component of the copper signaling pathway for appropriate expression and regulation of genes involved in high affinity copper transport.
The CuRE Is Differentially Occupied under Copper Deprivation and Repletion-Although Mac1p exhibits homology to the Ace1p and Amt1p copper-activated DNA-binding proteins, we  CTR1 and FRE1 promoters. A, CYC1-lacZ fusion genes containing wild type DNA fragments and CuRE mutants (M5 and M6) derived from CTR1 and FRE1 promoters were analyzed by RNase protection as described in Fig. 4A. B, quantitation of the RNase protection experiment as in Fig. 4B except that the nucleotide numbers refer to the position relative to the A of the ATG codon of the CTR1 (26) and FRE1 ORFs (24). C, sequences of the promoter elements from the CTR1 and FRE1 genes and mutant derivatives.
have been unable to demonstrate CuRE DNA binding activity for Mac1p expressed in E. coli or by in vitro transcription translation. To begin to understand the molecular responses to copper mediated by the CuREs, and the function of Mac1p in this pathway, we have examined the CTR3 promoter in cells starved for copper or grown in the presence of copper by in vivo dimethyl sulfate footprinting. In the wild type strain (DTY1) under copper deprivation conditions that parallel induced CTR3 gene expression, we observed that a region encompassing and flanking the two CuREs located in the CTR3 promoter is strongly protected from methylation by dimethyl sulfate relative to copper-treated cells in which CTR3 gene expression is repressed (Fig. 7). Upon copper treatment of the wild type strain with 10 nM CuSO 4 , we observed that all of the G residues encompassing and immediately flanking the CuREs were strongly modified by dimethyl sulfate. Specifically, the G residues corresponding to positions Ϫ182 and Ϫ184 inside of the downstream TTTGCTC element (oriented in the opposite direction relative to the direction of the transcription), and the G located at position Ϫ228 inside of the upstream TTTGCTC element were strongly protected from dimethyl sulfate modification. The protection detected for the G residue at position Ϫ184 correlates with the critical requirement for this G for copper-responsive gene regulation in CTR3-CYC1-lacZ fusion gene (see construct M4, Fig. 4). Furthermore, each G residue between the two CuREs, plus two additional G residues (at positions Ϫ172 and Ϫ165) located downstream of the repeat were found to be protected from dimethyl sulfate under conditions in which CTR3 is expressed, but not when CTR3 expression is repressed, suggesting the presence of one or more proteins occupying this region under condition of expression. Furthermore, a similar pattern as the one described for the wild type strain under copper deprivation conditions was observed when this strain was grown in the presence of BCS (data not shown). Consistent with the constitutive high level CTR3 gene expression observed in the MAC1 up1 strain, a constitutive protected pattern in the CTR3 promoter, comparable to that found for the wild type strain (MAC1) under copper deprivation conditions, was revealed by dimethyl sulfate footprinting. Conversely, the CTR3 promoter region in the isogenic mac1⌬ strain revealed no copper-dependent changes in dimethyl sulfate reactivity and strongly resembled the copper-repressed promoter configuration in the wild type strain (Fig. 7). Therefore, a functional MAC1 gene is required for copper-dependent changes in the CTR3 promoter region encompassing and flanking the CuREs.

DISCUSSION
Because of its essential yet toxic nature, all cells must maintain tight homeostatic regulation of the levels of bioavailable copper. Regulatory responses to copper include the activation or repression of gene transcription, copper-modulated turnover of proteins that transport or utilize copper, and alterations in the intracellular trafficking of copper transporting ATPases in response to fluctuations in copper concentrations (9 -13, 27). In this report, we have characterized the molecular mechanisms responsible for the copper-responsive regulation of CTR3, CTR1, and FRE1 genes that play a critical role in high affinity copper transport in yeast. Detailed studies of CTR3 gene regulation have revealed that half-maximal repression of CTR3 mRNA levels occurs in the presence of as little as 2.0 ϫ 10 Ϫ11 M CuSO 4 or 1.4 ϫ 10 Ϫ11 M AgNO 3 (Fig. 1). This represents an exquisitely sensitive and selective metal responsive system since at least 1000-fold higher concentrations of cadmium or mercury were required to achieve the same level of repression and no other metal tested was able to repress CTR3 mRNA levels. Since repression is observed with cadmium and mercury, however, we cannot at present eliminate the possibilities that either trace levels of copper, undetectable by atomic absorption, were present in our stock solutions or that there are significant differences in the efficiency with which copper and silver, versus cadmium and mercury, are transported or distributed in yeast cells. The similarity in the potency of silver in fostering the repression of CTR3 mRNA levels resembles previous results on the Ace1 and Amt1 copper MRTFs, which are known to form tetragonal coordinates with Cu(I) via cysteine thiolates (23). The electronic similarity of Ag(I) to Cu(I), but not Cu(II), suggests that the copper sensing machinery involved in the repression of yeast copper transport genes may sense Cu(I), rather than Cu(II).
Our studies have identified strictly conserved repeated CuREs present in two copies in each of the CTR1, CTR3, and FRE1 promoters. Inspection of the CTR1 promoter suggests the presence of a third CuRE, located between positions Ϫ529 and Ϫ523, however, the contribution of this element to copperresponsive regulation of CTR1 has not yet been determined. Since single or multiple point mutations within the CuREs abolish both copper-dependent repression and copper-starvation-induced expression of CTR1, CTR3, and FRE1, this suggests that yeast cells utilize a common mechanism to coordinately regulate the expression of the copper transport machinery. Examination of the CuRE sequences in CTR1, CTR3, and FRE1, suggest a consensus sequence WWWTTT-GCTCR (W ϭ A or T; R ϭ purine). The CuRE sequence bears an interesting sequence similarity to the binding sites for the Ace1 and Amt1 copper MRTFs (HTHXXGCTG, H ϭ A, C or T; X ϭ any residue), however, the inability to convert a CuRE to an Ace1p activation site by substituting the terminal C residue for a G (Fig. 4, plasmid derivative M4) suggests that other nucleotides within or flanking the CuRE confer specificity for copper repression and copper-starvation induction of gene expression which is independent of Ace1p (10). 2 Although the center-tocenter distances between the CuREs, for each of the CTR1, CTR3, and FRE1 promoters predict that they lie on opposite faces of the DNA, it is currently unknown whether this geometry plays a role in the regulation of copper transporter gene expression via the CuREs. The FRE1 promoter fragment, which has the shortest distance between the two CuREs, gave rise to the poorest repression by copper and the strongest activation by copper-starvation. Since this promoter fragment lacks binding sites for the Aft1 iron-responsive regulatory protein (31), this supports the notion that de-repression of the FRE1-CYC1-lacZ fusion gene represents a response to copper starvation rather than an indirect response to iron starvation. However, whether other promoter elements contribute to the magnitude of these regulatory responses must await a comprehensive dissection of each of the promoters of these genes encoding proteins that function in high affinity copper transport.
It is currently unknown how such exquisitely low extracellular copper concentrations are capable of signaling the repression of CTR1, CTR3, and FRE1 gene expression. Based on previous observations that S. cerevisiae cells harboring a deletion of the MAC1 gene display pleiotropic defects that are corrected by the addition of copper, and are defective in the regulation of FRE1 transcription (25,32), we tested the possibility that MAC1 plays a role in the regulation of all the genes known to be involved in high affinity copper transport in yeast. Indeed, we have demonstrated that cells bearing a disruption of the MAC1 gene are severely defective in the expression and regulation of mRNA levels from each of these genes (Fig. 6). Furthermore, isogenic cells harboring a dominant allele of MAC1, in which a His residue in the first Cys-rich carboxylterminal cluster repeat has been altered to Gln, fail to respond to copper administration to repress CTR1, CTR3, or FRE1 mRNA levels. Although Mac1p exhibits significant sequence similarity to the amino-terminal 40 amino acids of Ace1p and Amt1p, and a Mac1-␤-galactosidase fusion protein has been localized to the yeast cell nucleus (32), we have been unable to demonstrate specific CuRE binding activity for Mac1p either expressed in E. coli cells or produced by in vitro transcription and translation (data not shown). Therefore, Mac1p may function in the copper-signaling pathway to regulate copper transport gene expression through a number of potential mechanisms. First, Mac1p may require specific post-translational modifications or partner proteins for sequence-specific DNA binding that would not be present through expression in heterologous systems. Alternatively, Mac1p could be required for the synthesis or activity of another protein(s) which directly interacts with CuREs, or Mac1p may play a critical nuclear signaling role for copper that is upstream of a direct DNA binding activity. Consistent with all of these models, electrophoretic mobility shift experiments using extracts from yeast cells expressing a functional epitope-tagged MAC1 allele have revealed the presence of a specific CuRE-protein complex from control or BCS-treated cells which is absent in extracts from cells treated with 10 nM CuSO 4 (data not shown). Furthermore, the formation of this CuRE-protein specific complex was dependent on the presence of a functional MAC1 gene, and the complex was not abolished by the addition of copper to a strain bearing a MAC1 up1 allele. On the other hand, the inability of the anti-epitope monoclonal antibody to supershift this complex leaves open the possibility that Mac1p may not directly contact DNA sequences within the CuRE. Our in vivo footprinting studies clearly demonstrate that in the wild type MAC1 strains the CTR3 CuREs are occupied under conditions of active CTR3 gene expression (i.e. copper-starvation), and are highly accessible to dimethyl sulfate modification under conditions in which CTR3 gene expression is repressed (i.e. addition of 10 nM CuSO 4 ). In contrast, mac1⌬ mutants give rise to constitutive dimethyl sulfate modifications in the CTR3 promoter region encompassing the CuREs that resemble the repressed state in the wild type strain. Furthermore, MAC1 up1 mutants, which constitutively express the copper transport genes and are unresponsive to copper for repression, give rise to constitutive protection from dimethyl sulfate modification within and flanking the CTR3 CuREs. Taken together these results demonstrate that Mac1p is an essential component of the copper-signaling pathway that directly or indirectly modulates coordinated copper-responsive gene expression of yeast high affinity copper transport genes through the CuREs.