The fission yeast copper-sensing transcription factor Cuf1 regulates the copper transporter gene expression through an Ace1/Amt1-like recognition sequence.

Transcriptional regulation of genes encoding critical components of copper transport is essential for copper homeostasis and growth in yeast. Analysis of regulatory regions in the promoter of the ctr4(+) copper transporter gene in fission yeast Schizosaccharomyces pombe reveals the identity of a conserved copper-signaling element (CuSE), which is recognized by the transcription factor Cuf1. We demonstrate that CuSE is necessary for transcriptional activation in response to copper deprivation conditions. Interestingly, the CuSE element bears a strong sequence similarity to the recognition site, denoted MRE (metal regulatory element), which is recognized by a distinct class of copper sensors required for copper detoxification, including Ace1 from Saccharomyces cerevisiae and Amt1 from Candida glabrata. When a consensus MRE from S. cerevisiae is introduced into S. pombe, transcription is induced by copper deprivation in a Cuf1-dependent manner, similar to regulation by Mac1, the nuclear sensor for regulating the expression of genes encoding components involved in copper transport in S. cerevisiae. UV-cross-linking experiments show that the Cuf1 protein directly binds the CuSE. These results demonstrate that the Cuf1 nutritional copper-sensing factor possesses a module that functions similarly to domains found in the Ace1/Amt1 class of metalloregulatory factors, which allows the protein to act through a closely related MRE-like sequence to regulate copper transport gene expression in S. pombe.

Aerobic life requires trace amounts of transition elements like copper, which have the property to reversibly gain and lose electrons, thereby serving as catalytic centers of numerous proteins involved in a variety of critical enzymatic processes (1)(2)(3). Owing to its proclivity to engage in redox reactions, excess copper results in the production of detrimental hydroxyl radicals, which are extremely toxic to cells, causing lipid peroxidation, protein denaturation, and nucleic acid cleavage (4). Therefore, specialized proteins have evolved for the sensing, transport, and sequestration of copper within cells to maintain the delicate balance between essential and toxic levels (5)(6)(7).
Three genes in the bakers' yeast Saccharomyces cerevisiae are induced in response to copper excess: CUP1-and CRS5encoded metallothioneins and SOD1-encoded Cu/Zn superoxide dismutase (8 -10). Metallothioneins are known to counteract metal cytotoxicity by sequestration of cytosolic copper (11,12). Increased synthesis of these three proteins in response to copper is controlled at the transcriptional level by the Ace1 copper metalloregulatory transcription factor (MRTF) 1 (13)(14)(15). Furthermore, the promoter element necessary for copperinducible transcription of CUP1, CRS5, and SOD1 is denoted metal regulatory element (MRE) and is defined by the consensus sequence, 5Ј-HTHNNGCTGD-3Ј (D ϭ A, G, or T; H ϭ A, C, or T; N ϭ any residue). The GCTG region is termed the core sequence, whereas the region 5Ј to the core is known as the T-rich element (16,55). The core sequence of the MRE is recognized by Ace1 in the major groove, whereas the AT-rich 5Ј region is contacted in the minor groove (16,17). The first subdomain (residues 1-40) (denoted Zn 2ϩ module) of the DNA binding domain of Ace1 was found to bind a single Zn 2ϩ atom (18). The second subdomain of the DNA binding domain (residues , named copper-regulatory domain (17), harbors 9 Cys residues found as follows: one Cys-X 2 -Cys motif, three Cys-X-Cys motifs, and a single Cys residue. All of these Cys residues are necessary for copper response, except for the last, Cys 105 (17). The arrangement of the Cys residues is predicted to coordinate four Cu 1ϩ atoms through cysteine sulfur bonds (19). Upon copper activation, it appears that the formation of a tetra-copper cluster within the copper regulatory domain via cysteinyl thiolates is critical to make an active detoxifying factor, thereby making the cupro-Ace1 able to interact to MREs in a copper and conformational changes-dependent manner (11,(17)(18)(19)(20). The Candida glabrata ACE1 gene ortholog, AMT1, binds MREs with similar motifs and is essential for copper resistance (21,22). A copper biochemical pathway distinct from the detoxification system exists to allow S. cerevisiae cells to acquire trace amounts of copper from the environment (2,5,23). Yeast genetic studies have led to the identification of the FRE1 (24), CTR1 (25), and CTR3 (26) genes, which encode components involved in the high affinity copper uptake process. When grown under copper starvation conditions, the Fre1 plasma membrane protein reduces Cu 2ϩ to Cu 1ϩ (27), allowing copper ions to be recognized and transported into the cell by two separate, high affinity transport proteins, Ctr1 (28) and Ctr3 (29). A hallmark of these genes is the fact that they are tran-scriptionally expressed according to copper need; the transcription of FRE1, CTR1, and CTR3 is up-regulated under copper starvation conditions. This process is under the control of Mac1, another copper-sensing transcription factor harboring homology with the amino-terminal 40 amino acids found in the Ace1/Amt1 copper-activated metalloregulatory proteins but little homology outside of this region (30 -32). The carboxyl-terminal region of Mac1 harbors two Cys-rich repeats, REP-I and REP-II (also identified as C1 and C2) (39 -41). REP-I has been demonstrated to function in copper sensing (31,39,41). The copper-specific transcriptional regulation by Mac1 is mediated through cis-acting promoter elements, denoted CuREs (Curesponse elements) with the consensus sequence 5Ј-TTTGC(T/ G)C(A/G)-3Ј (32,33). Under low copper conditions, Mac1 binds to CuREs as a dimer (34,35) by making contacts in both the major (with 5Ј-GC(T/C)C(A/G)-3Ј sequence) and minor (with 5Ј-TTT-3Ј sequence) grooves (36). Interestingly, it has been shown that Cu 1ϩ ions can bind directly to Mac1 (37). Under elevated copper concentrations, Mac1 is released from the CuREs in vivo (32), and it is suggested that it undergoes intramolecular conformational changes to inactivate its transactivation domain (35,37,38).
The fission yeast Schizosaccharomyces pombe exhibits several features, such as genome organization, transcription initiation, signal transduction, post-translational modification, and cell division processing, which are more similar to those from mammalian cells than those of bakers' yeast S. cerevisiae (44). Recent studies of copper homeostasis in S. pombe have identified a high affinity copper transporter, denoted Ctr4, that exhibits more overall sequence homology to human Ctr1 than to S. cerevisiae Ctr1 (6,45,46). Likewise, the S. pombe high affinity copper transporter gene ctr4 ϩ is transcriptionally regulated by copper availability via the Cuf1 copper-sensing transcription factor (46). Cuf1 activates ctr4 ϩ gene expression under copper starvation conditions (46). Therefore, both the Ctr4 copper transporter and the Cuf1 nuclear copper-sensing transcription factor are essential for fission yeast cells to use copper ions from the environment. Although the Cuf1 protein is required for S. pombe high affinity copper transport, Cuf1 displays at its amino terminus an extended homology (amino acid residues 1-61) to the amino-terminal 63 and 62 amino acids of the S. cerevisiae Ace1 and C. glabrata Amt1 class of transcription factors and much less similarity to Mac1, its functional ortholog (46). Cuf1 possesses a cysteine-rich domain at its carboxyl terminus containing five Cys and one His residue that is absent in Ace1/Amt1 but found duplicated in both Mac1 and Grisea of Podospora anserina (43,46). With respect to copper detoxification in S. pombe, the only known molecules for sequestering excess of copper ions are the phytochelatins, since the fission yeast lacks metallothioneins (42). No precise molecular mechanism of how copper sequestration is regulated in S. pombe has yet been identified.
In this study, we identify a cis-acting element found upstream of the copper transport genes in S. pombe termed copper-signaling element (CuSE) that is necessary for activation in response to copper deprivation conditions. The CuSE sequence bears a strong sequence similarity to the MRE for the Ace1 and Amt1 copper metalloregulatory detoxifying transcription factors from the bakers' yeast (S. cerevisiae) and the yeast C. glabrata. When MREs from S. cerevisiae are expressed in the fission yeast S. pombe, transcriptional regulation occurs in the opposite direction from that observed for the S. cerevisiae CUP1, SOD1, and CRS5 genes, which are activated in an MRE-dependent manner in response to copper. Moreover, the transcriptional copper starvation-mediated activation that is observed when MREs of S. cerevisiae are introduced in S. pombe depends on Cuf1. UV-cross-linking experiments using extracts derived from a heterologous system expressing a functional cuf1 ϩ gene reveal that the Cuf1 protein binds the CuSE element. Taken together, these results show that the S. pombe Cuf1 nutritional copper-sensing factor acts through a closely related MRE-like element to regulate expression of fission yeast genes encoding components of the high affinity copper transport machinery.
Promoter Deletions and Site-directed Mutagenesis-The plasmid pSP1ctr4 ϩ -737 carries a 1607-bp XhoI-BamHI PCR-amplified fragment, which harbors the ctr4 ϩ promoter region up to Ϫ737 from the start codon of the ctr4 ϩ gene in addition to the gene itself. This pSP1 (52) derivative was digested with BamHI (accessible end) and SacI (protected end) to perform unidirectional nested 5Ј deletions of the ctr4 ϩ promoter. The plasmid pSKctr4 ϩ 213 containing nucleotides from position Ϫ737 to position Ϫ524 with respect to the A of the ATG codon of the ctr4 ϩ ORF was created to introduce mutations in the CuSEs by sitedirected mutagenesis. Precisely, the oligonucleotides 5Ј-Ϫ571 CTGTC-CCTAAAATTCATTCTCTTATCATGCACTAATACCTCGATTG Ϫ616 -3Ј and 5Ј-Ϫ702 CAGTCGGAATCTCATATCCCAAACAGTGACATTTC-C Ϫ737 -3Ј (letters that are underlined represent multiple point mutations in the CuSEs) were used either together or separately in conjunction with pSKctr4 ϩ 213 and the Chameleon mutagenesis kit (Stratagene, CA). The DNA sequence for each mutant was confirmed by sequencing and then used to replace the equivalent fragment from either pSP1ctr4 ϩ -737 or pCF83ctr4 ϩ -CYC1-lacZ. This latter plasmid consists of the ctr4 ϩ promoter region (position Ϫ737 to position Ϫ524) within which lie all three CuSEs (positions Ϫ721 to Ϫ712; positions Ϫ603 to Ϫ594; positions Ϫ592 to Ϫ583) followed by the minimal promoter of the CYC1 gene fused to lacZ. To assess the ability of the CuSEs to regulate a heterologous reporter gene in a copper-dependent fashion, two versions of the ctr4 ϩ promoter were tested, one containing 213 bp (positions Ϫ737 to Ϫ524) and the other harboring a shorter segment of 175 bp (positions Ϫ737 to Ϫ562). Both regions were PCR-amplified and inserted in their natural orientation into the SmaI and XhoI sites of a CYC1-lacZ fusion plasmid pCF83 (46). A similar approach was also used to create both the wild type and mutant pCF83ctr5 ϩ -CYC1-lacZ fusions. Furthermore, a DNA segment (positions Ϫ292 to Ϫ163) encompassing four MREs from the S. cerevisiae CUP1 regulatory region was also amplified by PCR to obtain both wild type and MRE mutant elements. Both types of fragments were inserted into pCFLEU2 using the SmaI and XhoI sites. Plasmid pCFLEU2 was constructed by insertion of the HindIII LEU2 fragment isolated from pINV1-spc1 ϩ -HA6His (50) into the HindIII sites of pCF83 (46), replacing the ura4 ϩ cassette for the LEU2 marker.
RNA Analysis-To conduct the RNase protection analyses as described previously (16), three plasmids, denoted pSKctr4 ϩ , pSKact1 ϩ (46), and pKSlacZ (32), were used for making antisense RNA probes, allowing the detection of steady-state levels of ctr4 ϩ , act1 ϩ , and lacZ mRNAs, respectively. pSKCUP1 was created by the insertion of a 149-bp EcoRI-BamHI PCR fragment from the S. cerevisiae CUP1 gene into the same sites of pBluescript II SK.
Protein Epitope Tagging and Detection-Both the low and high copy plasmid versions of URA3-based GPDcuf1 ϩ FLAG 2 were created using two oligonucleotides, FLAG 2 up (5Ј-GGCCGAGACTACAAGGACGAC-GATGACAAAGGCGACTACAAGGACGACGATGACAAAGAC-3Ј) and FLAG 2 low (5Ј-GGCCGTCTTTGTCATCGTCGTCCTTGTAGTCGCCTT-TGTCATCGTCGTCCTTGTAGTCTC-3Ј), encoding two copies of the FLAG epitope. These purified oligonucleotides were annealed pairwise to form double-stranded DNA with EagI overhangs and then ligated into the NotI site of pSKcuf1 ϩ NotI in which a NotI restriction site was previously engineered by PCR and placed in-frame just before the stop codon of the cuf1 ϩ gene (46). Subsequently, the cuf1 ϩ FLAG 2 was isolated from this pSK derivative with EcoRI and BamHI and then cloned into the pGEM-7Zf(ϩ) vector (Promega, WI). To release the cuf1 ϩ FLAG 2 gene from this pGEM derivative, both XbaI and BamHI were used, allowing unidirectional gene insertion into the SpeI (which produces compatible ends to that of XbaI) and BamHI sites of the p416GPD/p426GPD vector for ectopic expression in S. cerevisiae cells (51). For immunoblot analysis, S. cerevisiae DTY59 cells were transformed with p416GPD (as control) and p416GPDcuf1 ϩ FLAG 2 . Yeast cells were grown to A 600 ϭ 1.1 either with or without 10 M copper chelator BCS and with or without 100 M CuSO 4 . Total cell lysates were prepared by glass bead disruption in lysis buffer (20 mM HEPES (pH 7.9), 10% glycerol, 75 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 1 g/ml pepstatin A, and 0.5 g/ml leupeptin). The cells were lysed by vortexing for 1 min at top speed 5 times with 1-2-min intervals on ice. After centrifugation for 6 min at 4,800 rpm, the protein concentration of the supernatant was determined by the Bradford assay. Equal amounts of each sample were used to detect the Cuf1FLAG 2 protein using anti-FLAG M2 antibody (Eastman Kodak Co.) as described by the manufacturer. To carry out immunodetection of the 3-phosphoglycerate kinase, monoclonal anti-PGK antibody 22C5-D8 (Molecular Probes, OR) was used. When examined in S. pombe cuf1⌬ cells, the cuf1 ϩ FLAG 2 fusion allele complements the cuf1⌬ glycerol growth defect (data not shown). To accomplish this complementation assay, the NdeI-BamHI cuf1 ϩ FLAG 2 fragment was isolated from the p416GPDcuf1 ϩ FLAG 2 plasmid and then inserted into the NdeI and BamHI sites of pSP1cuf1 ϩ S-B/4.2 plasmid (46).
UV Cross-linking-Cultures of DTY59 (ace1⌬), expressing either the p416/426GPD vector alone or p416/426GPDcuf1 ϩ FLAG 2 , were grown to early log phase under conditions of low copper availability. Cell extracts were prepared as described above in the lysis buffer. This whole cell extract preparation was then separated into aliquots in the presence of binding buffer containing 12.5 mM HEPES (pH 7.9), 4 mM Tris-HCl (pH 7.9), 50 mM KCl, 2 mM MgCl 2 , 2.5 M ZnSO 4 , 0.6 mM dithiothreitol, 0.05% Nonidet P-40, 3% Ficoll, and 0.5 g of random double-stranded DNA (pdN250) and incubated for 20 min at 25°C in the presence of 1 ng of oligomer probe. Typically, the probe was prepared by hybridizing two oligomers denoted CuSE1up (5Ј-GTCACTGTTTGTGATGCTGGATTC-CGACT-3Ј) and CuSE1lo (5Ј-ATCAGTCGGAATCCAGCA-3Ј) that are partially complementary to each other by 15 bp. Once annealed, the oligomers were made completely double-stranded by incubation with the Klenow fragment of DNA polymerase I in the presence of the 5-bromo-2Ј-deoxyuridine triphosphate (BrdUrd) (7.7 mM, made up in 40 mM Tris-HCl (pH 9.0) and 280 mM NaOH) and [␣ 32 P]dCTP for 90 min at 25°C, and then the reaction was chased with dATP, dGTP, and dTTP for an additional 90 min. The gel-purified BrdUrd and [␣ 32 P]dCTP oligomer probe that was used corresponds to the CuSE found at positions Ϫ721 to Ϫ712. After DNA protein binding, reaction mixtures were irradiated for 40 min at 4°C with an average intensity of 600 microwatts/cm 2 . The reactions were stopped in Laemmli buffer and heated before SDS-polyacrylamide gel electrophoresis.

Identification of Copper-signaling cis-Acting Elements Necessary for Copper Starvation Activation of the ctr4 ϩ Transporter Gene Expression-
We have demonstrated previously that in S. pombe, expression of the high affinity copper transporter ctr4 ϩ is dependent on the Cuf1 copper-sensing transcription factor (46). To identify the cis-acting elements necessary for copper starvation activation of copper transport gene expression by Cuf1, we constructed nested 5Ј deletions of the ctr4 ϩ promoter (Fig. 1). As shown in Fig. 1, all constructs were analyzed in a ctr4⌬ strain, ensuring that the only ctr4 ϩ transcripts detected were from the recombinant plasmids. Removal of the ctr4 ϩ upstream region between Ϫ737 and Ϫ524 completely abolished the BCS up-regulation of the ctr4 ϩ gene. Similarly, further deletion to position Ϫ331 gave constitutive levels of gene expression with complete failure to repress or activate gene expression in response to copper ions. Owing to the observation that the integrity of the region between positions Ϫ737 and Ϫ524 is essential to down-regulate (ϳ3-fold) and up-regulate (ϳ4-fold) ctr4 ϩ mRNA expression in the presence of copper or BCS, respectively, we sought to identify the cis-acting elements responsible for copper-regulated gene expression within this promoter region. Interestingly, the ctr4 ϩ promoter region between Ϫ737 and Ϫ524 harbors three copies of a repeated sequence, 5Ј-D(T/A)DDHGCTGD-3Ј (D ϭ A, G, or T; H ϭ A, C, or T), that is similar to the binding sites for Ace1 and Amt1 copper MRTFs from the yeasts S. cerevisiae and C. glabrata, respectively (16). To examine whether these MRElike ctr4 ϩ promoter sequences could mediate gene expression in response to copper, we inserted multiple point mutations that mimic changes known to abolish binding of Ace1 to MREs in either the first element (positions Ϫ721 to Ϫ712), the two other elements found downstream (positions Ϫ603 to Ϫ594 and Ϫ592 to Ϫ583), or in all three elements (Fig. 2). Although the overall magnitude of the response to copper is optimal with the presence of all three elements, the presence of only one of the three elements is sufficient to confer regulation in response to copper levels (Fig. 2). As compared with the wild type promoter segment, ϳ70 and 60% of the response was still observed when either the first element or the two other elements were unaltered, respectively, suggesting that the first element (positions Ϫ721 to Ϫ712) is slightly favored for copper starvation induction. Although a low basal level of expression was observed when all three elements were mutated, there was a complete lack of either down or up-regulation of the ctr4 ϩ gene. Furthermore, deletion of 24 bp (from Ϫ737 to Ϫ713) of the ctr4 ϩ promoter already harboring mutated elements abolished even the low basal gene expression of ctr4 ϩ (Fig. 2). Taken together, these data show that a conserved element in the ctr4 ϩ promoter, which we term CuSE, with the sequence 5Ј-D(T/ A)DDHGCTGD-3Ј (D ϭ A, G, or T; H ϭ A, C, or T) is required, at least in a single copy, for both repression of basal expression in the presence of copper ions and activation under copper starvation conditions.
The CuSEs Confer Copper Responsiveness to the Minimal Promoter CYC1 Fused to lacZ-To examine whether the CuSEs could regulate a heterologous reporter gene in a copper-dependent manner, the ctr4 ϩ promoter segment of 213-bp (positions Ϫ737 to Ϫ524) was inserted into the minimal promoter of the CYC1 gene fused to lacZ (Fig. 3). In the presence of copper ions, this fusion promoter gene was repressed by ϳ9 -12-fold. Conversely, under copper starvation conditions, lacZ mRNA expression was induced (ϳ3-fold). Interestingly, CYC1-lacZ fusion genes containing the previously described CuSE mutants (Fig. 3) derived from the ctr4 ϩ promoter repress lacZ mRNA with an overall efficiency similar to that observed in the endogenous ctr4 ϩ gene, except for the magnitude of the response, which is more pronounced when a heterologous system is used. Importantly, the sequence 5Ј-D(T/A)DDHGCTGD-3Ј (D ϭ A, G, or T; H ϭ A, C, or T) was also found in multiple copies in the ctr5 ϩ promoter, which drives the expression of a novel gene encoding an indispensable partner of the Ctr4 copper transporter involved in high affinity copper uptake in S. pombe. 2 As illustrated in Fig. 4, fusion of short regions containing 175 bp from the ctr4 ϩ or ctr5 ϩ promoters encompassing CuSE elements were able to copper-regulate expression of lacZ mRNA. Although the ctr4 ϩ promoter harboring this shorter fragment (positions Ϫ737 to Ϫ562) of 175 bp is regulated in a copper-dependent manner, the overall magnitude of the response decreases by ϳ70% as compared with the longer fragment (positions Ϫ737 to Ϫ524) of 213 bp. For both promoter fusions, ctr4 ϩ -CYC1-lacZ and ctr5 ϩ -CYC1-lacZ, the integrity of the CuSEs is essential for copper-responsive gene expression since CuSE mutants abrogate any regulation by copper (Fig. 4). Therefore, these results suggest that the S. pombe copper transport genes share a common promoter element, denoted CuSE (5Ј-D(T/A)DDHGCTGD-3Ј), which is necessary for activation in response to copper limitation.
Cuf1 Acts through Metal Regulatory Elements-Although we have demonstrated previously that Cuf1 regulates the expression of essential copper transport genes in S. pombe, the Cuf1 amino-terminal 61 amino acids more closely resemble aminoterminal 63 and 62 amino acids of Ace1 (51%) and Amt1 (45%) copper-detoxifying factors than the corresponding region of Mac1 (39%) (46). Based upon this observation and because of our finding that the CuSE bears a strong sequence similarity to the Ace1 and Amt1 binding sites (MREs), we sought to assess whether MREs of the S. cerevisiae CUP1 gene could serve in S. pombe for gene regulation as a function of copper availability by Cuf1. In the cuf1 ϩ strain, when a 129-bp CUP1 promoter DNA fragment (positions Ϫ292 to Ϫ163) encompassing four MREs was inserted in its natural orientation into the minimal promoter of the CYC1 gene fused to lacZ, the expression of lacZ mRNA was indeed copper-regulated, exhibiting induction in the presence of the Cu 1ϩ chelator BCS and repression upon the addition of copper (Fig. 5). In the cuf1⌬ mutant strain, although a high basal level of lacZ mRNA was detected in the presence of the wild type MREs, perhaps owing to the action of another transcription factor, the steady-state levels of lacZ mRNA were unregulated by modulation of copper status (Fig.  5). Furthermore, for both the wild type and mutant strains, no regulation by copper was observed when the MRE elements were mutated (Fig. 5). Therefore, when MREs from S. cerevisiae are introduced in the fission yeast S. pombe, transcriptional regulation occurs in the opposite direction from that observed for the S. cerevisiae CUP1, SOD1, and CRS5 genes, in which the presence of the MRE fosters the activation of gene expression in response to copper.
Cuf1 Interacts Directly with CuSEs-Based on the gene expression data we obtained, we predicted that the Cuf1 factor 2 D. J. Thiele, personal communication.

FIG. 2. The CuSEs confer copper responsiveness to the ctr4 ؉ promoter.
A, wild type (WT) DNA promoter and mutants (M) of the conserved element found in the ctr4 ϩ promoter were analyzed by RNase protection assay. Although the overall magnitude of the response is optimal with the presence of all three CuSEs (WT1-3), the presence of only one of the three elements (WT1, M2-3) is sufficient to confer the regulation in response to copper levels. Results illustrated are representative of three independent experiments. B, the gray boxes depict the wild type elements, and the filled boxes represent the mutated versions of the CuSEs. ⌬1 indicates a promoter deletion to position Ϫ713. For each of these derivatives, values for fold repression by copper and activation by BCS are shown.
directly interacts with CuSEs to mediate copper regulation. To test this hypothesis, we generated a Cuf1FLAG 2 fusion protein, allowing its detection in DNA-protein complexes using commercially available anti-FLAG M2 antibody. Importantly, to test whether insertion of the FLAG tags interfered with Cuf1 function, the Cuf1FLAG 2 fusion protein was tested for its ability to complement the cuf1⌬ nonfermentable carbon source growth defect in S. pombe. The tagged Cuf1 protein fully complemented the respiratory deficiency (data not shown). Fur-thermore, as assessed by RNase protection experiments when transformed into the cuf1⌬ fission yeast strain, the cuf1 ϩ FLAG 2 allele accurately functions to regulate the ctr4 ϩ gene expression in response to copper ions in a manner indistinguishable from the cuf1 ϩ wild type allele (Fig. 6). To examine whether the Cuf1 protein interacts with CuSE, we used the cuf1 ϩ FLAG 2 fusion gene in a heterologous context as the only S. pombe gene expressed in a bakers' yeast system (51) lacking the endogenous ACE1 gene. The ace1⌬ strain was utilized to  Wild type (WT) DNA fragments and CuSE mutants derived from both ctr5 ϩ and ctr4 ϩ (as control) promoters were inserted in their natural orientation into the CYC1 minimal promoter fused to lacZ. Results illustrated are representative of three independent experiments. B, shown is a schematic representation of the plasmid derivatives assayed and quantitation of fold copper repression and BCS derepression. The nucleotide numbers refer to the position relative to the A of the ATG codon of the ctr4 ϩ and ctr5 ϩ ORFs. ensure that Cuf1 was the sole protein with the ability to recognize the DNA probe, which contains a GCTG core with a T located four nucleotides upstream of the first G in the GCTG core. When expressed, the cuf1 ϩ FLAG 2 gene gave rise to fulllength Cuf1 protein with an electrophoretic mobility on SDSpolyacrylamide gel electrophoresis corresponding to ϳ54-kDa (Fig. 7A). No Cuf1 protein was detected from bakers' yeast extract preparations containing the expression plasmid alone (Fig. 7A). As observed previously for S. pombe (46), no drastic change in the ability to detect Cuf1 was observed when the bakers' yeast cells were grown either with or without exogenous copper ions (Fig. 7A). Interestingly, in the S. cerevisiae ace1⌬ strain expressing Cuf1FLAG 2 , the CUP1-encoded metallothionein is regulated by copper in the opposite direction, down-regulated by copper instead of exhibiting a rapid induc-tion after exposure to copper as compared with the CUP1 gene expressed in the wild type S. cerevisiae ACE1 strain (Fig. 7B). This observation is consistent with the manner by which Cuf1 regulates MRE elements from S. cerevisiae when these elements are introduced in S. pombe (Fig. 5). Importantly, the Cuf1FLAG 2 protein appears to conserve its property of copperresponsive DNA binding when ectopically expressed in S. cerevisiae. Using extracts prepared from S. cerevisiae ace1⌬ cells expressing the Cuf1FLAG 2 protein, UV-cross-linking experiments revealed a Cuf1-dependent complex of ϳ64-kDa, which is consistent with the predicted ϳ10-kDa that the covalently attached oligomer is anticipated to add to the mass of the Cuf1 protein (Fig. 8A). Furthermore, this DNA-protein complex of ϳ64-kDa was absent when S. cerevisiae ace1⌬ cells expressing the Cuf1FLAG 2 protein were grown under copper-replete conditions (100 M CuSO 4 ) (Fig. 8A). Furthermore, no DNA-protein complex was observed when the CuSE oligomer probe contained multiple point mutations (5Ј-GTCACTGTTTGG-GATATGAGATTCCGACTGAT-3Ј instead of 5Ј-GTCACTGTT-TGTGATGCTGGATTCCGACTGAT-3Ј) (data not shown). To test further whether the DNA binding complex observed is due to the Cuf1FLAG 2 protein, duplicates of each of the irradiated samples shown in Fig. 8A were resolved simultaneously on an SDS-polyacrylamide gel and then transferred for analysis by immunoblotting using anti-FLAG M2 antibody. As shown in Fig. 8B, when UV irradiation was omitted, Cuf1FLAG 2 mi- Bakers' yeast ace1⌬ cells transformed with pGPD416Cuf1FLAG 2 were grown to early log phase in a modified SD media containing ϳ16 nM copper (32). After no treatment (Ϫ) or incubation in the presence of either 100 M CuSO 4 or 100 M BCS (B), protein extracts were prepared from the cells and then analyzed by immunoblotting using either anti-FLAG M2 or anti-PGK (as an internal control) antibody. We note that the tagged Cuf1 migrates at ϳ54-kDa despite a predicted molecular mass of 45-kDa. B, the ace1⌬ strain, transformed with either the low-copy plasmid expressing cuf1 ϩ FLAG 2 or the vector alone, was grown to mid-log phase and treated with CuSO 4 (100 M) or BCS (B) (100 M) for the indicated time. After total RNA isolation, shown is a representative RNase protection analysis of CUP1 and ACT1 mRNA levels, which exhibits repression of CUP1 mRNA expression in the presence of copper in a Cuf1FLAG 2 -dependent manner. As a control (right side), the S. cerevisiae isogenic wild type strain (ACE1) displays high steady-state levels of CUP1 mRNA in response to copper. M, reference marker.
grates at an apparent molecular mass of 54-kDa (unbound form (U)). When irradiated, in addition to the unbound form (U) of Cuf1FLAG 2 (ϳ54-kDa), a higher molecular mass band (ϳ64-kDa) of slower electrophoretic mobility (bound form (B)) was detected (owing to the covalently attached oligomer) with anti-FLAG M2 antibody. Importantly, this higher molecular mass band of ϳ64-kDa was only observed when extracts were derived from cells grown under copper-limiting conditions (Fig.  8B). No such band of slower electrophoretic mobility was observed when extracts were prepared from copper-replete cells expressing Cuf1FLAG 2 (Fig. 8B). Because of the identical molecular mass of ϳ64-kDa observed by both immunoblotting and autoradiography of the Cuf1-DNA binding complex found by UV cross-linking (Fig. 8A), the results strongly suggest the presence of the Cuf1FLAG 2 protein in this complex. Taken together, these data reveal that the S. pombe Cuf1 protein activates copper transport gene expression under copper deprivation conditions by directly interacting with the conserved element, 5Ј-D(T/A)DDHGCTGD-3Ј, denoted CuSE, which exhibits a striking similarity with the S. cerevisiae MRE (Fig. 9). DISCUSSION The cuf1 ϩ gene encodes a nuclear copper-sensing transcription factor that occupies a central role in the S. pombe high affinity copper transport system (46). Indeed, we have shown previously that deletion of the cuf1 ϩ gene results in at least three phenotypes associated with copper starvation in yeast cells: inability to use respiratory carbon sources, impaired superoxide dismutase activity, and defects in iron accumulation (46). All three phenotypes are restored by the addition of exogenous copper to the medium, consistent with cuf1⌬ strains being specifically defective in the expression of genes required for high affinity copper transport. Furthermore, the previous observation that in S. pombe, expression of the high affinity copper transporter Ctr4 is dependent on the Cuf1 protein defines a critical role for trans-activation of copper transport gene by Cuf1 (46).
To identify the cis-acting elements necessary for copper starvation activation of copper transport gene expression by Cuf1, we have conducted detailed studies on the ctr4 ϩ promoter. Although there are two copies of the Mac1 binding site, 5Ј-TTTGC(T/G)C(A/G)-3Ј, denoted CuRE (32,33), located between positions Ϫ323 and Ϫ381 in the ctr4 ϩ promoter, no effect on copper-responsive regulation of the ctr4 ϩ promoter was observed when these two elements were mutated. 3 This latter observation is also consistent with the inability for Cuf1 to suppress a number of copper-remedial phenotypes associated with mac1⌬ disruption when the S. pombe cuf1 ϩ gene is ectopically expressed in a S. cerevisiae mac1⌬ strain. 3 Our studies have revealed that the ctr4 ϩ and ctr5 ϩ promoters are regulated in response to fluctuations in copper concentrations by a consensus cis-acting sequence, 5Ј-D(T/A)DDHGCTGD-3Ј (D ϭ A, G, or T; H ϭ A, C, or T) termed CuSE, which is highly similar to the binding sites for the Ace1 and Amt1 copper MRTFs, MREs. This finding is consistent with the high percentage of homology within residues 1-61 of Cuf1 amino terminus to the amino-terminal 63 (51% identity) and 62 (45% identity) amino acids of Ace1 and Amt1, respectively (46). This amino-terminal region of Cuf1 harbors a Zn 2ϩ module domain known to be involved in minor groove DNA contacts with the AT-rich element (53,55) and also appears to share with Ace1 and Amt1 critical residues involved in recognition of the core sequence. Interestingly, only a single CuSE is sufficient for near-maximal copper deprivation activation of the ctr4 ϩ gene expression. Similarly, both the Ace1 and Amt1 copper MRTFs can mediate their action through a single MRE, which is not the case for the Mac1 protein, which requires at least two copies of the CuREs for copper-regulated gene expression (32,38). Although few 3 S. Labbé, unpublished data.
FIG. 8. Cuf1FLAG 2 protein ectopically expressed in bakers' yeast binds to CuSE. A, representative UV-cross-linking gel of the wild type BrdUrd-and 32 P-labeled oligomer CuSE used with whole cell extracts prepared from S. cerevisiae ace1⌬ cells expressing either the plasmid alone or Cuf1FLAG 2 (pGPDcuf1 ϩ FLAG 2 ) as their sole source of Cuf1. Only cells grown under copper-limiting conditions expressing cuf1 ϩ FLAG 2 allele exhibits the CuSE oligomer-protein complex (arrow), which is absent in extract preparations from cells expressing cuf1 ϩ FLAG 2 allele but treated with CuSO 4 (100 M). No oligomerprotein complex is observed with extracts prepared from cells expressing the plasmid alone. P, probe DNA irradiated in the absence of protein. B, five UV-irradiated reactions of each of the three binding reactions shown in A were pooled and resolved by SDS-polyacrylamide gel electrophoresis. By immunoblotting using anti-FLAG M2 antibody, a fraction of the Cuf1FLAG 2 protein was found shifted,m presumably because of its irreversible attachment to CuSE by UV irradiation. This slower electrophoretic form of Cuf1FLAG 2 (bound (B)) migrates at ϳ64 kDa, as observed for the CuSE oligomer-protein complex shown in A. Only one form (unbound (U)) of the Cuf1FLAG 2 protein was detected at ϳ54 kDa in extract preparations from cells treated with CuSO 4 (100 M). As a control (right side), UV irradiation was omitted. V, extracts from cells expressing the vector alone. M, reference marker.
FIG. 9. Comparison of CuSE sequences from S. pombe copper transport genes with the Ace1 and Amt1 consensus MRE binding site. The conserved residues shown in gray play a critical role in copper-regulated gene expression of the S. pombe ctr4 ϩ and ctr5 ϩ transport genes. As highlighted for MREs (16), CuSEs possess four nucleotides perfectly conserved, GCTG (also named core element) and a T (or A for one CuSE) located four nucleotides upstream of the first G in the GCTG. The nucleotide numbers refer to the position relative to the A of the start codon of the indicated ORF. differences are observed between the CuSE and MRE consensus sequences (Fig. 9), the binding sites appear to be interchangeable since when MREs from S. cerevisiae are introduced in the fission yeast S. pombe, Cuf1 can BCS-activate target gene expression through those MREs (Fig. 5). It is anticipated that some nucleotides within the CuSE should optimize copper starvation induction of gene expression. Which nucleotides contribute to the magnitude of this regulatory response must await a comprehensive dissection of the CuSE. With respect to the conserved T-positioned four nucleotides upstream of the GCTG core element (Fig. 9), it has been shown previously (16) that its substitution by adenine results in ϳ20% reduction in DNA binding by Amt1. Based on this observation, we would predict that the CuSE between positions Ϫ596 and Ϫ581 has a weaker strength for copper transcriptional regulation.
To investigate further the mechanism by which Cuf1 regulates target gene expression through CuSEs in response to copper, we have used a S. cerevisiae yeast system (51) in which the endogenous ACE1 gene was inactivated. By using this approach, we sought to ensure the presence of Cuf1 as sole protein with the ability to recognize MRE-like elements. A heterologous test is essential, because we have identified a second gene in S. pombe, denoted cuf2 ϩ (SPCC584.02). This putative MRTF exhibits 41% identity to the amino-terminal 61 amino acids of Cuf1. 3 When expressed in S. cerevisiae ace1⌬ cells under conditions in which copper is scarce, the Cuf1 protein fosters high basal levels of CUP1 mRNA. Conversely, when S. cerevisiae ace1⌬ cells were grown in the presence of copper, the CUP1 mRNA levels were repressed by Cuf1, consistent with the manner by which Cuf1 regulates copper transport gene expression in S. pombe. To determine whether Cuf1 binds to CuSE when ectopically expressed in S. cerevisiae ace1⌬ cells, a BrdUrd-double-stranded CuSE oligomer derived from the ctr4 ϩ promoter (positions Ϫ700 to Ϫ731) was used for UV-cross-linking analysis (54). Precisely, the BrdUrd was placed on the noncoding strand two nucleotides upstream of the GCTG core element (position Ϫ718) since the Cuf1 protein is predicted to interact with one face of the DNA double helix at adjacent minor (encompassing that position Ϫ718) and major (containing the GCTG core element) grooves. Indeed, when the reactive base (BrdUrd) was positioned into the CuSE, a DNA-Cuf1-dependent complex was observed (Fig. 8). Interestingly, this DNA-protein complex of ϳ64-kDa was only observed when the extract preparations were obtained from yeast cells grown under conditions of low copper availability. When cells were grown in the presence of copper, no such DNA-protein complex was detected.
How does Cuf1 function in copper deprivation activation through the Ace1/Amt1-like recognition sequence? It is intriguing that Cuf1 displays DNA binding specificity similar to that of S. cerevisiae Ace1 and C. glabrata Amt1, yet clearly this recognition sequence is used for copper transporter gene expression in S. pombe. Although the first 61 residues of Cuf1 exhibit a strong homology to this same region of Ace1 and Amt1, Cuf1 does not possess the second half of the Ace1/Amt1 copper regulatory domain in which two highly conserved Cys-X-Cys sequence motifs are found (46). These two pairs of Cys that are missing in Cuf1 make improbable the formation of Ace1/Amt1-like copper regulatory domain that consists of two lobes separated by a cleft in which the Cu 4 S 6 center takes place in the presence of copper ions (17). Interestingly, Cuf1 harbors a REP-like motif near its carboxyl terminus, which shows a high percentage of identity (48%) to a similar domain found in S. cerevisiae Mac1 (46). Because in Mac1, the REP-I domain is able to sense copper ions, fostering intramolecular conformational changes and thereby inactivating the Mac1 DNA binding and its transactivation function, the possibility exists that the Cuf1 REP-like domain acts in a similar manner. Therefore, Cuf1 resembles two distinct types of copper MRTFs, sharing recognition sequence specificity to bind DNA of the Ace1 and Amt1 proteins while playing an essential role in the regulation of the high affinity copper transporter protein gene expression in S. pombe as found for the Mac1 protein in S. cerevisiae.