Evidence for (Mac1p)2·DNA Ternary Complex Formation in Mac1p-dependent Transactivation at theCTR1 Promoter*

The Mac1 protein in Saccharomyces cerevisiae is required for the expression CTR1 andFRE1, which, respectively, encode the copper permease and metal reductase that participate in copper uptake. Mac1p binds to a core GCTC sequence present as a repeated unit in the promoters of both genes. We show here that Mac1p DNA binding required an intact N-terminal protein domain that includes a likely zinc finger motif. This binding was enhanced by the presence of a TATTT sequence immediately 5′ to the core GCTC, in contrast to a TTTTT one. This increased binding was demonstrated clearly in vitro in electrophoretic mobility shift assays that showed Mac1p·DNA complex formation to a single TATTTGCTC element but not to a TTTTTGCTC one. Furthermore, the fraction of Mac1p in a ternary (Mac1p)2·DNA complex in comparison to a binary Mac1p·DNA complex increased when the DNA included two TATTTGCTC elements. A similar increase in ternary complex formation was demonstrated upon homologous mutation of the FRE1Mac1p-dependent promoter element. The in vivoimportance of this ternary complex formation at the CTR1promoter was indicated by the stronger trans-activity of this promoter mutated to contain two TATTT elements and the attenuated activity of a mutant promoter containing two TTTTT elements that in vitrosupported only a weak ternary complex signal in the shift assay. The stronger binding to TATTT appeared due to a more favorable protein contact with adenine in comparison to thymine at this position. An in vivo two-hybrid analysis demonstrated a Mac1p-Mac1p protein-protein interaction. This Mac1p-Mac1p interaction may promote (Mac1p)2·DNA ternary complex formation at Mac1p-responsive upstream activating sequences.

The Mac1 protein is a 46-kDa polypeptide that is essential for the copper-regulated expression of high affinity copper uptake activity in the budding yeast Saccharomyces cerevisiae (1)(2)(3)(4)(5)(6)(7). 1 Several features of the regulatory activity of Mac1p have been described. For example, in vitro EMSA 2 has demonstrated that Mac1p binds to a sequence in upstream activating regions of two genes whose expression is known to be Mac1pdependent, CTR1 and FRE1 (4). These two genes encode a high affinity copper permease (9) and metal reductase (10), respectively. The latter activity is required for both copper and iron uptake in yeast, that is reduction of medium Cu(II) and Fe(III) to the lower valent species is an essential first step of the accumulation of both metal ions (2,11). The sequence element associated with Mac1p binding in these two loci is shown in Scheme I. Inspection of these two fragments in CTR1 (9) and FRE1 (10) and in other activating sequences thought to be associated with Mac1p regulation (3,7,12) indicates that the core binding site is given by GCTC (shown in boldface). For example, Mac1p-and copper-dependent protection of this region in one of these other loci, CTR3, which encodes a second copper permease not expressed in all yeast strains, has been demonstrated by in vivo DNA footprint analysis (7).
Ϫ255 5Ј ATTTTTGCTCACCTTTTTTTTTTGCTCATCGAAAA 3Ј 3Ј TAAAAACGAGTGGAAAAAAAAAACGAGTAGCTTTT 5Ј SCHEME I In addition to these studies on Mac1p DNA binding activity, use of Mac1p fusions to heterologous DNA binding domains (DBD), e.g. that from Gal4 (5) and from the LexA protein (6), has provided evidence that Mac1p also has an inherent transactivation activity that is copper-dependent. The copper dependence of DNA binding and transactivation activity is negative, that is both activities are expressed in copper-deficient cells but are suppressed in copper-replete ones. The downregulation of Mac1p activity occurs between 1 nM and 1 M; at [copper] medium Ն1 M little expression from the CTR1 promoter is observed (5,9).
As shown in Scheme I, the core GCTC-binding site in the CTR1 promoter is found within a nearly perfect palindrome; this element in the FRE1 locus is a direct repeat (sequences underlined in Scheme I). CTR3 also contains an inverted repeat which, however, is separated by 44 base pairs. That both GCTC sites are essential for transactivation by Mac1p has been indicated by previous studies that demonstrated loss of such activity in reporter promoter constructs derived from the FRE1 (10), CTR1 (4), and CTR3 (7) loci which contained only a single GCTC-containing element. The presence of this repeated motif suggested to us the possibility of Mac1p binding to both sites simultaneously if not cooperatively.
Thus, the objective of the work described herein was to delineate in vitro the gross structural features of the Mac1p⅐DNA complex that had been demonstrated by EMSA (4) and suggested by in vivo DNA footprint analysis (7) and to correlate these features with promoter activity in vivo. We first tested the hypothesis that the N-terminal domain of Mac1p, which contains a CCHC zinc finger motif (1,13) and which is homologous to the N-terminal regions of two other yeast transfactors, Ace1p (14 -16) and Amt1p (17)(18)(19)(20), was essential to the binding of Mac1p to DNA. We next evaluated the possible formation of ternary complexes, that is (Mac1p) 2 ⅐DNA species, and determined what sequence characteristics in the DNA promoted such complex formation. We show a correlation between the in vivo activity of a promoter sequence and its ability to support formation of a (Mac1p) 2 ⅐DNA ternary complex in vitro. Finally, we present data from a two-hybrid analysis that demonstrate that in this assay, at the least, a Mac1p-Mac1p interaction can be demonstrated. The data indicate that Mac1p can and does bind at both sites in the CTR1 promoter simultaneously and that formation of the resulting ternary complex correlates to downstream transcriptional activity.

EXPERIMENTAL PROCEDURES
Strains of S. cerevisiae and Growth Media-Three yeast strains were used in this study. For the one-and two-hybrid analyses, strain SFY526 (obtained from CLONTECH) was used as host (MATa ura3-52 his3-200 ade2-101 lys2-801 trp1-901 leu2-3, 112 can r gal4-542 gal80-538 URA3::GAL1-lacZ). For the promoter activity studies using CTR1 promoter-lacZ reporter plasmids, DEY1457 was used as host (MAT␣ ade6 can1 his3 leu2 trp1 ura3). This strain was obtained from David Eide (21). For test of the complementing activity of mutant forms of Mac1p (⌬ZFMac1p and ZF*Mac1p), DEY1457(mac1⌬) was used as host (DEY1457 mac1::TRP1) (4). The following media were used for culture growth: YPD was used for routine growth of wild type strains (2% yeast extract, 2% peptone, 2% glucose); SC medium was used for routine selective growth of transformed strains (1.67 g/liter yeast nitrogen base without amino acids, 2% glucose plus the appropriate drop-out (otherwise complete)) mixture of amino acids; for growth of copper-depleted cultures, a completely synthetic, Chelex-treated medium was used that was prepared as described previously (2). This medium contained an estimated 0.5 nM residual copper based on the amount of copper in the individual media components (as determined by flameless atomic absorption spectrophotometry) and the affinity of the Chelex resin for copper. This medium was subsequently supplemented with a trace metal mixture lacking copper to which copper was added as required for a given experiment. This medium was also used with glycerol replacing glucose as carbon source to test for the respiration competence of DEY1457 (mac1⌬) when performing functional tests of mutant forms of Mac1p. This is a standard test for mac1⌬-complementing clones that takes advantage of the iron (and, consequently, respiratory) deficiency in this background (1).
Plasmid Construction-All PCR-generated fragments, mutations, and reading frames in the plasmid constructs described below were confirmed by dideoxy sequencing using the T7 Sequenase 2.0 kit from Amersham Life Sciences, Inc. The DNA sequence for the Mac1p ORF was derived from the plasmid MAC16 that was obtained by inserting a MAC1 SacI (Ϫ162) to HindIII(ϩ1997) fragment (1) into the complementary restriction sites in vector pRS316 (22). The ORF for ⌬ZFMac1p (lacking residues 1-41) was constructed using the endogenous NcoI site (ϩ121) in MAC1 which includes the codon for Met-42. The NcoI (filled in)/SalI fragment from MAC16 was ligated into the SmaI/SalI sites in pGEM3Zf(ϩ) (Promega) generating pGEM-⌬ZF for in vitro synthesis of ⌬ZFMac1p and into the bait pGBT9 (giving pGB-⌬ZF) and catch pGAD424 (giving pGAD-ZF) vectors for one-and two-hybrid analyses (23). To construct whole Mac1p ORF fusions, the endogenous NcoI (ϩ121) site in MAC16 was destroyed by PCR mutagenesis, and a new NcoI site was engineered at the ϩ1 nucleotide (translation start). This modified NcoI(Ϫ1)/PstI(ϩ210) fragment from MAC16 was exchanged with the NcoI(ϩ121)/PstI(ϩ210) fragment in pGB-⌬ZF to generate pGB-MAC, comprising the whole Mac1p ORF. The recombinant NcoI(Ϫ1)/SalI(ϩ1927) fragment from pGB-MAC was then subcloned into the catch vector pGAD424 (giving pGAD-MAC) and pGEM3Zf(ϩ) (giving pGEM-MAC) as above. The double mutant ZF*Mac1p (C23S/ H25N) was engineered by two rounds of PCR mutagenesis using pGEM-MAC as the template. The NcoI(Ϫ1)/HindIII(ϩ1927) fragment from pGEM-ZF*MAC was filled-in and subcloned into the SmaI sites in pGBT9 and pGAD424 to make plasmids pGB-ZF*MAC and pGAD-ZF*MAC, respectively.
⌬ZF-Mac1p and ZF*Mac1p were functionally tested by complementation in DEY1457⌬mac1 in the following manner. The respective ORFs from pGEM-⌬ZF and pGEM-ZF*MAC were subcloned into vector pADH; the resulting recombinants were subsequently transformed into the Mac1p minus strain and the transformants tested for respiration competence as described above. pADH was constructed from pGAD424 by replacing the HindIII fragment in this vector that contained all sequences from GAL4 by a multiple cloning site containing an NcoI site. Thus, pADH contained the yeast 2 replication origin, the promoter and terminator from ADH1, and LEU2 for plasmid selection (23).
The wild type CTR1 promoter-lacZ fusion reporter plasmid was obtained from Andrew Dancis (see Ref. 9). The two promoter mutants, C5ЈAT-lacZ and C3ЈTA-lacZ, were derived from the wild type by sitedirected mutagenesis using the Quick-Change kit from Stratagene. For the former plasmid, the complementary primers used were 5Ј-GCAAA-TCATGGGATTTTTGCTCAAGAC and 5Ј-GTCTTGAGCAAAAATCCC-ATGATTTGC (A to T transition underlined), whereas for the latter the primers used were 5Ј-CGGTAAAATGAGCAAATATGGCACGATCC and 5Ј-GGATCGTGCCATATTTGCTCATTTTACCG (T to A transition underlined).
EMSAs-The double-stranded oligonucleotides used in the EMSAs are given in Table I. A typical binding reaction contained 2-10 fmol of radiolabeled probe, 0.5 g of salmon sperm DNA, 12% glycerol, 12 mM Hepes-NaOH (pH 7.9), 60 mM KCl, 5 mM MgCl 2 , 4 mM Tris-HCl (pH 8.0), 0.6 mM EDTA, 0.6 mM dithiothreitol, and competitor DNA when present (14). This 15-l binding mixture, in addition, contained 3-5 l of a wheat germ extract transcription-translation reaction (Promega) that had been programmed with either vector alone (typically pGEM3Zf(ϩ)) or with vector containing either wild type or mutated Mac1p-encoding sequences. Binding reactions were performed by preincubation of all components except labeled probe for 10 min at room temperature with or without competitor DNA; 2-10 fmol of radiolabeled probe was then added, and the mixture was incubated for another 10 min at room temperature. The mixture was chilled on ice and then electrophoretically resolved on a 6.0% polyacrylamide gel at 4°C. The gel was dried and exposed to a PhosphorImager screen and to Kodak Biomax TM MR film. The screen was then read using a Bio-Rad model GS-505 PhosphorImager, and the digitized intensity data were then quantitated using Molecular Analyst 1.5. The EMSA figures herein were imaged directly from these digitized intensities. In Figs. 2-6, the relative amount of a given complex with respect to the control is shown using a bar graph; the control was the radioactivity associated with the faster migrating complex formed with the WT CTR1 promoter probe which we have assumed represents a binary Mac1p⅐DNA complex. Note that in some experiments either CuCl 2 (up to 100 M) or bathocuproine disulfonate (100 M) were added to determine a potential effect of free [copper] on the EMSA results. None of these treatments had a specific effect as has been reported (4).
To determine if the amount of input Mac1p protein was independent of the MAC1 construct used to program the wheat germ extract, each construct was expressed in the presence of [ 35 S]methionine, the reaction mixture resolved on SDS-polyacrylamide gel electrophoresis, and the dried gel developed by autoradiography. In all cases, a single Mac1p product was detected by appropriate molecular weight. Based on quantitation of the film by image analysis (Bio-Rad Gel Doc system) equivalent amounts of protein were produced from all templates.
␤-Galactosidase Assays-Yeast transformants were grown in the appropriate Chelex-treated/copper-supplemented defined medium for at least 5 doublings until the cultures reached mid-log phase (O.D. 660 nm ϭ 1.5-2.0). Samples (2 ϫ 10 7 cells) were assayed in triplicate for ␤-galactosidase activity which was expressed in Miller units in the standard fashion (24). The data in Fig. 7 are from three independent experiments (samples in triplicate), and statistical significance of differences in mean values was determined with the use of InStat (Graph-Pad, San Diego, CA).

Mac1p N-terminal Zinc Finger Element Is Required for DNA
Binding-Two mutant forms of Mac1p were tested to demonstrate that DNA binding was due at least in part to the Nterminal portion of Mac1p (1) that was homologous to the zinc finger, DNA-binding domain in both Ace1p (15,16) and Amt1p (17). One mutant had an N-terminal 41-amino acid deletion (⌬ZFMac1p). The other was a C23S/H25N double mutant designated ZF*Mac1p; Cys-23 and His-25 are two of the five conserved residues in the zinc finger-like motif found in the three trans-factors. These two constructs, as well as wild type Mac1p, were then tested by EMSA as shown in Fig. 1 using the wild type CTR1 Mac1p-specific promoter element as probe (see Scheme I and Table I). Binding of wild type Mac1p to this probe was clearly evident (lane 4). In contrast, neither mutant protein yielded a detectable level of protein-DNA complex in this assay (⌬ZFMac1p, lane 6; ZF*Mac1p, lane 8). Furthermore, neither mutant construct was capable of complementing a deletion in MAC1 when expressed in a ⌬mac1-containing background (see "Experimental Procedures" and data not shown). These results most reasonably suggest that a major determinant of functional Mac1p binding to DNA resides in the Ace1p/Amt1p homology element in the N-terminal domain of the protein.
Characterization of Mac1p⅐DNA Binary and (Mac1p) 2 ⅐DNA Ternary Complexes at the CTR1 Promoter-The repeated nature of the core GCTC element suggested to us the possible formation of a transcriptionally active protein-DNA complex, (Mac1p) 2 ⅐DNA. We tested this model by first establishing the presence of such a complex by EMSA and then by delineating the sequence features of the DNA that promoted its formation.
In fact, two mass species were apparent in an EMSA in which the wild type CTR1 sequence (WT, Table I)  The more slowly migrating of these two complexes (shaded bar, quantitation below) was present with an abundance of 50% that of the faster migrating species (solid bar, quantitation below). We infer that the more slowly migrating species is larger in mass and arises from separate protein molecules binding to the two core elements. This inference is derived from the disappearance of this slowly migrating species when either core element was mutated indicating that two elements were required for the formation of this complex (lanes 6 and 9, probes R 5Ј and R 3Ј , respectively). In probes R 5Ј and R 3Ј , the core sequences in either the 5Ј or 3Ј element, respectively, were randomized (Table I). However, the data indicated that these two sites were not equivalent since the fraction of the probe found in what we will refer to as a binary Mac1p complex (the smaller, faster migrating species) was essentially absent if the 5Ј core sequence was mutated (in R 5Ј , Fig. 2, lane 6) in comparison to mutation of the 3Ј core (as in R 3Ј , Fig. 2, lane 9). This result suggested that a binary complex at the 5Ј element was more stable than one at the 3Ј sequence.
The primary structural difference between these two core elements resides in the sequence immediately 5Ј to both. The upstream element is 5Ј-TATTTGCTC and the downstream one is 5Ј-GAGCAAAAA (5Ј-TTTTTGCTC on the opposite strand). The data above suggested the possibility that the presence of the TTTTT sequence in the latter 3Ј element dampened Mac1p binding to the core or, alternatively, the A in the former 5Ј element was preferentially recognized by Mac1p. Furthermore, the EMSA indicated that binding of Mac1p to the apparently weaker 3Ј element in the wild type promoter, as must occur in a ternary complex, was linked to Mac1p binding to the apparently stronger 5Ј element. This latter conclusion derives from the fact that a single TTTTTGCTC site cannot support formation of a detectable binary complex (Fig. 2, R 5Ј probe, lane 6), whereas in the context of an intact upstream TATTT-containing element, ternary complex formation does occur which must involve Mac1p binding to the 3Ј element (WT probe, lane 3).
One test of this conclusion was to convert the 3Ј-TTTTT-GCTC element in R 5Ј (to which Mac1p binds weakly if at all, Fig. 2, lane 6) to 5Ј-TATTTGCTC (again, reading on the opposite strand). The binding of Mac1p to this mutant, designated R 5Ј -3Ј-TA, would be expected to increase in comparison to binding to R 5Ј itself. This prediction was tested by direct binding with R 5Ј -3Ј-TA as probe (data not shown; see however, Fig. 6, lane 3) and also by the more quantitative competition assay. Thus, unlabeled mutant oligonucleotides were used as competitor DNA of the Mac1p binding to the WT CTR1 promoter element as probe (Fig. 3, control, lane 3). Based on these competition data as in Fig. 3 we estimate that Mac1p has a 3-4-fold greater affinity for the R 5Ј -3Ј-TA oligonucleotide (stronger competitor as shown in Fig. 3, lanes 6 -9) than for R 5Ј itself (Fig. 3, lanes 4 -6; also, see quantitation).
A second test of this conclusion was to use as probe mutated wild type double-stranded oligonucleotides in which both elements contained either TATTT or TTTTT, followed by GCTC. The TATTT-containing repeat would be expected to support both stronger binding overall and, importantly, a larger fraction of complex in what we propose is (Mac1p) 2 ⅐DNA. In contrast, the mutant that had T 5 at both sites would be expected to have weaker binding overall and little or no ternary complex formation. Both of these predictions were confirmed by experiment as demonstrated by the EMSA shown in Fig. 4A. Specifically, the symmetrical probe containing repeated TTTTT-GCTC elements (Fig. 4A, probe 5Ј-AT, lane 6) supported 75% ternary complex formation compared with WT probe (WT, lane 3), whereas the symmetrical probe containing TATTTGCTC (probe 5Ј-TA, lane 9) supported 150% of this complex compared with WT (shaded bars, quantitation). Furthermore, competition experiments in which wild type was used as probe and these two mutant oligonucleotides were used as competitors (see Fig. 4B) yielded a pattern in complete agreement with the direct binding data shown in Fig. 4A. Thus, the T 5 -containing direct repeat (5Ј-AT, lanes 3 and 4) was a poorer competitor than the TAT 3 -containing species (3Ј-TA, lanes 5 and 6; data in bar graph presented as a percent of values for the WT probe alone, see Fig. 4A, lane 3, for this control). In summary, the data in Figs. 2-4 indicate that Mac1p binds more strongly to the sequence TATTTGCTC than to TTTTTGCTC, and this better binding appears to stabilize what is interpreted to be a ternary (Mac1p) 2 ⅐DNA complex.  Table I for probes). Wild type Mac1p (lane 4), ⌬ZFMac1p (lane 6), and ZF*Mac1p (lane 8) were produced in vitro using the coupled wheat germ extract system (WGE, negative control, lanes 2 and 3) and added to the binding reaction as indicated in the figure. When present, unlabeled wild type oligonucleotide derived from the CTR1 promoter (WT) was added at a concentration 20-fold greater than the probe (lanes 3, 5, 7, and 9). The image is obtained from a Bio-Rad PhosphorImager as described under "Experimental Procedures." The FRE1 promoter provides a natural test of this model. The EMSA shown in Fig. 5 demonstrates that little complex formation at the FRE1 promoter element was observed (Fig. 5, FRE1-WT probe, lane 6; compare with binding to the CTR1-WT probe, lane 3). Competition by this FRE1 element of Mac1p binding to the wild type CTR1 44-mer showed that, indeed, Mac1p has a weaker affinity for the former sequence (data not shown). To show that this weaker binding was likely due to the absence of even one TATTT-containing site in the WT FRE1 probe, a mutant FRE1 element was used in which the 3Ј site contained the requisite T to A transition. Consistent with our model, while the wild type FRE1 element gave only a weak signal in the EMSA (Fig. 5, lane 6), the mutant probe supported the formation of both binary and ternary complexes (Fig. 5,  FRE1-3Ј-TA probe, lane 9). Although the stability of these complexes appeared less than with the wild type CTR1 element (which also has one each of the two types of T-rich sequences), the relative amount of the two complexes at the mutant FRE1 oligonucleotide was similar to what was observed with the CTR1 one (see quantitation). Importantly, the fact that a single TATTT at the FRE1 promoter converted this element from one that supported only a very weak interaction overall to one that actually could support ternary complex formation suggests a model in which formation of the ternary complex could be cooperative, i.e. that it might involve a thermodynamically important Mac1p-Mac1p protein-protein interaction linked to Mac1p binding at a TATTT-containing site. 3 Origin of the Stronger Binding of Mac1p to TATTTGCTC-One simple explanation for the difference in Mac1p binding to the TATTTGCTC sequence is that a specific protein-DNA interaction could be more favorable with A in comparison to T at this Ϫ4 position (relative to the G in the core sequence). To test this possibility, three probes were constructed based on the random 5Ј sequence (R 5Ј , Fig. 2) in which one of the thymines in the wild type 3Ј element was replaced by adenine. Thus, this series consisted of oligonucleotides containing 5Ј-TTTTTGCTC (R 5Ј ), 5Ј-ATTTTGCTC (R 5Ј -A41T), 5Ј-TATTTGCTC (R 5Ј -A40T; this is identical to R 5Ј -3Ј-TA used as competitor in Fig. 3), and 5Ј-TTATTGCTC (R 5Ј -A39T; sequences given for bottom strand, Scheme I). If the specific placement of the A were important to Mac1p binding, Mac1p binding to only one of these probes would be observed, presumably to R 5Ј -A40T as indicated by the competition data in Fig. 3. As the EMSA results in Fig. 6 show, Mac1p appeared to make a more favorable contact with A in comparison to T at this Ϫ4 position in as much as binding to the single, 3Ј-GCTC in R 5Ј was observed only with the probe containing 5Ј-TATTTGCTC on the bottom strand (Fig. 6, R 5Ј -A40T, lane 3). In terms of binding affinity for Mac1p, R 5Ј -A40T should be equivalent to R 3Ј , since both mutant oligonucleotides have only one TATTTGCTC site. Comparison of the data for R 5Ј -40T (Fig. 6, lane 3) and R 3Ј binding (Fig. 2, lane 9) shows this to be the case.
(Mac1p) 2 ⅐DNA Complex Formation in Vitro Correlates to Stronger Promoter Activity in Vivo-To test whether the ternary complex formation indicated by the EMSA data might be functionally important, the WT CTR1 promoter, in a fusion to the lacZ gene, was mutated to contain either two TTTTT or two TATTT elements. The transcriptional activity of the wild type and two mutant promoters was then quantitated by standard ␤-galactosidase assay. To demonstrate also the relative dependence of the activity of these promoter constructs on the copper concentration in the medium, cultures were grown in a copper-depleted medium (estimated 0.5 nM residual copper, "Experimental Procedures") or in the presence of added copper. The results of these measurements are shown in Fig. 7; they demonstrate a positive correlation between transcriptional activity in vivo and the fraction of Mac1p in (Mac1p) 2 ⅐DNA in vitro. Specifically, while the 5Ј-AT mutant promoter (Fig. 7, squares) supported only 65% the expression seen from WT CTR1 (Fig. 7, circles), the 3Ј-TA mutant supported 115% of this expression (Fig. 7, triangles). These differences can be compared with those described for the EMSA results (Fig. 4A) in which the symmetrical T 5 -containing 5Ј-AT and symmetrical TAT 3 -containing 3Ј-TA probes supported 75 and 150% (Mac1p) 2 ⅐DNA formation, respectively (in comparison to the WT CTR1 probe, Fig. 4A, lane 3).
The mutant promoters were equivalently down-regulated by copper indicating their essential dependence on Mac1p function. However, the dependence of this down-regulation on [copper] medium was shifted to higher values for the more active promoter element, that is in our model, the greater the fraction of ternary protein-DNA complex. From the data in Fig. 7, one can estimate that the [copper] medium that resulted in the halfmaximal change in trans-activity for the WT CTR1 promoter was ϳ20 nM (Fig. 7, circles); for the 5Ј-AT (triangles) and 3Ј-TA (squares) mutant CTR1 promoter sequences it was 10 and 30 nM CuCl 2 , respectively. This pattern would be consistent with a differential Mac1p-DNA interaction in the three promoter constructs. The fact that a single Mac1p-binding site is incapable of supporting transcription in vivo has been demonstrated previously (4, 7, 12). Thus, the quantitative data shown in Fig. 7   are consistent with a model in which a (more stable) ternary Mac1p-containing DNA complex at the CTR1 promoter is active in the subsequent up-regulation of the expression of this gene. Furthermore, they indicate that the structural difference(s) between TATTT and TTTTT that modulate the stability of this complex in vitro apply in vivo as well. It is important to note that the copper-dependent down-regulation of Mac1p trans-activity that occurs between 1 and 100 nM (as in Fig. 7) is entirely independent of the turnover of Mac1 protein that occurs when cells are exposed to [copper] medium Ͼ10 M or under conditions of incipient copper toxicity (25). A Mac1p-Mac1p Interaction Can Be Demonstrated by a Twohybrid Analysis-We inferred above that formation of what we suggest is a ternary, (Mac1p) 2 ⅐DNA complex might involve a thermodynamic contribution due to a protein-protein interaction between two Mac1p molecules. However, the data do not exclude an alternative model in which two Mac1p molecules bind to the DNA completely independently of one another. We have been unable to produce and isolate Mac1p that is active in the DNA binding experiments necessary to distinguish between these two possibilities. Therefore, we chose to demonstrate a possible Mac1p-Mac1p interaction independent of DNA binding by the in vivo two-hybrid genetic approach (23). 1 To do so, fusions of wild type Mac1p, the ⌬ZFMac1p truncation, and the mutated zinc finger Mac1p, ZF*Mac1p, were made to both the Gal4 DNA-binding (DBD) and transactivation domains. These fusions were tested for in vivo interaction by the standard two-hybrid assay in cells grown in both the absence and presence of copper. The results of these several assays are summarized in Table II. Also included in this table (in parentheses) are representative values for the Mac1p-Gal4 DBD fusions alone in a one-hybrid assay. These values describe the inherent transactivation activity of Mac1p and are needed for comparative purposes in as much as this Mac1p trans-activity is copper-dependent (5, 6). 1 The data in Table II (Table I) were incubated with either 5 l of coupled wheat germ extract alone (WGE, negative control) or with 5 l of wheat germ extract containing wild type Mac1p (WGE-Mac1p). The binding reactions were electrophoretically resolved on a 6.0% gel, and the gel was dried and exposed to a PhosphorImager screen. The resulting bands were quantitated using Molecular Analyst, and the digitized data were plotted as bar graphs with respect to the control, the counts associated with the faster migrating complex formed with the WT CTR1 probe. The data shown are representative of three experiments; although the absolute amount of complex varied, the relative amounts of the complexes formed (as given by percent of wild type monomer species) varied by less than 5%. independent of medium copper (compare values in first and third rows, all columns). This was in contrast to the results from the one-hybrid analysis that showed that the inherent trans-activity of Mac1p was consistently 7-10-fold greater in copper-deficient cells in comparison to copper-replete ones. This copper dependence is illustrated best by the transcriptional activity due to the ⌬ZFMac1p fusion (Table II, third column, compare second and fourth rows). 4 Second, a Mac1p-Mac1p interaction was increased ϳ10-fold upon removal of the N-terminal sequences suggested above to have DNA binding activity as in the ⌬ZF truncation (Table II, compare first to subsequent columns). This fold increase paralleled the increase in the trans-activity of this truncated Mac1p noted directly above (Table II, third column). 1 The effect of N-terminal deletion was geometric, that is this domain in each Mac1p fusion contributed equally in energy terms to the masking of the strong Mac1p-Mac1p interaction indicated for the ⌬ZFMac1p fusion pair (compare columns 1, 2, and 4, first and third rows). Significantly, increase in neither trans-activity nor Mac1p-Mac1p interaction was observed when a fusion of the C23S/ H25N double mutant was used (ZF*Mac1p, last column), a mutant which lacks DNA binding activity (Fig. 1) but which still has an N-terminal domain. Interaction between two mol-  (Table I) were added to the binding reactions as probe DNA and analyzed as above. B, WT CTR1 oligonucleotide was used as probe DNA. Competition experiments were carried out using 5-and 10-fold excess of WT (lanes 1 and 2) or 5Ј-AT (lanes 3 and 4) or 3Ј-TA (lanes 5 and 6) mutant oligonucleotides derived from the WT CTR1 probe. Analysis of the PhosphorImages was as described.
ecules of this mutant form was identical to that between the wild type protein, i.e. it was weak, giving a value only slightly above the blank. DISCUSSION The experiments described here establish six new facts and/or inferences about the structure-function relationships pertaining to the Mac1p-DNA interaction. These are as follows: 1) Mac1p DNA binding activity requires at the least the CCHC Ace1p/Amt1p zinc finger homology domain; 2) a complex of Mac1p and its DNA target forms in an EMSA that is most reasonably described as a ternary, (Mac1p) 2 ⅐DNA species in addition to a Mac1p⅐DNA binary one; 3) Mac1p binding to any core GCTC element is modulated by the immediate 5Ј sequence, specifically, binding is stronger to TATTTGCTC than to TTTTTGCTC; 4) Mac1p appears to make a specific and favorable contact with adenine at the Ϫ4 sequence position (tAttt) in comparison to a thymine (tTttt); 5) a mutated CTR1 promoter that conforms to a perfect palindrome including the TATTT sequence supports a larger fraction of Mac1p in the ternary complex in vitro and a stronger Mac1p-dependent expression from the CTR1 promoter in vivo; and 6) a Mac1p-Mac1p protein-protein interaction occurs that appears negatively modulated by the N-terminal domain of the protein that is essential to DNA binding. Based on these new facts and on  (Table I) were used as probes in this EMSA. The binding mixtures were analyzed and quantitated as described. , the values for the three promoters were different from one another at p Ͻ 0.001 or better. The [copper] in the medium with no added copper was estimated to be 0.5 nM since this level was below the detectable limit of flameless atomic absorption spectrophotometry. This estimate was based on the residual [copper] in the individual components used to make up the medium and the affinity of Chelex for Cu(II).
previously reported ones, we propose the following model of how the interactions of Mac1p with itself and with its DNA binding site are linked to the activation of gene expression at Mac1p-dependent genes.
The components of this model are as follows. (a) (Mac1p) 2 ⅐DNA is, at the least, a transcriptionally active species. Mac1p-dependent transactivation requires two Mac1p DNA-binding sites, as in the FRE1 (10), CTR1 (4), and CTR3 promoters (7). A single core GCTC element cannot support Mac1p-and copper-dependent expression from these promoters. Our data indicate that this requirement is most reasonably explained by a model in which the Mac1p⅐DNA complex that is competent for transcriptional activation is a ternary one, (Mac1p) 2 ⅐DNA. (b) Mac1p can be in (minimally) two conformational states. The "resting" state is inactive with respect to any of the interactions necessary for Mac1p-dependent transactivation, i.e. DNA binding, Mac1p-Mac1p interaction, and recruitment of components of the pre-initiation complex. The "active" state can participate in each of these intermolecular interactions. (c) The interconversion between these two states is linked to these three interactions, that is Mac1p has the characteristics of a thermodynamically cooperative system: the intermolecular interactions in the active state (which include all interacting components) compensate for the loss of the intramolecular ones (in Mac1p) which stabilize the resting state. That the DNA binding activity of Mac1p requires at the least the N-terminal domain including an intact CCHC motif is not surprising. Previous work with Ace1p (16) and Amt1p (19) suggested that in both of those systems the homologous Nterminal domains made important DNA contacts, as, for example, in the major groove in the case of Ace1p binding to the metal-responsive element in the CUP1 promoter (16). However, in neither case are the core cis elements (GCTG in the case of the Ace1p-binding site) and the flanking regions, which also make protein contacts, repeated as they are in CTR1 and CTR3 (and FRE1, albeit as a direct rather than an inverted repeat). Furthermore, the C-terminal core cysteine-rich regions of Ace1p and Amt1p, which when bound to Cu(I) form what has been described as a "copper fist," also make specific DNA contacts (16,19). Indeed, it is the copper fist domain in Ace1p that makes the major contacts with the Ace1p-core binding site, TTTCCGCTG (16). Structure-function relationships in Mac1p appear to be quite different. Current data are most consistent with a more strict demarcation of function between the N-and C-terminal domains in the protein; the C-terminal domain expresses all of the transactivation activity (and its copper dependence, much as in Ace1p and Amt1p), whereas the N terminus is required for DNA binding. Note also that both Ace1p and Amt1p are positively regulated by copper, the inverse of the behavior of Mac1p in response to copper level.
The presence of the inverted repeat in the CTR1 (and CTR3) promoters (3,7,9), TTTGCTC, and the fact that promoter deletion analyses in the CTR1 (4) and CTR3 promoters (7) demonstrated the need for both repeats suggested to us that Mac1p might bind at both sites irrespective of the precise nature of the ternary complex. Our data are the first that support this possibility. Our EMSA results show clearly that two Mac1p-dependent complexes form and, furthermore, that the formation of both, particularly the more slowly migrating one, appears driven by a stronger Mac1p binding to the 5Ј element, TATTTGCTC. Importantly, this stronger binding in vitro can be correlated with stronger promoter activity in vivo in support of our model that a (Mac1p) 2 ⅐DNA complex is at the least more active transcriptionally if not the active complex. However, our data do not prove this latter constraint, leaving open the question of whether both, or only the ternary complex, can support downstream transcription initiation.
The stronger Mac1p binding to the TATTT-containing motif appeared due to a more favorable interaction with A in comparison to T at this Ϫ4 position specifically. Of interest is that of the six Mac1p elements analyzed (two each in CTR1, CTR3, and FRE1), only two contain an adenine at this position, one in CTR1 and one in CTR3. In contrast, the equivalent Ϫ4 position in the Ace1p and Amt1p core binding sites at various loci has an invarient thymine, or T(T/C)XXGCTG (8,19). This conservation suggests that this specific nucleotide base-protein contact, which has been confirmed by experiment (16,19), makes a more substantive contribution to the overall stability of the protein⅐DNA complex in the case of Ace1p and Amt1p than it does in the case of Mac1p. One explanation for this difference would be that in the case of Mac1p-DNA ternary complex formation, a Mac1p-Mac1p interaction provides an additional and significant driving force.
One caveat of our work is that we have not provided direct evidence for the presence of two Mac1p molecules in the more slowly migrating DNA complex. Although we have attributed this difference in mobility to a larger size, i.e. to a (Mac1p) 2 ⅐DNA complex, another possibility is that the more slowly migrating species is a conformer of the Mac1p⅐DNA binary complex. However, the appearance of this species in the binding of Mac1p to the FRE1 promoter element upon mutation of only one of the core sequences to include an upstream TATTT is more consistent with a ternary complex model. With  this mutant oligonucleotide as probe, the distribution of Mac1p-DNA species was equivalent to what was observed with the wild type CTR1 promoter, that is introducing one strong site led to the appearance of both mobility species. This can most simply be explained by a cooperative binding model in which the "good" site recruits Mac1p and that the resulting DNA-protein interaction stabilizes a second Mac1p binding (to an inherently weaker DNA-binding site) through a Mac1p-Mac1p interaction. However, while a cooperative dimer model is consistent with the EMSA results shown, it clearly remains to be tested more rigorously than it has been here.
The two-hybrid results indicated that Mac1p can, in effect, self-associate. One aspect of these data was particularly significant, namely the potential protein-protein interaction was strongly suppressed by an intact N-terminal domain. Thus, the ⌬ZF construct that lacked Mac1p residues 1-40 supported a 10 -40-fold stronger interaction in comparison to the fusions with Mac1p (1-417) as both bait and catch. We interpret this result to show that the N-terminal domain (whether wild type or mutated, e.g. as in ZF*Mac1p) sterically blocks self-association of Mac1p by an intramolecular contact with the region in Mac1p involved in the intermolecular protein-protein interaction. In our model, we postulate that this intramolecular interaction is replaced (energetically compensated for) by the intermolecular one the N-terminal domain has with the DNA. In the Mac1p⅐DNA complex, therefore, the DNA-bound Mac1p (conformationally equivalent to ⌬ZFMac1p) is active for binding to another molecule of Mac1p. The apparent strength of this latter interaction will be dependent on whether the second Mac1p is also active for binding (is bound to DNA, resulting in a ternary complex) or, in the two-hybrid case, whether the second Mac1p is intact or N-terminal deleted.
We interpret the one-hybrid data to indicate that the Nterminal domain also blocks the recruitment by Mac1p of components of the general transcription machinery (as it does the apparent Mac1p-Mac1p interaction) when Mac1p itself is not bound to DNA via its own N-terminal domain. Thus, the 10-fold increase in the (copper-dependent) expressed trans-activity in the N-terminal deleted construct in comparison to Mac1p-(1-417) was nicely consistent with the suggestion that the DNAbinding N-terminal domain masks elements in the C-terminal region involved in both types of protein-protein interaction.
The role of copper in the regulation of the inherent transcriptional activity of Mac1p was not clarified by the results presented here. Clearly, Mac1p, either by itself or as the TA domain in a DBD fusion, e.g. to Gal4 DBD (above and Ref. 5) 1 or to the LexA protein (6), is down-regulated by a copper concentration in the medium that is known to down-regulate the expression of CTR1 (cf. Fig. 7). The [copper] medium that causes a 50% suppression of the transcriptional activity of Mac1p is ϳ20 nM. Furthermore, in vivo DNA footprinting at the CTR3 promoter (Ref. 18, and at the CTR1 promoter as well) 5 has demonstrated a Mac1p-dependent protection at the Mac1pbinding sites that is negatively modulated by copper. On the other hand, neither the Mac1p-Mac1p protein-protein interaction indicated by our two-hybrid data nor the in vitro DNA binding experiments employing a full-length Mac1p (EMSA) exhibited a copper dependence (described under "Experimental Procedures"). Whereas the latter results could be ascribed to the use of, for example, in vitro generated protein, the former result cannot be so easily disregarded. The fact that the same fusion exhibited a copper-dependent trans-activity in a onehybrid assay but a copper-independent protein-protein interaction in a two-hybrid one most reasonably suggests that one specific protein-protein interaction, self-association, is, in fact, not modulated by the alteration in Mac1p that is caused by copper.
Clearly, the model that we propose here requires additional rigorous experimental evaluation. For example, one could predict differential protection of the two GCTC elements in the wild type CTR1 promoter in a DNA footprinting experiment, whether in vivo or in vitro. The one published in vivo footprint, at the CTR3 promoter, was relatively symmetric (7). However, additional footprinting experiments using altered Mac1p-binding sites as used here in the EMSA along with Mac1 protein that either has or does not have the potential to self-associate could provide new and important insight into the protein-protein and protein-DNA equilibria involved in the Mac1p-DNA interaction whether or not they specifically confirm the present model. These experiments are in progress.