Mapping of the DNA Binding Domain of the Copper-responsive Transcription Factor Mac1 from Saccharomyces cerevisiae *

Mac1 from Saccharomyces cerevisiaeactivates transcription of genes, including CTR1 in copper-deficient cells. N-terminal fusions of Mac1 with the herpes simplex VP16 activation domain were used to show that residues 1–159 in Mac1 constitute the minimal DNA binding domain. Mac1-(1–159) purified from Escherichia coli contains two bound Zn(II) ions. Electrophoretic mobility shift assays showed direct and specific binding by Mac1-(1–159) to a DNA duplex containing the copper-responsive element TTTGCTCA. The DNA binding affinity of Mac1-(1–159) for a duplex containing a single promoter element or an inverted repeat was 5 nm for the 1:1 complex. The N-terminal 40-residue segment of Mac1 is homologous to the DNA binding zinc module found in the copper-activated transcription factors Ace1 and Amt1. A MAC1 mutation yielding a Cys11→ Tyr substitution at the first candidate zinc ligand position relative to Ace1 resulted in a loss of in vivofunction. Two TTTGCTCA promoter elements are necessary for efficient Mac1-mediated transcriptional activation. The elements appear to function synergistically. Increasing the number of elements yields more than additive enhancements in CTR1 expression.

Copper is an essential cofactor for a variety of proteins and is required for normal cellular functions (1,2). In enzymes, copper ions participate in catalytic functions or fulfill structural roles (3). Abnormally low levels of copper can impair the function of many copper-dependent enzymes. However, abnormally high levels of copper can be toxic. The toxicity of copper may arise, in part, from cellular damage caused by reactive oxygen intermediates. Copper and iron ions can catalyze the formation of highly reactive hydroxyl radicals through the Fenton reaction (4). Hydroxyl radicals can react with biomolecules, resulting in polypeptide bond cleavage as well as DNA base and sugar oxidation (5)(6)(7). Homeostatic mechanisms exist in all cells to regulate the cellular concentration of copper ions, maintaining copper balance and minimizing the deleterious effects of excess copper ions.
The budding yeast Saccharomyces cerevisiae has been used as a model system to study the mechanisms of copper detoxification, transport, and distribution. The mechanism of high affinity copper uptake in S. cerevisiae has been determined and includes cell surface permeases as well as a metal ion reductase (8 -11). In an oxygen-containing environment copper ions are predominantly in the form of Cu(II). Cu(II) ions are reduced to Cu(I) prior to transport into S. cerevisiae cells (10,11). Cu(II) reduction is mediated at the cell surface by the NADPH-dependent ferric and cupric reductases Fre1 and Fre2 (10 -12). Cu(I) is then transported across the plasma membrane by high affinity copper transporters Ctr1 and Ctr3 (9,13).
Under copper starvation FRE1, CTR1, CTR3, and FRE7 (a homologue of FRE1) are highly expressed. However, in the presence of elevated concentrations of copper these genes are down-regulated (10, 14 -16). Copper-dependent inhibition of expression occurs at the level of transcription and is mediated by Mac1 (metal binding activator) (15)(16)(17). The role of Mac1 in the expression of FRE1 and CTR1 has been studied in cells with mutations in the MAC1 locus (10,(15)(16)(17). Expression of FRE1 and CTR1 is not copper-inhibited in cells containing a semidominant gain of function MAC1 mutation, resulting in a His to Gln substitution at residue 279 designated MAC1 up1 (17). MAC1 up1 cells exhibit a copper-hypersensitive phenotype, since copper uptake is unregulated (17). In contrast, a frameshift mutation in MAC1, designated mac1-1, resulted in substantial loss of both Cu(II) and Fe(III) reduction as well as uptake (17). The observed phenotypes of mac1-1 cells include respiratory deficiency and sensitivity to a myriad of stresses. The phenotypes are reversed upon addition of exogenous copper salts.
The mechanism of copper attenuation of Mac1 function appears to involve copper repression of both the DNA binding function and transactivation (16,18). Fusions of the Gal4 DNA binding domain with the Mac1 transactivation domain revealed copper attenuation of transcriptional activation of a GAL1/lacZ reporter gene (18). A repeated Mac1-responsive element consisting of at least two repeats of the sequence TTTGC(T/G)C(A/G) has been identified in 5Ј sequences of CTR1, CTR3, FRE1, and FRE7 (15,16,19). The presence of two repeats of this sequence appears to be critical for high level expression of Mac1-regulated genes (16,19). In vivo DNA footprinting has shown a loss of protection of the TTTGCTCG repeats in the promoter of CTR3 upon addition of copper salts (16). The loss of protection in the in vivo DNA footprint suggests the loss of DNA binding in response to copper treatment. Whether the loss of protection of this sequence is due to the loss of DNA binding activity or nuclear export is unclear.
In this study the DNA binding domain of Mac1 was mapped to within the N-terminal 159 residues. Mac1-(1-159) was purified from E. coli as either a His 6 tag or a glutathione Stransferase fusion and contained two bound Zn(II) ions. The purified zinc-Mac1-(1-159) complex was capable of specific and high affinity binding to a single TTTGCTCA element. Sequential binding of the Mac1-(1-159) protein to a DNA probe containing two TTTGCTCA elements was observed, suggesting that Mac1-(1-159) is a monomer. The requirement for at least two TTTGC(T/G)C(A/G) elements in the promoters of Mac1-regulated genes appears to be due to synergism between the Mac1 binding sites.
Modified CTR1/lacZ Reporter Plasmids-CTR1 promoter/lacZ hybrid genes containing either one, two, three, or four TTTGCTCA elements were constructed in a derivative of the ⌬UAS vector pNB404 with a LEU2 selectable marker (pNB404L) (22). CTR1 promoter sequences were amplified by PCR and subcloned into pNB404L to generate vectors pC1 sequences Ϫ323 to Ϫ299 (relative to the translation start site) and pC2 and Ϫ339 to Ϫ299. A promoter containing three TTTGCTCA elements (pC3) was generated by PCR mutagenesis using a mutagenic oligonucleotide, which inserted a copy of sequences Ϫ339 to Ϫ323 5Ј to the repeated TTTGCTCA element in pC2. The spacing between the second and third TTTGCTCA elements was 13 nucleotides, which is similar to that seen in the CTR1 promoter of 14 nucleotides. pC4 containing four TTTGCTCA elements was constructed in a similar manner to pC3. The spacing between the third and fourth TTTGCTCA elements was 12 nucleotides.
Mac1/Myc Epitope Fusions-Six copies of the Myc epitope from vector CS2ϩMT (23) were excised as a BamHI/StuI fragment and subcloned into pHolly (generously supplied by Richard Palmiter) a pBluescript derivative with a small insert creating sites for NcoI, NdeI, and StuI, to pick up a stop codon. The Myc epitope was excised as a BamHI/SalI fragment and cloned into the yeast expression vector pYeF2 generating pYeF2-Myc. The Mac1 Myc fusions were transformed into yeast strain YJJ1 (generously provided by Stefan Jentsch). YJJ1 (mac1-1) has a frameshift in the MAC1 gene, resulting in a loss of function of MAC1. Yeast cultures were grown in synthetic complete medium lacking uracil or in low copper synthetic complete medium lacking uracil with either 30 M BCS or 100 M CuSO 4 added. All cultures had 2% galactose added as a carbon source. The cells were harvested by centrifugation and resuspended in lysis buffer (10 mM Tris-HCl, pH 8, 10 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride) and lysed with glass beads. The lysed cells were centrifuged, and the extract was filtered through a 0.45-m membrane. Protein concentration was determined by the method of Bradford (24).  was excised form Mac1-(1-194)/VP16 as a AflIII/NotI fragment and subcloned into a derivative of pET21d, which contained sequences coding for a single copy of the HA epitope (pET21HA) generating a C-terminal fusion of the HA epitope with Mac1- . Mutations in the two Cys pairs in the N-terminal 194-residue, C69S,C70S and C172S,C173S, and all four cysteines C69S,C70S,C172S,C173S were created in the Mac1-(1-194)pET21HA. These mutations were initially generated using the altered sites mutagenesis system (Promega) (18). Residues 42-194 were amplified by PCR with the mutant Gal4DBD Mac1-(42-417) in pAlter as templates (18). The PCR products were digested with NcoI/NotI and cloned into Mac1- pET21HA, replacing the wild type sequence from amino acid 42 to 194. DNA coding for the first 159 residues of Mac1 was amplified by PCR and was digested with NdeI/HinDIII and subcloned into pAED4-His or the Cterminal glutathione S-transferase fusion vector pAED4-GST; the first 24 codons in the open reading frame were highly biased E. coli codons to optimize expression (25). The correct sequence in all the constructs was verified by restriction digestions and nucleotide sequencing.

E. coli Expression Vectors and Purification of Mac1-(1-159)-Mac1-
Mac1-(1-159) as either a C-terminal His 6 tag or GST fusion was expressed in E. coli strain BL21 (DE3). Cells were grown at 37°C to an A 600 nm of 0.6 and isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 0.4 mM to induce production of the protein. Thirty min after induction, 1 mM ZnSO 4 was added, and the cells were grown for an additional 3 h. The cells were collected by centrifugation and washed once in 0.25 M sucrose and finally resuspended in sonication buffer (20 mM NaH 2 PO 4 , pH 7.5, 300 mM NaCl, 10% glycerol, and 1 mM dithiothreitol). The cells were lysed by freeze-thawing and repeated sonication. The lysate was centrifuged at 4°C for 30 min at 100,000 ϫ g, and the supernatant was collected and filtered through a 0.45-m filter. The Mac1-(1-159)-His tag protein was purified using nickelnitrilotriacetic acid agarose (Qiagen). The extract was loaded onto the nickel-nitrilotriacetic acid agarose column and washed with 10 column volumes of sonication buffer. The Mac1-(1-159)-His tag protein was eluted with a 0 -500 mM imidazole linear gradient in sonication buffer. The Mac1-(1-159)-GST fusion protein was purified using glutathione-Sepharose (Amersham Pharmacia Biotech). Triton X-100 was added to 1%, and the extract was loaded onto the glutathione-Sepharose column and washed with 10 column volumes of sonication buffer. The Mac1-(1-159)-GST fusion protein was eluted with 20 mM glutathione in 50 mM Tris-HCl, pH 8. The eluted proteins were monitored by SDS-polyacrylamide gel electrophoresis, and a single band at the correct molecular weight for each fusion was observed. ␤ , and pVT102-U were cotransformed with a CTR1/lacZ reporter construct containing an inverted repeat of the TTTGCTCA element into yeast strain YJJ1. Plasmids pC1, pC2, pC3, and pC4 were cotransformed into YJJ1 with Mac1 pVT102-U. Yeast transformants were grown in low copper synthetic complete medium with 2% dextrose lacking uracil and leucine. The cells were grown to an A 600 of 1.0 with the addition of either 30 M BCS or 100 M CuSO 4 added to the growth media. The cells were pelleted and washed with Z-buffer (85 mM Na 2 HPO 4 , 45 mM NaH 2 PO 4 , 10 mM KCl, 85 mM ␤-mercaptoethanol, and 1 mM MgSO 4 ) and frozen at Ϫ70°C. Cells were resuspended in the same buffer and lysed by vortexing with glass beads. ␤-Galactosidase activities were assayed using o-nitrophenyl ␤-D-galactopyranoside as a substrate. The absorbance at 420 nm was recorded, and protein concentrations were determined by the method of Bradford (24). Each sample was assayed in triplicate.
Western Blot Analysis-Mac1- )/Myc, Mac1-(1-159)/Myc, Mac1-(1-125)/Myc, and Mac1-(1-101)/Myc in pYeF2 were grown in low copper synthetic complete medium lacking uracil to an A 600 nm of 1.0 in medium containing 2% raffinose in place of dextrose. Protein expression was induced by the addition galactose to 2%. Cells were grown for an additional 6 h. The cells were harvested by centrifugation, washed with water, resuspended in lysis buffer, and lysed by vortexing with glass beads. The extract was clarified and filtered. SDS-polyacrylamide gel electrophoresis was carried out on a 12% gel. Proteins were transferred to nitrocellulose and the blots incubated with the 9E10 anti-c-Myc (Boehringer Mannheim) in blocking buffer (10% non-fat dry milk, phosphate-buffered saline, and 0.1% Tween 20). Detection by enhanced chemiluminescene was performed after incubation with horseradish peroxidase-conjugated secondary antibody.
Electrophoretic  )/HA were also tested for binding to this duplex. Mac1-(1-159) purified from E. coli was tested for binding to oligonucleotides containing the two TTTGCTCA copper-responsive sites in CTR1 or mutant duplexes. The oligonucleotides were end-labeled with [␥-32 P]ATP, boiled for 3 min, and slowly cooled to anneal. Protein extracts or purified Mac1-(1-159) was added to a 26nucleotide duplex containing a half-site with one copy of the repeated element (cat ggg ata TTT GCT Caa gac gac gg). Mac1-(1-159) was also tested with two mutant duplexes containing multiple changes in the Mac1 binding site. Duplex CTR-G had nucleotides at positions 10 -13 changed to G (cat ggg ata GGG GCT Caa gac gac gg), and duplex CTR-A had nucleotides at positions 15-17 changed to A (cat ggg ata TTT GAA Aaa gac gac gg). Protein titrations were carried out with two 26-nucleotide duplexes containing each of the two CTR1 elements and a 48-base pair probe containing the inverted repeat of the TTTGCTCA sequence (cat ggg ata TTT GCT Caa gac gac ggt aaa atG AGC AAA aat ggc acg). The protein and DNA were incubated for 15 min at 25°C prior to electrophoresis of the protein-DNA complexes on 6% polyacrylamide gels and were visualized by autoradiography.

MAC1-VP16 Activation Domain
Fusions-Many fungal transcription factors have distinct and separable DNA binding domains and transactivation domains (26,27). Thus, one strategy to map the DNA binding domain of Mac1 is to evaluate the in vivo function of Mac1 truncates fused with a heterologous activation domain. The candidate DNA binding domain of Mac1 was expected to be located near the N-terminal end of the protein as the N-terminal 40 residues are homologous to the conserved DNA binding zinc module in Ace1 and Amt1 (17,28,29). The minimal DNA binding domain of Ace1 consists of two independent modules, the N-terminal 40-residue zinc motif and an adjacent domain stabilized by a polycopper cluster (28,30). Only the N-terminal zinc module is conserved in Mac1. If Mac1 resembles Ace1 in having two independent domains participating in DNA binding, one would expect the minimal DNA binding domain of Mac1 to consist of a region greater than just the conserved N-terminal zinc module.
The minimal AD from the herpes simplex virus-1 VP16 has been used extensively as a heterologous activator (31)(32)(33)(34). Fusion of the N-terminal 194 residues of Mac1 to residues 417-491 of VP16 generated a functional transcription factor. Transformation of the MAC1-(codons 1-194)/VP16 fusion reversed the methionine auxotrophy of mac1-1 cells (14, 17) (Fig. 1). Cotransformation with both the MAC1- /VP16 hybrid and a CTR1/lacZ reporter fusion resulted in lacZ expression (Fig. 2). The Mac1/VP16 fusion gave greater CTR1/lacZ expression than did an episomal full-length Mac1 containing its own activator. The VP16 AD is a strong activator and may be more potent than the Mac1 AD. The Mac1-(1-194)/VP16 AD fusion was not modulated by the addition of CuSO 4 as is the full-length Mac1. The lack of copper modulation was expected as the candidate copper binding motif resides in the C-terminal activation domain (18). To further narrow the DNA binding domain of Mac1, three other C-terminal truncations were constructed, terminating at residues 159, 125, and 101. The activity of the Mac1-(1-159)/VP16 AD fusion was identical to that of Mac1-(1-194)/VP16. However, truncation to residue 125 markedly reduced the activity of the CTR1/lacZ reporter gene, and truncation to residue 101 essentially eliminated activity (Fig.  2).
Yeast extracts from mac1-1 cells expressing Mac1-(1-194)/ Myc were tested for binding to a DNA duplex containing a single TTTGCTCA element (Fig. 3A). Addition of Mac1-(1-194)/Myc to the probe resulted in the appearance of a gelretarded complex. The addition of a monoclonal anti c-Myc antibody resulted in a supershift, demonstrating that the gelretarded complex contained Mac1-(1-194)/Myc (Fig. 3A). The  (Fig. 4B). These results suggested that the minimal DNA binding domain was within the N-terminal 159 residues.
Activity of MAC1 Mutants-The N-terminal 194 residues of Mac1 contain the 40-residue candidate zinc module conserved in Ace1 and two downstream Cys-Cys sequences. To determine whether the conserved zinc module was important for function, a mutation in codon 11 was engineered, resulting in a Cys 11 3 Tyr substitution. A corresponding Cys 11 3 Tyr mutation in Ace1 abolished in vivo function and attenuated DNA binding activity. The mutation was engineered in the full-length MAC1 open reading frame, and the mutant gene was transfected into mac1-1 cells to test for function. Cells harboring a mutant Mac1 with the Cys 11 3 Tyr substitution were unable to activate expression of CTR1/lacZ (Fig. 5).
Two Cys-Cys sequence motifs exist in Mac1. To determine whether the Cys-Cys motifs were important in the function of Mac1, double Cys-to-Ser substitutions were engineered in fulllength Mac1 at residues 69 and 70 and 172 and 173. These mutant Mac1 molecules were active in vivo and yielded wild- type copper regulation of CTR1/lacZ expression (Fig. 5). The presence of the same double substitutions in Mac1-(1-194)/HA from E. coli extracts did not impair the ability of Mac1-  to bind to the half-site duplex containing a single TTTGCTCA element (data not shown). These experiments suggest that the two Cys-Cys pairs are not important for the formation of the Mac1⅐DNA complex.
Purification of Mac1-(1-159)-Mac1-(1-159) was expressed in E. coli as either a C-terminal His 6 tag or as a GST fusion. The Mac1-(1-159)/His tag protein was affinity-purified using nickel-nitrilotriacetic acid-agarose, and the Mac1-(1-159)/GST fusion protein was affinity-purified using glutathione-Sepharose. The purified fusion proteins showed single stained bands on SDS-polyacrylamide gel electrophoresis consistent with the calculated molecular mass of 20 kDa for Mac1-(1-159)/His tag and 45 kDa for Mac1-(1-159)/GST (data not shown). Each fusion had the correct amino acid composition by amino acid analysis. Metal analysis by atomic absorption spectroscopy revealed that each Mac1-(1-159) fusion protein bound two zinc ions. A similar zinc binding stoichiometry for the two fusion proteins suggests that zinc binding is not occurring within the poly(His) tag.
Purified Mac1-(1-159) was tested for its ability to bind to DNA duplexes containing either the copper-responsive element (TTTGCTCA) or two mutant duplexes with changes in the element as well as a longer duplex containing the inverted repeat of the TTTGCTCA element separated by 14 nucleotides derived from the CTR1 promoter. The addition of Mac1-  to the half-site duplex containing the wild-type sequence resulted in the appearance of a gel-retarded complex (Fig. 6). No complex was observed in gel shift assays with mutant DNA duplexes containing sequences GGGGCTCA or TTTGAAAA (Fig. 6). The DNA binding affinities of Mac1-(1-159) for each TTTGCTCA site within the CTR1 promoter were near 5 nM (Fig. 7, A and B). A 5 nM concentration of Mac1-(1-159) resulted in 50% of the DNA probe being present as a protein-DNA complex. Sequential binding of Mac1 was observed to a DNA duplex containing the inverted repeat. The affinity for occupancy of the first site was similar to the isolated half-site at 4 nM. The binding affinity of the second Mac1-(1-159) molecule was approximately 11 nM (Fig. 7C).
Synergism between Mac1 Binding Sites-Repeated TTT-GCTC elements do not appear to be necessary for high affinity DNA binding based on the electrophoretic mobility shift assay data. Since the Mac1-(1-194)/VP16 fusion was able to stimulate transcription to a higher level than wild-type Mac1, we tested whether the fusion was active on a promoter with a single TTTGCTCA element. The fusion stimulated transcrip- tion of lacZ from a promoter containing a single element 100fold compared with a promoter lacking a TTTGCTCA element, but lacZ expression was an additional 4-fold greater with two TTTGCTCA promoter elements compared with one element (data not shown). The 4-fold difference in activity between one and two elements with Mac1-(1-194)/VP16 is suggestive of synergism between the sites. The four known Mac1-regulated genes in yeast require two TTTGC(T/G)C(A/G) elements for Mac1 responsiveness.
To determine whether the requirement for two elements was due to synergism between the Mac1 binding sites, promoter/ lacZ hybrid genes containing either one, two, three, or four TTTGCTCA elements were engineered and tested for activity in cells containing an intact Mac1 (Fig. 8). Compared with the control vector containing no upstream activation sequences, low level activation is seen with a single TTTGCTCA element. However, a promoter lacZ fusion containing two elements yielded approximately 8-fold greater lacZ expression than that for a single element. The presence of three elements increased expression nearly 5-fold more than the expression level of the promoter/lacZ fusion with two elements. The promoter fusion containing four elements was about 2-fold higher than the promoter fusion containing three elements. The increase in expression in each case is more than additive, suggesting that the requirement for two TTTGC(T/G)C(A/G) elements in the promoter of Mac1-regulated genes is due to synergism between the sites.

DISCUSSION
The N-terminal 40 residues of Mac1 are highly homologous with a corresponding segment in the copper-activated Ace1 transcription factor from S. cerevisiae (17). The conserved region is one of two domains forming the DNA binding segment of Ace1 (28,30). The homologous 40-residue domain in Ace1 and Amt1 (the Ace1 ortholog in Candida glabrata (36)) was shown to comprise a nonclassical zinc module (28,37). The high level of homology between Mac1 and Ace1 in this region suggested that the DNA binding domain of Mac1 would contain at least the N-terminal 40 residues. Mac1/VP16 fusions were used to map the DNA binding domain of Mac1. The activities of the Mac1/VP16 and Mac1/Myc fusions were not modulated by the addition of copper salts. Whereas the DNA binding activity of Mac1-(1-159) is insensitive to copper, DNA binding of the full-length Mac1 protein appears to be copper-dependent (16). The reported loss of an in vivo footprint in copper-treated cells may arise from a variety of mechanisms, including copper-dependent nuclear export, copper-dependent degradation of Mac1, or alternatively masking of DNA binding activity through an intramolecular interaction between the N-terminal DNA binding domain and the C-terminal copper binding activation domain. As mentioned, Mac1 contains a copper-regulated transactivation domain in the C-terminal region of the protein (11,18).
The N-terminal 159 residues of Mac1 bind two Zn(II) ions. One Zn(II) ion is presumably bound within the conserved zinc module found in Ace1 and Amt1 (28). The N-terminal zinc motifs in Ace1 and Amt1 appear critical for A/T-rich minor groove DNA contacts (29, 38 -39). The solution structure of the Amt1 zinc module reveals a compact conformation with 3 cysteines and 1 histidine as zinc ligands (37). These residues, Cys 11 , Cys 14 , Cys 23 , and His 25 are conserved in Mac1. This is consistent with the presence of the nonclassical, Ace1-like zinc motif in Mac1-(1-159). By homology, Mac1 is expected to contain the same zinc binding motif and be responsible for the DNA contacts within the TTT segment of the TTTGCTCA site. Analogous to Ace1, a Cys 11 3 Tyr substitution in Mac1 abolishes in vivo function. The DNA binding segment of Mac1, therefore, appears to consist of two domains, an N-terminal 40-residue Ace1-like module and a contiguous domain (residues 41-159). The ligands for the second bound Zn(II) ion in Mac1 are unresolved. If the Cys-Cys pairs at positions 69 and 70 are involved, the Zn(II) would not be critical for DNA binding, as Mac1-  containing substitutions of C69S,C70S or C172S,C173S were competent to bind DNA containing the consensus TTTGCTCA site. Furthermore, these Cys-to-Ser substitutions in the full-length Mac1 did not alter Mac1 function. The Mac1 segment from residues 41-159 is expected to contact the GCTCA subsite of the Mac1-responsive element.
Purified Mac1-(1-159) binds to a DNA probe containing a single repeat of the TTTGCTCA sequence with an affinity of 5 nM, but not to mutant duplexes. This confirms that Mac1-(1-159) was binding specifically to the TTTGCTCA element. Although Mac1 binds to a single TTTGCTCA element in vitro, two repeats of this element are necessary for in vivo activity (16,19). Addition of Mac1-(1-159) to a DNA probe containing the inverted repeat of the TTTGCTCA element resulted in two gel-retarded complexes consistent with binding of two protein molecules. The DNA binding affinity of the 1:1 complex with DNA containing the inverted repeat was the same as with either DNA half-site, suggesting that the two sites are independent with no cooperativity in binding. If binding cooperativity occurred between the two sites, the DNA binding affinity of Mac1-(1-159) to the duplex containing the palindromic element would be expected to be greater for the 2:1 complex than the additive affinity for duplexes containing a single element.
Repeated TTTGCTCA elements are not necessary for high affinity DNA binding based on the electrophoretic mobility shift assay data. However, two elements are necessary for high level in vivo transcriptional activation of genes regulated by Mac1. The requirement for two Mac1 binding sites appears to be due to synergism between the sites. This is seen with both the Mac1/VP16 fusion and intact Mac1. Increasing the number of TTTGCTCA elements in promoter/lacZ fusions from one to four showed a more than additive effect on lacZ expression levels consistent with synergism. Synergistic activation through multiple sites is well established with a number of yeast factors, including Gal4 (35,40,41). Increasing the number of Gal4 sites gave much more than additive effects in promoting transcription (35). Furthermore, the observed synergism in Gal4 is dependent on the strength of the activator (35). A Gal4 fusion with a strong activator such as the VP16 AD resulted in activation from a single site, whereas a fusion with a weak activator did not. We observe a similar effect using the VP16 activator. Using the weak activator Gal4 fusion, Carey et al. (35) observed a 66-fold stimulation in transcription by increasing the number of Gal4 sites from 1 to 5. This is compared with a 75-fold enhancement in transcription by increasing the number of Mac1 sites from 1 to 4.
Synergism in Gal4 and other transcriptional activators has been suggested to be a consequence of interaction of each of the bound activators with some component or components of the transcriptional machinery rather than with each other (35,40). Synergism between the Mac1 molecules bound to multiple sites may be important for high level expression of Mac1-regulated genes. FIG. 8. Synergism between Mac1 DNA binding sites. CTR1 promoter/lacZ hybrid genes containing 0 -4 TTTGCTCA elements in the ⌬UAS vector pNB404 were tested in YJJ1 cells harboring the Mac1 pVT102-U vector. The spacing between the second and third TTT-GCTCA elements was 13 nucleotides, which is similar to that seen in the CTR1 promoter of 14 nucleotides. The construct with four TTT-GCTCA elements had a spacing between the third and fourth TTT-GCTCA elements of 12 nucleotides. Quantitation of ␤-galactosidase specific activity is shown along with S.D. values of triplicate assays.