Metalloregulation of FRE1 and FRE2Homologs in Saccharomyces cerevisiae *

The high affinity uptake systems for iron and copper ions in Saccharomyces cerevisiae involve metal-specific permeases and two known cell surface Cu(II) and Fe(III) metalloreductases, Fre1 and Fre2. Five novel genes found in theS. cerevisiae genome exhibit marked sequence similarity to Fre1 and Fre2, suggesting that the homologs are part of a family of proteins related to Fre1 and Fre2. The homologs are expressed genes inS. cerevisiae, and their expression is metalloregulated as is true with FRE1 and FRE2. Four of the homologs (FRE3-FRE6) are specifically iron-regulated through the Aft1 transcription factor. These genes are expressed either in cells limited for iron ion uptake by treatment with a chelator or in cells lacking the high affinity iron uptake system. Expression of FRE3-FRE6 is elevated inAFT1–1 cells and attenuated in aft1null cells, showing that iron modulation occurs through the Aft1 transcriptional activator. The fifth homolog FRE7 is specifically copper-metalloregulated. FRE7 is expressed in cells limited in copper ion uptake by a Cu(I)-specific chelator or in cells lacking the high affinity Cu(I) permeases. The constitutive expression of FRE7 inMAC1 cells and the lack of expression in mac1–1 cells are consistent with Mac1 being the critical transcriptional activator of FRE7 expression. The 5′ promoter sequence of FRE7 contains three copper-responsive promoter elements. Two elements are critical for Mac1-dependent FRE7 expression. Combinations of either the distal and central elements or the central and proximal elements result in copper-regulated FRE7 expression. Spacing between Mac1-responsive sites is important as shown by the attenuated expression of FRE7 and CTR1 when two elements are separated by over 100 base pairs. From the three Mac1-responsive elements in FRE7, a new consensus sequence for Mac1 binding can be established as TTTGC(T/G)C(A/G).

The significance of cell surface metalloreductases in iron and copper homeostasis in yeast is that iron and copper ions in the environment are largely present in insoluble, oxidized valent states. Bioavailability of these ions increases through reduction to the ferrous and cuprous states. Fre1 and Fre2 appear to elaborate a diffusible reductant to accomplish this solubilization and mobilization (5). An alternative mechanism of scavenging iron in certain other species is to mobilize iron by secreting Fe(III)-chelating siderophores into the environment (9). Besides metalloreductases, metal ion-specific permeases are important for high affinity iron and copper ion uptake. The iron transport system involves the Ftr1 permease and the Fet3 ferro-oxidase (4,10). Fet3 is a trinuclear copper oxidase whose ferro-oxidase activity is critical for iron transport (10 -12). The mechanistic role of Fet3 ferro-oxidation in iron transport is unresolved. Two distinct copper permeases, Ctr1 and Ctr3, are important for cuprous ion uptake (2,13). No accessory proteins, analogous to Fet3, are known for copper transport.
Components of the high affinity iron and copper uptake systems are regulated in their biosynthesis in a metal-dependent manner. Genes encoding the proteins are expressed in cells exposed to iron-or copper-deficient conditions (14 -16). Irondeficient conditions result in induced expression of FTR1, FET3, FRE1, and FRE2. The regulation of expression of these genes occurs at the transcriptional level and is mediated by Aft1 (14,15). Copper-deficient conditions lead to elevated expression of CTR1, CTR3, and FRE1 (8, 13, 16 -18). The transcriptional regulation of copper uptake gene products is mediated by Mac1 (6,8,(17)(18)(19). FRE1 is unique in its regulation by both iron and copper ions through Aft1 and Mac1, respectively (6,8,15). These genes are down-regulated in iron-or copperreplete conditions. Fre1 is also homologous to the gp91 phox subunit of the NADPH oxidase that is critical for production of microbicidal oxidants in human neutrophils (20,21). The gp91 phox subunit combines with a smaller subunit p22 to generate a heterodimeric, transmembrane oxidase complex (22). Redox centers in the oxidase include a cytochrome b, FAD, and NADPH binding domains (20,23). Fre1 is a flavocytochrome, has NADPH oxidase activity, and shares sequence similarity with gp91 phox in regions believed to be FAD and NADPH binding sites (5,24). Four critical histidines have been identified in Fre1 as candidate cytochrome b heme ligands, and these histidines are conserved in gp91 phox (25).
The disruption of FRE1 and FRE2 resulted in cells lacking surface ferric reductase activity but retaining residual cupric reductase activity (6). With the sequence of the S. cerevisiae genome completed, we found five unidentified, candidate ORFs 1 that share considerable sequence similarity to Fre1 and especially Fre2. Four homologs exhibit more than 30% sequence identity to Fre2 but less than 30% identity to Fre1. The similarities between Fre1, Fre2, and the Fre2 homologs are so considerable, it is likely that these molecules will be structurally and functionally related. Sequence identity in excess of 30% typically predicts similar tertiary folds (26).
Our approach was to initially determine whether the five candidate yeast genes were metalloregulated as an initial criterion of whether the candidate gene product may be important in copper or iron ion uptake. Our results show that all five Fre2 homologs are metalloregulated. Four are specifically modulated by iron salts through Aft1 and one uniquely regulated by copper salts in a Mac1-dependent manner.

MATERIALS AND METHODS
Strains-Strains used in the studies are listed in Table I. Strains CM3260, M2 (AFT1-1 up ) (14), and Y18 (aft1⌬) in the CM3260 background were kindly provided by A. Dancis. The ⌬fet3 strain DEY1397-6A in a W303 background (10) was kindly provided by J. Kaplan. Mac1 up1 strain UPC31 and its BR10 parental strain were provided by Dean Hamer and mac1-1 strain YJJ1 and parental YWO1 (27) were provided by S. Jentsch. Standard methods were used for culturing yeast.
Plasmid Construction-Promoter-lacZ fusion plasmids were constructed from FRE1-FRE7 using PCR to amplify promoter sequences (approximately Ϫ900 bp upstream from ATG to ORF codon 10) from yeast genomic DNA and subsequent cloning into YEp354 to create the Fre-lacZ fusions (28).
The PCR amplified FRE7 5Ј fragment was subcloned into pAlter-1 for mutagenesis using the Promega Altered Sites Mutagenesis System. Mutations were generated in the TTTGCnC sequences resulting in TTTCAXC at positions Ϫ169 (pF7A), Ϫ128 (pF7B), and Ϫ345 (pF7C), respectively. Plasmids containing double mutations were generated by subcloning the mutagenized FRE7 promoter from pF7A-Alter and pF7B-Alter back into pAlter-1 generating vectors pF7AЈ-Alter and pF7BЈ-Alter, and the third mutagenic primer was then used to change the TTTGCnC sequence at position Ϫ345 to TTTCAnC in pF7AЈ-Alter and pF7BЈ-Alter generating vectors pF7AC-Alter and pF7BC-Alter. Mutagenesis was carried out as described by the Promega instructions. Restriction mapping and sequence analysis confirmed the inclusion of the mutagenic sequence. The mutagenized FRE7 promoters were excised and subcloned into YEp354.
The minimal copper-responsive region of CTR1 was mapped to a segment Ϫ339 to Ϫ299 (numbers indicate bp upstream of translation start). This segment was inserted into SalI/XbaI sites in the ⌬UAS vector pNB404 (29), which uses the lacZ reporter gene. A BglII site was introduced between the TTTGCTCA repeats in the CTR1 promoter by PCR. The PCR product was digested with SalI/XbaI and subcloned into vector pNB404 generating pCTR109. DNA for the insertional mutagenesis was obtained by PCR from CTR1 sequences 5Ј to the repeats. This segment contributed no activation of lacZ expression when cloned into the pNB404. The PCR fragments having BglII sites at both the 5Ј and 3Ј ends were ligated into pCTR109 digested with BglII and treated with calf alkaline phosphatase. The spacing between the TTTGCTCA repeats was increased from 14 nucleotides to 60 (pCTR110), 114 (pCTR111), or 164 (pCTR112) nucleotides.
␤-Galactosidase Assays-Plasmids pCTR109 -112 and pNB404 were transformed into strain DY395. Mutated FRE7-lacZ plasmids were transformed into strain S288c. Yeast transformants were grown in low copper synthetic medium (BIO101) lacking uracil to an A 600 nm value of 1. The cells were then diluted 100-fold into medium containing either no added copper or 100 M CuSO 4 for DY395 and 30 M bathocuproine disulfonate (BCS) or 100 M CuSO 4 for S288c. The cells were harvested at stationary phase and washed with buffer (85 mM Na 2 HPO 4 , 45 mM NaH 2 PO 4 , pH 7.4, 10 mM KCl, 1 mM MgSO 4 , and 5 mM ␤-mercaptoethanol). Cells were resuspended in 200 l of the mentioned buffer and lysed by vortexing with glass beads. The extract was filtered clarified, and ␤-galactosidase was assayed using -nitrophenyl ␤-D-galactopyranoside as the substrate (30). Assays were incubated at 30°C until the A 420 nm was between 0.1 and 0.4, at which time 1 M Na 2 CO 3 was added to stop the reaction. The A 420 nm was recorded, and protein concentrations were determined by the method of Bradford (31). Activity is quantified as A 420 nm (ϫ1000)/min/mg protein. Each sample was assayed in triplicate.
mRNA Quantitation by S1 Nuclease Analysis-S1 nuclease assays were carried out as described previously and quantified by Phospho-rImaging (32). RNA was isolated from mid-logarithmically grown cells.
For certain experiments cells were incubated in BCS or bathophenanthroline sulfonate (BPS) prior to isolation of RNA. Details of each experiment are outlined in the figure legends. DNA probes used consisted of 60 nucleotides for FRE1, 60 for FRE2, 55 for FRE3, 66 for FRE4, 65 for FRE5, 69 for FRE6, and 72 for FRE7. The sequences used for S1 probes arise from 5Ј ORF sequences. In the experiment presented in Fig. 5, different DNA probes were used for FRE6 (55 nucleotides) and FRE7 (50 nucleotides). The CMD1 calmodulin probe was 40 nucleotides in length arising from 5Ј ORF sequences.

RESULTS
Fre1 and Fre2 share only 25% sequence identity, yet sequences involved in the candidate NADPH and FAD binding sites are conserved in the two as are the four candidate hemeliganding histidyl residues (25). The five homologs exhibit sequence conservation in the same regions, and four of the homologs show Ͼ30% sequence identity to Fre2 (Fig. 1). The homologs also share similar polypeptide lengths, pI, and the predicted number of transmembrane segments from hydropathy plots (Fig. 1). Fre7 shows the weakest similarity to Fre2 and Fre1 yet exhibits sequence conservation in the NADPH, FAD, and heme binding sites. The sequence similarities suggest that these gene products are part of a protein family with Fre1 and Fre2. Family names are suggested based on sequence relatedness to Fre2 (Fig. 1).
The initial goal was to determine whether the five FRE2 homologous genes were expressed genes. Because FRE1 and FRE2 are expressed under copper and iron limiting conditions, respectively, we carried out mRNA quantitation using the S1 nuclease RNA protection assay on RNA extracted from logarithmically growing CM3260 cells cultured in the presence or absence of metal ion chelators (Fig. 2). The Cu(I)-selective chelator, BCS, or the metal chelator, BPS, was added to the culture medium to decrease the availability of metal ions. The addition of the copper-specific BCS chelator to the culture medium led to a prominent enhancement (20-fold) in mRNA levels for only FRE1 and FRE7 relative to the calmodulin mRNA control (Fig. 2A). In contrast, the addition of BPS, which exhibits a preference for Fe(II) over Cu(I) ions resulted in a marked enhancement of five Fre genes FRE1-FRE5 and limited enhancement (2.8-fold) in FRE6 mRNA levels (Fig. 2B). FRE6 exhibited high basal expression in cells cultured in the  Because BPS is not specific for iron ions, we sought to determine whether the BPS-induced expression of FRE3-FRE6 was a result of iron deprivation. Iron deprivation occurs in cells lacking a functional FET3 (strain DEY1397, fet3⌬), which encodes a critical protein involved in high affinity iron uptake (10). Quantitation of mRNA levels of the chromosomal Fre genes in fet3⌬ cells showed clear iron-inhibited mRNA levels for FRE1-FRE6 (Fig. 3). No changes in FRE7 expression levels were observed. Only limited iron inhibition of FRE6 expression was observed, consistent with the minimal BPS induction observed in Fig. 2B. FRE1 is metalloregulated by both iron and copper salts through two independent sensory pathways (8,15). Quantitation of mRNA levels in stationary phase fet3⌬ cells revealed marked diminution in FRE3 and FRE4 levels but a 2-fold increase in FRE6 mRNA levels relative to that in logarithmically growing cells (data not shown).
To determine whether the iron regulation of FRE3-FRE6 expression was mediated by the iron-responsive factor Aft1, mRNA levels were quantified in wild-type, aft1⌬, and AFT1-1 up cells using the S1 nuclease assay (Fig. 4). FRE3, FRE4, FRE5, and FRE6 mRNA levels were attenuated 5-, 2.6-, 3.9-, and 1.3-fold in aft1⌬ cells (strain Y18) cultured in BPS medium compared with wild-type CM3260 cells (Fig. 4A). Iron inhibition of gene expression is abrogated in AFT1-1 up cells (strain M2) resulting in elevated expression (14). Likewise, FRE3, FRE4, FRE5, and FRE6 were elevated in AFT1-1 up cells with respect to wild-type CM3260 cells (Fig. 4B). The Aft1 modula-  BPS (B). RNA was extracted from cells at A 600 nm ϭ0.5, incubated with DNA probes for the various Fre genes, and digested with the S1 nuclease. Calmodulin (CMD1) mRNA levels were quantified as the loading control. Different length DNA probes (55-72 bp as specified under "Materials and Methods") were used for the Fre genes for clarity. In each case the Fre gene probe exceeds the CMD1 probe in length, and therefore the S1 protected fragments for the Fre genes are retarded in mobility. Note: The FRE5 and FRE6 S1 probes yield doublets, and both are specific bands. The doublets may arise from hybridization instability at the 3Ј ends of the probes. tion of FRE4 was the most modest of the Fre homolog genes in the CM3260 background. In two different experiment the ratios of FRE4 mRNA levels in AFT1-1 up versus null cells were 6 and 10. Thus, Aft1 mediates iron regulation of transcription of the four new iron-regulated Fre genes (FRE3, FRE4, FRE5, and FRE6) in addition to its known effect on FRE1 and FRE2.
FRE7 and FRE1 were the only genes up-regulated by the addition of the copper-specific BCS chelator to the culture medium suggestive of specific copper regulation. The BCS-induced expression of FRE1 and FRE7 is consistent with FRE7 being copper-regulated as is known for FRE1 (8,17,18). If FRE7 is copper-metalloregulated, the prediction is that FRE7 expression would be elevated in copper-deficient ctr1, ctr3 cells but inhibited in copper-treated cultures. The absence of either high affinity copper permease is known to limit copper uptake into cells (13). DY395 cells lacking functional Ctr1 and Ctr3 copper ion permeases revealed copper-inhibited expression of FRE7 (Fig. 5A). Cu(II) can be transported in ctr1,ctr3 cells by low affinity transporters (11,16). Copper salts had no effect on expression levels of FRE2 and FRE4-FRE6. FRE3 mRNA levels were slightly attenuated in copper-treated cells, but it is known that iron uptake can be compromised in ctr1,ctr3 cells by virtue of the requirement of copper ions in Fet3 for its ferro-oxidase activity (12). The inhibition of FRE7 expression was specific for copper salts, because the addition of iron salts did not attenuate FRE7 expression.
To document whether the copper metalloregulation of FRE7 involves Mac1, the known copper-regulated activator of FRE1, CTR1, and CTR3, FRE7 expression was quantified in cells containing the semidominant MAC1 up1 allele, which precludes the normal copper inhibition of expression (27). FRE7 and, as expected, FRE1 were clearly induced in MAC1 up1 cells (strain UPC31) relative to the isogenic wild-type cells (strain BR10) (Fig. 5B). The MAC1 up1 allele had no effect on expression of FRE2-FRE6. No BCS-induced expression of FRE7 was observed in mac1-1 cells (data not shown).
Curiously, FRE3 and FRE4 are well expressed in BR10 wildtype cells but not in CM3260 wild-type cells. To test whether the elevated expression of FRE3 and FRE4 in BR10 cells resulted from a iron-deficient state, BR10 cells were cultured in the presence of added iron salts in the growth medium. FRE3 and FRE4 were attenuated in their expression 6-and 11-fold, respectively, relative to standard unsupplemented medium (data not shown). Thus, BR10 cells cultured in minimal medium are iron-deficient.
Mac1-responsive genes (CTR1, CTR3, and FRE1) are known to contain two TTTGCTCA repeats, both of which are critical for transcriptional initiation (Fig. 6B) (17, 18). The 5Ј sequence of FRE7 contains one consensus copper-regulated Mac1-specific UAS elements (TTTGCTCA Ϫ128 inverted) and two imper- 3. Effect of iron treatment on expression of FRE1, FRE3,  FRE4, FRE5, FRE6, and FRE7 in the ⌬fet3 strain (DEY1397). Cells were pregrown and incubated for 3 h in the presence or absence of 20 M ferric ammonium citrate. Cells were harvested at a final A 600 nm ϭ 0.3. RNA was quantified by the S1 nuclease assay. With normalization to CMD1 mRNA levels, the addition of Fe(III) resulted in a 3-, 4-, 7-, and 1.6-fold attenuation in FRE3, FRE4, FRE5, and FRE6 mRNA levels, respectively. The FRE5 changes were readily apparent in a longer exposure. FIG. 4. Quantitation of mRNA levels of FRE3-FRE6 in wildtype cells, AFT1 up1 and aft1⌬ null cells. mRNA levels of each gene relative to calmodulin mRNA levels were quantified by PhosphorImaging after S1 analysis. The ratio of radioactivity is presented for pairwise comparisons of wild-type (WT) CM3260 and aft1⌬ null cells (strain Y18) cultured in the presence of 0.05 mM BPS (A) and wild-type versus AFT1 up1 cells cultured in the absence of BPS (B). C shows the candidate Aft1 binding sites in FRE1-FRE6. 5. Copper metalloregulation of FRE1 and FRE7. The effect of copper treatment on expression of Fre genes was assessed in ctr1,ctr3 cells (strain DY395) (A). Cells were cultured in low copper medium or medium containing 20 M Cu(II) to a final A 600 nm of 0.4. RNA quantitation was carried out by S1 analysis. Expression of the Fre genes was assessed in wild-type (wt, BR10 cells) versus MAC1 up1 (up) cells (strain UPC31) (B). Cells were cultured in minimal medium to a final A 600 nm of 0.3. FRE1 and FRE7 mRNA levels were elevated 15-and 20-fold greater in MAC1 up1 cells compared with the wild-type control.
fect elements (TTTGCGCA Ϫ169 and TTTGCGCA Ϫ345 inverted). The site centered at Ϫ172 is palindromic TTTGCGCAAA. To determine whether these sequences were copper-regulated UAS elements, mutations were engineered in each of the candidate elements converting a consensus TTTGCXCA sequence to TTTCAXCA (Fig. 6A). The mutations were introduced into the FRE7-lacZ fusion gene. A mutated proximal element (Ϫ128-bp element) attenuated BCS-induced lacZ expression by about 3-fold in S288c wild-type cells. In contrast, the mutations within the palindromic element (Ϫ172-bp element) abolished lacZ expression. The mutations in the distal element (Ϫ345-bp element) was without effect by itself, but mutations in both the distal and proximal elements reduced lacZ expression dramatically. Thus, the Mac1-responsiveness of FRE7 requires at least two elements. The preferred pair are the two proximal sites, but in the absence of the proximal element at Ϫ128, the distal element is partially functional.
The lack of lacZ expression in the FRE7-lacZ fusion containing double mutations in both proximal and distal elements may relate to either a dominance of the middle palindromic element or to a spacing constraint. To test whether spacing of a pair of Mac1-responsive elements is important, we carried out inser-tional mutagenesis with the Mac1-responsive CTR1 5Ј sequences. Insertional mutagenesis was then carried out on a minimal sequence (Ϫ339 to Ϫ299) of CTR1 containing the two elements in inverted orientation (17,18). Elimination of the upstream element abolished lacZ expression, consistent with the known importance of two elements for Mac1-mediated regulation (18). The spacing between elements was increased from 14 to 60 to 164 bp (Fig. 7). Increasing the spacing between the two elements resulted in a gradual attenuation in lacZ expression. The spacing of 164 bp reduced lacZ expression nearly 4-fold. Thus, spacing between Mac1-responsive elements is important in CTR1. This probably accounts for the lack of Mac1responsiveness in FRE7 containing only proximal and distal elements (construct with the Ϫ172-bp element mutated) (Fig.  6A). DISCUSSION The Fre1 and Fre2 metalloreductases are well known to function in increasing the bioavailability of iron and copper complexes in metal deficient S. cerevisiae cells (3,6). These metalloreductases are believed to be NADPH oxidases based on their homology to the human neutrophil gp91 phox and based on FIG. 6. A, mutations were introduced into the putative Mac1 binding sites of the FRE7 promoter. The TTTGCXC was altered to TTTCAXC in each case. The constructs were transformed into strain S288c and grown in low copper medium with either 30 M BCS or 100 M CuSO 4 added. Unlike other genes regulated by Mac1, the FRE7 promoter contains three repeats of the TTTGCnC sequence. Cells harboring mutant FRE7-lacZ fusions with mutations in TTTGCnC elements were assayed in triplicate for ␤-galactosidase activity. Activity is quantified as A 420 nm (ϫ1000)/min/mg protein. B, the Mac1 sites in the four genes known to be regulated by Mac1. Sites in CTR1 and CTR3 were mapped previously (18).

FIG. 7. Effect of spacing between TTTGCTC elements.
Regions of the CTR1 promoter 5Ј to the repeats were amplified by PCR and subcloned into a BglII site engineered between the two TT-TGCTC repeats. The spacing between the repeats was increased from 14 nucleotides to 60, 114, or 164 nucleotides. The constructs were assayed as stated above in ctr1,ctr3 cells (strain DY395) cultured in the presence and absence of 100 M Cu(II). Note: ctr1,ctr3 cells cultured in low copper medium are copper-deficient, so the addition of BCS is not required to attain a copper-deficient state. the fact that Fre1 binds flavocytochrome and exhibits NADPH oxidase activity (5,21,24,25). The five homologs to Fre1 and Fre2 in S. cerevisiae exhibit sequence similarity in regions expected to bind FAD, NADPH, and the two hemes (25). The combination of the similarities in sequences, pI, and hydropathy plots suggests that the homologs are part of a family of proteins related to Fre1 and Fre2.
In the present study it is shown that the homologs are expressed genes in S. cerevisiae and that their expression is metalloregulated as is true with FRE1 and FRE2. Four of the homologs (FRE3-FRE6) are specifically iron-regulated through the Aft1 transcription factor. The evidence for ironspecific regulation through Aft1 is 3-fold. First, these genes are induced in cells treated with the BPS chelator that is known to lower the availability of Fe(II) ions in the growth medium. Because BPS is not specific for Fe(II) ions, the BPS result is not conclusive. Second, the genes are expressed in fet3 cells that are impaired for high affinity iron uptake (10). Third, the genes are not induced in cells lacking the iron-responsive Aft1 transcriptional activator (aft1⌬ cells) even in the presence of the chelator BPS. As expected, FRE3 and FRE6 contain at least one consensus Aft1 binding site in their 5Ј sequences (Fig. 4C). FRE4 and FRE5 contain candidate Aft1 sites that deviate slightly from the consensus at the 3Ј base. These two genes contain a T as the 3Ј base of the candidate Aft1 sites. FRE4 contains only a single candidate Aft1 site, and its expression is the lowest of the homologous genes as detected by S1 analyses.
The fifth homolog FRE7 is specifically copper-regulated in its expression by Mac1. Copper modulation of FRE7 was demonstrated by the induced expression of FRE7 in copper-deficient cells resulting from either the addition of a Cu(I)-specific chelator to the growth medium or genetic disruption of the high affinity permease Ctr1. The Mac1 dependence of FRE7 expression was shown by the elevated expression in MAC1 up1 cells and absence of expression in copper-deficient mac1-1 cells.
Four important features emerge for Mac1-responsive promoter elements. First, two Mac1-responsive elements are critical for FRE7 expression as well as Mac1-dependent CTR1 and CTR3 expression (17,18). FRE7 contains three Mac1-responsive promoter elements in its 5Ј sequences. Two promoter elements yield Mac1-responsive expression of FRE7 as long as the central element is present. The distal element can be deleted without compromising FRE7 expression. Although the distal element is a functional UAS, it does not act synergistically with two functional proximal elements. Second, the two functional promoter elements can be either direct or inverted repeats (Fig.  6B). The two elements in CTR1 and CTR3 are inverted repeats but direct repeats in FRE1. The proximal two elements in FRE7 are inverted repeats. FRE1 may also have additional elements. Although the two direct repeats in FRE1 are effective for Mac1-dependent expression, additional candidate elements may exist upstream. Third, the spacing between Mac1-responsive sites is of limited importance because CTR1 expression remains high when the spacing between the two elements varies from 14 to 60 bp. Expression is attenuated when the spacing is increased to 164 bp. Likewise, the combination of only the proximal and distal sites in FRE7 resulting in a 209 bp spacing yields only marginal expression. Fourth, from the three Mac1-responsive elements in FRE7, a new consensus sequence for Mac1 binding can be established as TTTGC(T/G)C(A/G). The importance of two elements with loose requirements of spacing and orientation in each of the Mac1-responsive genes is suggestive that synergism but not dimerization of Mac1 is required for transactivation.
Nothing is currently known of the function of the Fre homologs. All genes known to function in high affinity iron uptake are transcriptionally regulated by Aft1 (15). This fact is consistent with Fre3-Fre6 functioning in iron homeostasis. The three known Mac1-responsive genes are important in copper ion uptake, so Fre7 may be important in copper homeostasis. The PROSITE algorithm predicts a plasma membrane localization for each homolog, and each appears to have a signal sequence. However, cells lacking Fre1 and Fre2 have no apparent Fe(III) reductase activity and only marginal Cu(II) reductase activity (6). A fre1,fre2,fre7 null strain retained the residual Cu(II) reductase activity, suggesting that Fre7 was not responsible for the weak Cu(II) reductase activity observed in the fre1, fre2 strain (data not shown). It is conceivable that the Fre homologs may function as internal metalloreductases. Alternatively, the observed homology may not accurately predict a metalloreductase function for these molecules. Future studies on localization of the homologs may provide clues for possible function.