Aft2p, a novel iron-regulated transcription activator that modulates, with Aft1p, intracellular iron use and resistance to oxidative stress in yeast.

The yeast, Saccharomyces cerevisiae, contains a transcription activator, Aft1p, that regulates the transcription of the high affinity iron transport system genes. This report describes the properties of Aft2p, a protein 39% homologous to Aft1p. Aft2p was found to activate transcription. Overproduction of Aft2p activates the transcription of the AFT1 target gene FET3. The double aft1aft2 mutant was unable to grow in iron-deprived conditions. Because a fet3 mutant does not show this deficiency, the defect is not solely caused by mis-regulation of iron transport but also involves defective iron use by the cells. The aft1 cells were unable to grow in aerobic conditions on plates containing raffinose as the sole carbon source. The inability to grow on raffinose is not caused by the cell iron content being too low to sustain respiratory metabolism, because the oxygen consumption of aft1 mutants showed that their respiratory activity is 2-fold higher than that of controls. The double aft1aft2 mutant also has many phenotypes related to oxidative stress such as H(2)O(2) hypersensitivity, oxygen-dependent copper toxicity, and oxygen-dependent methionine auxotrophy, which are suppressed in anaerobiosis. These results suggest that Aft2p and Aft1p have overlapping roles in the control of iron-regulated pathway(s) connected to oxidative stress resistance in yeast.

Iron is required by all organisms; it is used as a cofactor in key redox enzymes involved in such diverse biological processes as cell respiration and the synthesis of metabolic intermediates. However, iron can damage cells by reacting with hydrogen peroxide to form the hydroxyl radical. This highly toxic compound damages DNA, proteins, and lipids (1). Prokaryote and eukaryote cells therefore have evolved various systems for the close regulation of iron transport and its intracellular use (2,3). The yeast Saccharomyces cerevisiae has several pathways of iron transport including a low affinity transporter encoded by FET4 (4) and a number of high affinity systems. One of these high affinity systems is encoded by FET3 and FTR1 (5,6). It requires the reduction of Fe 3ϩ to Fe 2ϩ by plasma membrane reductases (FRE1 or FRE2) (7,8). The other high affinity transport systems are siderophore-mediated and depend on several homologous transporters of the major superfacilitator (MSF) family encoded by ARN1, SIT1, TAF1, and ENB1 (9 -12). All the genes involved in the high affinity iron transport systems are under the control of the iron-dependent transcription activator Aft1p (12)(13)(14). However, Aft1p does not affect the regulation of FET4 (15). In addition to the high affinity transport genes, Aft1p regulates the transcription of other iron-responsive genes such as those encoding Ccc2p (the intracellular copper transporter responsible for delivering copper to Fet3p) (16), Fth1p (which forms an iron transport complex on the vacuole with Fet5p) (17), Fre3p/Fre4p (which are potential siderophore-iron reductases) (18), and Fre5p/Fre6p of unknown function (19). Finally, Aft1p seems to be involved in, but is not required for, the transcription of genes such as ATX1, which encodes the copper chaperone that delivers copper to Ccc2p, and ISU1/ISU2, the products of which are involved in the mitochondrial assembly of iron-sulfur clusters (20,21). In conditions of iron deprivation, Aft1p actives the transcription of target genes by binding directly to the consensus sequence T/CG/ACACCC, a motif in its 5Ј-upstream regions. Although the functional DNA binding and activation domains of Aft1p have not yet been characterized, its similarities to other transcription factors suggest that the N-terminal region rich in basic residues is involved in recognizing DNA, and the Cterminal glutamine-rich domains may be required for the transcriptional activation function. The dominant allele AFT1 up , which leads to Phe-291 in place of Cys-291 in the N-terminal region of the protein, constitutively binds to DNA (13,14). Just how Aft1p senses iron and activates gene expression is not understood, and recent data indicate that the Aft1p-mediated response is not restricted to iron deprivation conditions but can be induced by the concentrations of molecular oxygen or copper in the culture medium (22,23). We have now attempted to obtain further information about the regulation of iron homeostasis by inspecting the yeast genome by blast analysis. We found a 416-amino acid protein encoded by the YPL202c gene that presents 39% overall identity to Aft1p, which we designated Aft2p (Fig. 1). This paper describes some of the characteristics of Aft2p. It is a second iron-regulated transcriptional activator in yeast. Aft2p is required for iron homeostasis and resistance to oxidative stress when Aft1p is absent. Molecular data and phenotypic analysis suggest that Aft1p and Aft2p have overlapping functions.  (MAT␣, trp1-63, leu2-3, 112 gcn4-101, his3-609, aft1::TRP1,  aft2:: kanMX4) strains were constructed by integrating kanMX4 at the AFT2 locus in strains CM3260 and Y18 as described in Ref. 24 . The  following primers were used to amplify the open reading frame replacement cassette containing the kanMX4 marker with long flanking homology regions of the AFT2 promoter and terminator: 5Ј-GGGTATA-AGGAGTGTCAAAG-3Ј, 5Ј-GGGGATCCGTCGACCTGCAGCGTACCA-TTTCTTGGGGTCGCTTTC-3Ј, 5Ј-AACGAGCTCGAATTCATCGATGA-TATAATTATTTAGTTTTCAACTC-3Ј, and 5Ј-GATGCCTTATTTGTGG-TCTG-3Ј. The sequences underlined are homologous to the kanMX4 marker. Kanamycin-resistant clones were selected on YPD plates containing G418 (200 g/ml). Deletions were confirmed by polymerase chain reaction using primers flanking the insertion region.

Yeast Strains and
The plasmid pEG202 and the derivative plasmids pEG202-AFT1 and pEG202-AFT2 contained the DNA binding domain of LexA. Plasmid pEG202-AFT1 expressing LexA-Aft1p and plasmid pEG202-AFT2 encoding LexA-Aft2p were constructed by integrating the open reading frames of AFT1 or AFT2 in frame with the DNA binding domain of LexA. Open reading frames were amplified with the following oligonucleotides: AFT1, 5Ј-CAGAAGAATTCACGACAATGGAAGGCTTC-3Ј and 5Ј-TTCATCTCGACTAATCTTCTGGCTTCAC-3Ј; and AFT2, 5Ј-A-AGCGGGATCCAAGAAATGAAAGCAAAGTCGA-3Ј and 5Ј-GAAAACT-CGAGAATTAATATTTTGATATTAAGGC-3Ј. The amplified fragments were digested by EcoRI and XhoI for AFT1 and by BamHI and XhoI for AFT2 and ligated to the DNA binding domain of LexA derived from the plasmid pEG202 (25) digested with the appropriate restriction enzymes. The plasmid pSH18 -34 contained the GAL1-lexop-lacZ reporter gene from which the upstream activation sequence GAL1 had been deleted and replaced by four lexA operators (26). Plasmid pFC-W, generously supplied by Yamaguchi-Iwai, contained a 30-base pair cassette of the FET3 upstream activation sequence (Ϫ263 to Ϫ234) fused to LacZ (14).
Media, Growth Conditions, and Plate Assays-Yeast strains were grown in rich medium (1% yeast extract/2% peptone) containing either 2% glucose (YPD) or 2% raffinose, or in copper and iron limiting yeast nitrogen base (Bio101) plus the required amino acids and 2% glucose. Ferric ammonium sulfate (100 M) was added to the defined medium to create the iron-repleted medium. Copper ammonium sulfate (10 -100 M) was added to the defined medium to create the copper-repleted medium in experiments where copper toxicity was assessed. However, because of the hypersensitivity of the double aft1aft2 mutant to this ion (see "Results"), copper was omitted from the regular synthetic medium. Cells were grown anaerobically in a Forma Scientific anaerobic chamber. The cells were grown in liquid cultures in 50-ml Falcon tubes placed on a gyroshaker at 30°C. For plate assays, the cells were suspended in water at 2 ϫ 10 6 cells/ml, plated in serial 10-fold dilutions, and incubated at 30°C for 3-4 days prior to imaging.
RNA Isolation and Northern Analysis-Total RNA was isolated by the hot phenol method (27). Northern blotting was performed as described previously (28). The DNA fragments used as probes corresponding to the open reading frames of each gene were as follows: for FET3 a 0.8-kilobase BamHI-ScaI fragment, for ACT1 a 1.2-kilobase BamHI-HindIII fragment, and for ATX1 a 0.26-kilobase fragment that was isolated from a polymerase chain reaction-amplified product corresponding to the open reading frame of ATX1 by using the primers 5Ј-CAAGAGGATCCTAGCGAAAAGATGGCAGAG-3Ј and 5Ј-ATTTCTCTGCAGTTTCATTCACAATTGTTTGCC-3Ј.
Measurement of Total Iron-The designated isogenic yeast strains were grown in the iron and copper limiting yeast nitrogen base medium (Bio101) supplemented with 10 M iron for 18 h, collected by centrifugation, and washed twice in 10 mM EDTA and once in metal-free water. The total iron was measured by inductively coupled plasma atomic emission spectroscopy at the Microanalysis Facility of the Centre National de la Recherche Scientifique (Vernaison, France) on cells suspended in 0.1 N nitric acid.
Oxygen Consumption-The designated isogenic yeast strains were grown in the YPD medium without (ϪFe) or with 100 M iron (ϩFe). The respiratory activity of whole cells was evaluated by an oxypolarographic method. The rate of oxygen consumption was measured using a 2.2-ml thermostated oxypolarographic cell equipped with a Clark-type electrode. The respiratory medium was a 0.1 M potassium phosphate buffer, pH 7.2 (30°C), saturated with air (236 M dissolved O 2 ) containing 2 mM glucose. The cells were prepared as 50% (w/v) suspensions in 0.1 M potassium phosphate buffer, pH 7.2.

Aft2p, a Protein Homologous to Aft1p-
The open reading frame of AFT2 predicted a protein with a molecular mass of 47 kDa and a pI of 9. The N-terminal parts of Aft2p and Aft1p were the most similar regions in the proteins. This region is rich in basic amino acids and contains four cysteine residues, which are potential iron ligands, at residues 143, 215, 291, and 293 in Aft1p and at 86, 109, 187, and 189 in Aft2p. But we found notable differences between the two protein sequences despite their similarity. Aft2p contained no N-or C-terminal histidine-rich domains; these have been predicted to coordinate iron in Aft1p (13). Aft2p lacked the C-terminal glutamine-rich domain of Aft1p, which may be required for the transactivation function of Aft1p (31). The homologous domains in the conserved N-terminal region were separated by inserted sequences in Aft1p. For example, the very highly charged segment of Aft1p (amino acids 147-168) was absent from Aft2p ( Fig. 1). Finally, the pSORT program indicated that there are several nuclear localization signals such as basic clusters and a bipartite sequence signal in the N-terminal region of Aft1p. No such signals could be identified in the Aft2p sequence ( Fig. 1). Therefore, there is a 78% possibility that Aft1p is nuclear but only a 56% chance that Aft2p is a nuclear protein. But there is also a 35% possibility that Aft2p is a mitochondrial protein.

AFT2 Encodes an Iron-regulated Transcription Activator-
The complete open reading frame of the Aft2p protein was fused to that of the bacterial lexA DNA binding domain. This lexA-Aft2p chimera was then produced in the wild-type strain CM3260 carrying a Gal1-lacZ reporter gene with four lexA operators replacing the upstream activation sequence GAL1. The lexA-Aft1p fusion protein was also produced in these cells. The lexA-Aft2p fusion protein activated the transcription of the Gal1-lacZ reporter gene, indicating that Aft2p has a transactivation function (Table I). Cells were grown in synthetic medium with or without 100 M ferrous ammonium sulfate; irondepleted cells and iron-repleted cells produced the same amount of ␤-galactosidase activity in the presence of lexA-Aft1p, whereas iron-repleted cells produced only half as much as iron-depleted cells in the presence of lexA-Aft2p (Table I).
The same results were obtained with iron citrate as a source of iron (data not shown).
Effect of Aft2p on the Transcription of Genes Involved in Iron Metabolism-We examined the role of AFT2 in the transcription of the AFT1-regulated genes by constructing isogenic wildtype and aft1 mutant strains lacking the AFT2 gene. We also constructed a strain overproducing Aft2p by introducing a high copy number plasmid carrying the AFT2 gene into the double aft1aft2 mutant. We measured the amounts of FET3 and ATX1 mRNAs in the various aft mutants grown under iron-depleted or iron-repleted conditions. The aft2 and wild-type strains contained similar amounts of FET3 transcripts ( Fig. 2A). The amount of FET3 mRNAs in the aft1 strain was much lower, and the low level of transcription was repressed by iron. No signal was detected in the double aft1aft2 mutant even in iron-deprived conditions. The overproduction of Aft2p in the aft1aft2 mutant restored the iron-regulated transcription of FET3. There were fewer ATX1 transcripts in iron-repleted conditions than in iron-depleted conditions in the wild type or in single aft1 and aft2 mutants. However, the levels of ATX1 mRNA seemed unaffected by the iron status in the double aft1aft2 mutant. Moreover, overproduction of Aft2p in the aft1aft2 cells was accompanied by a significant increase in ATX1 mRNA concentration. We investigated the capacity of AFT2 to activate the transcription of FET3 by acting at the Aft1p DNA binding site. Cells of the same strains were transformed with a plasmid containing the AFT1-responsive ele-ment of FET3 placed in front of a ␤-galactosidase reporter gene. This construct made ␤-galactosidase activity regulated by AFT1 (14). Thus, the ␤-galactosidase activity gave a measure of the activation of transcription at the AFT1 binding site. Both aft2 and wild-type strains grown under iron-deprived conditions had similar ␤-galactosidase activities (Fig. 2B). In contrast, no ␤-galactosidase activity was detected in the aft1 and aft1aft2 strains. Thus, overproduction of Aft2p by the aft1aft2 cells significantly increased the ␤-galactosidase activity, which is in agreement with the results shown in Fig. 2A. This AFT2dependent transcription activity was also repressed strongly by adding iron.
Regulation of Intracellular Iron Use by the Aft Proteins-The wild-type strain and the aft2 and fet3 mutants grew well on agar plates supplemented with the specific iron chelator ferrozine, whereas the aft1 strain grew less well under these condition (Fig. 3). The growth of the aft1 mutant was restored by increasing the concentration of iron from 10 to 100 M. The growth of the double aft1aft2 mutant was completely abolished under iron-deprived conditions. It was only restored by adding 100 M iron to the ferrozine-containing plates. Thus, the aft1 strain was sensitive to iron deprivation, and this sensitivity was exacerbated by the absence of Aft2p. The inability of the aft1 and the aft1aft2 strains to grow on iron-poor medium was not caused by a deficiency in the high affinity iron transport system, because the growth of the isogenic fet3 mutant was not affected under these conditions. We next determined whether the requirement of more external iron by the aft1 and aft1aft2 mutants was because of a smaller intracellular iron pool or a defect in the iron availability in the mutant cells. We measured the intracellular iron content of the cells before growing them in iron-deficient medium. The aft mutants had iron pools similar to those of the wild-type strain before the shift (Fig. 4A). However, the maximal cellular density reached in the medium without iron by the aft1 strain was lower compared with the wild-type and aft2 strains (Fig. 4B). The growth defect was exacerbated for the aft1aft2 mutant. The growth of the aft1aft2 strain was restored completely by adding a limited concentration of iron (Fig. 4C).
Effects of AFT1 and AFT2 on Growth Under Respiratory Conditions-The rates of oxygen consumption of the two aft1 mutants were much higher than those of the wild-type and aft2 strains (Table II). The oxygen consumption was sensitive to cyanide and was nearly the same with or without high iron concentrations in the growth medium. We investigated this further by analyzing the phenotypes of cells grown in a medium containing raffinose as the sole carbon source. Unlike glycerol, raffinose is a not a strictly respirable carbon source. We thus could analyze the growth of the cells in the presence or absence of oxygen. The growth of the fet3 mutant on raffinose under aerobic conditions was slightly affected compared with the aft2 and wild-type strains (Fig. 5). In contrast, aft1 mutant strains were unable to use this carbon source under aerobic conditions. The aft1-related growth deficiency was suppressed by overproducing Aft2p. The aft1 mutant phenotype could also be rescued by adding iron and also copper to the medium. The aft1aft2 mutant phenotype was not rescued by the addition of iron or copper. Similar results were obtained when glycerol was used instead of raffinose (data not shown). Finally, the growth of both the aft1 and the aft1aft2 strains was restored completely under anaerobic conditions without adding iron or copper.
Hypersensitivity of aft1 and aft1aft2 Mutants to Oxygen Stress-We investigated the sensitivity of the aft mutants to , which is marked by an asterisk. The highly charged N-terminal sequence of Aft1p, which is absent from Aft2p, is also boxed. Sequences that are underlined indicate the potential nuclear localization signals in Aft1p as predicted by pSORT.

TABLE I LexA-Aft2p as a transcription activator
Wild-type strain CM3260 was transformed with both the plasmid pSH18-34 and plasmids pEG202, pEG202-AFT1, or pEG202-AFT2. Appropriate transformants in the exponential phase of growth were placed in copper and iron limiting medium without (ϪFe) or with 100 M iron (ϩFe). ␤-galactosidase activity was measured as described under "Experimental Procedures." The averages of three independent experiments are shown with standard deviations.

LexA fusion present
LexAop-GAL1-lacZ expression oxygen toxicity by analyzing their growth in a medium containing hydrogen peroxide (H 2 O 2 ). The breakdown of hydrogen peroxide by metals such as iron generates toxic hydroxyl radicals by the Fenton reaction (1). The aft1 mutant grew less well on YPD medium containing 1 mM H 2 O 2 than did the wild-type, aft2, and fet3 strains (Fig. 6A). The aft1aft2 mutant was unable to grow under the above conditions. The overproduction of Aft2p by the double aft1aft2 mutant allowed this strain to grow in the presence of H 2 O 2 . The growth of the aft1 strain was completely restored by supplementing the medium with either iron or copper, whereas the growth of the aft1aft2 strain was restored only by adding iron. These results indicate that the aft1, and particularly the aft1aft2, mutants are hypersensitive to oxidative stress. We confirmed these results by phenotype analysis for stress phenotypes related to metal redox dysregulation. The aft1aft2 mutant was hypersensitive to 10 -100 M copper in YPD medium (Fig. 6B). This strain was also unable to grow in minimal medium unless we added methionine or homocysteine. It also had a partial lysine auxotrophy ( Fig. 6C and data not shown). All these oxidative stress-related phenotypes were suppressed under anaerobic growth conditions and partially suppressed by adding iron to the aerobic growth medium.

DISCUSSION
AFT1 controls the transport of iron via the reductive and the siderophore-mediated high affinity iron uptake systems in response to iron deprivation conditions. We have now shown that the Aft1p-homologous protein, Aft2p, is a second mediator of iron-regulated control of transcription in yeast. The most similar domains of Aft1p and Aft2p are in the basic N-terminal regions of the proteins. This region has four conserved cysteine residues including the CDC sequence supposed to ligate iron in Aft1p (13). Aft2p was found to possess a transactivation activity that is lower relative to that of Aft1p (Table I). The transactivation activity of Aft2p was reduced by 50% when there was iron in the medium, whereas the transactivation activity of Aft1p was not affected. This confirms a previous report that indicates that the transcriptional activation function of Aft1p is independent of iron (31). In contrast, the transcriptional activation function of Aft2p may be iron-dependent.
Our data confirm previous studies that indicate that a residual iron regulated transcription of FET3 is present in an AFT1 deleted strain (31). We show that Aft2p may be responsible for the remaining transcription of FET3. We have also demonstrated that overproduction of Aft2p activates, in an iron-regulated manner, the transcription of both the FET3 gene and the LacZ reporter gene placed under the control of the AFT1responsive element. The modulation by iron of the AFT2-dependent regulation in aft1 strains seems to be greater than 2-fold, suggesting that iron could regulate both the transactivation and DNA binding functions of Aft2p. These results suggest that Aft2p is recruited to a sequence identical or similar to the T/GG/ACACCC AFT1-responsive element and may take part in the transcription of genes that overlap AFT1 target genes. Our present findings show that the amount of ATX1 mRNA is regulated by iron. This regulation appears to be lost in the double aft1aft2 mutant. Moreover, the amount of the ATX1 mRNA is increased with the overproduction of Aft2p. These results support a previous report that indicates that ATX1 mRNA is unaffected in strains containing an aft1⌬ null mutation but is increased in the hyperactive AFT1-up allele (20). It indicates that the Aft proteins in addition to other transcription factors may be involved in the iron-regulated transcription of ATX1.
We also compared the growth of the aft strains with those of the fet3 and wild-type strains under iron-deprived conditions to identify the role of Aft2p in iron metabolism. Previous studies FIG. 2. Iron-and AFT-dependent expression of FET3 and ATX1. The strains CM3260 (WT), Y18 (aft1), CM3260aft2⌬ (aft2), Y18aft2⌬ (aft1aft2), and Y18aft2⌬ harboring the plasmid pEG202-AFT2 (aft1aft2/2AFT2) were grown for 18 h in an iron and copper limiting medium with either 1 or 100 M iron and then diluted to an A 600 nm (O.D.) of 0.3 in the same medium without (ϪFe) or with 100 M iron (ϩFe) and grown further to an A 600 nm of 0.8. A, total RNA was prepared from the designated strains grown without (Ϫ) or with (ϩ) added iron. 10 g of total RNA from each culture was analyzed by Northern blot hybridization. Northern blot analysis was performed with sequential hybridization of the indicated probes. B, ␤-galactosidase activities were measured as described under "Experimental Procedures." Black boxes, no added iron; white boxes, 100 M iron. The average of three independent experiments is shown with standard deviations. have indicated that the aft1 mutant is more sensitive to iron deprivation than is the fet3 mutant despite their comparable measured ferrous iron uptakes. It has been suggested that Aft1p mediates some intracellular iron use in addition to its role in the regulation of iron transport (13). Our phenotypic analysis of growth on agar plates containing an iron-depleted medium indicates that the growth of the aft1 mutant is affected by iron starvation, whereas that of the isogenic fet3 mutant is not. The double aft1aft2 mutant is totally unable to grow under these conditions. We have also shown that aft1 and aft1aft2 cells both have normal iron contents similar to those of wildtype cells. Thus, the inability of aft mutants to grow in an iron-poor medium is not solely caused by defective iron uptake. It confirms the predicted role for AFT1 in the control of intracellular iron use and suggests that Aft2p is critical, with Aft1p, for the modulation of iron homeostasis in yeast.
A previous report indicated that the aft1 mutant is unable to grow in medium with glycerol as a respiratory carbon source (31). The authors attributed this to a defective high affinity iron uptake. This defect makes the intracellular iron content too low to sustain respiratory growth. Our data do not support this. First, the aft1 mutants all have functional mitochondrial respiration that is 2-fold higher than in the wild-type and aft2 strains whatever the iron concentration in the medium (Table  II). This result is consistent with the measurements of cytochrome absorption spectra and cytochrome c oxidase activities (data not shown) that indicate that the aft1 and aft1aft2 mutants have a normal capacity to synthesize cytochromes and a normal respiratory cytochrome c oxidase activity. Second, the aft1 mutant is unable to grow under aerobic conditions on agar plates containing raffinose as carbon source, whereas the isogenic fet3 mutant does grow. Third, this oxygen-dependent phenotype is not specific to an iron deficiency, because adding copper restores the growth of the aft1 mutant to the same extent as does iron (Fig. 5). But, adding iron or copper does not suppress the aft1-mediated mutant phenotypes in the absence of Aft2p. This suggests that elevated iron or copper concentration in the medium could induce some other pathway(s) to compensate for the deficiency of aft1 and that this putative metal-mediated compensatory pathway requires the presence of Aft2p. Thus, these results do not support the notion that the inability of aft1 mutants to grow on respiratory medium is due entirely to a defective high affinity iron uptake.
The aft1, and particularly the aft1aft2, mutants are hypersensitive to H 2 0 2 , and the aft1aft2 mutant has several oxygendependent and iron-dependent phenotypes such as a hypersensitivity to copper and a methionine auxotrophy. Mutants, the copper detoxification of which is affected such as the ace1 and cup1 mutants, are hypersensitive to copper under aerobic con-   5. Oxygen-dependent inhibition of the growth of aft1 and aft1aft2 mutants on raffinose. Strains CM3260 (WT), Y18 (aft1), CM3260aft2⌬ (aft2), Y18aft2⌬ (aft1aft2), Y18aft2⌬ harboring the plasmid pEG202-AFT2 (aft1aft2/2AFT2), and Y19 (fet3) were plated in serial dilutions on raffinose medium with or without added 100 M iron (ϩFe) or 100 M copper (ϩCu). The plates were placed in aerobic (ϩoxygen) or anaerobic (Ϫoxygen) conditions as described under "Experimental Procedures." ditions (32,39). Oxygen-dependent methionine auxotrophy has been reported for the Cu/Zn superoxide dismutase sod1 mutant and for the glucose-6-phosphate dehydrogenase zwf1 mutant, both of which are critical for maintaining the redox state of cells (33). We suggest therefore that the aft1aft2-mediated misuse of iron renders the cells hypersensitive to a variety of metal redox stresses. This agrees with recent data showing a relationship between superoxide stress mediated by a sod1 mutant and the dysregulation of iron homeostasis (34,35). The phenotypic similarities between the aft1aft2 and the sod1 mutants could occur, because the aft1aft2 mutant lacks Sod1p activity. However, the degree of SOD1 transcription and Sod1p activity in aft1aft2, aft1, aft2, and wild-type strains are not significantly different (data not shown). Thus, the aft1aft2mediated stress phenotypes are not caused by a lack of Cu/Zn superoxide dismutase per se.
Copper transport and copper detoxification is under the control of two related transcription activators, Mac1p and Ace1p, in yeast (36). The parallel between the metabolisms of iron and copper suggests that the two homologous proteins Aft1p and Aft2p may coordinate the transcription of the genes involved in the homeostasis of iron and protection against iron toxicity. In support of this, Aft1p was shown recently to interact in a two-hybrid assay with Grx3p, a glutaredoxin involved in the protecting of proteins from oxidative damage, and with Yap5p, a yeast AP-1 homolog (37). Exposure to the alkylating agent methyl methanesulfonate, which induces the transcription of many stress response or detoxification genes, also resulted in a 3.4-fold increase in the amount of AFT2 transcripts (38).
The phenotypic analysis presented herein indicates that the aft1-mediated mutant phenotypes are exacerbated by the deletion of AFT2 and are suppressed by the overproduction of Aft2p. Moreover, some phenotypes related to oxidative stress do not occur in the simple aft1 and aft2 mutants but are clearly present in the double aft1aft2 mutant. These results suggest that Aft2p has functions that overlap those of Aft1p in the regulation of iron metabolism pathways. We are presently developing a global analysis to identify the putatives AFT target genes that are involved in these pathways.
Aknowledgements-We thank Dr. Yuko Yamaguchi-Iwai for the generous gift of the plasmid pFC-W. The English text was edited by Aman Sci.