Regulation of Iron Homeostasis Mediated by the Heme-binding Protein Dap1 (Damage Resistance Protein 1) via the P450 Protein Erg11/Cyp51*

Fungal infections arise frequently in immunocompromised patients, and sterol synthesis is a primary pathway targeted by antifungal drugs. In particular, the P450 protein Erg11/Cyp51 catalyzes a critical step in ergosterol synthesis, and the azole class of antifungal drugs inhibits Erg11. Dap1 is a heme-binding protein related to cytochrome b5 that activates Erg11, so that cells lacking Dap1 accumulate the Erg11 substrate and are hypersensitive to Erg11 inhibitors. Heme binding by Dap1 is crucial for its function, and point mutants in its heme-binding domain render Dap1 inactive for sterol biosynthesis and DNA damage resistance. Like Dap1, the human homologue, PGRMC1/Hpr6, also regulates sterol synthesis and DNA damage resistance. In the present study, we demonstrate that the Dap1 heme-1 domain is required for growth under conditions of low iron availability. Loss of Dap1 is suppressed by elevated levels of Erg11 but not by increased heme biosynthesis. Dap1 localizes to punctate cytoplasmic structures that co-fractionate with endosomes, and Dap1 contributes to the integrity of the vacuole. The results suggest that Saccharomyces cerevisiae Dap1 stimulates a P450-catalyzed step in sterol synthesis via a distinct localization from its homologues in Schizosaccharomyces pombe and mammals and that this function regulates iron metabolism.

Fungal infections are important clinically because they contribute to the mortality of patients with human immunodeficiency virus/AIDS, cancer, and other diseases associated with immunosuppression. Mammalian hosts combat fungal infections via the immune system and by sequestering free iron in the bloodstream. Fungal infections can be suppressed with the azole group of antifungal drugs, a commercially important group of drugs that includes fluconazole, itraconazole, and miconazole. These drugs inhibit Erg11/Cyp51/lanosterol demethylase (1,2), which catalyzes a critical step in the synthesis of ergosterol, a key component of the fungal cell membrane. Erg11 is one of a large class of monooxygenases that are called P450 proteins due to the spectral absorbance of a cysteine-linked heme molecule in their active site (3).
In Saccharomyces cerevisiae, Erg11 is activated by Dap1 (damage resistance protein 1) (4), which is related to cytochrome b 5 (5), a heme-binding protein that activates P450 reactions (3,6). Cells lacking Dap1 partially arrest sterol synthesis at the stage catalyzed by Erg11 (4), and dap1⌬ cells are hypersensitive to the Erg11 inhibitors itraconazole and fluconazole (4). According to microarray databases, DAP1 expression is induced by azole antifungal drugs (7), but this has not been independently confirmed. Azole sensitivity in dap1⌬ cells is suppressed by overexpressing Erg11 (8), and dap1⌬ cells have ϳ4-fold lower levels of Erg11 than wild-type cells (8), although this regulation occurs primarily at the post-transcriptional level. Notably, the effect of Dap1 on sterol synthesis is conserved with its human homologue (9), called PGRMC1 (for progesterone receptor membrane component 1) or Hpr6 (for heme-1 domain protein/human progesterone receptor) (10).
Dap1 binds to heme (8,11), as does its homologues in Schizisaccharomyces pombe, mice, and humans (9,12,13). The heme-binding activity of Dap1 is critical for its function, and heme-binding defective mutants are inactive in sterol synthesis or damage resistance (8). Furthermore, the damage sensitivity and sterol synthesis phenotypes of dap1⌬ mutants can be suppressed by adding exogenous heme (8), suggesting a role for Dap1 in maintaining heme metabolism. One appealing model for Dap1 family proteins is that they utilize their heme-binding activity to directly activate P450 proteins. However, unlike cytochromes and related proteins, Dap1 homologues lack the histidine residues that coordinate heme binding in cytochrome b 5 (5), suggesting that Dap1 may participate in intracellular heme trafficking. In addition to regulating ergosterol synthesis, Dap1 is required for resistance to the alkylating agent methyl methanesulfonate, and dap1⌬ cells have decreased mitochondrial function (4).
Heme (iron protoporphyrin IX) is synthesized in an eightstep pathway that is subject to regulation at various steps (14). Two key steps in heme synthesis are catalyzed by Hem1/5-aminolevulinate synthase, a mitochondrial protein, and Hem2/␦aminolevulinate dehydratase/porphobilinogen synthase, which localizes to the cytoplasm and nucleus. Heme and ergosterol share the same upstream precursors (15), and the synthesis of heme and ergosterol are closely synchronized. The Hap1 transcription factor (16) is directly regulated by heme through a series of heme-regulated sequence motifs that control multi-protein complex formation and DNA binding (17). Heme is also required for the transcription of iron transport and sterol synthetic genes (18). Iron levels, in turn, regulate the transcription, post-transcriptional stability, and post-translational levels of numerous genes and gene products (19 -21). One of the ironregulated transcripts is DAP1, which is post-transcriptionally regulated by proteins that respond to low iron conditions (19).
Because of the close relationship between iron metabolism and the synthesis of heme and ergosterol, we have examined the role of Dap1 in these processes. We have found that Dap1 regulates growth under low iron conditions through a mechanism that requires its heme-1 domain, and Dap1-mediated growth on low iron is mediated by Erg11. We have also shown that Dap1 localizes to endosomes and regulates the structure of the vacuole, which is characteristic of other sterol biosynthetic proteins. The results represent a novel function for the Dap1 family proteins, which include homologues in mammals that are important for sterol synthesis and in cancer.
The strains RCY409-2a (DAP1) and RCY409-4b (dap1⌬:: LEU2) and the RCY409-2a and RCY409-4b derivatives harboring the control (pRS313, YEplac181, or YEplac195), DAP1 (pRC41), DAP1-D91G (pRC39), or ERG11 (pRH4) plasmids have been described previously (8). The pSCCC1 plasmid has also been described previously (24). The dap1⌬::HIS3 strain RCY456 was constructed by one-step transplacement of the entire DAP1 gene with a PCR product containing HIS3 and DAP1-flanking homology. Oligonucleotide sequences are available as supplemental Fig. 1. For Dap1 expression studies, the DAP1 open reading frame was fused in frame with three copies of the HA epitope tag sequence by one-step integration of the epitope tag fused to the KanMX reporter gene. The HA tag sequence was amplified by PCR with the primers DAP1-TAGF and DAP1-TAGR, using the plasmid pFA6a-3HA-KanMX6 (25) as a template. W303a cells were transformed with the resulting PCR product, and Geneticin-resistant isolates were tested for insertion of the epitope tag by PCR.
Plasmids-The plasmid pJM63, encoding HEM1, was prepared by amplifying the entire HEM1 gene using the primers HEM1-300F-Bam and HEM1ϩ1722R-Eco and subcloning the product into the plasmid YEplac195. Similarly, HEM2 was amplified with the primers HEM2-300F-Hind and HEM2ϩ1028R-Eco and subcloned into YEplac195, forming the plasmid pJM64. For NSG1 overexpression, the entire NSG1 gene was amplified using the primers NSG1-300F and NSG1ϩ873R-Myc and was initially subcloned into the PCR cloning plasmid pCR2.1, forming the plasmid pJM76. The NSG1 fragment was the liberated as a XhoI-BamHI fragment and subcloned into the SalI and BamHI sites of YEplac195, forming the plasmid pJM77.
FET3 Promoter Assays-The FET3-lacZ plasmid, consisting of the FET3 promoter fused to lacZ in the plasmid YEp354, was the kind gift of Dr. Jerry Kaplan and has been described previously (27). For lacZ measurement, cells were grown in synthetic medium, with or without 100 M bathophenanthroline disulfonic acid (BPS) for 3 h. The A 600 was measured, and the cells were centrifuged and lysed with 250 l of the Y-PER solution (Pierce), and the lysate was incubated with 700 l of Z buffer (60 mM Na 2 HPO 4 -7H 2 O, 40 mM NaH 2 PO 4 -H 2 O, 10 mM KCl, 1 mM MgSO 4 , and 50 mM ␤-mercaptoethanol) containing 1 mg/ml o-nitrophenyl ␤-D-galactopyranoside. The reaction was stopped with 500 l of 1 M Na 2 CO 3 , clarified by centrifugation, and measured at 420 nm, using the same concentrations of Y-PER, Z buffer, and Na 2 CO 3 as a blank. Miller units ((A 420 ϫ 1000)/(A 600 /min/ml)) were calculated for each point. FET3 transcription was assayed independently by reverse transcription-PCR as described previously (8) with the FET3 primers FET3ϩ100F (ACAGGAACGTTGATGGGCTA) and FET3ϩ380R (GAATGGTACCAGTAGGTGCC). Primers for the SCS2 transcript (8) served as a control for loading and were included in the same reaction as for FET3, and the products were separated in 2% agarose. Assays for ferric reductase activity were performed as described previously (28).
Protein Analysis-Log phase yeast cultures were lysed in Y-PER lysis solution (Pierce) containing 1 mM phenylmethylsulfonyl fluoride and 10 g/ml aprotinin and analyzed essentially as described (8). The antibody to HA (HA11) was from BabCo, and the antibody to ␣-tubulin was developed by J. Frankel and was obtained from the Developmental Studies Bank at the University of Iowa under the auspices of the NICHD, National Institutes of Health.

Dap1 Regulates Iron Metabolism
then centrifuged at 700 ϫ g for 10 min and resuspended in 5 ml of lysis buffer (0.8 M sorbitol, 10 mM MOPS, pH 7.2, and 1 mM EGTA) containing 1 mM phenylmethylsulfonyl fluoride. The cells were then lysed with three 10-s pulses from a Polytron PT1200 homogenizer, and the homogenate was centrifuged at 2500 ϫ g for 10 min to remove unlysed cells. One ml of the lysate was separated on a discontinuous 12-60% sucrose gradient by centrifugation at 100,000 ϫ g for 16 h. at 4°C. Fractions were collected and analyzed by Western blot using antibodies to the peroxisome marker Ypt7 (a kind gift from Dr. William Wickner), the mitochondrial marker Por1 (Molecular Probes), the lipid particle marker Erg6 (a kind gift from Dr. Gunther Daum), the endosomal marker Pep12 (Molecular Probes), the plasma membrane ATPase Pma1 (a kind gift from Dr. Ramon Serrano), and Erg11 (8).
Fluorescence-Staining was performed largely as described (29). Log phase diploid cells were fixed with 3.7% formaldehyde at 37°C for 30 min, centrifuged, resuspended in 1 M sorbitol containing 3.7% formaldehyde, and rotated at 4°C overnight. Cells were spheroplasted in 1 M sorbitol containing 10 g/ml zymolase and 70 mM ␤-mercaptoethanol at 30°C for 1 h, washed once in PBS, and applied to a poly-L-lysine-coated slide. Cells were then permeabilized with ice-cold methanol, blocked with phosphate-buffered saline containing 1 mg/ml bovine serum albumin, and stained with the HA11 monoclonal antibody (BabCo) for Dap1-HA and an anti-Myc tag antibody (Genscript) for Erg11-Myc, followed by fluorescein isothiocyanate-conjugated secondary antibodies. For FM4 -64 staining, cells were incubated with 20 M FM4 -64 (Molecular Probes) in YPD medium for 10 min, washed, and incubated for an additional 10 min in YPD before microscopic analysis. In all cases, cells were examined using a Zeiss microscope, and images were captured using Axioskop software.
Sterol Analysis-200 ml of early log phase cells were grown in YPD medium and treated with 100 M BPS for 3 h. Cells were pelleted and washed once with distilled water and extracted with potassium hydroxide-ethanol. Sterols were subsequently extracted with hexane, as described previously (4,8), and analyzed by gas chromatography at the University of Kentucky Mass Spectrometry Facility.

Strains Lacking Dap1
Are Sensitive to Iron Depletion-Wildtype and dap1⌬ strains were maintained on iron-depleted medium by culturing in 100 M BPS. Wild-type cells grew normally, whereas the dap1⌬ strains did not (Fig. 1A, columns 3 and 4). A small zone of residual growth in the dap1⌬ strain was dark red and consisted of nonbudded cells with a disrupted morphology. The BPS sensitivity of dap1⌬ cells was complemented by the wild-type DAP1 gene (Fig. 1B, row 7) but not by the DAP1-D91G mutant (Fig. 1B, row 8), which does not bind to heme (8). The results suggest that the heme binding function of Dap1 is required for growth under iron-depleted conditions.
The expression of the multicopper oxidase Fet3 is induced under iron-depleted conditions (30). A construct containing the FET3 promoter fused to the bacterial lacZ gene was used to measure FET3 expression. FET3 increased 22-fold in wild-type cells and 31-fold in dap1⌬ cells following iron depletion (Fig.   1C), a difference that was statistically significant (p ϭ 5 ϫ 10 Ϫ6 , two-tailed t test). The difference in FET3 transcription in BPStreated wild-type and dap1⌬ cells was confirmed using using reverse transcription-PCR, where we reproducibly detected a 25% increase in FET3 levels (Fig. 1D). Despite increased FET3 transcription, we did not detect any change in the uptake of radiolabeled iron or iron reductase activity (50.5 Ϯ 2.8 nmol of Fe 2ϩ /min/A 600 for wild type versus 52.6 Ϯ 5.0 for dap1⌬). The results suggest a disparity between iron-regulated transcription and iron uptake in dap1⌬ cells.
Because Dap1 regulates sterol synthesis, we tested the effect of the Dap1 mutation on sterol biosynthesis under low iron conditions. As expected, dap1⌬ cells had elevated lanosterol and episterol relative to wild-type cells (Fig. 2, compare A and C). Treatment of wild-type cells with BPS caused a modest accumulation of episterol (Fig. 2B), whereas dap1⌬ cells accumulated increased levels of squalene (Fig. 2D), suggesting an iron-related defect in Erg1 function in dap1⌬ cells. The sterol synthetic pathway is diagrammed in Fig. 2E. Although the Erg1 substrate was elevated under iron-depleted conditions, multicopy expression of Erg1 did not suppress BPS sensitivity in dap1⌬ cells (Fig. 2F, bottom, compare bottom rows), indicating that altered Erg1 function is not directly related to iron metabolic defects in dap1⌬ cells.
Erg11 Suppresses the Requirement for Dap1 in Iron Metabolism-Because Dap1 requires heme binding for sterol synthesis, damage resistance, and viability in low iron, we devel-FIGURE 1. Dap1 is required for growth under iron-depleted conditions. A, wild-type (RCY409-2a) or dap1⌬ (RCY409-4b) cells were serially diluted 1:10 and tested for viability on normal growth medium (rows 1 and 2) or the same medium containing 100 M BPS, an iron-chelating agent. The dap1⌬ strain grew poorly, and the residual colonies exhibited a dark red color. B, wild-type cells harboring a control vector (rows 1 and 5) or dap1⌬ cells harboring the control plasmid pRS313 (rows 2 and 6), the DAP1 expression plasmid pRC41 (rows 3 and 7), or the DAP1-D91G expression plasmid pRC39 (rows 4 and 8) were tested for growth on control medium (rows 1-4) or medium containing BPS (rows 5-8). DAP1-D91G encodes a point mutant of Dap1 that is incapable of binding to heme. The wild-type DAP1 gene complemented the dap1⌬ mutation, whereas the DAP1-D91G mutant did not, demonstrating a requirement for heme binding for growth on BPS. C, Dap1 suppresses the expression of iron-regulated genes. Wild-type or dap1⌬ cells harboring a FET3-lacZ construct were untreated (light gray columns) or were treated with 100 M BPS (dark gray columns), and ␤-galactosidase activity was determined in triplicate. Error bars, the S.D. between individual measurements. D, FET3 transcripts analyzed by reverse transcription-PCR in which wild-type cells (lanes 2 and 3) or dap1⌬ cells (lanes 4 and 5) were untreated (lanes 2 and 4) or treated with 100 M BPS (lanes 3 and 5) for 3 h. Primers directed to the SCS2 transcript, which is not regulated by iron, were included as an internal control for cDNA loading (lower band, labeled CON for "control"), and lane 1 is a negative control in which the RNA template from the sample in lane 2 was amplified without added reverse transcriptase. The numbers below each lane indicate the FET3/control ratio relative to the untreated wild-type cells in lane 2. For each assay, the results shown represent experiments that were performed at least in triplicate. DECEMBER 14, 2007 • VOLUME 282 • NUMBER 50 oped a genetic system for increasing heme levels. The HEM1/ ␦-aminolevulinate synthase and HEM2/␦-aminolevulinate dehydratase genes were expressed at high copy numbers using the multiple copy plasmid YEplac195, and this suppressed the elevated susceptibility of dap1⌬ cells to the alkylating agent methyl methanesulfonate (MMS) (Fig. 3A). The effect was more pronounced for HEM2 than HEM1 (Fig. 3A, rows 11 and  12). However, HEM1 and HEM2 high copy expression did not suppress itraconazole sensitivity, suggesting that the expression of these genes is not limiting for sterol synthesis (Fig. 3A,  rows 17 and 18). Nsg1 is the yeast homologue of the human Insig-1 protein, which binds to the human Dap1 homologue (31). However, a high dosage of NSG1 did not suppress MMS, itraconazole, or BPS susceptibility in dap1⌬ cells (data not shown).

Dap1 Regulates Iron Metabolism
Although the heme-binding domain of Dap1 was required for MMS resistance, high dose expression of HEM2 did not suppress BPS sensitivity in dap1⌬ cells (Fig. 3B, row 11). In contrast, high dosage of ERG11 completely suppressed BPS sensitivity (Fig. 3B, row 12). This effect was not generalized to other sterol biosynthetic genes, because high dosage of ERG1 and ERG5 had no effect on BPS sensitivity in dap1⌬ strains (data not shown). Dap1 binds to heme and is related to cytochrome b 5 , encoded by CYB5 in S. cerevisiae, but high copy expression of CYB5 did not suppress loss of Dap1 (data not shown).
Because ERG11 suppressed loss of DAP1, we tested the extent to which Erg11 inhibitors affected growth on iron-depleted medium. We detected partial growth of wild-type cells on 2 M itraconazole and 100 M BPS (Fig. 3C, rows 2 and 3, respectively) but essentially no growth when the two compounds were combined (Fig. 3C, row 4). Taken together, the results suggest that sterol synthesis mediated by Erg11 and Dap1 is required for survival in iron-depleted medium, and this Dap1 function is not a general property of cytochrome-like proteins.
Characterization of Dap1-We fused three copies of the HA epitope tag sequence through one-step integration to the 3Ј end of the DAP1 gene. As a result, DAP1 was transcribed from its own promoter and synthesized from a single copy of its gene. The tagged Dap1 protein was readily detected as a 25-kDa protein by Western blot (Fig. 4A, top, lanes 2 and 3). In contrast, strains lacking the in frame epitope tag did not have a detectable 27-kDa protein (Fig. 4A, top, lane 1). Thus, the migration of Dap1 was similar to that of related rat and human proteins. In all cases, blots were probed with an antibody to tubulin as a control for protein loading (Fig. 4A, bottom).
Dap1 expression was reported previously to change upon treatment with the antifungal triazole drug itraconazole (7), and dap1⌬ mutants are sensitive to itraconazole (4,8). The expression of Dap1 increased 14-fold in cells treated with 0.1-1 M itraconazole in a dose-dependent manner (Fig. 4B). In contrast, Dap1 expression did not change significantly following treatment with MMS, heat shock, or hydroxyurea (supplemental Fig. 2A). In addition, we did not detect any changes in Dap1 expression following treatment with BPS (supplemental Fig.  2B) or in strains with elongated or shortened telomeres (tel1⌬ or rif1⌬ rif2⌬, respectively; supplemental Fig. 2C) or in strains with acute damage sensitivity (mec1-21, rad9⌬, or dun1⌬).
Dap1 Localizes to Punctate Cytoplasmic Sites and Co-fractionates with Endosomal Markers-A diploid strain expressing the tagged Dap1 protein (Fig. 4A, lane 4), which was not expressed in the control untagged diploid strain (Fig. 4A, top, lane 5) was stained by immunofluorescence with the HA antibody. Dap1 localized to bright, clearly defined spots within the cytoplasm ( Fig. 4C; bright field in Fig. 4D). The spots varied in size and number between the cells and were detectable in both mother and daughter cells. There was no evidence for Dap1 staining within the mother-bud neck, within the nucleus, or at the cell periphery. In contrast, the control diploid strain did not stain with the HA antibody after the same procedure ( Fig. 4E; bright field shown in Fig. 4F). Similar spots of Dap1 staining were detected in haploid strains but were more difficult to visualize due to the smaller cell size.
Punctate cytoplasmic staining is characteristic of several types of subcellular structures. Thus, the subcellular fractions of a Dap1-tagged strain were separated on a sucrose gradient  (4), and iron depletion led to a further accumulation of squalene. E, the sterol biosynthetic pathway is diagrammed, showing the positions of the sterol intermediates that are highlighted in A-D, as well as the relevant activators and inhibitors. F, wild-type and dap1⌬ cells containing a control multicopy plasmid or a multicopy plasmid harboring ERG1 were tested for growth on control medium (top) or medium containing BPS (bottom). High copy expression of ERG1 did not suppress iron sensitivity in dap1⌬ cells.

Dap1 Regulates Iron Metabolism
and probed for Dap1 or various markers. Dap1 co-fractionated with Ypt7 (Fig. 5, A and B), a GTP-binding protein that regulates vacuolar transport (32), indicating that Dap1 co-localizes with endosomal proteins. Dap1 also partially co-fractionated with Pep12, a Q A -type SNARE/syntaxin (33) that is a marker for late endosomes (Fig. 5C). In contrast, Dap1 did not co-fractionate with Erg6/sterol C-24 methyltransferase, a marker for lipid particles (Fig. 5D) (34,35) or with Por1/porin, a voltage-gated anion channel and marker for mitochondria (Fig. 5E). The finding that Dap1 did not co-fractionate with Erg6 is consistent with an earlier mass spectrometric analysis of abundant proteins in purified lipid particles, which did not identify Dap1 (34,36).
Mass proteomic screens for interacting proteins have identified the plasma membrane ATPase Pma1 as a putative Dap1binding protein (37). However, our gradient detected Pma1 in a distinct portion of the gradient from Dap1 (Fig. 5F). Dap1 interacts genetically with Erg11 and stabilizes Erg11 expression (8). Erg11 localization was diffuse throughout the sucrose gradient (Fig. 5G), and Erg11 staining by immunofluorescence was granular and excluded from the vacuole (Fig. 5H). Previous studies with GFP-tagged Erg11 derivatives localized Erg11 to the endoplasmic reticulum (38), cytoplasm (35), or granular sites (39).
Ypt7 is related to the mammalian Rab7 protein and regulates the movement of proteins from the prevacuolar compartment to the vacuole (32). We tested the role of Dap1 in vacuolar structure by incubating wild-type and dap1⌬ cells with the fluorescent dye FM4-64, which is internalized and stains the yeast vacuole (40). In wild-type cells, vacuoles stained as circular structures (Fig. 6C; the bright field image is shown in Fig. 6A), whereas in dap1⌬ cells, a significant fraction of vacuoles consisted of multiple smaller structures that resembled unfused vesicles (Fig. 6D), and others contained irregular edges or a central septum. The color version of this figure is supplemental Fig. 3. The percentage of cells with aberrant vacuoles was elevated in dap1⌬ cells (23 Ϯ 5%) compared with wild type (7 Ϯ 3%) to a significant extent (p ϭ 0.006, two-tailed t test).
Defective vacuolar structure could inhibit vacuolar iron storage. In wild-type cells harboring a CCC1 (cross-complements Ca 2ϩ phenotype of csg1 (41); csg is an abbreviation for calcium-sensitive growth) high copy plasmid, FET3-lacZ reporter activity was highly induced (Fig. 6E), most likely due to iron being sequestered in the vacuole. In dap1⌬ cells harboring the same CCC1 plasmid, FET3-lacZ activity was diminished by 40% (Fig. 6E), a significant difference (p ϭ 0.011, two-tailed t test). Decreased FET3 transcription in CCC1-expressing dap1⌬ cells was almost identical when FET3 transcripts were assayed using an independent PCR technique (Fig. 6F, compare lanes 3 and 5). We detected a 38% decrease in FET3 transcripts in dap1⌬ cells expressing CCC1 compared with comparable wild-type cells when calculated from the FET3/control transcript ratios (shown at the bottom of Fig. 6F). In triplicate analyses, the difference in FET3 levels between CCC1-overexpressing wild-type and dap1⌬ cells was statistically significant (p ϭ 0.02, two-tailed t test). This finding is consistent with a diminished ability of CCC1 to direct iron storage in dap1⌬ cells. We note that the CCC1 high copy plasmid did not affect growth of dap1⌬ cells under low iron conditions, consistent with an additional defect in iron uptake.

DISCUSSION
Iron utilization in fungi is important clinically because the availability of free iron limits the growth of pathogenic yeast. In dap1⌬ cells depleted of iron, ERG11 activity is inhibited, causing toxic levels of lanosterol to accumulate. This effect is reversed by ERG11 overexpression but not by elevated heme. It is unclear why iron depletion is toxic to cells containing elevated levels of lanosterol. One potential explanation is that the pathway that detoxifies lanosterol requires iron. Thus, deletion of DAP1 increases lanosterol levels, and iron depletion increases its toxicity (diagrammed in Fig. 7). This model is supported by our finding that BPS increases susceptibility to azole drugs and by similar findings in C. albicans (42).
In cells lacking Dap1, iron deprivation triggers increased induction of FET3 expression, but increased FET3 levels do not lead to increased iron uptake or iron reductase activity, suggesting that the sterol imbalance in dap1⌬ cells interferes with iron uptake. The model that sterol synthesis promotes iron metabolism is consistent with the findings of Li and Kaplan, in which . MMS resistance is maintained by a Dap1-mediated pathway that is distinct from resistance to azole antifungal drugs and BPS. A, wild-type and dap1⌬ cells harboring the control vector YEplac195 (VEC), a multicopy expression plasmid encoding HEM1 (pJM63), or a multicopy expression plasmid encoding HEM2 (pJM64) were tested for growth on synthetic dextrose (SD) medium lacking uracil (rows 1-6) or the same medium containing 0.015% MMS (rows 7-12) or 10 M itraconazole (rows [13][14][15][16][17][18]. High copy expression of HEM1 and HEM2 suppressed the susceptibility of dap1⌬ cells to MMS but not itraconazole. B, wild-type (rows 1-3 and 7-9) and dap1⌬ (rows 4 -6 and 10 -12) cells harboring the control vector YEplac195 or multicopy plasmids encoding HEM2 (pJM64) or ERG11 (pRH4) were tested for growth on synthetic dextrose medium lacking uracil (rows 1-6) or the same medium containing 100 g/ml BPS (rows 7-12). High level Erg11 expression suppressed loss of Dap1, whereas Hem2 did not. C, an azole Erg11 inhibitor increases susceptibility to BPS. Wild-type RCY409-2a were serially diluted and tested for growth on YPD plates containing 2 M itraconazole (row 2), 100 M BPS (row 3), or the two compounds combined (row 4). DECEMBER 14, 2007 • VOLUME 282 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 36547 a mutation in ERG25/sterol C-5 methyloxidase causes sensitivity to low iron (43). However, it is unlikely that dap1⌬ cells are sensitive to low iron solely because of low levels of ergosterol. Because conditions that have a modest effect on ergosterol levels (i.e. overexpression of ERG11) can suppress low iron sensitivity in dap1⌬ mutants, it is likely that the balance of sterol intermediates affects iron metabolism rather than ergosterol levels per se.

Dap1 Regulates Iron Metabolism
Deletion of DAP1 causes accumulation of the sterol intermediates lanosterol (the Erg11 substrate) and episterol (the Erg3 substrate) and of the sterol ergosta-5,7-dienol (4). Lanosterol accumulation can be suppressed by overexpressing ERG11, which restores azole resistance to dap1⌬ mutants and triggers a marked accumulation of episterol and ergosta-5,7-dienol (8). Episterol presumably accumulates because Dap1 is required for the activity of Erg3/sterol C-5 desaturase (Fig. 7) (44), which is a highly conserved protein that is activated by cytochrome b 5 (6,45), with which Dap1 shares structural homology and heme binding activity. In addition, Erg3 is also proposed to bind to iron through a series of coordinated histidine residues (43, 46), Dap1 was detected by Western blot in haploid and diploid strains. Lanes 1 contained the untagged haploid strain RCY408-1d, whereas lanes 2 and 3 contained the epitope-tagged strains RCY408-1b and RCY408-5b, respectively. Two untagged strains were mated to yield the strain RCY410 (lane 5), whereas the epitope-tagged strains in lanes 2 and 3 were mated to yield the diploid strain RCY411 (lane 4). HA-tagged Dap1 was detected as a 27 kDa band only in the strains containing the epitope-tagged DAP1 alleles. A parallel Western blot of the same samples was probed with an antibody to tubulin as a loading control (bottom). B, Dap1 is induced by itraconazole. No Dap1 was detected in the untagged strain RCY408-1d (lane 1). The Dap1-HA-expressing strain RCY408-1b was untreated (untr) (lane 2) or treated with 0.1, 0.5, or 1 M itraconazole for 90 min, and extracts were analyzed by Western blot for HA (top) or tubulin (bottom). Dap1 expression increased 14-fold relative to tubulin following itraconazole treatment. C-F, Dap1 localizes to punctate sites in the cytoplasm. Diploid cells expressing Dap1-HA (C and D) or untagged Dap1 (E and F) were stained with an anti-HA antibody and detected with a fluorescein isothiocyanate-labeled secondary antibody. Cells were visualized by fluorescence (C and E) or bright field microscopy (D and F). C, Dap1 localized to bright cytoplasmic spots. E, in the RCY410 strain, containing the native, untagged DAP1 gene, punctate cytoplasmic staining was not detected. The strains RCY410 and RCY411 differed only by the HA epitope tag cassette adjacent to DAP1 and were stained at the same time using the same conditions. The punctate staining pattern observed for RCY411 was detected in at least five separate experiments. FIGURE 5. Dap1 does not co-fractionate with mitochondrial or lipid particle markers. RCY411 cells were lysed and separated on a 0 -60% discontinuous sucrose gradient, and fractions from the gradient were collected and analyzed by Western blot. A, HA-tagged Dap1 was detected in fraction 2, which overlapped with the late endosomal marker Ypt7 (B) and partially overlapped with the late endosomal marker Pep12 (C). The lipid particle protein Erg6 was detected in fractions 7-9 (D), whereas the mitochondrial marker Por1 (E) was detected in fractions 7 and 8. F, the putative Dap1-interacting protein Pma1 localized to fraction 6. G, the Erg11 protein, a target of Dap1, localized in multiple subcellular fractions and overlapped partially with Dap1. H, in immunofluorescence studies, Erg11 was fused to a single Myc epitope tag sequence and expressed from the plasmid pRH7 in the diploid strain RCY411 and showed a granular staining pattern.

Dap1 Regulates Iron Metabolism
and its substrate accumulates slightly under conditions of low iron (Fig. 2). It is intriguing to speculate that altered iron metabolism in dap1⌬ cells may inhibit Erg3. One potential clue to the activation of Erg3 by Dap1 is that treatment of dap1⌬ cells with heme suppresses lanosterol accumulation (i.e. activates Erg11) but has little effect on episterol accumulation (8), suggesting a nonheme role for activating Erg3.
Under low iron conditions, dap1⌬ cells accumulate squalene, suggesting a defect in Erg1/squalene epoxidase. Erg1 is an oxygen-requiring enzyme that is inhibited by the allylamine class of antifungal drugs (1). Although squalene was one of the primary sterols to accumulate in iron-deprived dap1⌬ cells, overexpression of ERG1 did not restore viability to dap1⌬ cells under low iron conditions, suggesting that defective Erg1 activity is not solely responsible for their defective growth.
The common target of Dap1 and its homologues is P450 protein activation, and this activity is shared among the S. cerevisiae, S. pombe (9), and rodent (12) Dap1 homologues. Mammalian Dap1 homologues are highly expressed in the liver, kidney, and adrenal, which are sites of P450 activity (10,47,48), and mammalian Dap1 homologues localize to microsomal fractions, including the endoplasmic reticulum (49,50), which are also common sites for P450 proteins. The exception to this rule is in neuronal cells, where some mammalian Dap1 homologues also localize to the cell membrane (51,52). Both the S. pombe Dap1 and human PGRMC1 bind directly to Cyp51 and other P450 proteins (9). In contrast, we have not detected a stable interaction between Dap1 and Erg11 in S. cerevisiae, and our analysis suggests only a partial overlap in their localization. Furthermore, Dap1-Erg11 complexes have not been detected using genome-wide proteomic screens or two-hybrid analyses. S. cerevisiae Dap1 differs from its mammalian and S. pombe homologues in that it lacks a membrane-spanning sequence, and it is possible that this sequence is required for a stable interaction with Erg11. Based on the model proposed above, Dap1 probably regulates vacuolar structure via sterol synthesis. Ergosterol is nec-  D). The inverted images of C and D are shown for improved contrast. In wild-type cells, vacuoles stained as round shapes, whereas a significant number of dap1⌬ cells exhibited irregular vacuoles, including multiple smaller vacuoles. E, wild-type (wt) (RCY409-2a) or dap1⌬ (RCY456) cells harboring a control plasmid, YEplac181 (VEC.), or the CCC1 high copy plasmid pSCCC1 (CCC1) and a FET3-lacZ reporter construct were tested for ␤-galactosidase activity. High copy CCC1 expression increased ␤-galactosidase activity in wild-type cells, with a 40% decrease in CCC1-directed activity in dap1⌬ cells. Error bars, S.D. between individual measurements. F, FET3 transcription was assayed by reverse transcription-PCR in wild-type (RCY409-2a) or dap1⌬ (RCY456) cells harboring a control plasmid, YEplac181, or the CCC1 high copy plasmid pSCCC1. Primers for the SCS2 transcript were included in the same reaction (the product is the lower band) and served as a loading control (marked CON). The changes in the FET3/CON ratio relative to lane 2 are shown below the gel. Lane 1, a control for lane 3, in which reverse transcriptase was omitted from the cDNA synthesis reaction. All of the assays were performed at least in triplicate. FIGURE 7. A model in which Dap1 regulates heme and ergosterol synthesis and iron metabolism. Iron is imported into the cell and is incorporated into heme, among other molecules. Dap1 binds to heme from its endosomal locale and activates Erg11 and Erg3, promoting sterol synthesis. Erg11 catalyzes the demethylation of lanosterol, an essential step in ergosterol synthesis, and some of the downstream metabolites of Erg11 are also detected in endosomes. Erg1 is also activated by Dap1 under low iron conditions. Our results do not exclude the possibility that iron plays a role in detoxifying sterol intermediates that accumulate in cells lacking wild-type Erg11 function (dashed line, left). There are a number of regulatory interactions between heme, the sterol pathway, and the uptake of iron, and these are not indicated. essary for vacuole formation (53)(54)(55)(56), and yeast microsomes contain a significant fraction of the sterols fecosterol, episterol, and ergosta-5,7,9(11),22-tetraenol (57)(58)(59). Our results suggest a model in which the apparent vacuolar defects in dap1⌬ cells limit the extent to which CCC1 sequesters iron in the vacuole. CCC1 is postulated to act as an ion transporter, and CCC1 overexpression triggers increased cellular iron uptake (24) and vacuolar iron storage (60). These functions are partially dependent on Dap1, although the mechanism linking these proteins is unknown.
The localization of Dap1 to intracellular vesicles and the broad distribution of Erg11 suggest that some steps in P450 activation and sterol synthesis are distributed throughout the cell. Indeed, green fluorescent protein-tagged forms of Erg2, Erg3, and Erg25 all localize to a vesicle fraction (35). Furthermore, Dap1 has been previously implicated in endocytosis (4), and the localization of Dap1 to an endosomal fraction suggests a role in intracellular transport. Although most sterol-defective mutants are hyperpermeable to dyes (61), dap1⌬ mutants have defective dye uptake, particularly at low temperatures (4), suggesting that Dap1 is required for the endocytosis of some compounds (4). In addition, dap1⌬ cells do not display abnormalities in membrane integrity (4) that are typical of erg mutants (61). The model in Fig. 7 is based on the assumption that the primary role of Dap1 is in sterol synthesis, although it is possible that Dap1 contributes to endosome function or vacuole structure through other, uncharacterized mechanisms.
Although heme is required for sterol synthesis, our results suggest that heme levels are not limiting for azole drug resistance but are limiting for MMS resistance. In contrast, we found that iron levels are limiting for azole drug resistance. Although the presumed target for MMS is DNA, Lum et al. (62) previously identified HEM1 and HEM2 in a screen for MMS targets using a genome-wide pool of heterozygous deletion strains. One interpretation of these findings is that MMS targets Hem1 and/or Hem2 directly. However, MMS sensitivity in HEM1/ hem1 and HEM2/hem2 strains could also arise from diminished heme levels due to haploinsufficiency of 5-aminolevulinate synthase and ␦-aminolevulinate dehydratase, respectively. The results are important, because MMS closely resembles chemotherapeutic drugs that are used clinically, and our results suggest a role for heme biosynthesis in susceptibility to chemotherapeutic drugs. Indeed, the human Dap1 homologue Hpr6/ PGRMC1 is overexpressed in clinical tumors and cancer cell lines (50), and inhibition of Hpr6/PGRMC1 causes increased susceptibility to chemotherapeutic drugs (13).
Iron deficiency is the most prevalent nutritional disorder in the world, affecting almost two billion people (63), and abnormal iron metabolism contributes to a number of diseases, including Friedrich's ataxia, hemochromatosis, aging, and microbial infections (64). Dap1 has homologues in pathogenic fungi, including Candida species, and our results suggest that Dap1 homologues may have a role in iron metabolism in those organisms. If so, Dap1 homologues may be useful as targets for therapeutics, particularly in combination with other antifungal agents or under conditions leading to deprivation of available iron. There are also Dap1 homologues in higher eukaryotic species, including mam-mals (10,47,65), and the rodent and human Dap1 homologues bind to heme (12). However, the role of these proteins in iron metabolism has not been examined.