Proc. Natl. Acad. Sci. USAVol. 89, pp. 3869-3873, May 1992Biochemistry

Ferric reductase of Saccharomyces cerevisiae: Molecularcharacterization, role in iron uptake, andtranscriptional control by ironANDREW DANCIS*, DRAGOS G. ROMAN*, GREGORY J. ANDERSON*, ALAN G. HINNEBUSCHt,AND RICHARD D. KLAUSNER**Cell Biology and Metabolism Branch, and tSection on Molecular Genetics of Lower Eukaryotes, Laboratory of Molecular Genetics, National Institute ofChild Health and Human Development, Bethesda, MD 20892

Communicated by Helmut Beinert, January 3, 1992 (received for review October 18, 1991)

ABSTRACT The principal iron uptake system of Saccha-romyces cerevisiae utilizes a reductase activity that acts on ferriciron chelates external to the cell. The FREI gene product isrequired for this activity. The deduced amino acid sequence ofthe FRE1 protein exhibits hydrophobic regions compatiblewith transmembrane domains and has significant similarity tothe sequence of the plasma membrane cytochrome b558 (theX-CGD protein), a critical component of a human phagocyteoxidoreductase, suggesting that FREl is a structural compo-nent of the yeast ferric reductase. FREI mRNA levels arerepressed by iron. Fusion of 977 base pairs of FREI DNAupstream from the translation start site of an Escherichia coilIacZ reporter gene confers iron-dependent regulation on ex-pression of fi-galactosidase in yeast. An 85-base-pair segmentof FREI 5' noncoding sequence contains a RAPi binding siteand a repeated sequence, TTTTTGCTCAYC; this segment issufficient to confer iron-repressible transcriptional activity onheterologous downstream promoter elements.

Iron is an essential nutrient that is required by many enzymesfor cellular functions such as DNA synthesis and respiration.Ferric iron forms insoluble ferric hydroxide complexes in thepresence of oxygen and water, making iron availability abiological problem under these conditions. Ferrous iron ismuch more soluble but is rapidly converted to ferric iron inthe presence of oxygen (1). Cells have developed two majorsolutions to the problem ofassimilation offerric iron from theenvironment. In Escherichia coli, specific ferric bindingcompounds termed siderophores are secreted that bind andsolubilize environmental iron and deliver it to specific recep-tors for transport across the outer and inner membranes (2).In the yeast Saccharomyces cerevisiae, the major iron uptakesystem under aerobic conditions depends on a transplasmamembrane electron transport system (3, 4) that reduces ferriciron external to the cell. Multicellular eukaryotes may alsouse an externally directed ferric reductase prior to transmem-brane movement of ferrous iron (5-8). Our previous resultsindicated that the FREI gene product in S. cerevisiae isrequired for external ferric reduction and for ferric ironutilization, thus linking these two processes (9). However, amutation in FREI did not affect the uptake offerrous iron (9),suggesting that the ferrous transport system is an indepen-dent component of the iron utilization apparatus in yeast.Many organisms control their rate of iron assimilation by

ascertaining their requirement for iron and expressing alimiting component of the uptake system accordingly (10).Our previous results suggest that during aerobic growth of S.cerevisiae, changes in iron availability are reflected in alter-ations in the levels ofFREI mRNA and of ferric iron uptake

(9). In this paper we present the structure of the FREI geneand identify DNA sequences that mediate its regulation byironA

MATERIALS AND METHODSYeast Strains. The following yeast strains were derived

from the wild-type strain F113 (MATa, can), inol-13, ura3-52): W103 (MATa, can], inol-13, ura3-52,frel-1) was derivedby chemical mutagenesis of F113 and selection for a reduc-tase-negative phenotype (9); W126 was derived by transfor-mation of F113 with the plasmid pWDC20, which carries theFREI gene on the high-copy-number vector YEp24 (11). Weused one-step gene disruption (12) to create strain N1 (MATa,can), inol-13, Afrel:: URA3), in which an 800-base-pair (bp)Xho I fragment internal to the FREI coding region wasreplaced with a URA3 marker gene, and strain N2 (MATa,can, inol-13, Afrel:: URA3), in which a 2.7-kilobase (kb) ClaI fragment containing the entire FREI coding region wassimilarly replaced. Disruption/deletion strains of the FREIlocus were also constructed in the wild-type strain H1085,yielding strain W218 (MATa, leu2-3,112, Afrel:: URA3) andstrain W258 (MATa, leu2-3,112, Afrel:: URA3) by replace-ment of the FREI Xho I or Cla I fragment, respectively. Thestructure of the Afrel:: URA3 alleles was confirmed by PCRamplification of genomic sequences (13). The FREI disrup-tion/deletion strains were similar in their ferric reductaseactivity and sensitivity to iron deprivation.

Plasmids. Plasmid pWDC6 conferred reductase activity onthe W103 mutant strain. It contains the FREI gene andflanking sequences on a low-copy-number URA3 vector (9).Plasmid pCL4.3 contains the FREI gene on a low-copy-number LEU2 vector. FREI sequences were amplified byPCR with the boundaries specified in Fig. 3A (13) andinserted in-frame 5' of an E. coli lacZ reporter gene on thelow-copy-number URA3 plasmid pRS416 (14), yieldingpJ105, pJ111, pJ106, and pJ107. The FREI sequence -977 to+9 was similarly amplified and inserted in-frame at the 5' endof lacZ on the high-copy-number plasmid YEp24 (11), yield-ing pEG2.The low-copy-number URA3 plasmid p1087, generously

provided by C. Moehle (15), contains 250 nucleotides ofCYC) 5' noncoding sequences and the ATG start codon fusedto the E. coli 1acZ gene (16). FREI sequences with theboundaries specified in Fig. 3B were inserted upstream of theCYC) sequence between Bgl II and Xho I restriction sites,creating plasmids pGC7-8, pGC7-8.0, pGC7-8.5, pGC7-9,pGC8-9, pGC6-9, pGCR, and pGCRM. The point mutation in

Abbreviations: ORF, open reading frame; UAS, upstream activatingsequence.*The sequence reported in this paper has been deposited in theGenBank data base (accession no. M86908).

3869

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

3870 Biochemistry: Dancis et al.

plasmid pGCRM is a C -- A transversion in FREI at position-307 relative to the ATG.Growth Conditions. Yeast cultures were grown in standard

media (SD or YPD) or iron-depleted minimal defined medium(MD) according to published methods (9, 17). For analyses offerric reductase activity, ferric and ferrous uptake, andf3-galactosidase activity, three individual colonies trans-formed with each construct were picked from an SD plate andgrown to saturation in YPD medium. These cultures werethen diluted 1:200 into MD lacking uracil with or without 1mM added ferric chloride and grown for an additional 12-16hr. For the analysis of iron-dependent growth, three trans-formed yeast colonies were initially grown to stationaryphase in MD and then diluted 1:1000 into fresh MD medium.

Assays. Ferric reductase activity and uptake of ferric andferrous iron were measured as described (9). The concentra-tion of iron was 20 ,uM for the measurement of ferric uptakeand 1 AM for the measurement of ferrous uptake. fB-Galac-tosidase was measured from washed cells that were perme-abilized with 0.1% SDS and chloroform (18, 19).

Proc. Natl. Acad. Sci. USA 89 (1992)

Genetic and Microbiological Methods. Methods for DNAmanipulation were as described (20). The DNA sequence ofpWDC6 and its subclones was determined (21) with a Se-quenase kit (United States Biochemical). Primer extension ofradiolabeled oligonucleotide primers corresponding to thecomplement of FREI sequences from +8 to -12 and from+60 to -33 was performed by a published protocol (22).Oligonucleotides were synthesized on an Applied Biosys-tems model 381A DNA synthesizer.

RESULTSLocalization of FREI. The FREI gene was cloned by

complementation of the ferric reductase deficiency of afrel-lmutant strain (W103), using a low-copy-number yeast ge-nomic library (9). A 2.7-kb Cla I fragment of plasmid pWDC6(bp -575 to 2139, Fig. 1) exhibited frel-J complementingactivity, while deletion to the EcoRI site at the 5' end (bp 192,Fig. 1) or the BstEII site at the 3' end (bp 1658, Fig. 1)eliminated this activity (data not shown). Sequence analysis

attcggtcoatttccct -961

-841

-721

-601

-481

ctcttattggtggcttaggttgatagt tcacaataaatttaggaacctttggtgtgctaacaaattcaagtttttgactggcgactaatgcaggtataggagcacagctttcctcgtcga

aagatgacctcacgttgcagccaccgacacatatcggtccagcgccacat tcattatacg'gcgagcagcacggccattctgcgggacaagggt tatttggagaacaagaaatggagt ttttgggtttgtataattctcccagtgcaaatccgatgtagccaacaaccggcaccagctgtaatattgatattctcatattttitatttcgtgsaaaagecccttgatgatactstaaactaaa

clacgactgttcttgtatt tcgttattaatcgataattattagctggtattgtcttttttttttttttcctacatgatcgtcgcgaggctttacataatt tttgtgacgcctttaaaagtata

gtagaacgagcacataagaagtaaatgaaaagt acggcagatgcaat tgacagtaaagagcggtacaatagt tgaagaatacccgataaaaatgtat ttaggttgcttgacgggtttgca -361

gcaattgccaagaacactaaciEgtggcaat cttggtgagaatct actcccaacccaaacattt tcgccgatatttttgctcaccttttttttttgctcat'cgaaait-tgt tatagcggct -241

cgactttgatttactaatacacccaatttctaatatcctc~aggctagatcgttctctcaaggaacttaaagtgcctgatettgcgatgat taatctaca'gcgatggatacttaaatcat -121

gtaaaaatctcagt tttgaagtcgtttgctctcttccatgcttcagttcccttttggaaggtaatataatcatc taatttetcgcatattacagccgacg'aagaacgagccggatcaat -1

ATGGTTAGAACCCGTGTATTATTCTGCTTATTTATATCmTGr7rCTACGGTTCAATCGAGTGCTACACTTATTAGCACTTCATGTATTTCCCAAGCTGCGCTATACCAATTTGGATGT 120M V R T R V L F C L F IS F F _A T V Q S S A T L I S T S C I S Q A A L Y Q F G C 40

EwTCTAGTAAATCTAAAAGTTGCTACTGTAAAAACATCAATTGGCTGGGTTCAGTGACAGCATGTGCCTATGAGAATTCCAAATCTAACAAAACACTAGACAGCGCCTTAATGAAGTTAGCA 240S S K S K S C Y C K N I N W L G S V T A C A Y E N S X S N K T L D S A L M K L A 80

TCCCAATGTTCAAGCATCAAAGTTTATACTTTAGAGGACATGAAGAATATTTATTTAAATG;CGTCAAATTATTTGAGAGCACCTGAGAAAAGTGATAAAAAAACCGT&GTTAGTCAACCG 3 60S Q C S S I K V Y T L E D M K N I Y L N A S N Y L R A P E K S D K K T V V S Q P 120

CTCATG&CGAACGAGACAGCGTATCATTATTATTATGAGGAAAATTAT&GTATCCATCTTAACCTAATGCGCTCTCAAT&GTGCGCTT&.GG&TCTCGTCTTCTTCTG.GGT&&CTGT&CTT 480L M A N E T A Y H Y Y Y E E N Y G I H L N L M R S Q W C A W G L V F F W V A V L 160

ACTGCAGCCACTATCTTGAACATTCTGAAAAGGGTGTTTGGTAAGAACATCATGGCAAACTCCGTCAAAAAATCACTTATTTATCCTTCTGTTTACAAAGATTATAATGAACGAACTTTT 600T A A T I L N I L K R V F G K N I M A N S V K K S L I Y P S V Y K D Y N E R T F 200

TATmATGGAAGCGTCTACCATTTAATTTTACAACTCGAGGCAAGGGTCTCGTCGTATTAATTTTTGblTTATTTTG'ACTATATTATCTCTCAGTTTTGGTCATAATATTAAACTTCCACAC 720Y L W K R L P F N F T T R G K G L V V L I F V I L T I L S L S F G H N I K L P H 240

CCATATGATAGGCCCAGAT&GAGAAGAAGTAT&GCCTTTG;TGAGTCGTAGAGCAGACTTGAT&&CCATTGCACTTTTCCCAGTAGTCTATCTATTCGGAATAAGAAATAATCCCTTCATC 840P Y D R P R W R R S M A F V S R R A D L M A I A L F P V V Y L F G I R N N P F I 280

CCTATAACA&GGCTTTCCTTTTCTACATTTAATTTCTATCATAAAT&GTCT&CCTACGTTTGTTTCATGTTG&CCGTTGTACACTCAATTGTCAT&ACCGCCTCGGGAGTGAAAA&AG&T 9 60P I T G L S F S T F N F Y H K W S A Y V C F M L A V V H S I V M T A S G V R R G 320

GTGTTTCAAAGTCTGGTTAGGAAATTTTACTTAGGTr.GGGTATAGTGGCAACGATATTAATGTCTATTATTATTTTCCAAAGTGAAAAAGTATTTAGAAATAGAGGGTATGAGATATTC 1080V F Q S L V R K F Y F R W G I V A T I L M S I I I F Q S E X V F R N R G Y E I F 360

CTTCTTATTCATAAAGCGATGAATATTATGTTCATTATTGCCATGTACTACCATTGTCACACCCTGGGCT&GAT&&GGTT&GATTTGGTCAAT&GCTGGTATTTTATGCTTTGATAGATTC 1200L L I H K A M N I M F I I A M Y Y H C H T L G W M G W I W S M A G I L C F D R F 400

TGCAGGATTGTTrAGAATAATCATGAATGGT&GCTTGAAAACTGCTACTTT&AGTACCACTGATGATTCTAATGTTATTAAAATTTCAGTAAAAAAACCAAAGTTrTTTCAAGTACCAAGTA 1320C R I V R I I M N G G L X T A T L S T T D D S N V I X I S V K K P K F F K Y Q V 440

GGAGCTTrrCGCATACATGTAT'TTCTTATCACCAAAAAGTGCATGG&TTCTATAGTTTCCAATCACATCCATTACAGTATTATCGGAACGACACCGTGATCCAAACAATCCAGATCAATTG 1440G A F A Y M Y F L S P K S A W F Y S F Q S H P F T V L S E R H R D P N N P D Q L 480

ACGATGTACGTAAAGGCAAATAAAG&TATCACTCGAGTTTTGTTATCGAAAGTTCTAAGTGCTCCAAATCATACTGTTGATTGTAAAATATTCCTTGAAGGCCCATAT&GTGTAACGGTT 1560T M Y V K A N K G I T R V L L S K V L S A P N H T V D C K I F L E G P Y G V T V 520

BstCCACATATCGCTAAGCTAAAAAGAAATCTGGTAGGTGTAGCCGCTGGTTTGGGTGTTGCGGCTATTATCCGCACTTTGTCGAATGTTACGGTTACCATCTACTGATCAACTTCAGCAT 1680P H I A X L K R N L V G V A A G L G V A A I Y P H F V E C L R L P S T D Q L Q H 560

AAA7=TACTGGATTGTTAATGACCTATCCCATTTGAAATG&TTTAAAATGAATTGCAATGGTTAAAGGAGAAAAGTTGTGAAGTCTCAGTCATATATACTGGTTCCAGTGTTGAGGAC 1800K F Y W I V N D L S H L K W F E N E L Q W L K E K S C E V S V I Y T G S S V E D 600

ACAAATTCAGATGAGAGTACAAAAGGTTTTGMAT&ATAAAGAAGAAAGCGAAATCACTGTTGAATGTCTCAATAAAAGACCTGATTTGMAAAGAACTAGTGCGCTCGGAAATAAAACTCTCA l192 0T N S D E S T K G F D D K E E S E I T V E C L N K R P D L K E L V R S E I K L S 640

GAACTAGAGAATAATAATATTACCTTTTATTCCTGCGGGCCAGCAACGTTAACGACGATTTTAGAAATGCAGTGGTCCAAGGTATAGACTCTTCCTTGAAGATTGACGTTGAACTAGAA 2040E L E N N N I T F Y S C G P A T F N D D F R N A V V Q G I D S S L K I D V E L E 680

GAAGAAAGTTTTACATGGTAAggcccc ttgttaataattcttgcacgcatacttttgttattttgttgc tttatcggataaaagttaatataaatcgatgtaaataatt ttatataatca 2160E E S F T W * 686

tacaacagggtaataaaaagacgaatatat tcaaatgcgggttctaaatctaagtttcat tctcactatcactgctttcttcctttgcctgttgagagagtagt tctttccatttttgag 2280

tgagctc 2287

FIG. 1. DNA sequence of FREI. Nucleotides are numbered with respect to the translation start so that the A of the initiator ATG is +1.Restriction enzyme sites for Cla I, EcoRI, and BstEII are indicated by the enzyme name over the first nucleotide of the enzyme recognitionsequence. The transcription start is shown by an arrow over the A at position -49. The sequence of the 85-bp fragment (-339 to -255) ableto confer iron-repressible transcription on downstream elements of a CYCJ-IacZ fusion is enclosed in brackets. The coding region is shown inuppercase letters. Amino acids, indicated in one-letter code below each codon, are numbered beginning with the initiator methionine. A candidateleader peptide is underlined. Potential sites for N-linked glycosylation are underlined. Hydrophobic stretches of amino acids that are candidatesfor transmembrane domains occur at positions 147-169, 216-236, 258-277, 2%-316, 329-348, 369-397, and 529-550. Figure annotations weremade with the DNADRAW (23) program.

Proc. Natl. Acad. Sci. USA 89 (1992) 3871

150 160 170 180 190 200

X-CGD ESY LNFARKR IKNPEGGLYLAVTLLAGITGVVITLCLIIT'!S'STKTIR~RW1YFEVEIV8NWTHFREl IRNNPF IP ITGLSFSTFNFYHKWSAYVCFM.LAVVHSIV A:G

280 290 300 310 320 330

210 220 230 240 250 260

X-CGD H:LF.V.::T..:FF.OLA DHGAMWRV GOQ.T.:AWS't..:A:VHN:I:TVCEQK USEWG:~KIK~ELPI:P Q.F.AGNPP R

FREl G': AT:ILM:1:IDFQSL tKV13NRY I: L:LIHKA MNIMFDI -AMYYHL LGw.--

340 350 360 370 380

X-CGD 1fK1TIV1.GPM L:Y:- :L:MEWL: F:WR-S:QQKV:V:IDKVVUH PF:K~~TD.~E-~LQ~M -1GMEdI~~SMAG &''F' :KS:VffP F M YQj~jAF'AI390 400 410 420 43 04440

330 340 350 360 370 380

X-CGD VK J-:r:EE :R LF*:ACGOOFRE1 MYF L:SIM3SAWF Y:SFO:P:QS V1V SERHR .PNNP0TTMY V:K'ANXZ.:TRVL L:S.:KVLSAPNHT

450 40 470 480 490 50 0

390 400 410 420 430

X-CGD KLP AVDMF A DYFSYS 9 F tSWYKYCNRX-TNIKL1:: .: Q Hu

510 520 530 540 550 560

440 450 460 470 480 490

X-CGD IYGMCn f'A::MJ[EQQRNNX .::'.YLMVM A~~ A VH.HD0:EWFREl O M I.VNI NEM I DD1(-:E:E

570 580 590 600 610

500 510 520 530 540 550

X-CGD KON I TG LIBQK.TIIYGR PNWD.N::F:K~T:I-A:$:Q:HPNTR UG:ViFL WEAlE:A:A:E:T:L:$.'K:QS:I~SN-S-EMGPFRE1 I TI- JECLN IZP 1RPKD ELVRSS KLS: AT

620 630 640 650 660 670

560 57 0

X-CGD kGVHF IF7:NKW2NUFREl K -:I6VE L::E:E

of the Cla I fragment revealed a single long open reading the X-CGD protein (Fig. 21

frame (ORF) (bp 1-2061, Fig. 1). To confirm the correct amino acids of FREl and

reading frame and location of the ORF, we fused a fragment identity and 62.2% similaril

including FREI 5' noncoding sequences and the first three clusters of much higher i4codons to an E. coli lacZ reporter gene (bp -977 to +9, Fig. beginning at amino acid pi1). When lacZ was fused in the same reading frame as the GP--G at 509, and V-AG-C

predicted FREI ORF, the construct (pEG2) gave rise to Characterization of FREIf8-galactosidase activity that was regulated by iron in yeast used one-step gene replac,

(see below), whereas constructs with lacZ fused at nearly the strains in which the FREI gisame position in each of the other two reading frames gave no and/or deleted. FREI is ni

detectable enzyme activity. An iron-regulated 3-kb transcript rate of growth of the haplo~was detected previously by RNA blot-hybridization analysis was indistinguishable fronf(9). The 2-kb EcoRI-Sac I fragment used as the probe strain (compare H1085 wit]

includes the ORF from bp 192 to its 3' end (Fig. 1). The size of FREI led to growth retaof this transcript is sufficient to encode the FREl protein. (Table 1) and impaired ferricThe major transcription start ofFREI was mapped by primer uptake (Table 2), as describextension to the A nucleotide at -49 with respect to the ATG (9). The presence of con

initiation codon (data not shown), predicting a 49-nucleotide wild-type and AfreJ]:: URA3iuntranslated region of FREI mRNA that is devoid of AUG uptake system is independtriplets. more, expression of FREI)Homology Between FREl and the Large Subunit of Human mid pWDC2O in strain WE4

Cytochrome b5m. The predicted FREl protein is 686 amino 3-old higher than that seeiacids long with a calculated molecular mass of 78.8 kDa. whereas the rate of iron ulThere are six potential sites for the addition of N-linked the wild type (Table 2). The

sugars (24) (Fig. 1). The first 22 amino acids conform to the disruption/deletion retaine

von Heijne consensus for the leader peptide of a membrane ferric reductase activity an(or secreted protein (25) (Fig. 1). Hydrophobicity analysis (26) residual ferric reductase actrevealed two amino-terminal hydrophobic regions that are iron content of the growth

strong candidates for transmembrane domains and five other 20-fold by iron (Table 1).

hydrophobic regions that may also cross the membrane Transcriptional Regulatio(coordinates in Fig. 1 legend). bp of the 5' noncoding regiComparison of the FREl sequence with the protein data FREI in-frame with the E.

bank (National Biomedical Research Foundation, June 1991) number plasmid pEG2. Thiby the FA5TA algorithm (27, 28) revealed similarity to the tosidase activity in wild-tyllarge subunit of the human cytochrome b558, also known as iron content of the growth

FIG. 2. Homology of FRE1and X-CGD protein. The FRE1and the X-CGD protein aminoacid sequences were aligned byusing the FASTA program (27). Theoptimal alignment score of 171 is>16 standard deviations above themean optimal score determined byaligning the FRE1 sequence with500 randomly generated aminoacid sequences of the same aminoacid composition as FREl (28).The DNADRAW program (23) wasused to depict the alignment of theFRE1 and the X-CGD proteins,with black background indicatingidentical residues and shading in-dicating conserved residues.

(29). The carboxyl-terminal 402the X-CGD protein share 17.9oity. In addition, there are severallentity, such as KK-K-FK--VG)osition 431, HPFT at 462, KI---'V at 533 of the FREl sequence.I Disruption/Deletion Strains. Weement (12) to construct haploidrenomic sequence was interruptedot required for viability, and the)id strains in iron-replete mediumn that of the isogenic wild-typeIh W218 in Table 1). The deletionardation in iron-depleted mediumc reductase activity and ferric ironted previously for thefrel- allelenparable ferrous iron uptake instrains indicates that the ferrous

lent of the FREI gene. Further-from the high-copy-number plas-26 led to ferric reductase activityn in the parental wild-type cells,ptake was indistinguishable fromstrains carrying the AfreJ:: URA3d residual (5-20% of wild-type)Id ferric iron uptake activity. Thistivity was clearly regulated by themedium, being repressed about

)n of FREI by Iron. We fused 977ion and the first three codons ofcoli lacZ gene on the high-copy-lis construct gave rise to P-galac-pe cells that was regulated by the[medium, showing 55-fold higher

Table 1. Effect of FREI deletion on ferric reductase activity and iron-dependent growth

Iron-depleted medium Iron-containing mediumRelevant Ferric reductase, Doubling Ferric reductase, Doubling

Strain genotype nmol/hr per 106 cells time, min nmol/hr per 106 cells time, minH1085 FREI 12 + 0.3 271 ± 11 0.32 ± 0.01 274 ± 5.0W218 AfreI::URA3 2.1 ± 0.17 449 ± 8.5 0.11 ± 0.01 295 ± 3.0W218/pCL4.3 Afrel::URA3/FREJ 13 ± 0.12 282 ± 20 0.28 ± 0.02 293 ± 20

Cells were grown in iron-depleted medium or iron-supplemented MD medium prior to assay of ferric reductase activity.

Biochemistry: Dancis et al.

Proc. Natl. Acad. Sci. USA 89 (1992)

Table 2. Effect of overexpression and disruption of FREI onfemc reductase activity and iron uptake

Ferric Ferric Ferrousreductase, uptake, uptake,nmol/hr fmol/hr fmol/hr

Relevant per 106 per 106 per 106Strain genotype cells cells cells

F113 FREI 30 ± 1.1 2200 ± 540 79 ± 7.6W126 FREI (high-copy) 93 ± 7.5 2200 ± 320 79 ± 16N1 4frel::URA3 3.5 ± 0.24 120 ± 0.2 120 ± 3.7

Cells were grown in iron-depleted MD medium prior to assay.

expression in MD medium depleted of iron than in MDcontaining iron (420 units and 7.6 units, respectively). Therange of regulation of 83-galactosidase activity from thisconstruct was similar to that seen for ferric reductase. Theseresults suggest that regulation of FREI expression by ironoccurs at the level of transcription. To locate more preciselythe regulatory sequences, we examined the effects of 5'deletions of FREI sequences beginning at position -977 inthe low-copy-number plasmid pJ105 (Fig. 3A). For reasonsnot well understood, the range ofregulation by pJ105 was lessthan that by pEG2. The high activity present under low-ironconditions in the construct pJ105 was essentially unaffectedby removal ofsequences 5' to -341 (pJ106) but was abolished

by deletion to -279 (pJ107). Examination of the FREIsequence in the interval between -341 and -279 revealed asequence (-311 to -298) that matches at 11 of 14 positionswith the consensus binding site for RAPi (30), a multifunc-tional yeast protein known to serve as a transcriptionalactivator in some sequence contexts (31). While the expres-sion of (3-galactosidase activity in pJ106 was significantlyrepressed by iron, the loss of expression with deletion to-279 precluded the ability to assess whether an iron-regulatory element is present between -341 and -279. Wetherefore asked whether FREI sequences from this regioncould serve as an upstream activating sequence (UAS) in aCYCI-lacZ fusion from which the endogenous UAS wasdeleted. The parental construct, which contained 250 bp ofCYCI 5' noncoding sequence (including the transcriptionstart site and the ATG start codon) fused in-frame to the E.coli lacZ gene, gave negligible (<1 unit) 8-galactosidaseactivity. Addition of the FREI sequence -317 to -280 inpGCR conferred UAS activity but not iron regulation. Thissequence includes the consensus binding site for the RAPiprotein. Introduction of a C -* A transversion at position-307 (Fig. 3C), altering the invariant C at the fifth nucleotideof the RAPi consensus binding site (30) in construct pGCRM,decreased expression of (-galactosidase by a factor of >6.These results suggest that RAP1 activates FREI transcriptionbut is not sufficient for iron-mediated regulation.

A PLASMID FREl SEQUENCEP-Galactosidase (U) Ratio

-Fe +Fe - Fe/+ Fe

-600

- --900 -800 -700 -600

-341

-279R Ernni0 I -0 -0 1

-500 -400 - 30 - 200 - 100

FREl SEQUENCE-317 -280

-317 X -280

-339 -278

-339 -255

-254 -176

R E

- 4 -4 - 2 - 200-440 -400 -360 280 240 -200

I3-Galactosidase (U) Ratio-Fe +Fe - Fe/+ Fe

9.8 13 0.74

1.5 1.9 0.74

13 13

99 29

141 26

88 12

1.0

3.4

5.4

6.9

0.3 0.7 0.43

-176 29 4.4 6.5

C3-39 -25TGTGGCAATCTTGGTGAGAATCTACTC AA AAACATT CGCCGATATTTTTGCTCACCTTTTTTTTTTGCTCATCGAAAAACACCGTTAGAACCACTCTTAGATGAG TTaTTTGTAA GCGGCTATAAAAACGAGTGGAAAAAAAAAACGAGTAGCTTTT

FIG. 3. Localization of iron-responsive transcriptional regulatory elements in the FREI promoter. (A) Deletion of 5' noncoding sequencesfrom FREJ-IacZ constructs. pJ105, pJ111, pJ106, and pJ107 are low-copy-number plasmids containing FREJ-4acZ fusions. FREI sequencesare represented by solid lines, and the 5' and 3' boundaries of these sequences are indicated by the numbers adjacent to the lines. The scaleindicates the position of these sequences relative to the A of the ATG codon of the FREI ORF. The locations of the RAPi consensus bindingsite (R) and repeat element TTTTTGCTCAYC (E) are shown on this scale. The FREJ-4acZ constructs were introduced into strain F113 andassayed for 3-galactosidase expression (units, U) in iron-depleted (-Fe) or iron-supplemented (+Fe) medium. (B) FREI sequences confer ironregulation upon CYCJ-4acZ expression. The plasmids are low-copy-number plasmids containing FREI sequences fused to downstream elementsof a CYCIJ-acZ fusion from which the UAS has been deleted. The boundaries of these FRE1 sequences are indicated by the numbers adjoiningthe solid lines. The X represents the location of a point mutation in the RAPi consensus binding site. The scale indicates the position of thesesequences relative to the A of the ATG codon of the FREI ORF. The RAPi consensus binding site (R) and repeat (E) are indicated. Plasmidswere introduced by transformation into wild-type strain F113 and analyzed as in A. (C) Minimal FREI sequence able to confer iron-repressibletranscription on a UAS-deleted CYCJ-IacZ fusion. The FREI sequence contained in pGC7-8.0 is depicted. The RAP1 consensus binding siteis boxed. The point mutation in this site shown to impair function is shaded. The repeat is indicated by arrows.

pJ 105 -977

pJ111

pJ 106

pJ107

B

+3 5.9 0.8

+3 7.9 1.9

+3 10 1.4

+3 0.3 0.4

PLASMID

7.4

4.1

7.2

0.8

pGCR

pGCRM

pGC7-8

pGC7-8.0

pGC7-8.5

pGC7-9

pGC8-9

pGC6-9

3872 Biochemistry: Dancis et al.

Proc. Natl. Acad. Sci. USA 89 (1992) 3873

We next studied the effect of adding flanking FREI se-quences to the pGCR construct containing the RAP1 bindingsite from FREI. Addition of FREI sequence from between-317 and -339 resulted in negligible changes in derepressedactivity or repression by iron (PGC7-8 in Fig. 3B). However,further addition of FRE1 sequence, -277 to -255, led to a7.5-fold increase in activity under low-iron conditions and theacquisition of iron-repressible expression (compare pGC7-8and pGC7-8.0 in Fig. 3B). The 85-bp sequence in pGC7-8.0that confers iron-repressible transcription on downstreamelements of a CYCI-lacZ fusion contains a 12-bp directrepeat, TlTIT'GCTCAYC, located 3' to the RAPI bindingsite. A search of sequence motifs for the recognition sites ofDNA-binding proteins failed to reveal any match with thisrepeated sequence (32). Inserting sequences to positions-208 or -176 resulted in an increased range of regulation byiron similar to that seen for the larger promoter fragments inpGC6-9 and pJ105, and thus a second iron-control elementmay be present 3' of -255.

DISCUSSIONThe mechanism(s) by which eukaryotic cells acquire ironfrom the environment has remained one of the unsolvedproblems in the biology of this essential nutrient. Although anexternally oriented ferric reductase has been proposed to beimportant for iron uptake by many eukaryotes, until recently,no molecular evidence for, or characterization of, such asystem had been reported. We initially provided geneticevidence suggesting that the FREI gene of S. cerevisiae isrequired for high-level ferric reductase activity, ferric ironuptake, and the ability to grow in iron-limited environments.Here we characterize that gene at the molecular level.Complete deletion of FREI in wild-type cells led to reducedferric reductase activity, deficient ferric iron uptake, andimpaired growth on iron-depleted medium. Ferrous assimi-lation was unimpaired in strains containing a disruption ordeletion of FREI, demonstrating that the ferrous uptakesystem is independent of FREI.The DNA sequence of FREI suggests that it encodes a

structural component of the externally directed reductase.The deduced FRE1 protein sequence of 686 amino acidscontains a potential leader sequence and several hydrophobicregions consistent with transmembrane domains. In addition,the FRE1 amino acid sequence shows significant similarity tothe large subunit of cytochrome b558, also known as theX-CGD protein, an essential component of a multisubunitreductase present in the plasma membranes ofhuman phago-cytic cells. Cytochrome b558 is thought to be the terminalcomponent of a respiratory chain that transfers single elec-trons from cytoplasmic NADPH across the plasma mem-brane to molecular oxygen on the exterior (29).FREI may not encode the only reductase used by S.

cerevisiae in iron reduction and iron uptake, as suggested bythe residual ferric reductase and ferric iron uptake activitiesof the Afrel:: URA3 strains. Ferric iron in the growth mediumrepresses the residual reductase activity of these strains, afeature consistent with an activity involved in iron uptake.The existence of an alternative uptake system, perhaps oneof lower affinity than that involving FRE1, is also suggestedby the ability of high concentrations of ferric iron in themedium to completely correct the growth deficiency of thestrains carrying the FREI disruptions/deletions.As with other iron uptake systems, in S. cerevisiae,

increased availability of iron leads to reduced expression ofan essential component of the uptake system. FREI expres-sion is controlled by alterations in the rate of transcription ofits mRNA. This was demonstrated by the ability of FREI 5'nontranscribed sequences to transfer this regulation to the E.coli lacZ gene. The range of iron regulation by the completeFREI-lacZ fusion in the high-copy-number plasmid pEG2

closely reflects the iron-induced changes in ferric reductase.An 85-bp FREI sequence element sufficient for transferringiron-dependent activity to a CYCJ-IacZ construct lacking itsnative UAS includes a consensus binding site for RAP1 andthe repeated sequence, T-`TTTTGCTCAYC. This repeatedsequence constitutes a candidate binding site for an iron-responsive regulatory protein. The RAP1 binding site byitself is not sufficient for this regulatory effect, but a proteinmediating the iron-dependent changes in transcriptional ini-tiation might interact with RAP1.

M. Shapiro assisted with the design of Figs. 1 and 2. P. Fitzgerald,V. Zenger, and R. Blasco patiently answered questions related tocomputer sequence analysis. C. Moehle provided plasmid p1087 andhelpful advice and discussions. G.J.A. is supported by a C. J. MartinFellowship from the National Health and Medical Research Councilof Australia.

1. Williams, R. J. P. (1990) in Iron Transport and Storage, eds. Ponka,P., Schulman, H. M. & Woodworth, R. C. (CRC, Cleveland), pp.2-15.

2. Nielands, J. B. (1981) Annu. Rev. Nutr. 1, 27-46.3. Crane, F. L., Roberts, H., Linnane, A. W. & Low, H. (1982) J.

Bioenerg. Biomembr. 14, 191-205.4. Yamashoji, S. & Kajimoto, G. (1986) Biochim. Biophys. Acta 852,

25-29.5. Chaney, R. L., Brown, J. C. & Tiffin, L. 0. (1972) Plant Physiol.

50, 208-213.6. Bienfait, F. (1987) in Iron Transport in Microbes, Plants and

Animals, eds. Winkelmann, B., van der Helm, D. & Nielands, J. B.(VCH, Weinheim, F.R.G.), pp. 337-349.

7. Buckhout, T. J., Bell, P. F., Luster, D. G. & Chaney, R. L. (1989)Plant Physiol. 90, 151-156.

8. Seligman, P. A., Klausner, R. D. & Huebers, H. A. (1987) in TheMolecular Basis ofBlood Diseases, eds. Stamatoyannopoulos, G.,Nienhuis, A. W., Leder, P. & Majerus, P. W. (Saunders, Philadel-phia), pp. 219-244.

9. Dancis, A., Klausner, R., Hinnebusch, A. G. & Barriocanal, J. G.(1990) Mol. Cell. Biol. 10, 2294-2301.

10. Nielands, J. B., Konopka, K., Schwyn, B., Coy, M., Francis,R. T., Paw, B. H. & Bagg, A. (1987) in Iron Transport in Microbes,Plants and Animals, eds. Winkelmann, B., van der Helm, D. &Nielands, J. B. (VCH, Weinhein, F.R.G.), pp. 3-33.

11. Botstein, D., Falco, S. C., Sterwart, S. E., Brennan, M., Scherer,S., Stinchcomb, D. T., Struhl, K. & Davis, R. W. (1979) Gene 8,17-24.

12. Rothstein, R. J. (1983) Methods Enzymol. 101, 202-211.13. Innis, M. A., Gelfand, D. H., Sninsky, J. J. & White, T. J. (1990)

PCR Protocols (Academic, San Diego).14. Sikorski, R. S. & Hieter, P. (1989) Genetics 122, 19-27.15. Moehle, C. M. & Hinnebusch, A. G. (1991) Mol. Cell. Biol. 11,

2723-2735.16. Guarente, L. & Ptashne, M. (1981) Proc. Natl. Acad. Sci. USA 78,

2199-2203.17. Sherman, F., Fink, G. R. & Hicks, J. B. (1986) Laboratory Course

Manualfor Methods in Yeast Genetics (Cold Spring Harbor Lab.,Cold Spring Harbor, NY).

18. Guarente, L. (1983) Methods Enzymol. 101, 181-191.19. Ruby, S. W., Szostak, J. W. & Murray, A. W. (1983) Methods

Enzymol. 101, 253-269.20. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular

Cloning: A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY).

21. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad.Sci. USA 74, 5463-5467.

22. Caroline, D., Favreau, M., Dunsmuir, P. & Bedbrook, J. (1987)Nucleic Acids Res. 9, 133-148.

23. Shapiro, M. (1990) Binary 2, 187-190.24. Marshall, R. D. (1972) Annu. Rev. Biochem. 41, 673-702.25. Von Heijne, G. (1983) Eur. J. Biochem. 133, 17-21.26. Kyte, J. & Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132.27. Pearson, W. R. & Lipman, D. J. (1988) Proc. Natl. Acad. Sci. USA

85, 2444-2448.28. Devereux, J., Haeberli, P. & Smithies, 0. (1984) Nucleic Acids Res.

12, 387-395.29. Orkin, S. H. (1989) Annu. Rev. Immunol. 7, 277-307.30. Nieuwint, R. T. M., Mager, W. H., Maurer, K. C. T. & Planta,

R. J. (1989) Curr. Genet. 15, 247-251.31. Shore, D. & Nasmyth, K. (1987) Cell 51, 721-732.32. Ghosh, D. (1990) Nucleic Acids Res. 18, 1749-1756.

Biochemistry: Dancis et al.