(Received for publication, May 13, 1996, and in revised form, February 25, 1997)
From the Department of Biochemistry and Molecular Biology, University of Minnesota, Duluth, Minnesota 55812
The low affinity Fe2+ uptake system of Saccharomyces cerevisiae requires the FET4 gene. In this report, we present evidence that FET4 encodes the Fe2+ transporter protein of this system. Antibodies prepared against FET4 detected two distinct proteins with molecular masses of 63 and 68 kDa. In vitro synthesis of FET4 suggested that the 68-kDa form is the primary translation product, and the 63-kDa form may be generated by proteolytic cleavage of the full-length protein. Consistent with its role as an Fe2+ transporter, FET4 is an integral membrane protein present in the plasma membrane. The level of FET4 closely correlated with uptake activity over a broad range of expression levels and is itself regulated by iron. Furthermore, mutations in FET4 can alter the kinetic properties of the low affinity uptake system, suggesting a direct interaction between FET4 and its Fe2+ substrate. Mutations affecting potential Fe2+ ligands located in the predicted transmembrane domains of FET4 significantly altered the apparent Km and/or Vmax of the low affinity system. These mutations may identify residues involved in Fe2+ binding during transport.
In many organisms, iron uptake is a two-step process in which extracellular Fe3+ is reduced to the more soluble Fe2+ form by plasma membrane Fe3+ reductases. The Fe2+ product is then taken up by Fe2+-specific transport systems. This strategy of iron uptake is found in the yeast Saccharomyces cerevisiae (1-3), some bacteria (4, 5), other fungi (6, 7), and many plant species (8, 9). Mammalian cells may use a similar mechanism for uptake of iron across the mucosal membrane of the intestine (10-12) and for the uptake of free iron in blood plasma (13, 14). Mammalian cells acquire most of their iron from transferrin. Fe3+-transferrin complexes bind to transferrin receptors on the cell surface. These receptor-ligand complexes are endocytosed to an endosomal compartment; the iron is dissociated and then transported across the endosomal membrane. Some studies have suggested that transferrin-delivered Fe3+ is reduced to Fe2+ in the endosome and transported into the cytoplasm by Fe2+-specific transporters (15-17). Clearly, Fe2+ transporters play a prominent role in iron acquisition by a wide variety of organisms.
In S. cerevisiae, extracellular Fe3+ is reduced to Fe2+ by the plasma membrane Fe3+ reductases encoded by the FRE1 and FRE2 genes (18, 19). The Fe2+ product is then taken up by either of two transport systems. One system has a high affinity for iron (apparent Km of 0.15 µM), is necessary for iron-limited growth, and requires the products of the FET3 and FTR1 genes for activity (20-23). The high affinity system is induced in iron-limited cells, and its components are transcriptionally regulated by the product of the AFT1 gene (24). AFT1 is an iron-responsive DNA binding protein that activates transcription of the target promoters to which it binds (25).
Iron-replete yeast cells obtain iron through a second, low affinity uptake system with an apparent Km of 30 µM. This system requires the FET4 gene for activity. Our previous results suggested that FET4 is the low affinity Fe2+ transporter (26). First, overexpression of the FET4 gene increased activity of an iron uptake system that was indistinguishable from the low affinity system. Second, disruption of the FET4 gene eliminated low affinity uptake activity but did not diminish high affinity activity. Finally, the sequence of the FET4 gene suggested that its product is a transporter protein. The predicted FET4 amino acid sequence is 552 residues in length and contains over 50% hydrophobic amino acids. Many of these hydrophobic residues are arranged in six regions that may be transmembrane domains. FET4 has no homology to any known protein including FTR1, the feoB Fe2+ transporter of Escherichia coli (27), and the IRT1 Fe2+ transporter from Arabidopsis thaliana (28). Therefore, while the hydrophobic character of FET4 suggested that it is a transporter, we could not rule out other models of FET4 function. The central goal of the experiments described in this report was to further test the hypothesis that FET4 encodes the Fe2+ transporter of the low affinity system.
Yeast strains used were
DY1457 (MAT ade6 can1 his3 leu2 trp1 ura3),
DEY1394 (MAT
ade6 can1 his3 leu2 trp1 ura3
fet3-2::HIS3), DEY1422 (MATa can1 his3 leu2 trp1
ura3 fet4-1::LEU2), DEY1446 (MATa can1 his3 leu2
ura3 fet4-1::LEU2 trp1::YIpGAL1-FET4), DDY4 (MATa ade6 can1 his3 leu2 trp1 ura3 fet3-2::HIS3
fet4-1::LEU2), DEY1514 T1 (MATa/MAT
ade2/+
ade6/+ can1/can1 his3/his3 leu2/leu2 trp1/trp1::YIpGAL1-FET4
ura3/ura3), and DEY1515 (MATa/MAT
ade2/+ can1/can1
his3/his3 leu2/leu2 trp1/trp1 ura3/ura3
fet4-1::LEU2/fet4-1::LEU2). Cells were grown
in 1% yeast extract, 2% peptone (YP) or synthetic defined (SD) medium
(6.7 g/liter yeast nitrogen base) supplemented with any necessary
auxotrophic requirements and either 2% glucose or 2% galactose. Cells
were also grown in a modified iron-limited medium (LIM, Ref. 29)
prepared without EDTA (i.e. LIM-EDTA) and supplemented with
FeCl3 to the stated concentrations. LIM-EDTA is iron
limiting for growth of fet3 mutant strains when supplemented with less than 10 µM FeCl3 (data not shown)
because of its high concentration (20 mM) of citrate, an
iron-binding chelator. Yeast and E. coli transformations
were performed using standard methods (30, 31).
The locations of
potential transmembrane domains and the orientation of FET4 were
predicted using TOP-PREDII software (32). Three segments of the FET4
protein, i.e. amino acids 1-60 (pGEMEX-N), 120-220
(pGEMEX-L1), and 410-460 (pGEMEX-L5), were selected as antigens. DNA
fragments corresponding to these regions were obtained using polymerase
chain reaction primers with appropriate restriction sites added to
their 5 ends and cloned into pGEMEX-1 (Promega). In-frame cloning of
these inserts into this vector produced genes in which the
bacteriophage gene 10 protein is fused to the FET4 peptide. The fusion
proteins were expressed in E. coli strain BL21 (DE3) pLysS
as described by Studier et al. (33). Cells were harvested by
centrifugation, boiled in SDS sample buffer, and centrifuged at
12,000 × g for 1 min to remove cell debris. The
supernatant was fractionated by SDS-polyacrylamide gel electrophoresis in a Bio-Rad 491 Prep Cell. Rabbits were injected subcutaneously with
100 µg of semi-purified fusion protein in adjuvant. Anti-FET4 antibodies were affinity purified against their corresponding gene
10-FET4 fusion protein by column chromatography (34).
A BamHI-SacI fragment bearing the FET4 open reading frame was generated by polymerase chain reaction and inserted into pLO-LB (L. Opresko, University of Utah) to generate pLO-LBFET4. In vitro transcription/translation was performed using the TnT system (Promega).
Preparation of Protein Extracts and Fractionation on Sucrose Density GradientsCells were grown to exponential phase (100 ml, A600 of 2-4), spheroplasts were prepared (35), resuspended in 10 ml of 0.6 M mannitol, 20 mM HEPES-KOH, pH 7.4, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM pepstatin A, and disrupted in a Dounce homogenizer. Total cell homogenates were obtained by centrifuging these samples at 3000 × g for 5 min at 4 °C and discarding the pellet of unbroken cells. The homogenates were then centrifuged at 123,000 × g for 30 min at 4 °C to yield the soluble (supernatant) and particulate/membrane (pellet) fractions. Sucrose density gradient fractionation was performed as described previously (36). One ml of total cell homogenate (approximately 1 mg of protein) was loaded onto the top of linear sucrose gradients (20-55% w/w). The gradients were centrifuged for 16 h at 110,000 × g in an SW41 rotor at 4 °C. Fractions (700 µl each) were collected sequentially from the top of the gradients beginning with fraction 1.
Immunoblot AnalysisImmunoblots were performed as described previously (34) using primary antibodies specific to FET4, PMA1 (37), HMG1 (38), VPH1 (Molecular Probes, Inc.) and OMP2 (G. Schatz, Basel). Unless stated otherwise, anti-L1 was used for detection of FET4. Horseradish peroxidase-conjugated goat anti-rabbit antibody (Pierce) was used as the secondary antibody; protein-antibody complexes were detected with enhanced chemiluminescence (Amersham Corp.). Densitometric scanning was performed using a CCD camera and IMAGE 1.44 software (National Institutes of Health).
Indirect Immunofluorescence MicroscopyIndirect
immunofluorescence microscopy was performed essentially as described by
Pringle et al. (39) with the following modifications. Cells
were fixed in 10 volumes of cold methanol (20 °C) for 30 min.
Fixed cells were treated with glusulase to remove the cell wall and
bound to polylysine-treated coverslips. Primary antibody staining was
performed by incubating the cells with affinity-purified anti-FET4
antibodies (1:200 dilution in phosphate-buffered saline) at room
temperature for 16 h. Following washing of the cells, goat
anti-rabbit IgG secondary antibody (Pierce) was applied (1:200 dilution
in phosphate-buffered saline), and the cells were incubated at 37 °C
for 1 h. This was followed by incubation at 37 °C for 1 h
with streptavidin-conjugated fluorescein isothiocyanate (Zymed) diluted
1:400 in phosphate-buffered saline.
YIpGAL1-FET4 was
constructed by inserting the 2.7-kilobase
KpnI-SacI fragment from pCB1 (26) into pRS304
(40). This plasmid was digested with XbaI and transformed
into DEY1422 to generate DEY1446 (41). Plasmid pCB2 is a derivative of
pCB1 that contains a 66-base pair deletion in the FET4
5-untranslated region. YIpGF4d1 was constructed by cloning the
GAL1-FET4 KpnI-NotI fragment of pCB2 into pRS304.
Mutations were generated in YIpGF4d1 by site-directed mutagenesis using
the Transformer system (CLONTECH) and verified by
DNA sequencing. The resulting plasmids were linearized by digestion with XbaI and transformed into DDY4. Trp+
colonies were isolated and confirmed to contain the
GAL1-FET4 fusion gene by polymerase chain reaction.
The Fe2+ uptake and Fe3+ reductase assays were performed at 30 °C as described previously (3) except that 55Fe was substituted for 59Fe, and radioactivity was measured by liquid scintillation counting. Iron accumulation by wild type and mutant cells at 30 °C was found to be linear over the entire time of the uptake rate determination. 55Fe accumulation due to cell surface binding was estimated by incubating parallel samples at 0 °C for the same period as the assay. These values were then subtracted from the 30 °C samples before calculation of uptake rates. The 0 °C values, which never exceeded 5% of the 30 °C samples, were similar to the level of iron accumulation observed with a fet3 fet4 mutant at 30 °C indicating that this was an appropriate method for measuring cell surface iron binding. Determinations of apparent Km and Vmax values were made by fitting the data directly to theoretical curves using KINETASYST software (Intellikinetics, Princeton, NJ).
Based on the
"positive-inside" rule (42), a model was devised describing the
topology of FET4 in a lipid bilayer membrane (Fig. 1).
Three hydrophilic regions of the protein were selected for use as
antigens to generate antibodies against FET4. These regions were the
amino-terminal 60 amino acids (anti-N), 101 amino acids located between
transmembrane domains 1 and 2 (anti-L1), and 51 amino acids located
between transmembrane domains 5 and 6 (anti-L5). These portions of FET4
were expressed in E. coli as fusions to the phage T7 gene 10 protein and purified. The fusion proteins were then injected into
rabbits and affinity-purified antisera were prepared.
On immunoblots, while the preimmune serum did not detect any proteins
in total cell homogenates (Fig. 2A,
lane 1), anti-L1 antibody detected three proteins of 68 kDa
("p68"), 63 kDa ("p63"), and 46 kDa ("p46") molecular mass
(Fig. 2A, lane 2). The predicted molecular mass
of FET4 is 63 kDa. Detection of p68, p63, and p46 could be blocked by
preincubation of the antibody with the gene 10-FET4 fusion protein but
not by preincubation with gene 10 protein alone (Fig. 2A,
lanes 3 and 4). Thus, these three proteins are detected by antibodies that recognize the FET4 portion of the fusion
protein. Further experiments demonstrated that both p63 and p68 are
products of the FET4 gene, whereas p46 is encoded by another
gene. First, the two other anti-FET4 antibodies, anti-N and anti-L5,
detected p68 and p63 but not p46 (Fig. 2A, lane 5 and 6). Furthermore, p68 and p63 levels were altered by
overexpression and deletion of the FET4 gene; p68 and p63
levels were low in wild type cells, undetectable in fet4
mutant cells, and very high in cells overexpressing the FET4
gene from the GAL1 promoter (Fig. 2B). The level
of p46 was unaffected by differential FET4 expression.
What is the relationship between the p68 and p63 forms of FET4? Neither p68 nor p63 is N-glycosylated or phosphorylated; no change in electrophoretic mobility was observed when membrane proteins were treated with endoglycosidase H, peptide N-glycosidase F, or alkaline phosphatase (data not shown). To determine which form was the primary translation product, we synthesized FET4 in vitro (Fig. 2C). Although no protein product was detected in a control reaction with the vector alone, a FET4 expression plasmid directed the synthesis of a FET4 protein with the same electrophoretic mobility as p68. These results suggested that p68 is the primary translation product of the FET4 gene and p63 is an altered form generated, perhaps, by proteolytic cleavage of p68. This proteolysis probably occurs in vivo because it was not prevented by the addition of protease inhibitors to the homogenization buffers nor was it prevented when proteins were prepared from a strain, BJ2168, that is defective for several vacuolar proteases (43). The physiological significance of this modification is unclear; the abundance of p63 relative to p68 was variable and did not correlate with the level of iron in the growth medium, the level of FET4 expression, or the carbon source on which the cells were grown (data not shown).
FET4 Is an Integral Membrane ProteinCellular proteins were
separated into soluble and particulate/membrane fractions, and these
fractions were examined for the presence of FET4 by immunoblotting
(Fig. 3A). In either FET4
overexpressing (lanes 1-3) or wild type (lanes
4-6) cells, both p63 and p68 were highly enriched in the
particulate/membrane fraction. A small amount of the p63 form was also
found in the soluble fraction in FET4-overexpressing cells.
This may be due to the presence of p63 in small vesicles that sediment
slowly during ultracentrifugation. The p46 protein was found only in
the soluble protein fraction.
The enrichment of FET4 in the particulate/membrane fraction suggested
that this protein was associated with membranes. When this fraction was
treated with high pH, NaCl, NaBr, or urea, i.e. agents that
disrupt protein-protein interactions (44), FET4 remained associated
with the particulate/membrane fraction (Fig. 3B). Treatment
with the detergents Triton X-100 and
n-octyl--D-glucopyranoside released FET4 into
the soluble fraction. These results indicate that FET4 is an integral
membrane protein. Slower migrating forms of FET4 were observed in the
detergent-solubilized fractions that may be dimeric FET4. Consistent
with this hypothesis, the molecular mass of this complex was 130-140
kDa, i.e. twice the monomeric FET4 mass.
The subcellular location
of FET4 was first assessed by fractionation of cellular membranes on a
20-55% (w/w) sucrose gradient. The density of the isolated fractions
increased linearly from 1.08 to 1.22 g/ml, and protein was most
abundant in the lowest density fractions where soluble proteins are
found (Fig. 4A). Equal volumes of alternate
fractions were analyzed by immunoblotting for the presence of FET4 and
several marker proteins specific to particular subcellular compartments
(Fig. 4B), and these blots were quantitated by densitometric
scanning (Fig. 4C). FET4 was most abundant in fraction 15 (d 1.22 g/ml; 55% sucrose). PMA1, the plasma membrane
marker protein, was also most abundant in fraction 15 as was the
product of an epitope-tagged CTR1 allele (data not shown).
CTR1 is the high affinity copper transporter and has also been
localized to the plasma membrane (45). The presence of PMA1, CTR1, and
FET4 in the bottom fraction of the gradient was not due to protein
aggregation. When fraction 15 was treated with
n-octyl--D-glucopyranoside and reloaded onto a sucrose gradient, these proteins were found in the low density fractions (data not shown). Marker proteins specific for intracellular compartments, i.e. the vacuolar VPH1 protein, the
mitochondrial OMP2 protein, and the endoplasmic reticulum HMG1 protein
were most abundant in lower density fractions. In a similar experiment, a Golgi marker protein (dipeptidyl-aminopeptidase A) showed a distribution in the gradient like that of OMP2 (36). Thus, FET4 co-fractionated with plasma membranes in these gradients.
By indirect immunofluorescence microscopy, the FET4 protein could be
visualized as a bright rim of fluorescence at the periphery of cells
overexpressing the FET4 gene (Fig. 5).
Similar results were obtained with anti-N antibody (data not shown). In
contrast, a fet4 mutant strain did not show this peripheral
staining. These results also indicate that FET4 is a plasma membrane
protein. Attempts to detect FET4 when expressed at wild type levels
were unsuccessful probably due to the protein's normally low level of
synthesis (Fig. 2B).
Correlation of FET4 Levels, Low Affinity Uptake, and Regulation by Iron
FET4 overexpression increased low affinity uptake
activity, whereas disruption of the gene eliminated that activity (26). This correlation supported the hypothesis that FET4 encodes
the Fe2+ transporter of the low affinity system. To test
this correlation more rigorously, we used the fusion gene in which
FET4 is expressed under the regulation of the
GAL1 promoter. Cells overexpressing FET4 in
galactose-containing medium were split into two cultures, and glucose
was added to one culture to shut off expression of the GAL1
promoter; GAL1 promoter activity is reduced to less than 10% of the induced level within 5 min of glucose addition (46). Cells
were harvested periodically and assayed for Fe2+ uptake
activity (Fig. 6A) and FET4 levels (Fig.
6B). The activity of the low affinity system in untreated
cells increased slightly as did the level of FET4. In glucose-treated
cells, low affinity uptake decreased approximately 40-fold. FET4 levels
declined to a similar degree, and its profile was almost superimposable
with the loss of uptake activity. Thus, FET4 levels and uptake activity of the low affinity system closely correlated over a broad range of
expression levels.
To determine if the low affinity system is iron-regulated, we measured
this activity in iron-replete and iron-limited cells. A fet3
mutant was used for this analysis to allow measurement of low affinity
activity in the absence of the high affinity system. The low affinity
system is iron-regulated; uptake activity increased approximately
3-fold in iron-limited cells (Fig. 7A).
Fe3+ reductase activity was also induced approximately
3-fold in these cells, confirming that this medium was iron-limiting
(Fig. 7B). To assess if AFT1 plays a role in this
regulation, we measured uptake activity in a strain bearing the
AFT1-1up allele. AFT1-1up causes
constitutively induced expression from AFT1-responsive promoters (24). Activity of the low affinity system was not increased
by the AFT1-1up allele, whereas Fe3+
reductase activity was constitutively active in this strain. These
results indicate that the low affinity activity is regulated in
response to iron by a mechanism distinct from AFT1 transcriptional activation.
Regulation of low affinity uptake and FET4
levels by iron. fet3 cells (DEY1394, filled
columns) or fet3 cells transformed with the
AFT1-1up plasmid pT14 (hatched columns)
were grown to exponential phase in LIM-EDTA supplemented with 1000 µM
(+Fe) or 10 µM (Fe)
FeCl3. Cells were harvested and assayed for
Fe2+ uptake activity (A) and Fe3+
reductase activity (B). Shown are the results of a
representative experiment, and each value was derived from four
samples. The error bars represent 1 S.D. C,
kinetic analysis of Fe2+ uptake in iron-replete and
iron-limited cells. DEY1394 was grown to exponential phase in LIM-EDTA
supplemented with 1000 µM (+Fe, closed
squares) or 10 µM (
Fe, open
squares) FeCl3. The data shown are the means of two
separate experiments each performed in duplicate, and the standard
deviation within each experiment was <10% of the mean. D,
particulate/membrane fractions were prepared from DEY1394 cells grown
as in C and analyzed by immunoblotting using anti-PMA1 and
anti-FET4 antibodies. The intensities of the FET4 bands were measured
by densitometry (arbitrary units).
The increased uptake activity in response to iron limitation was caused by a change in Vmax, whereas the apparent Km was unaffected (Fig. 7C). The apparent Km and Vmax of iron-replete cells was 41 ± 7 µM Fe2+ and 110 ± 15 fmol/min/106 cells, respectively. In iron-limited cells, the apparent Km was 33 ± 8 µM Fe2+, and the Vmax was 372 ± 27 fmol/min/106 cells. Immunoblots demonstrated that while the level of the PMA1 plasma membrane ATPase was unaffected by iron limitation, FET4 levels increased approximately 3-fold (Fig. 7D).
Characterization of FET4 Mutant AllelesThe experiments described above support the hypothesis that FET4 encodes the low affinity transporter protein. The identification of mutations in FET4 that alters intrinsic kinetic properties of the low affinity system also supports this hypothesis. Especially informative are mutations that alter the affinity (i.e. apparent Km) of the system for Fe2+ because this parameter is determined by the direct interaction of the transporter with its substrate (47). We anticipated that critical ligands for Fe2+ binding would be located in or near the predicted transmembrane domains. Nine potential ligands were chosen for site-directed mutagenesis (Fig. 1) (see "Discussion"). In each case, the amino acid was replaced with an alanine residue because such mutations have been demonstrated to minimize structural alterations in the protein (48). The mutant alleles were expressed from the GAL1 promoter in a fet3 fet4 mutant strain and assayed for Fe2+ uptake.
The effects of these mutations on the concentration dependence of
FET4-mediated Fe2+ uptake were determined (Fig.
8), and the apparent Km and
Vmax values derived from these data are
summarized in Table I. Two of the nine alleles, Y222A
and Y408A, had little effect on the apparent Km for
Fe2+. A third allele, Y392A, increased the
Km approximately 2.5-fold. Three other alleles
increased the apparent Km even higher. These
increases ranged from 13-fold for Y352A to more than 30-fold for D271A.
Vmax values were also altered for several of
these alleles, ranging from 30 to 200% of the wild type rates. No
saturability of D271A-dependent uptake was observed in
assays conducted with Fe2+ concentrations as high as 3 mM (data not shown), preventing an accurate determination
of the Vmax of this allele. Clearly, mutations in FET4 can greatly alter the kinetic properties of the low
affinity system.
|
Three mutations, D354A, D400A, and E406A, completely eliminated low affinity uptake activity. While D354A and D400A produced wild type levels of FET4, no protein was detected in the E406A-expressing strain (data not shown). Subcellular fractionation of proteins from D354A and D400A on sucrose density gradients indicated that these forms were properly localized to the plasma membrane (data not shown). When overexpressed in a wild type FET4 strain, however, D354A and D400A were both found to be recessive.
The experiments described in this report test the hypothesis that FET4 is the Fe2+ transporter protein of the low affinity system. Consistent with this role, FET4 is an integral membrane protein and localized to the plasma membrane. Additional supporting evidence was provided by the close correlation between FET4 levels and uptake activity. This correlation was demonstrated in studies where the FET4 gene was overexpressed under the control of the GAL1 promoter as well as when it was expressed from its own promoter.
These experiments also indicated that the low affinity system is regulated by iron. The Vmax of the low affinity uptake increased approximately 3-fold in iron-limited cells relative to iron-replete cells, and a similar degree of induction was observed for FET4 levels. Fe3+ reductase activity was also induced by iron limitation in these cells. Despite this similarity, FET4 and the Fe3+ reductase activity are probably not regulated by the same mechanism because they respond differently to the AFT1-1up allele. AFT1 encodes a transcriptional activator that controls the expression of several iron-responsive genes including the Fe3+ reductase genes FRE1 and FRE2. The AFT1-1up allele causes constitutive expression of all of the genes known to be regulated by this protein (24, 25). This allele had no effect on the regulation of the low affinity system, suggesting that an additional system of iron-responsive regulation exists in S. cerevisiae.
Strong evidence for the role of FET4 as an Fe2+ transporter was also obtained from characterizing mutations in the FET4 gene. We hypothesized that if FET4 was the transporter, it would be possible to isolate mutant alleles of FET4 that alter the kinetic properties of the low affinity system. Of particular interest were mutations that changed the affinity (i.e. apparent Km) of the system for Fe2+ because this parameter is determined by the direct interaction of the transporter with its substrate. An examination of the amino acid sequence of FET4 did not reveal any obvious metal-binding motifs in the hydrophilic regions of the protein. Such motifs, which have been observed for other metal transporters such as CTR1 (45), CCC2 (21), FTR1 (23), and IRT1 (28), may be involved in substrate binding during transport. The observation that FET4 lacks such sequences suggested that initial binding of Fe2+ by this protein could be mediated by ligands located within the transmembrane domains.
Fe2+ is a borderline hard-soft Lewis acid, so potential ligands include oxygen-containing hard Lewis bases as well as sulfur-containing soft Lewis bases (49). Thus, each transmembrane domain contains several potential Fe2+ ligands. For this analysis, we mutagenized aspartate and glutamate residues and the "hydrophobic anion" (50) tyrosine. Aspartates and glutamates were chosen because their negative charge makes them likely candidates for interaction with a cationic substrate. Such amino acids have been implicated in substrate binding by other cation transporters. Based on an analysis similar to ours, the substrate-binding site of the sarcoplasmic reticulum Ca2+-ATPase has been proposed to utilize three glutamates and an aspartate (51, 52). Furthermore, we considered the iron-binding protein ferritin as a paradigm for how an Fe2+ transporter might bind its substrate. Ferritin is a cytoplasmic protein that assembles into a hollow, spherical complex that is capable of taking up Fe2+. This complex has an outer diameter of 130 Å and an inner diameter of 75 Å and the channels through which Fe2+ passes are lined with glutamates. These negatively charged amino acids, present in the consensus sequence RE(G/H)AE, have been implicated in the transport of iron into the protein shell of ferritin (53). Recently, glutamates in a similar sequence motif (REGLE) were found in a potential transmembrane domain of FTR1 and demonstrated to be critical for iron uptake by this permease (23). Although a similar motif is not found in FET4, these observations suggested that negatively charged residues may be important for FET4 Fe2+ binding.
Tyrosine residues in the potential transmembrane domains were also of
interest because these residues can bind iron through interaction of
the dipole moment of the electronegative oxygen in the hydroxyl group
or through a cation- interaction involving the quadrupole moment of
the aromatic ring. It was recently proposed that the relatively
hydrophobic side chains of amino acids with quadrupole moments
(i.e. tyrosine, phenylalanine, and tryptophan) would be
well-suited to serve as cation ligands while embedded in the
environment of a transmembrane domain (50). Furthermore, tyrosines have
been proposed to play a role in the ion selectivity of voltage-gated
K+ channels like the Shaker channel of
Drosophila (54).
Based on these criteria, nine amino acids were chosen for mutagenesis. Measurable Vmax values in this collection of mutants ranged from 30 to 200% of wild type rates. Three alleles, D271A, Y276A, and Y352A, had 13- to >30-fold higher apparent Km values than the wild type. These effects strongly suggest that there is a direct interaction between FET4 and the substrate of the low affinity system. It is very possible that one or more of these residues are ligands for Fe2+ binding during transport. It seems unlikely that all three amino acids are ligands given that D271A and Y276A are predicted to be near the outer surface of the plasma membrane and Y352A is predicted to be near the inner surface of the membrane. Reconciling these data in terms of a single Fe2+ binding site will require a careful analysis of the membrane topology of FET4.
We thank Jon Holy for assistance with the indirect immunofluorescence microscopy, Lee Opresko for pLO-LB, and Gottfried Schatz, Jon Leighton, Andre Goffeau, and Jasper Rine for their generous gifts of some of the antibodies used in this study. We also thank Andy Dancis and Rick Klausner for providing the epitope-tagged CTR1 gene and pT14, and Hui Zhao and Ann Thering for critical reading of the manuscript.