(Received for publication, November 12, 1996, and in revised form, March 24, 1997)
From the Department of Pathology, University of Utah School of Medicine, Salt Lake City, Utah 84132
The FET3 gene product of Saccharomyces cerevisiae is an essential component of the high affinity iron transport system. Based on FET3 sequence homology to the multicopper oxidase family and iron oxidation studies in spheroplasts (De Silva, D. M., Askwith, C. C., Eide, D., and Kaplan, J. (1995) J. Biol. Chem. 270, 1098-1101), it was hypothesized that the Fet3 protein (Fet3p) was a cell surface ferroxidase. To further characterize the protein, we have isolated Fet3p from yeast membranes and purified the protein to apparent homogeneity. Consistent with its localization at the plasma membrane, Fet3p is a glycosylated protein. SDS-polyacrylamide gel electrophoresis analysis showed that the protein was present in two differentially glycosylated forms of approximately 120 and 100 kDa. Purified Fet3p is a copper-containing protein that is able to catalyze the oxidation of a variety of organic compounds in addition to ferrous iron. Azide and metal chelators strongly inhibited enzyme activity. Iron appeared to be the best substrate for the enzyme, and the apparent Km for ferrous oxidation was 2 µM. Interestingly, Fet3p was able to effectively catalyze the incorporation of iron onto apotransferrin. We conclude that Fet3p is a ferro-O2-oxidoreductase in yeast, homologous to the human plasma protein ceruloplasmin.
Saccharomyces cerevisiae utilizes multiple pathways for iron acquisition. The low affinity iron transport system comprising the FET4 gene product also transports other metals in addition to iron (1). The high Km (30 µM) of the FET4 transporter for iron restricts the activity of this system to iron-rich growth conditions. A high affinity (Km = 0.15 µM) iron transport system is induced in response to iron deprivation and is extremely specific for iron. This transport system is coded for by two genes, FET3 and FTR1. Expression of these genes is necessary and sufficient for iron transport. Mutations in either gene abrogate transport, and forced iron-independent expression of both genes allows transport (2, 3). Both the high and low affinity transport systems depend on an active plasma membrane ferrireductase complex (4, 5) to reduce ferric to ferrous prior to iron transport.
Studies indicate that Fet3p1 and Ftr1p act as a ferroxidase-permease complex. The role of Fet3p as an oxidase is suggested by two lines of data. First, sequence analysis of FET3 shows homology to the multicopper oxidase family of proteins that includes ascorbate oxidase, plant and fungal laccases, and the mammalian enzyme ceruloplasmin (6). These proteins catalyze the 4-electron reduction of dioxygen to water coupled to the oxidation of a variety of organic substrates. Enzymatic catalysis of the oxidation of ferrous to ferric ions is unique to ceruloplasmin. This ferroxidase activity of ceruloplasmin is essential for its function in vivo, which is to mobilize tissue iron stores and load iron onto apotransferrin for distribution throughout the body. It was shown recently that the genetic disorder aceruloplasminemia caused severe tissue iron overload in affected individuals (7, 8). Although Fet3p has greater homology to ascorbate oxidase than ceruloplasmin, the essential role of Fet3p in iron transport led us to hypothesize that it has ferroxidase activity (6). The second line of evidence indicating that Fet3p is a ferroxidase is based on the observation that yeasts exhibit a FET3-dependent oxygen consumption in the presence of iron (9). Oxygen consumption was dependent upon functional Fet3p and could be reduced by copper depletion or by the addition of anti-Fet3p antibodies to yeast spheroplasts. Calculation of the stoichiometry of iron oxidation and oxygen consumption indicated a ratio of approximately 4:1 that is consistent with the action of a multicopper oxidase.
In an effort to further define the mechanism of high affinity iron transport, Fet3p has been purified to homogeneity, and its properties have been studied. Purified Fet3p is an oxidoreductase similar to other multicopper oxidases. We further demonstrate that in contrast to most multicopper oxidases, Fet3p has ferroxidase activity. Thus, Fet3p and ceruloplasmin (the only two multicopper oxidases known to be ferroxidases) may be functional homologues. The ability of ceruloplasmin to load iron onto apotransferrin is critical for its in vivo function. We demonstrate that the oxidation of ferrous ions by Fet3p can also be coupled to the loading of apotransferrin. This observation suggests that a stable physical complex between the ferroxidase and the ferric acceptor may not be required for transfer of ferric iron.
The yeast strain used in this study was F113:MATa ura3-52 inol1-1 (4). Yeast extract-peptone-dextrose (YPD) medium contained 1% yeast extract, 2% Bacto-peptone (Difco Laboratories), and 2.5% glucose. Iron-limited medium YPD-bathophenanthrolinedisulfonate (5) was composed of YPD containing 80 µM bathophenanthrolinedisulfonate (Sigma) and 5 µM ferric chloride (Sigma).
Buffers and SubstratesTris-HCl and NaCl buffers were used throughout the purification of Fet3p and were supplemented with the following protease inhibitors: leupeptin (20 µM) and pepstatin A (10 µM), which were obtained from Sigma, and AEBSF (1 mM) from ICN. For Fet3p separation on Cu HiTrap columns and for oxidase assays, sodium acetate buffer (100 mM) was used at the indicated pH with no protease inhibitors present. Ferrous ammonium sulfate, p-phenylenediamine dihydrochloride, epinephrine, dopamine, and apotransferrin were obtained from Sigma and used as described in the legends to the figures.
Protein Purification and Characterization40-70 liters of
the F113 strain of S. cerevisiae were grown to saturation
overnight in iron-limited medium, YPD-bathophenanthrolinedisulfonate (5), containing 2.5% glucose. Cells were harvested and washed with
lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM
NaCl). A volume of glass beads equal to the volume of the pellet was
added to the cells, which were frozen at 20 °C. Typically, between
500 and 800 g of yeast cells (wet weight) were used per
purification. Frozen cells were thawed on ice and disrupted by
vortexing for 30 s at maximum setting, followed by cooling on ice
for 30 s. This procedure was repeated seven times. Homogenates
were diluted 5-fold in lysis buffer and centrifuged 5 min at 4000 × g to pellet glass beads and cell debris. Supernatants
were removed to fresh tubes, and membranes were pelleted by
centrifugation at 4 °C for 30 min at 20,000 × g.
Pellets were washed once with lysis buffer and recentrifuged for 10 min
at 20,000 × g. The pellets were solubilized in
approximately 3 volumes of lysis buffer containing 1% Triton X-100
overnight on ice at 4 °C. Samples were clarified by
ultracentrifugation at 4 °C for 30 min at 100,000 × g using a SW28 rotor in a Beckman Ultracentrifuge.
Supernatants were pooled, filtered (0.45-µm cellulose acetate filter,
Costar Corp.), and concentrated using an Amicon concentrator through a
YM30 membrane. The concentrated sample was chromatographed on a
Superdex 200 gel filtration column (HR10/30, Pharmacia Biotech Inc.)
using 25 mM Tris-HCl, pH 7.4, 50 mM NaCl, 1%
Triton X-100 as the mobile phase. Fractions were analyzed for Fet3p by
monitoring oxidation of p-phenylenediamine dihydrochloride (pPD). Briefly, 20 µl of 1.5-ml fractions were added to 200 µl of
100 mM sodium acetate buffer, pH 5.0, containing 0.05% pPD in a 96-well microtiter plate. The increase in absorbance was monitored
at 570 nm using a kinetic microplate reader (Molecular Devices).
Fractions containing pPD activity were pooled and concentrated and then
applied to a fast protein liquid chromatography ion exchange Mono Q
column (HR 5/5, Pharmacia) equilibrated with 25 mM
Tris-HCl, pH 7.4, 50 mM NaCl, 1% Triton X-100. Fractions
were eluted using a salt gradient from 50 mM to 500 mM, with holding steps at 100 mM NaCl and 200 mM NaCl. Fet3p eluted during the 200 mM NaCl
holding step. Again, pPD oxidation was used to identify
Fet3p-containing fractions that were pooled, concentrated (Amicon
Centriprep 30), and applied to a HiTrap affinity column (Pharmacia)
charged with CuSO4 according to the manufacturer's
instructions and equilibrated with 100 mM sodium acetate
buffer, pH 7.2, 500 mM NaCl, 1% Triton X-100. Fractions
were eluted using the same buffer at pH 3.5 and immediately neutralized
with 10 µl of 1 M Tris-HCl, pH 7.0. Fet3p-containing fractions were desalted using a Sephadex G-25 column (Pharmacia) and
used for further assays. Protein concentrations were determined using
the bicinchoninic microassay (Pierce). For enzyme inhibition studies
EDTA was obtained from Sigma, and bathocuproinedisulfonic acid was
obtained from Aldrich. Amino acid analysis was performed to quantify
Fet3p levels after hydrolysis of samples in 5.7 M HCl at
110 °C in vacuo for 24 h on a Beckman model 6300 analyzer. Atomic absorption analyses were performed on a Perkin-Elmer
model 305A instrument.
SDS-PAGE was performed using 4-20% gradient gels (Bio-Rad). Deglycosylation was carried out by incubating samples in endoglycosidase H (New England Biolabs) according to the manufacturer's instructions. For Western blot analysis, proteins separated by SDS-PAGE were transferred to polyvinylidene difluoride membrane (Gelman Instrument Co.) using a Bio-Rad Trans-Blot cell transfer apparatus. Fet3p was detected using an antibody generated to a predicted peptide sequence at the C-terminal end of Fet3p as described previously (9).
Fet3p Oxidase AssaysAll oxidase assays were performed in
100 mM sodium acetate, pH 5.0. As an assay of ferroxidase
activity Ferrozine, a ferrous specific chelator, was employed.
Fet3p-catalyzed Fe(II) oxidation was monitored as follows: 0.2-ml
reaction mixtures containing purified Fet3p were incubated in 96-well
microtiter plates at room temperature. Reactions were started by the
addition of ferrous ammonium sulfate at specified concentrations and
quenched at 2-min intervals by the addition of 50 µl of 15 mM Ferrozine. Fe(II) remaining in solution was determined
using the molar absorption constant 570 nm = 27.9 mM
1. p-Phenylenediamine oxidation
was assayed in 0.2-ml reaction volumes, and oxidation of pPD was
monitored by following an increase in absorbance at 570 nm as described
above. Detection of Fet3p oxidase activity in nonreducing
polyacrylamide gels was performed as described previously (10).
Partially loaded Fet3p was obtained as described above without
supplementing the lysis buffer with copper, and 2 µg of purified
protein was incubated with 50 µM CuSO4 in the
presence or absence of 1 mM sodium ascorbate in a 20-µl
volume at room temperature for 10 min. Samples were denatured in 1%
SDS and loaded onto a 4-20% SDS-polyacrylamide gel (Bio-Rad), and
renatured gels were incubated for 1 h in 100 mM sodium
acetate, pH 5.0, containing 0.05% pPD. No azide was included in the
buffer.
Yeast cells grown to saturation in iron-limited medium (YPD-bathophenanthrolinedisulfonate (5)) were harvested and disrupted by vortexing with an equal volume of glass beads. It was important to keep protease inhibitors present in all buffers and to carry out purification steps at 4 °C as rapidly as possible. Membrane fractions were obtained by high speed centrifugation, and Fet3p was solubilized using 1% Triton X-100. Fractions containing Fet3p were followed by assaying for pPD oxidase activity as described under "Materials and Methods." The steps involved in the purification of Fet3p from S. cerevisiae are summarized in Table I. Quantification of total units of Fet3p activity in crude homogenates is relatively imprecise due to difficulties in detecting activity in this fraction. We estimate, however, that the enzyme was purified approximately 15,000-fold relative to the crude extract with a recovery of approximately 9.0%. The purity of the enzyme was confirmed by silver staining (Fig. 1A). It was important that samples denatured in SDS were not heated above 55 °C. Boiling resulted in aggregation of Fet3p, which prevented the protein from entering polyacrylamide gels. Gel electrophoresis of the purified enzyme revealed two closely spaced bands with molecular masses of approximately 100 and 120 kDa. Western analysis using an antibody generated to a predicted peptide sequence at the C-terminal of Fet3p showed that both bands represented Fet3p (Fig. 1B). Treatment of the sample with endoglycosidase H reduced both bands to one band with a molecular mass of approximately 85 kDa, indicating that these bands represented differentially glycosylated Fet3p (Fig. 1B). Analysis of Fet3p activity in detergent-solubilized membrane extracts (by either molecular sieve chromatography or glycerol gradients (not shown)) indicated that Fet3p eluted at a molecular mass between 100 and 150 kDa. This is similar to the molecular mass observed by SDS-PAGE and is consistent with the size of the predicted glycosylated protein indicating that (at least in the presence of detergent) Fet3p is a monomer and is not tightly associated with any other proteins.
|
The final homogeneous preparation of the enzyme lacked the
characteristic blue color of type 1 copper-containing enzymes, and the
absorbance spectra of the Fet3p did not exhibit the typical type 1 copper absorbance peak at 610 nm. The low concentration of our purified
Fet3p preparation (1-5 µM) may explain the lack of
absorbance in the 600 nm range. Typical extinction coefficients for
type 1 copper centers are between 2 and 8 mM1
cm
1 (11). The predicted copper ligands for the Fet3p
mononuclear site are two histidines, a cysteine, and a leucine. Typical
type 1 "blue" copper centers use methionine instead of leucine
(12). Although the blue color of these proteins is attributed to a
charge transfer between copper and cysteine, lack of methionine as a mononuclear copper ligand in Fet3p may weaken the absorption in the 600 nm range.
To determine if the purified protein did contain copper, the preparation was subjected to amino acid analysis, and the copper content was determined using atomic absorption spectroscopy. A copper to protein ratio of 1.5 ± 0.1 copper atoms/molecule was determined. Assuming a copper:protein ratio of 4:1 as predicted by sequence analysis, this result suggests that less than 40% of the purified preparation contains holoFet3p. Loss of copper could have occurred during the purification of Fet3p or could reflect the ratio of apo- to holoprotein in vivo. In an attempt to determine whether the low copper:protein ratios reflect a mixture of apo- and holoFet3p in our preparations, an excess of copper sulfate (50 µM) was included in the initial homogenization step during purification of the protein. Addition of copper to cell homogenates resulted in a purified Fet3p preparation with increased activity and a copper:protein ratio of 3.79 ± 0.3. Incubation of partially copper-loaded purified Fet3p samples with copper sulfate and ascorbate increased the activity of the protein significantly, and this increase in Fet3p activity was not observed if ascorbate was omitted (Fig. 1C). These results indicate that holoFet3p contains four copper atoms/molecule as predicted for a multicopper oxidase and that reduced copper is easily incorporated into the purified apoprotein to generate the active enzyme without the need for any additional molecules.
Ferroxidase Activity of Purified Fet3The oxidation of Fe(II)
by Fet3p was assayed by monitoring the rate of disappearance of ferrous
ammonium sulfate using the ferrous chelator Ferrozine. The reactions
were carried out in 100 mM sodium acetate at pH 5.0 in the
presence of 100 µM citric acid. The low pH of the buffer
minimized the autoxidation of Fe(II), and citrate was included to
prevent the formation of insoluble ferric hydroxides that could bias
the rates of Fe(II) oxidation. Ferroxidase activity is not seen with
ascorbate oxidase (data not shown) or heat-inactivated Fet3p. Initial
kinetic experiments were done with preparations of partially loaded
Fet3p isolated from cells without copper supplementation. As indicated
in Fig. 2A, Fet3p exhibited Michaelis-Menten
type kinetics for the oxidation of ferrous ions. When an identical
concentration of fully loaded Fet3p was used in the same assays, the
apparent Km (estimated to be 2 µM)
remained unchanged; however, the Vmax was
increased from 1 µM/min for the partially loaded samples
to 1.8 µM/min for the fully loaded samples (Fig.
2B).
Substrate Specificity
Fet3p is capable of oxidizing a number of organic substrates. Typically, multicopper oxidases can oxidize a variety of amines, catechols, and phenolic substrates such as phenylenediamine (13). p-Phenylenediamine oxidase-activity was used as an assay for Fet3p throughout our purification protocol. The apparent Km values for o-phenylenediamine, p-phenylenediamine, and epinephrine are indicated in Table II. The rate of Fet3p-catalyzed epinephrine oxidation was extremely slow, and we were unable to detect any Fet3p-dependent dopamine oxidation. It has been suggested that the ability of ceruloplasmin to catalyze the oxidation of amines reflects contamination of preparations with adventitious transition metals (14). The lack of adventitious metal contamination of the Fet3p preparation may explain the lack of oxidation of these substrates. In comparison with other compounds tested, it is clear that Fe(II) is by far the best substrate for Fet3p.
|
Azide has been shown
to inhibit the activity of multicopper oxidases. This inhibition has
been attributed to interference with the catalytic copper sites of
these enzymes, specifically inhibition of type 1 or mononuclear copper
sites (14). The ferroxidase activity of Fet3p was inhibited 75-80% by
the addition of 1 mM azide to the assay buffer (Fig.
3A), and complete inhibition of enzyme
activity was observed with 10 mM azide (Table
III).
|
The addition of either 1 mM EDTA or 1 mM bathocuproinedisulfonic acid (a copper chelator) inhibited the Fet3p-catalyzed oxidation of pPD by 60% and 40%, respectively (Table III). This was not a result of the chelation of adventitious metal, since heat inactivation of the enzyme preparation abolished all oxidase activity. We conclude that the loss of a labile, catalytically active copper by chelation may be responsible for the loss of activity. Fet3p exhibited maximum ferroxidase activity at pH 5.0, which is similar to the typical growth environment of yeast (pH 4.5-5.0). Although the autoxidation of ferrous iron is higher at alkaline pH, the enzyme-catalyzed rate is easily distinguished (Fig. 3B).
Incorporation of Fe(II) into Apotransferrin by Fet3pAn
important function of ceruloplasmin in vivo is the
incorporation of Fe(II) into apotransferrin, an activity dependent on the ferroxidase activity of the protein. The interaction between FET3 and FTR1 in S. cerevisiae
suggests that the ferroxidase activity of Fet3p may be required to load
iron onto Ftr1p. We examined the ability of Fet3p to catalyze the
incorporation of Fe(II) onto apotransferrin. This experiment was
performed to determine if ferroxidase-dependent iron
loading required a specific interaction between the ferroxidase and
acceptor. The data in Fig. 4 indicate that Fet3p is
capable of iron loading transferrin in a specific concentration-dependent manner, albeit at a much slower
rate than ceruloplasmin, (7-10 µM Fe3+
transferrin/min/mM enzyme versus 550 µM Fe3+ transferrin/min/mM
ceruloplasmin (15)). The amount of Fe(III)-transferrin formed was
calculated using the molar absorption constant of
460 nm = 2500 M
1
cm
1/Fe(III)-transferrin formed (15). The inclusion of
ascorbate in the transferrin-loading assay did not interfere with the
enzymatic rate of Fe(III)-transferrin formation by Fet3p.
The essential components of the S. cerevisiae high affinity iron transport system consist of the products of the FET3 and FTR1 genes (2, 6). It was hypothesized that the protein products of these genes function as a ferroxidase and a membrane transporter, respectively. Evidence that suggested that Fet3p is a ferroxidase included its homology to the multicopper oxidase family (6), a copper requirement for high affinity iron transport (6, 15), and the demonstration that a functioning Fet3p confers iron-dependent oxygen consumption in yeast (9). The present study confirms the identity of Fet3p as a ferroxidase; Fet3p purified to apparent homogeneity contained copper and demonstrated ferroxidase activity in vitro.
Two discrete differentially glycosylated forms of Fet3p were identified in our purified enzyme preparation. These two bands can also be detected in whole cell preparations analyzed by SDS-PAGE. This observation suggests that the two forms seen in our purified preparation were not generated during the isolation process. Examination of the copper content of purified Fet3p indicates that, assuming a 4:1 stoichiometry between copper and protein, under normal conditions only about 40% of the protein is in the holoform. While loss of copper could have occurred during the isolation procedure, there is another alternative, which is that the apoform represents an intermediate in the biosynthesis of Fet3p. Fet3p could be easily reconstituted by the addition of reduced copper to the purified preparation to yield a fully active protein. Therefore, copper incorporation into Fet3p probably does not require the participation of any other molecules except reduced copper. The facile addition and removal of copper from Fet3p is in contrast to other multicopper oxidases such as ceruloplasmin where copper removal and reconstitution require stringent conditions (16). Differential glycosylation of Fet3p may reflect localization of this protein in different parts of the secretory pathway. The addition of copper to apoFet3p is thought to occur in a Golgi or post-Golgi compartment (17). Unless Ftr1p is present, apoFet3p is thought to be retained within this intracellular compartment (2). Our data are consistent with a significant amount of apoFet3p being retained in a subcellular compartment and may account for the low copper content of purified Fet3p isolated without copper supplementation. It is possible, therefore, that high affinity iron transport is limited by the amount of Ftr1p available to form a complex with Fet3p. Two observations support this hypothesis. First, while FET3 mRNA concentration increases with the severity of iron limitation, iron transport per se reaches a maximum and does not increase further (6). Second, we observed that galactose-dependent expression of FET3 and FTRI will result in an increase in high affinity iron transport independent of iron availability. Higher levels of transport were observed if endogenous FTR1 and galactose-driven FTR1 expression were combined, suggesting that iron transport is limited by Ftr1p availability (3).
Among the multicopper oxidases, Fet3p and ceruloplasmin have so far proved unique in possessing the ability to catalyze the oxidation of ferrous to ferric. Although the oxidation of ferrous iron occurs spontaneously in solution, the involvement of a multicopper oxidase in the enzymatic conversion of Fe(II) to Fe(III) is important because it prevents the formation of toxic reactive oxygen intermediates. In S. cerevisiae, functional Fet3p is localized to the plasma membrane with the copper domain on the extracellular surface. In mammals, ceruloplasmin catalyzes the oxidation of ferrous iron and incorporation onto apotransferrin in plasma. Analysis of the phenotype of aceruloplasminemic individuals indicates that ceruloplasmin is involved in removal of iron from cells. Given the same enzymatic activity in an extracellular domain, what are the factors that determine vectoral transport? Clearly, the direction of iron uptake results from differences in the acceptor molecule, Ftr1p or transferrin. Two models for the interaction of the multicopper oxidase and the iron acceptor can be proposed. In the first model the specificity of transport is defined by the physical interaction of the ferroxidase and the acceptor. For example, the substrate for the ferroxidase is not iron per se, but Ftr1p-Fe(II). In this scenario the multicopper oxidase oxidizes either the protein or the protein-bound substrate. In an alternative model the ferroxidase may oxidize iron independent of the acceptor; the specificity of the interaction is determined simply by the stochastic availability of the acceptor. Previous studies have demonstrated that ceruloplasmin can exhibit ferroxidase activity in the absence of a specific acceptor (13). Our studies on Fet3p extend that observation by showing that highly purified Fet3p also has ferroxidase activity in the absence of a specific acceptor. Thus, the inherent ferroxidase activity of the enzyme may allow the delivery of iron to a variety of ferric acceptors. As demonstrated here, Fet3p is capable of iron loading apotransferrin in vitro. In an analogous experiment, ceruloplasmin was able to catalyze the incorporation of iron into ferritin in vitro (18, 19). Of interest is that an essential REGLE domain in FTR1, the permease partner of FET3 in S. cerevisiae, is similar to a conserved domain, REGAE, in the ferritin light chain that is essential for iron nucleation and core formation (2). Thus, the interaction between Fet3p and the yeast permease need not be strong, but merely the proximity of the two proteins within the membrane may allow effective transfer of ferric iron between the two. It may be hypothesized that the interaction between ferroxidase and ferric acceptor may be a general physiological mechanism for translocation or storage of iron.
On the basis of genetic experiments, Stearman et al. (2) concluded that a complex between Fet3p and Ftr1p is required for delivery of both molecules to the cell surface. Nevertheless, attempts to demonstrate a physical complex between these two cell surface molecules have not been fruitful. For example, Fet3p behaves as a monomer when analyzed by either molecular sieve chromatography or velocity sedimentation. We recognize that the caveat to this conclusion is that these analytical approaches involve the use of detergent, and perhaps even mild detergents destabilize the complex. Attempts to cross-link the two proteins in situ, while not exhaustive, have also been unfruitful. Our results suggest that a stable high affinity complex between cell surface Fet3p and Ftr1 may not be necessary for iron transport. That is, the transfer of Fe(III) from Fet3p to Ftr1p may occur simply by diffusion-mediated collision between Fet3p and Ftr1p. Examination of the literature has not provided any evidence for a complex between transferrin and ceruloplasmin.
In conclusion, purification and analysis of the yeast Fet3 protein demonstrates that this membrane-bound protein required for high affinity iron acquisition in S. cerevisiae is a ferro-O2-oxidoreductase. The functional homology between Fet3p and ceruloplasmin suggest that these two proteins comprise a unique subset of multicopper oxidases that possess essential ferroxidase activity required for maintenance of iron homeostasis in vivo.