Cation diffusion facilitator transporters are found in all three Kingdoms of life and are involved in transporting transition metals out of the cytosol. The metals they transport include Zn2+, Co2+, Fe2+, Cd2+, Ni2+ and Mn2+; however, no single transporter transports all metals. Previously we showed that a single amino acid mutation in the yeast vacuolar zinc transporter Zrc1 changed its substrate specificity from Zn2+ to Fe2+ and Mn2+ [Lin, Kumanovics, Nelson, Warner, Ward and Kaplan (2008) J. Biol. Chem. 283, 33865–33873]. Mutant Zrc1 that gained iron transport activity could protect cells with a deletion in the vacuolar iron transporter (CCC1) from high iron toxicity. Utilizing suppression of high iron toxicity and PCR mutagenesis of ZRC1, we identified other amino acid substitutions within ZRC1 that changed its metal specificity. All Zrc1 mutants that transported Fe2+ could also transport Mn2+. Some Zrc1 mutants lost the ability to transport Zn2+, but others retained the ability to transport Zn2+. All of the amino acid substitutions that resulted in a gain in Fe2+ transport activity were found in transmembrane domains. In addition to alteration of residues adjacent to the putative metal- binding site in two transmembrane domains, alteration of residues distant from the binding site affected substrate specificity. These results suggest that substrate selection involves co-operativity between transmembrane domains.
- cation diffusion facilitator
- transition metal
Transition metal ions (Cu+/2+, Fe2+, Zn2+, Co2+ and Mn2+) are essential for all organisms as they serve as cofactors for various proteins. They can also be toxic in excess owing to their participation in redox reactions or by their competing with other metals for protein binding sites. Organisms have evolved mechanisms to regulate uptake, delivery, storage and detoxification of these metals. One mechanism that prevents metal toxicity is metal transport out of the cytosol by CDF (cation diffusion facilitator) transporters. These transporters mediate metal resistance by either exporting metals out of cells or into intracellular organelles, thus reducing cytosolic metal concentration. Some examples include Escherichia coli ZitB  and mammalian ZnT1(Slc30a1), which mediate zinc efflux from cells  and Saccharomyces cerevisiae Zrc1 and Cot1, which transport zinc and cobalt into vacuoles [3,4].
The majority of CDF transporters have six TMDs (transmembrane domains) with the N- and C-termini extending into the cytosol. A recent study classified CDF family members into three major groups (Zn2+, Fe2+/Zn2+ and Mn2+) based on substrate specificity . Structural studies of the putative Zn2+ transporters CzrB from Thermus thermophilus  and YiiP (also known as FieF) from E. coli  indicate that the metal-binding sites of the transporter are formed by charged residues in TMDs II and V. The importance of amino acids in determining substrate specificity has been examined by sequence comparisons of different transport groups and by site-specific mutagenesis. Mutations that abrogate metal transport provide little information on substrate selection, because such mutations may affect transport activity independently of substrate selection. For example, loss of function may be due to alteration of residues involved in metal binding, as demonstrated by the study of YiiP(D157A) , or due to residues involved in stabilizing the protein structure, as implicated by the study of ZitB(E214A), CzcD(H237R) and CzcD(H280A) [8,9]. In contrast, amino acid mutations that lead to a gain of function by changing the metal transported, by their very existence, indicate the involvement of that amino acid in substrate selection. In this regard, gain-of-function mutations are informative in identifying residues involved in substrate selection.
We identified a mis-sense mutation in the yeast vacuolar Zn2+ transporter Zrc1 that resulted in the loss of Zn2+ transport and the acquisition of Fe2+ and Mn2+ transport activity . The finding that one amino acid substitution can change transport specificity provides strong evidence for a role of that residue in substrate selection. Mis-sense mutations that result in a gain, rather than a loss, of function provide a robust approach for examining substrate specificity in CDF transporters. We combined a strong selection, the ability of cells with a deletion in the gene that encodes the vacuolar iron transporter CCC1 to grow on high iron medium, with random and directed mutagenesis of ZRC1 to identify residues that alter the transport specificity of Zrc1. We identified gain-of-function mutations that give rise to altered metal specificity. Our studies show that amino acids near the putative metal-binding amino acids in TMD II, as well as amino acids in other TMDs that are distant from the metal-binding residues, are important in substrate selection.
Yeast strains and growth media
The following yeast strains (W303 background) were used: wild-type DY150 [Mat a ade2-1 his3-11 leu2-3,112 trp1-1 ura3-52 can1-100(oc)]. Δccc1 and Δzrc1 strains were generated in the DY150 background by double-fusion PCR, using the HIS3 gene as a selectable marker as described in . Strains with a FET3-lacZ reporter integrated at the HO locus were constructed as described in . Wild-type BY4743 (Mat a/α his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 lys2Δ0/+met15Δ0/+ura3Δ0/ura3Δ0) and Δpmr1::KanMX strains in the BY4743 background were obtained from Research Genetics.
Yeast strains were grown in YPD medium (1% yeast extract, 2% peptone and 2% dextrose) or in CM medium (0.67% yeast nitrogen base without amino acids, 2% dextrose and 0.13% amino acid drop-out mix). Plates with high concentrations of metals were made by adding either ferrous ammonium sulfate, zinc sulfate or manganese chloride. Liquid media was supplemented with FeSO4 to the indicated concentrations, with a final concentration of 1 mM ascorbic acid.
Transposon mutagenesis of Δccc1
A mTn-lacZ transposon insertion library from Professor Michael Snyder (Department of Molecular, Cellular and Developmental Biology, Yale University, New Haven, CT, U.S.A.) was used for mutagenesis. Three separate library pools were digested with NotI and transformed into Δccc1 cells. Cells were replicate plated on to medium containing 3 mM iron, and mutants able to grow on high iron medium were selected for further study. The transposon target site was identified by sequencing as described in .
Error-prone PCR and isolation of mutants
Wild-type ZRC1 with its own promoter and 3′ end was generated by PCR from genomic DNA using Pri54 and Pri60 as primers (Table 1). The PCR fragment was digested with BglII and KpnI and inserted into pRS426 (a yeast URA3-containing episomal vector) to generate ZRC1/pRS426. This plasmid was used as the template for error-prone PCR with PR0064 and Pri56 as primers. Mutations were introduced by the PCR error of Taq DNA polymerase (Fisher Scientific). The PCR conditions were: 95 °C for 2 min, 35 cycles of (95 °C, 30 s; 55 °C, 30 s; 72 °C, 2 min) and 72 °C for 10 min. Four independent PCR products were pooled from the first round of PCR and used to seed six independent PCRs for the second round of PCR. The PCR products from the second round were pooled and transformed together with a gap plasmid (ZRC1/pRS426 digested with HindIII and KpnI and the large fragment was purified) into Δccc1 cells. Cells were grown for two days and then replica plated to CM-Ura (complete minimal mix minus uracil) plates containing 5 mM ferrous ammonium sulfate. Cells able to grow on high iron were selected for further study. Plasmids rescued from the high iron-resistant colonies were sequenced using primers pri86, pri87 and pri88, which covered the ZRC1 coding region.
The generation of ZRC1/pYES2, a His6-tagged version of ZRC1 under the control of the galactose-inducible promoter GAL1, was described previously . Site-directed mutagenesis using ZRC1/pYES2 as a DNA template was performed using the QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. Primers used in the present study are listed in Table 1. All mutations were confirmed by DNA sequencing.
Western blot, β-galactosidase activity and iron content
Cells were disrupted with glass beads in the presence of protease inhibitors (1.0 mM PMSF, 10 μM pepstatin A and 20 μM leupeptin) (Sigma). The supernatant from a low speed centrifugation was solubilized using SDS/PAGE sample buffer. Samples (20 μg protein) were subjected to SDS/4–20% PAGE, transferred to nitrocellulose, and probed with rabbit anti-(His6 tag) (1:2000, Abcam) or mouse anti-CPY (carboxypeptidase Y) antibody (1:4000, Invitrogen), followed by peroxidase-conjugated goat anti-rabbit or goat anti-mouse antibody (1:10000, Jackson ImmunoResearch). Membranes were developed using chemiluminescence reagents (PerkinElmer Life Sciences). β-Galactosidase activity was performed in 96-well plates using o-nitrophenyl β-galactoside as a substrate . Protein concentration was determined by the bicinchoninic acid method (Pierce) using BSA as standard. Vacuoles were isolated as described previously in . The iron content of isolated vacuoles was determined using a PerkinElmer ICP (inductively coupled plasma)-Optical Emission Spectrometer with a standard curve generated from mixed metal standards.
Immunofluorescence was performed as described in . Cells were incubated with a rabbit anti-(His6 tag) (1:100, Abcam) overnight at room temperature (20 °C) in a humid chamber, and then incubated with an Alexa-594-conjugated goat anti-rabbit antibody (1:750; Invitrogen) for 2 h at room temperature.
Random mutagenesis results in Zrc1 mutations that have gained iron transport activity
Zrc1 is a yeast vacuolar metal transporter that protects cells from the toxic effect of high concentrations of Zn2+ [3,16]. We showed that a single amino acid change (N44I) in Zrc1 dramatically altered its substrate specificity from Zn2+ to Fe2+ and Mn2+ . As a consequence, expression of Zrc1 (N44I) was able to rescue the high iron growth defect of Δccc1 cells by transporting and storing iron in the vacuoles. Zrc1(N44I) was identified through a UV-mutagenesis screen that selected for Δccc1 mutant cells able to grow on high iron medium. We also utilized a transposon-generated library  to identify high iron-resistant mutants of Δccc1 cells. From 15000 colonies, two of the transposon-containing mutants, 22-3 and 25-1, were able to grow on high iron (Figure 1A). These mutants were found to have an insertion of the transposon upstream of ZRC1 (Figure 1B). The insertion point of the transposon in each mutant was different, demonstrating that they were independent events. The mutants, although iron resistant, were unable to grow on high Zn, suggesting an alteration in Zrc1 substrate specificity (results not shown). Sequencing of ZRC1 in each of the mutants revealed base changes resulting in an amino acid substitution, although different in each case. Mutant 22-3 showed an A to T nucleotide change at position 131, resulting in a N44I amino acid substitution in Zrc1. The second mutant, 25-1, showed a G to T nucleotide change at position 99, resulting in a L33F amino acid substitution in Zrc1. Thus two different ZRC1 mutations lead to a high iron-resistant phenotype. The topology of Zrc1 and the location of the amino acid substitutions within Zrc1 are shown in Figure 2.
Other mutations in Zrc1 that resulted in a change in metal substrate specificity were discovered through a second genetic screen. Previously, we identified MRS3 and MRS4 as high-copy suppressors of iron toxicity in a Δccc1 strain . To identify other high-copy suppressors of poor growth of Δccc1 on high iron and circumvent the identification of MRS3 and MRS4, we generated a high-copy genomic library from Δmrs3Δmrs4 cells. The genomic library was transformed into Δccc1 cells, and cells capable of growth on high iron medium were selected. As expected, we identified plasmids containing CCC1 (identified by colony PCR) as suppressors that conferred high iron resistance (results not shown). We also identified two high copy suppressors that contained ZRC1 with T to C nucleotide mutations at positions 119 and 120, resulting in the same mis-sense mutation F40S in Zrc1.
Directed evolution of Zrc1 reveals that different mutations lead to acquisition of iron transport activity
To identify other Zrc1 mutants that were able to transport iron, the ZRC1 coding sequence was randomly mutagenized through error-prone PCR. The PCR products were transformed together with a gap plasmid into Δccc1 cells. We performed PCR mutagenesis on two independent preparations of ZRC1. Transformed colonies were replica plated on to medium containing 5 mM ferrous ammonium sulfate. Colonies (50000) were screened and resistant colonies (250) were identified. Plasmids from 62 resistant colonies were rescued, their ability to support the growth of Δccc1 cells reconfirmed and rescued plasmids were sequenced. Many mutant ZRC1 plasmids showed more than one base change, but often the second base change was silent, as it did not result in a change in amino acid sequence (Table 2). We identified a number of mutants that had the same amino acid substitution (e.g. F40S). We know that those mutants were independently generated, as some mutants had a silent base change in addition to the mis-sense mutation (compare pZm165 with pZm225). We also identified the same amino acid substitution occurring by itself and in conjunction with other amino acids (compare pZm165 with pR3 and pR26). We also identified different substitutions of the same amino acid (L33S/L33F and I275N/I275F). These observations suggest that the mutagenesis was extensive; however, as shown below, we do not think it was saturated.
We first focused on plasmids in which there was only one amino acid substitution. All the ZRC1 mutants were able to protect Δccc1 cells from high iron toxicity (Figure 3A). For comparison purposes we included ZRC1(D45A), which we had generated previously . Aspartic acid in the corresponding position in the E. coli YiiP is a critical zinc-binding residue . Zrc1(D45A) does not have Zn2+ or Fe2+ transport activity. To further demonstrate that the mutant proteins altered cellular iron metabolism, they were overexpressed in wild-type cells containing an iron-sensitive reporter FET3-lacZ. Overexpression of Ccc1, by transporting iron into the vacuoles, lowers cytosolic iron, resulting in induction of FET3-lacZ . Overexpression of wild-type Zrc1 did not change the expression level of this reporter; however, all Zrc1 mutants were able to induce FET3-lacZ, indicating their ability to lower cytosolic iron, although less efficiently than Ccc1 (Figure 3B). Consistent with loss of transport activity, Zrc1(D45A) had no effect on expression of FET3-lacZ. Previously, we demonstrated that Zrc1(N44I) protects Δccc1 from high iron toxicity by transporting iron into vacuoles . Overexpression of selected mutant Zrc1 proteins increased vacuolar iron content in Δccc1 cells relative to cells transformed with a vector control or with wild-type ZRC1 (Figure 3C).
Two amino acid substitutions in Zrc1 are less efficient in transporting iron than a single amino acid substitution
To test whether combining amino acid substitutions might result in an additive effect generating a more efficient Fe2+ transporter, three independently isolated mutants in TMD II (L33F, F40S and N44I) were chosen to generate double mutants. Zrc1 with two amino acid substitutions was not more efficient in Fe2+ transport activity than Zrc1 with a single substitution, as assayed by growth on high iron medium (Figure 3A) and induction of FET3-lacZ (Figure 4A). We note that Zrc1(L33F, F40S) showed a low induction of FET3-lacZ, yet conferred high iron resistance when assayed on plates (Figure 3A). This difference may be ascribed to a difference in affinity, as the induction of FET3-lacZ was performed in moderate iron medium, whereas the plate assay utilized high iron medium. Zrc1(L33F, N44I) conferred less protection from high Fe2+ toxicity than the single substitution mutants, and its ability to induce the low iron sensor FET3-lacZ was also reduced. It is possible that some of the double amino acid substitutions give rise to an unstable Zrc1. To test this, wild-type Zrc1, single mutants and double mutants were epitope tagged at the C-terminus with His6 and expressed using the GAL1 promoter. The expression levels of single mutants and double mutants were similar to that of the wild-type Zrc1 (Figure 4B). We also tested the possibility that the double-mutant proteins were mistargeted and did not accumulate in the vacuole. Immunofluorescence showed that all Zrc1 mutants were targeted to the vacuole (Figure 4C). These data suggest that combining some amino acid substitutions in Zrc1 did not have an additive effect on iron transport.
Gain of iron transport activity in Zrc1 mutants does not necessarily mean the loss of intrinsic zinc transport activity
As described previously, Zrc1(N44I) gained Fe2+ transport activity but lost its inherent Zn2+ transport activity . Loss of Zn2+ transport is not, however, sufficient to protect Δccc1 from high concentrations of iron. Zrc1(D45A) does not have Zn2+ transport activity and does not protect Δccc1 cells from high iron . In our previous study , we used a low copy plasmid to test Zn2+ transport activity of mutant Zrc1. In the present study we used a high copy plasmid to determine if overexpressed mutant Zrc1 retained any Zn2+ transport activity, as assayed by protection of Δzrc1 cells from high zinc toxicity. Some mutants (L33F and L87H) clearly lost Zn2+ transport activity, however, others (F40S, A52T, G79S, F86S, R101G and I275F) did not show significant differences from the wild-type Zrc1 in protecting Δzrc1 cells from zinc toxicity (Figure 5). Other mutants of Zrc1 (L33S, N44I, S272P and I275N) showed decreased Zn2+ transport activity compared with wild-type Zrc1. It is interesting that the double mutant Zrc1(F40S, N44I) lost Zn2+ transport activity, whereas individually each of the single mutants had Zn2+ transport activity.
Zrc1 mutants that transport iron also transport manganese
The Golgi membrane P-type ATPase Pmr1 is responsible for transporting Ca2+ and Mn2+ into the Golgi . Δpmr1 cells accumulate Mn2+ in the cytosol and show poor growth on high concentrations of Mn2+ . Overexpression of Ccc1, by transporting Mn2+ into the vacuoles, protects Δpmr1 cells from toxic amounts of manganese in the environment. To test whether the Zrc1 mutants were able to transport Mn2+ into vacuoles and detoxify it, Δpmr1 cells were transformed with these mutant constructs and grown on plates with 3 mM Mn2+. All the mutants showed protection of Δpmr1 cells compared with wild-type Zrc1 (Figure 6), although some mutants were less efficient than others. Surprisingly, although the D45A substitution lost both Fe2+ and Zn2+ transport activity, it was able to partially rescue the growth defect of Δpmr1.
Effect of amino acid substitutions in TM2 on substrate selection
The sequence adjacent to the putative metal-binding residues in TMD II is different in different CDF family members. These amino acids were hypothesized to determine the substrate specificity of CDF families. We identified two residues, Phe40 and Asn44, adjacent to the putative metal-binding residues His41 and Asp45 in TMD II, which, when mutated, resulted in iron transport activity (F40S and N44I). The F40S mutation occurred quite often and no other mutations of Phe40 were found. One interpretation of this result is that the hydroxy group of the introduced serine residue is required for Fe2+ transport. To test this possibility, we determined whether other substitutions in this amino acid would affect substrate selection. We generated Zrc1(F40T) and Zrc1(F40A) and examined their effect on metal resistance. Zrc1(F40S) and Zrc1(F40T) were able to protect Δccc1 cells from high iron toxicity (Figure 7A). Complementation by F40A was less efficient than for F40S or F40T, but was still noticeable (note vector alone). This result indicates that the hydroxy group at position 40 enhances, but is not absolutely required for, Fe2+ transport. All the Phe40 mutants retained Zn2+ transport activity, with Zrc1(F40S) having slightly decreased activity (Figure 7B). Zrc1(F40S) protected Δpmr1 cells from Mn2+ toxicity to a much greater extent than F40T (Figure 7C). The fact that F40T did not lead to Mn2+ transport shows that it is not simply the presence of a hydroxy group that permits transport.
A hydrophobic amino acid substitution in Zrc1 at position 44 (N44I) resulted in Fe2+ transport activity. This residue is adjacent to the putative metal-binding residue Asp45 and may alter the structure of the metal-binding site. To examine whether the size of the side chain at this position affects substrate selection, we generated N44I, N44A and N44V mutations. All Asn44 substitutions were able to protect Δccc1 cells from iron toxicity (Figure 7A). Zrc1(N44A) retained Zn2+ transport activity, whereas Zrc1(N44I) and Zrc1(N44V) did not (Figure 7B). All three mutants were able to transport Mn2+, as shown by their ability to protect Δpmr1 cells from Mn2+ toxicity (Figure 7C). A summary of phenotypes for all mutant Zrc1 isoforms is detailed in Table 3.
Residues that impart metal specificity to CDF transporters have been identified through site-specific mutagenesis, sequence comparisons and structural studies . Sequence comparisons of CDF transporters implicated residues in TMDs II and V as critical determinates of metal specificity. The crystal structure of the only CDF member resolved, YiiP, shows a substrate-binding site formed by residues from TMDs II and V . Recombinant YiiP has Zn2+ transport activity in vitro and the crystal structure shows bound zinc . Confusing the issue is that YiiP appears to protect cells from iron toxicity , but no iron was found bound to the protein. Mutation of aspartic acid residues in TMD II (Asp49) and TMD V (Asp157) affects Zn2+ transport activity and metal binding . These data suggest that these aspartic acid residues are important in metal binding, but it is not clear what their role is in metal selectivity.
Gain-of-function mutations in ZRC1 that alter substrate selection provide insight into the role of specific amino acids in metal specificity. Gain-of-function ZRC1 mutations were identified through their ability to suppress the phenotype of poor growth of Δccc1 cells on high iron medium. Previously, we demonstrated that suppression of the high iron toxicity phenotype resulted from the transport of iron into the vacuole . We confirmed that result with surrogate assays for cytosolic iron, including the expression of iron-sensitive reporter constructs FET3-lacZ. We have not been able to successfully measure the transport of 59Fe into vacuoles. To date there are only two published papers that measured transition metal uptake into vacuoles in vitro: one on cadmium  and one on zinc . There are no published studies of direct measurements of iron transport into isolated organelles, such as mitochondria or vacuoles. Measurement of direct iron transport into organelles is problematic, as iron is ‘sticky’ and non-specific binding of iron to organelles is so high that specific transport cannot be assessed. Published studies measuring iron transport into mitochondria have utilized surrogate assays such as iron incorporation into haem or the activity of iron-sulfur-containing enzymes for just that reason (for example [22–24]). The measurement of iron-dependent toxicity is an indirect measure of iron transport activity. It is of interest that the only high copy suppressors of iron-dependent growth toxicity we identified are organelle iron transporters: the vacuolar iron transporter Ccc1 , the yeast mitochondrial transporters Mrs3 and Mrs4 , Rim2 (L. Li, unpublished work) and mutant vacuolar zinc transporters . No other genes have been found to suppress high iron toxicity, giving us confidence that our toxicity assays are indeed measuring the activity of iron transporters.
The strength of this phenotype permitted facile identification of rare mutations. For example, we identified mutations that were generated through use of a transposon mutagenesis screen. The transposon mutagenesis occurs by activating transposon movement in bacteria carrying a yeast genomic library . The library is then transformed into yeast and the transposon-carrying chromosomal fragments are selected to replace the endogenous piece of chromosome by homologous recombination. Using this procedure, we identified two independent ZRC1 missense mutations, even though the transposon was inserted 5′ to the coding region. Generation of a genomic library from yeast with a wild-type ZRC1 also resulted in gain-of-function ZRC1 mutants. Both results suggest that transformation into bacteria or yeast is mutagenic, or that some level of expression of ZRC1 in bacteria is toxic and leads to the selection of mutations.
The finding of different mutations in ZRC1 that led to a gain of function was the basis for performing random mutagenesis of ZRC1. The mutagenesis was extensive, as the same substitution (for example F40S) was identified independently more than once. As shown by site-directed mutagenesis of Asn44 (Figure 7), there were substitutable amino acids we did not find in the screen, leading us to conclude that, although the mutagenesis was extensive, it was not saturating. All amino acid substitutions occurred in, or immediately adjacent to, the TMDs (Figure 2). No mutations in the cytosolic side of Zrc1 were identified. We did not find mutations distributed throughout the protein, suggesting that the cytosolic domain is not involved in substrate selection. This portion of CDF transporters is thought to be involved in sensing metals or in the docking of metal chaperones . The fact that a single mutation can change the substrate specificity calls into question the nature or existence of a chaperone. If chaperone binding is critical to metal transport activity, how can a single amino acid change affect chaperone selectivity and transport activity? Would a zinc chaperone be promiscuous and permit iron binding?
Structural studies on YiiP demonstrated that four residues from TMD II and TMD V form a zinc-binding site, and Zn2+ may cross the membrane through this site . The corresponding aspartate residue in Zrc1, when mutated to an alanine, results in a loss of Zn2+ transport activity, suggesting a conservation of function of the corresponding residues in Zrc1. Consistent with this finding, we found that mutations in four critical residues in TMD II (Leu33, Phe40, Asn44 and Ala52) affected metal specificity. It is of interest that we did not identify mutations in TMD V, as this domain is expected to comprise the metal-binding domain. We did, however, identify mutations in TMD III (G79S, L86S, L87H, E97G and R101G) and TMD VI (S272P, I275F and I275N) that affected metal specificity.
Sequence alignment of YiiP and Zrc1 shows significant homology within the TMDs (Figure 8). The crystal structure of YiiP has been solved to 3.8 Å (1 Å=0.1 nm) resolution . The protein functions as a homodimer. Using the structure of YiiP (PDB entry 2QFI) as a model for Zrc1, we generated a homologous model of the Zrc1 dimer. There may be differences in the number of amino acids in the TMDs between YiiP and Zrc1, as the TMDs are predicted and have not been experimentally verified. The model for Zrc1, based on YiiP, provides a tool to discuss structure–function relations. In this model, many of the Zrc1 residues shown to affect metal transport specificity were distant from the predicted tetraco-ordinate metal-binding residues (site identified in YiiP) and by homology analysis in other CDF members . By realigning the monomer subunits to approximate the assumed biological dimer, we generated a model that shows that many of the residues that alter metal specificity of Zrc1 are located at the dimer interface (Figure 9). This model suggests that interaction between subunits is as important in determining metal specificity as amino acids found near the essential metal-binding site. We believe that these mutations make the active site flexible so that it can accommodate metal ions of different co-ordination geometries. Mutation of Asn44 to a smaller hydrophobic group, N44A, led to both Zn2+ and Fe2+ transport activity, suggesting that the size of the amino acid side chain affects the geometry of the active site. There are mutations that affect substrate specificity that are not adjacent to the active site or in the dimer interface. This result suggests that the metal-binding site is altered through allosteric interactions.
Zrc1 mutations that conferred Fe2+ transport activity also conferred Mn2+ transport activity, as demonstrated by complementation of the growth defect of Δpmr1 cells under high concentration of Mn2+. Many transporters that transport Fe2+ also transport Mn2+, such as Nramp1 (natural-resistance-associated macrophage protein 1) , Smf1 (suppressor of mitochondrial import function 1) , Ccc1  and IRT1 (iron-regulated transporter 1) . Mutations in the plant transition-metal transporter IRT1 affect substrate specificity, although the mutations do not result in a gain of new transport capabilities, but rather a selective loss of transport for some metals and increased activity for others . It was noted that the ability to transport Fe2+ tracks with that of Mn2+, but not with that of Zn2+, as loss of Fe2+ transport was concomitant with loss of Mn2+ transport. There are, however, transporters that are specific for Mn2+. Pmr1 is an ATP cassette transporter that transports Mn2+ into the Golgi. Overexpression of PMR1 in Δccc1 cells did not confer iron resistance. There are CDF members that are specific to Mn2+ . Our selection system focused on mutations that led to resistance to high iron toxicity. It might be of interest to repeat the selection process but focusing on resistance to Mn2+ toxicity to determine if there are mutations that can discriminate between Fe2+ and Mn2+.
Huilan Lin designed the genetic screen and performed most of the experiments. Damali Burton performed the experiments that identified the mutation in ZRC1 resulting from the transposon screen. David Warner assisted in carrying out most of the experiments. John Phillips was consulted on the structural studies and provided the models showing the relationship between ZRC1 and YiiP. Diane McVey Ward and Jerry Kaplan were responsible for the overall direction of the project and for writing the manuscript.
This work was supported by the National Institutes of Health [grant number DK030534 (to J. K.)]. Support for use of the Core Facilities was provided through the National Institutes of Health [grant number NCI-CCSG P30CA 42014] and NIDDK Center of Excellence [award number 5P30KD72437].
We express our appreciation to members of the Kaplan laboratory for critically reading the manuscript prior to submission. We also acknowledge Dr Dax Fu (Brookhaven National Laboratory) for helpful discussions.
Abbreviations: CDF, cation diffusion facilitator; CM-Ura, complete minimal mix minus uracil; CPY, carboxypeptidase Y; ICP, inductively coupled plasma; TMD, transmembrane domain
- © The Authors Journal compilation © 2009 Biochemical Society