The Mutation F227I Increases the Coupling of Metal Ion Transport in DCT1*

Metal ion transport by DCT1, a member of the natural resistance-associated macrophage protein family, is driven by protons. The stoichiometry of the proton to metal ion is variable, and under optimal transport conditions, more than 10 protons are co-transported with a single metal ion. To understand this phenomenon better, we used site-directed mutagenesis of DCT1 and analyzed the mutants by complementation of yeast suppressor of mitochondria import function-null mutants and electrophysiology with Xenopus oocytes. The mutation F227I resulted in an increase of up to 14-fold in the ratio between metal ions to protons transported. This observation suggests that low metal ion to proton transport of DCT1 resulting from a proton slippage is not a necessity of the transport mechanism in which positively charged protons are driving two positive charges of the metal ion in the same direction. It supports the idea that the proton slippage has a physiological advantage, and the proton slip was positively selected during the evolution of DCT1.

Metal ion transport by DCT1, a member of the natural resistance-associated macrophage protein family, is driven by protons. The stoichiometry of the proton to metal ion is variable, and under optimal transport conditions, more than 10 protons are co-transported with a single metal ion. To understand this phenomenon better, we used site-directed mutagenesis of DCT1 and analyzed the mutants by complementation of yeast suppressor of mitochondria import function-null mutants and electrophysiology with Xenopus oocytes. The mutation F227I resulted in an increase of up to 14-fold in the ratio between metal ions to protons transported. This observation suggests that low metal ion to proton transport of DCT1 resulting from a proton slippage is not a necessity of the transport mechanism in which positively charged protons are driving two positive charges of the metal ion in the same direction. It supports the idea that the proton slippage has a physiological advantage, and the proton slip was positively selected during the evolution of DCT1.
Metal ions are vital elements for all living cells. The NRAMP 1 family of metal ion transporters apparently plays a major role in metal ion homeostasis (1)(2)(3)(4). Most of the information available about the mechanism of these transporters has resulted from studies of the family members NRAMP2 (DCT1) from mammalian and SMF1 from yeast expressed in Xenopus oocyte (5)(6)(7)(8)(9). Xenopus oocytes have a very low metal ion uptake background, which makes them the ideal heterologous expression system for metal ion transporters. In addition to the uptake measurements of the various divalent cations, it is possible to analyze the electrophysiological parameters generated by imposed potentials in the transporter-expressing oocyte as compared with control (10). The studies of DCT1 and Smf1p demonstrated that both of them function as general divalent metal ion transporters and that a proton gradient is the driving force for the metal ion transport (8,9). However, the transport of protons and metal ions is "loosely coupled" because the process exhibits a variable stoichiometry. At pH 7 and membrane potentials of Ϫ90 to Ϫ30 mV, DCT1 transports one Fe 2ϩ ion with one H ϩ . At high proton concentration (low pH), the number of H ϩ ions transported with one Fe 2ϩ ion increased to 10 (8). Moreover, on changing the membrane potential from ϩ10 to Ϫ80 mV at this low pH, the number of H ϩ ions transported with one Fe 2ϩ ion increased from 3 to ϳ18 (9). In DCT1, this phenomenon was defined as a metal ion-dependent proton slippage (1,9,11). In contrast to DCT1, Smf1p showed a metal ion-independent sodium slip through the proton-translocating pathway (9,12). The mechanism of this phenomenon is not well characterized, and the sites on the transporters that generate it are not known.
In this work, we addressed the coupling between proton and metal ion transport, and report on a mutation in DCT1 that exhibits an ϳ14-fold increase in the ratio of metal ion to proton transport.

MATERIALS AND METHODS
Site-directed Mutagenesis of DCT1-Oligonucleotide-directed, sitespecific mutagenesis was performed by overlapping nucleotides with the mutation using the polymerase chain reaction (PCR) method (13). The DCT1 cDNA was cloned into pSPORT1 plasmid and was used as a template for PCR mutagenesis. NdeI and SpeI restriction sites were introduced at positions 145 and 1719 of the DCT1 reading frame (12). This resulted in the substitutions G50M, S573T, and V574S, which did not change the properties of DCT1. The mutation was confirmed by sequence analysis. The cRNAs of DCT1 and the mutants were synthesized by in vitro transcription from their cDNAs. Oocytes were injected with 50 l of water containing 50 ng of cRNA.
Oocyte Preparation and Uptake Measurements-Xenopus laevis oocytes were handled as described previously (12). Uptake experiments were performed 1-4 days following injections. The uptake solution for radiotracer experiments contained 100 mM NaCl, 10 mM Hepes, 2 mM Mes, 2 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 2 mM L-ascorbic acid, and the pH was adjusted to 5.5 with Tris base. Usually, 20 oocytes were incubated in 0.5 ml of a solution containing 55 FeCl 2 , 54 MnCl 2 , 65 ZnCl 2 , or 60 CoCl 2 . The radioactive tracer was usually mixed with 7-20 M unlabeled metal ion, and the uptake was followed for 30 min. The numbers given under "Results" are background-corrected and expressed in specific activity (pmol/oocyte/h).
Electrophysiological Experiments-Experiments utilizing the twomicroelectrode voltage clamp technique were performed as described previously (14 -16). Data analysis was performed as described previously (12). To analyze the pre-steady-state currents, the current traces were fitted to I(t) ϭ I 1 exp(Ϫt/ 1 ) ϩ I 2 exp(Ϫt/ 2 ) ϩ I ss , where I 1 is a capacitative current with time constant 1 associated with the oocyte plasma membrane ( 1 is also observed in non-injected control oocytes), I 2 is a transient current associated with DCT1 expression with time constant t 2 , and I ss is the steady-state current. The parameters 1 , 2 , I 1 , I 2 , and I ss were allowed to vary in all fits. The transient charge movements Q were obtained from the time integrals of I transient (t) ϭ I 2 exp(Ϫt/ 2 ) during the on and off responses for all depolarizing and hyperpolarizing potentials and fitted by the Boltzmann equation, where Q max represents the total charge movement, z represents the effective valence, V 0.5 represents the midpoint of the charge distribution, and Q hyp represents the charge movement for extreme hyperpolarizing potentials (14,15). R, T, and F are the usual thermodynamic constants.
Simultaneous Voltage-clamped Tracer and Current Measurements-Before the start of tracer uptake, the oocyte was clamped at Ϫ50 mV and perfused (washed) with substrate-free solution (that described above). After the perfusion was stopped, 50 l of the uptake solution * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. containing 55 Fe 2ϩ ( 55 FeCl 2 mixed with 20 M unlabeled FeCl 2 to a final concentration of 26 M) were added manually using a pipette, which washed out the substrate-free solution. The uptake lasted 3 min in the chamber, which volume was ϳ50 l. To avoid dehydration, 1.5 min after the uptake began, another 50 l of uptake solution was added. The uptake was terminated by perfusing the oocyte with the substrate-free solution. Currents were continuously measured during uptake. The oocyte was then dissolved in 50 l of 10% SDS mixed with 4.5 ml of scintillation mixture, and the uptake activity was measured. The transported charge was calculated by integrating the Fe 2ϩ -evoked current over the uptake period (3 min). Next, the charge:uptake and the metal ion to proton ratios were calculated. The transported charge (C, coulomb) was converted to the number of electric charges (1 C ϭ 6.24 ϫ 10 18 e Ϫ ) and the uptake (pmol/oocyte/3 min) was converted to the number of Fe 2ϩ ions that were transported (multiplying by Avogadro's number).

RESULTS
It has been shown that DCT1 and Smf1p function as general divalent metal ion transporters in mammals and yeast, respectively. Both use protons as their motive force to transport a broad-range of the same divalent metal ions and exhibit similar affinities for their various substrates (8,12,16). The mammalian DCT1 complements the phenotype of SMF1 null mutation in yeast (5,(17)(18)(19). The null mutant smf1⌬ is unable to grow in the presence of EGTA. Expression of Smf1p or the mammalian transporter DCT1 suppresses the above phenotype and allows growth of the yeast mutant in the presence of EGTA (18). Despite the relatively high homology, there are a few known differences between DCT1 and Smf1p. The most striking one is in the slips generated by the transporters, H ϩ slip through DCT1 and Na ϩ slip through Smf1p (11). The differences between the two homologues have to be due to the non-conserved amino acids in both proteins. Consequently, to locate the sites on the transporters responsible for this phenomenon, we substituted amino acids in DCT1 for the corresponding ones residing in Smf1p (19). It has been found that a mutation in the conserved glycine at position 216 (G216R) in the putative TM4 of DCT1 causes microcytic anemia in mkϪ/Ϫ mice and Belgrade rats (20). This mutation was subsequently shown to result in loss of Fe 2ϩ transport ability (21). This suggests that the TM4 may have an important biological function. Therefore, we mutated the Phe-227 of DCT1, located in the same TM, to the corresponding isoleucine residue of Smf1p. The properties of the mutant were analyzed by expression in Xenopus oocytes, where measurements of uptake activity and electrophysiological parameters of the oocytes expressing the mutated transporter were recorded. In addition, the DCT1 mutant was tested for complementation of EGTA sensitivity of the Smf1p null mutants.
Metal Ion Uptake Activity of DCT1 and F227I Mutant-The uptake activity of Zn 2ϩ , Fe 2ϩ , Mn 2ϩ , and Co 2ϩ into the Xenopus oocytes expressing DCT1 and the F227I mutant is exhibited in Fig. 1. The data are representative of three experiments with oocytes from different frogs. The mutation did not result in any significant changes in uptake activity or specificity to the different metal ions, except for a small decrease in Co 2ϩ uptake. Therefore, it was not surprising that the F227I mutant retained the capability to suppress the growth arrest of the yeast null mutant smf1⌬ in the presence of EGTA. 2 We generated more than thirty amino acid substitutions in DCT1, none of which exhibited the properties of F227I substitution described below.
Pre-steady-state Currents of DCT1 and the Mutation F227I-At pH 5.5 and a millisecond time scale, oocytes expressing DCT1 exhibited pre-steady-state currents, mainly at positively imposed potentials (8,19). Pre-steady-state currents are an excellent indication of the expression level of transporters in Xenopus oocytes (10). The magnitude of the pre-steadystate currents recorded in oocytes injected with cRNA of native DCT1 or the mutant F227I (Fig. 2, A and C) indicated similar expression levels of the two transporters. To characterize the features of the F227I mutant, we subtracted the capacitative and steady-state currents to yield the transient current (Q) (Fig. 3B). At each applied voltage, time integration of the off transients yielded the charge moved. The charge-voltage (Q-V) relationship obtained (data not shown) was fitted to a single Boltzmann function (see "Materials and Methods"), and Q max of 43 nC and V 0.5 of 42 mV were determined. Similar Q max of 43 nC and a little higher V 0.5 of 49 mV were determined for native DCT1 (Fig. 3A). The time constant of the pre-steady-state current decay in F227I, in the potential jump from Ϫ25 to ϩ50 mV, was 35 ms for the on response and 23 ms for the off response, and in DCT1, was 43 and 26 ms, respectively. According to the Q max , ϳ2.5 ϫ 10 11 transporters were expressed on the plasma membrane of each oocyte expression DCT1 or F227I mutant. This calculation was made based on the mechanistic concept of transient charge movements in ioncoupled transporters that was proposed by Wright and coworkers (14,22) for the Na ϩ -coupled glucose transporter SGLT1 and the similarity between the pre-steady-state properties of SGLT1 and DCT1. Essentially, the pre-steady-state current results from imposed potential induction of occlusion and release of protons by DCT1 in the absence of substrate (1). The small decrease in V 0.5 and in on of the mutation could indicate a small decrease in the affinity of the transporter to protons.
Metal Ion Induced Steady-state Currents in DCT1 and the Mutation F227I-The addition of Fe 2ϩ , Mn 2ϩ , and Co 2ϩ to oocytes expressing DCT1 resulted in the disappearance of the pre-steady-state currents and appearance of large steady-state currents, especially at the negatively imposed potentials. At pH 5.5 and Ϫ80 mV, the coupling between H ϩ and Fe 2ϩ is interrupted, and the charge:uptake ratio is ϳ18 (9). The steadystate currents in those conditions are therefore mainly because of the protons transported (8,9). The addition of Mn 2ϩ to oocytes expressing the mutation F227I did abolish the presteady-state currents and generated only small steady-state currents compared with DCT1 (Fig. 2, B and D). At Ϫ125 mV, 2 A. Cohen, unpublished results. Simultaneous Measurement of Iron-induced Currents and Uptake Activity in DCT1 and the Mutation F227I-To further investigate these findings, we simultaneously measured the iron-induced currents and 55 Fe 2ϩ uptake. First, the pre-steadystate currents were recorded in DCT1 and F227I expression oocytes that were bathed in substrate-free solution, pH 5.5 ( Fig. 4A and C). The wild-type DCT1 and F227I mutant exhibited similar pre-steady-state currents. Addition of solution containing 20 M of FeCl 2 resulted in the disappearance of the pre-steady-state currents and induced a steady-state current (Fig. 4, B and D). Fig. 5D shows the currents generated at different imposed potentials on DCT1 and its F227I mutant by the addition of FeCl 2 . At Ϫ125 mV, the addition of Fe 2ϩ induced a steady-state current of ϳ650 nA for DCT1 and only 75 nA for F227I mutant. Similar phenomenon was recorded between Ϫ25 and Ϫ125 mV. To analyze the amounts of charge movement in relation to metal ion uptake, we simultaneously traced the charge movement and 55 Fe 2ϩ uptake in oocytes clamped at Ϫ50 mV. Fig. 5 shows the results of five different oocytes injected with DCT1 or F227I mutant. The charges that were transported during the 3-min exposure to 26 M Fe 2ϩ were 25 C and 2.3 C, and the uptake activities were 7.25 pmol/oocyte/3 min and 5.4 pmol/oocyte/3 min for DCT1 and F227I, respectively. The charge:uptake ratio under these conditions (pH 5.5 and membrane potential of Ϫ50 mV) was calculated (see "Materials and Methods") to be 36 for DCT1 and 4.4 for the F227I mutant. Because each Fe 2ϩ contributes two charges and each proton one charge, the number of protons transported with each metal ion is 34 for DCT1 and 2.4 for the F227I mutation. Hence, the mutation exhibits a 14-fold increase in the ratio of metal ion to proton transported. These observations suggest that the mutation F227I almost abolished the proton slip through DCT1 without statistically significant change in the metal ion transport activity.
Zinc Uptake Activity Exhibited by DCT1-At pH 5.5 and at Ϫ50 mV, divalent metals, including Zn 2ϩ , generated similar inward currents of Ϸ100 nA in oocytes expressing DCT1 (1,12). Moreover, Zn 2ϩ inhibits Fe 2ϩ and Mn 2ϩ uptakes in oocytes expressing DCT1 or Smf1p and in yeast cells expression Smf1p (5,8,12). However, in previous studies, we could not detect 65 Zn 2ϩ uptake into Xenopus oocytes expressing DCT1 or Smf1p (12,19). These observations were in line with results from other studies with DCT1 expressed in Caco-2 TC7 cells, which suggest that Zn 2ϩ is not transported by DCT1 (23). We therefore concluded that Zn 2ϩ binds to the transporter, probably to the same site as the other metal ions, and generates the proton slip through DCT1 but does not transport. In this work, we contradicted our previous findings, since we recently detected 65 Zn 2ϩ uptake into oocytes expressing DCT1, in agreement with the recently published data by Marciani et al. (24). As depicted in Figs. 1 and 6A, the values of 65 Zn 2ϩ uptake activity were ϳ10 pmol/oocyte/h. We measured the uptake activity of Fe 2ϩ and Zn 2ϩ during various days after the injection of DCT1 cRNA to the oocytes (Fig. 6A). As shown in Fig. 6C, the ratio between Zn 2ϩ and Fe 2ϩ uptake was altered during the various days, from ϳ3% on the first day to ϳ30% on the fourth day. The ratio between Mn 2ϩ and Fe 2ϩ was also altered, from 30% on the first day to 116% on the fourth day (data not shown). Although Fe 2ϩ uptake activity reached steady-state 2 days after injection, Zn 2ϩ and Mn 2ϩ uptake activities continued to rise even 4 days after injection. The data are representative of four experiments with oocytes from different frogs. Reduction of the amount of cRNA that was injected into the oocytes, to an amount that is under the saturation level, did indeed reduce the uptake activity but did not change these ratios (data not shown). This phenomenon did not occur when we measured the uptakes of two substrates, GABA and betaine, by the GABA transporter GAT3. 3 These results can be explained either by oligomerization or by conformation changes in DCT1 that occur during the time of its presence in the plasma membrane, which altered the specificity of the transporter to the various substrates.

DISCUSSION
The family of NRAMP metal ion transporters plays a major role in metal ion homeostasis (2)(3)(4)25). We developed a concerted approach to the study of Smf1p and DCT1. It employs 3 E. Kfir, unpublished results. The second addition of 50 l of uptake solution is marked with an arrow. A, the iron-induced currents recorded for oocytes expressing DCT1. B, the iron-induced currents recorded for the F227I mutant. C, the iron uptake activity measured for the oocyte expressing DCT1 (black bar) and the F227I mutant (gray bar). D, the differential currents in the presence and absence of the Fe 2ϩ measured in Fig. 4 were plotted against membrane voltage. Diamonds, DCT1; squares, F227I mutant. The data for each point and bar represent means (after subtraction of the control values) ϩ S.E. (for A, B, and C, n ϭ 5; and for D, n ϭ 3). S.A., specific activity. yeast null mutants lacking the SMF metal ion transporters (18) and Xenopus oocytes as an expression system for Smf1p and DCT1 (12). Using this approach, we investigated the first external loop of DCT1. The properties of various mutants in this region suggest that loop 1 is involved in the metal ion binding and its coupling to the proton-driving force (19). One of the prominent mechanistic properties of these transporters is the proton slip through DCT1 and the Na ϩ slip through Smf1p (11). Under conditions of neutral pH and/or neutral membrane potential, DCT1 would be in the coupled state; that is, the stoichiometry of proton:metal ion transport would be close to 1:1. Increasing the driving force, by either reducing the pH or increasing the negative potential, generates a proton slip, which maintains the levels of substrate transport at rates that are close to the "normal" rates of the coupled state. However, it simultaneously increases the rate of proton movement through the transporter. A similar mechanism applies to Smf1p-mediated slippage, except that here the ion that slips is Na ϩ . Metal ion transport by Smf1p is also driven by proton motive force, and inhibition of the latter by Na ϩ results in reduced metal ion uptake. The physiological significance of this phenomenon in metal ion transporters could be protection of cells against the excessive transporting of these elements. Such a mechanism is necessary when cells are exposed to conditions of abundant metals and acidic environment. The mechanism of this phenomenon is not well characterized, and the sites on the transporters generating this phenomenon are not known. To locate these sites, we attempted to gain control over the slippage magnitude by reciprocal amino acid substitution between DCT1 and Smf1p. We managed to diminish the proton slip through DCT1 by substituting a single amino acid (F227I). This mutation did not generate any Na ϩ slip, and the transporter properties were not influenced by the removal of Na ϩ from the medium (not shown).
The amino acid Phe-227 is located in the putative TM4 of DCT1 that may be situated in close proximity to the proton pathway of the transporter. This notion is in line with a previous study that demonstrated the importance of TM4 for the function of the transporter (20,21). The mutation F227I did not result in any significant changes in the uptake activity or specificity of the different metal ions, except for a small decrease in Co 2ϩ uptake. It was therefore not surprising that the F227I mutant retained the capability to suppress the growth arrest of smf1⌬ in a medium containing EGTA. When expressed in Xenopus oocytes, the Q max value of the F227I mutant was comparable with that of native DCT1, indicating that the expression levels of the mutant are similar to those of the native DCT1.
The coupling between the driving force and substrate transport may be modulated by different physiological conditions (11). However, for most of the transporters studied so far, a fixed number of driving ions are reported to cross the membrane per the single substrate molecule that is transported (26 -29). Even for proliferated families of transporters, such as the Na ϩ /Cl Ϫ neurotransmitter transporters, stoichiometric amounts of Na ϩ , Cl Ϫ , and substrate are reported (30 -33). Different transporters in this family operate with unique stoichiometries, and some of them even utilize K ϩ as part of the driving force (34 -36). In contrast, transporters such as the Escherichia coli lactose permease were reported to operate with a fixed stoichiometry of 1 for the symport of proton and lactose (37). Studies involving site-directed mutagenesis yielded numerous mutants in which the properties of proton and/or the substrate transport were modified (19,38,39). One of these mutants, E269D, is of particular interest to this discussion (40 -42). This delicate mutation, which changed the position of a carboxylic acid by 1.5 A, led to a few changes in the transport properties of lactose permease. The changes included reduction of the coupling between proton to lactose transport, such that ϳ10 protons had to cross the transporter for each lactose. This proton slippage is comparable with the native slip of DCT1.
The mild mutation F227I in DCT1 resulted in marked reduction in the proton slip through DCT1 and essentially mimicked the reversal of the E269D mutation in lactose permease (40 -42). The effect of the F227I mutation may also be analogous to the one of E269D, except that in the former it may induce a conformational change that cancels the slip that was evolutionarily selected for the benefit of the physiology of metal ion transport in the mammalian duodenum (11). Further studies will be required to identify the proton translocation pathway that was influenced by the F227I mutation in DCT1.