Sealed Reticulocyte Ghosts AN EXPERIMENTAL MODEL FOR THE STUDY OF Fez+ TRANSPORT*

Sealed right-side-out reticulocyte ghosts transported and accumulated iron offered as 6eFez+-ascorbate (K,,, = 1.1 pM). The uptake of iron by ghosts presented the characteristics of a transporter-mediated process: it responded to osmotic challenge, the rate of transport increased when iron was present in the opposing side, and the transport rate showed the temperature dependence typical of membrane-mediated processes. The transport of iron was dependent on an associated influx of C1- in order to keep electroneutrality. Other transition metals, such as Cu2+, Zn2+, and Co2+, inhib- ited the transport of Fez+. The overall characteristics of the system make reticulocyte sealed ghosts a very useful model in determining the basic mechanisms of membrane iron transport. with 10 pM 'Ve, 200 p~ ascorbate in saline buffer (pH 7.0). The 'Ve- loaded ghosts were divided into aliquots and centrifuged for 10 min at 25,000 X g. "Fe efflux was initiated resuspending aliquots of the ghosts in saline buffer containing 200 p~ ascorbate (pH 7.0). At specified times, the reaction was stopped by filtration as described above. Reagents-Solutions were prepared with distilled deionized water further treated with Chelex-100 (Bio-Rad) to remove traces of divalent metals. Similarly, stock solutions of monovalent salts and buffers were routinely filtered through Chelex-100. "Fe, as FeCI3 in 0.5 M HCI, was from Du Pont.

Knowledge of the membrane transport of iron is incipient. Transferrin-independent iron uptake by a variety of cell systems has been reported (Muir et al., 1984;Brissot et al., 1985;Wright et al., 1986Wright et al., , 1988Craven et al., 1987;Egyed, 1988;Morgan, 1988Morgan, ,1991Sturrock et al., 1990;Kaplan et al., 1991). Although valuable information has been obtained, the interpretation of results has often been obscured due to the inherent complexities of whole cell systems.
Working under the hypothesis that some of the basic characteristics of the membrane iron transport should be better explored using an experimental model that circumvents the complications found in whole cells studies, we decided to search for such a model. Two possible systems were considered isolated endocytic vesicles and right-side-out sealed reticulocyte ghosts. The former system (NGiiez et al., 1990) was not amenable to experimental manipulations; the difficulties were most derived from the small intravesicular volume of the endosomes and from the need to load the vesicles with iron prior to study its transport. Sealed ghosts proved to be a highly superior experimental system, yielding considerable amounts of material with relatively simple manipulations. In this experimental system the forward direction of transport moves iron towards the ghost lumen, which minimizes the manipulations needed to measure transport.
Most cell systems studied transport both Fez+ and Fe3+. We chose to study the transport of Fez+, because the ionic concentrations of Fez+ can be assessed more easily. Additionally, the concentration of ascorbic acid in plasma, 40-150 PM (Evans et al., 1982;Washko et al., 1989), makes the 2+ the most possible oxidation state, under condition in which iron in serum exceeds the total 3erum iron binding capacity. * This work was supported by FONDECYT 1080, Chile. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We found that sealed right-side-out reticulocyte ghosts transported and accumulated Fez+ in a process that was concentration and time-dependent and presented all the characteristics of being membrane-mediated.
We estimate that sealed reticulocyte ghosts are a very good model to determine the basic characteristics of membrane iron transport.

EXPERIMENTAL PROCEDURES
Reticulocyte Sealed Ghosts-Reticulocyte-rich rabbit red cells (30-80% reticulocytes) were lysed with 20 volumes of 5 mM sodium phosphate (pH 8.0) and sedimented for 20 min at 25,000 X g (Steck and Kant, 1974). The stromal pellet was washed three times with 20 volumes of lysis buffer and the plasma membrane fraction was purified by sucrose density centrifugation as described (Lodish and Small, 1975;N S e z and Glass, 1983). The material banding in the 20-40% sucrose interface was collected, suspended in distilled water, washed three times by centrifugation, and stored at -20 "C. Purified ghosts were pelleted and sealed by incubation for 30 min at 37 "C with 50 mM NaC1, 50 mM KCl, 1 mM MgSO,, 20 mM MOPS' (pH 7.0). The sealed ghosts were repeatedly washed with saline buffer: 50 mM NaC1,50 mM KCl, 20 mM MOPS (pH 7.0).
Microscopy-Sealed or unsealed ghosts were fixed for 60 min at room temperature in 3% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). After fixation, the ghosts were rinsed with phosphate buffer, mounted under 30% (v/v) acetone on aluminum plaquettes covered with 1% gelatine, gradually dehydrated with acetone, and dried to critical point with liquid COz. The samples were then coated with a 30-nm-thick layer of gold (Polaron E 5000 apparatus) and examined with the scanning attachment of a Phillips 300 electron microscope (Anderson, 1951).
Ghosts Internal Volume-The internal volume of the sealed ghosts was determined by horseradish peroxidase entrapment. Ghosts (1 mg/ml) were sealed in sealing buffer supplemented with 1,000 units/ ml of horseradish peroxidase (Type 11, Sigma). The sealed ghosts were repeatedly washed with saline buffer (pH 7.0) and the peroxidase activity entrapped into the ghosts was determined as described in the Sigma Technical Bulletin. To control for nonspecific binding, ghosts were sealed without peroxidase but this was added at the end of the sealing period. Peroxidase activity in control ghosts was between 6 and 8% that of test ghosts. The internal volume of the sealed ghosts determined by this method was 12.5 & 1.1 pl/mg of protein (mean & S.D. of three determinations).
"Fe Uptake-Sealed ghosts (20-40 pg/ml) in saline buffer (pH 7.0) were equilibrated at the assay temperature and the reaction was started by the addition of 10 p~ 'Te, 200 p~ ascorbate in saline buffer. At the specified times, aliquot8 of the assay medium (corresponding to 2-5 pg of protein) were filtered through Millipore filters (GSWP 0.22 pm) pretreated with saline buffer supplemented with 2% bovine serum albumin and 10 mM ascorbate. The trapped ghosts were quickly washed four times with cold saline, and the radioactivity trapped in the filters was measured in a Phillips y counter. Under these conditions, nonspecific binding of "Fe to the filters was less than 1% of input.
with 10 p M 'Ve, 200 p~ ascorbate in saline buffer (pH 7.0). The 'Veloaded ghosts were divided into aliquots and centrifuged for 10 min at 25,000 X g. "Fe efflux was initiated resuspending aliquots of the ghosts in saline buffer containing 200 p~ ascorbate (pH 7.0). At specified times, the reaction was stopped by filtration as described above.
Reagents-Solutions were prepared with distilled deionized water further treated with Chelex-100 (Bio-Rad) to remove traces of divalent metals. Similarly, stock solutions of monovalent salts and buffers were routinely filtered through Chelex-100. "Fe, as FeCI3 in 0.5 M HCI, was from Du Pont.

RESULTS
Morphology of Sealed Reticulocyte Ghosts-Sealed ghosts presented a distinctive morphology when compared with unsealed ghosts (Fig. 1). In scanning electron microscopy and in light microscopy, the sealed ghost showed the typical biconcave morphology of red cells with a diameter of 1-2 pm, whereas the unsealed ghost showed the appearance of empty bags. Substrate accessibility to glyceraldehyde-3-phosphate dehydrogenase and cytochrome c oxidoreductase indicated enzyme latencies of 89 and 98%, respectively. These values are in good agreement with published values for sealed erythrocyte ghosts (Steck and Kant, 1974).
69Fe Uptake by Ghosts-Reticulocyte ghosts incorporated Fe2+ as a function of incubation time with pseudo-first order kinetics (Fig. 2). In agreement with whole cell experiments (Egyed, 1988;Morgan, 1988), ghosts derived from erythrocytes did not incorporate iron (Fig. 2). The sealed reticulocyte ghosts effectively concentrated iron; considering an internal volume of 12.5 pl/mg of protein (see "Experimental Procedures"), the concentration of iron in the ghosts at equilibrium was 40-150-fold higher than the iron concentration in the incubation medium. The accumulation capacity of the sealed reticulocyte ghosts, expressed as moles of iron/mg of protein, was proportional to the reticulocyte content of the original blood.
Since a significant iron accumulation was observed, at least two processes have to be considered the transport of iron into the intravesicular space and the sequestration of the transported iron in order to maintain the concentration gradient necessary for further transport. The observed kinetic behavior will be primarily a reflection of the slower process.
To determine if the observed kinetics reflect the transmembrane iron transport process, 59Fe uptake was measured in the presence of the ionophore A23187, an effective membrane transporter of Fe2+ (Young and Gomperts, 1977;NGiiez and Glass, 1983). If the membrane transport process is ratelimiting, the addition of extra transporters should increase the rate of 59Fe accumulation by the ghosts. A23187 (0.3 pM) increased by 1.7-fold the rate constant for 59Fe accumulation by reticulocyte ghosts, with a slight increase (6.7%) in the maximal accumulation (Fig. 2). Larger increases in the rate constant were observed at higher A23187 concentrations. In contrast, sealed erythrocyte ghosts did not transport iron, although in the presence of A23187 they showed a considerable uptake of ''Fez+ (Fig. 2). These results indicate (a) that erythrocyte ghosts have lost the membrane iron transport system, ( b ) that the increase in the maximal amount of iron accumulated by reticulocyte ghosts produced by A23187 can be ascribed to the recruitment of erythrocyte ghosts, and (c) that although reticulocyte ghosts accumulate iron and reach internal concentrations well above the external concentration, the measured time courses reflect the kinetics of the transmembrane iron transport.
Effect of Temperature on the Rate Constants of Influx and Efflux-Transport processes through biological membranes often present break points in their rates values as a function of temperature. This criterium is used to differentiate trans- Ghosts were incubated in saline buffer and the osmolarity was increased by addition of mannitol. Maximal 59Fe accumulation was determined from kinetic curves such as shown in Fig. 2. From the composition of the sealing buffer, the osmolarity of the internal medium was estimated as 220 mosM.

TABLE I Effect of metals in the incorporation of Fez+
Ghosts were incubated a t 37 "C with 10 PM 'Te, 200 p M ascorbate in saline buffer and the metals at the concentrations indicated. The metals were in the chloride form except for ZnSO.. The maximal amount of iron incorporated was determined from curve fitting of "Fe uptake kinetics and expressed as percent of the uptake measured in the absence of the metals. Values are means f S.D. of three experiments. In the conditions of the experiment, the redox state of copper may be l+. port processes from simple binding of the substrate. Fig. 3 shows the dependence on temperature of the 59Fe2+ influx and efflux rate constants. Break points at approximately 15 "C were found for both iron influx and efflux. The rate constants for influx were considerably larger than those for efflux (Fig.   3), and, remarkably, iron efflux at temperatures below 15 "C was negligible. Therefore, iron flux data obtained below this temperature are a measure of the influx components, without significant contributions from the efflux components. Effect of Anions on Iron Uptake-The observed accumulation of iron indicates that the ghosts have systems to dissipate the electrical gradient generated by the iron influx. One way in which this may be accomplished is by the concurrent influx of an anion. The anion composition of the external buffer affected the amount of accumulated iron (Fig. 4). The preference in supporting maximal accumulation was: C1-= NOT > Br-= SO:-> I-. Gluconate, even at low concentrations, inhibited the incorporation of iron, probably through chelation (not shown).

Metal added
"Fe Efflux from the Ghosts-In agreement with kinetic models for carrier-mediated systems (Devks and Krupka, 1979;Devks and Boyd, 1989), the rate constant for "Fe efflux from the ghosts almost doubled when Fez+ was added to the external solution (Fig. 5).
"Fe Uptake Is Sensitive to Osmotic Challenge-An approach t o differentiate between transport and binding of a substrate is the sensitivity of the putative transport system to variations of the medium osmolarity. The uptake of "Fe by ghosts was affected by changes in the osmotic pressure: increased osmolarity in the external solution resulted in a marked decrease of the maximal amount of iron incorporated by the ghosts (Fig. 6). The above experiment was followed by determining the effect of osmolarity on iron efflux. A 3-fold increase in the efflux rate constant was obtained when the osmolarity of the external solution was approximately doubled (Fig. 7).
Divalent Transition Metals Inhibit Iron Uptake by Reticulocyte Ghosts-Several authors have reported that transition metals inhibit iron uptake in whole cells (Morgan, 1988;Wright et al., 1988, Sturrock et al., 1990Kaplan et al., 1991). Incorporation of Fez+ by reticulocyte ghosts was strongly inhibited by Cu2+. Variable degrees of inhibition were also observed with Mn2+, Cd2+, Zn2+, and Co2+ (Table I). A13+ concentrations up to 1 mM, did not inhibit Fez+ uptake by the ghosts.

DISCUSSION
Reticulocyte-derived sealed ghosts accumulate Fez+ in a concentration and time-dependent fashion. In agreement with whole cell studies (Egyed, 1988;Morgan, 1988), sealed ghosts derived from erythrocytes were unable to accumulate iron, but this property was generated when an artificial membrane iron transporter was added, indicating that erythrocyte ghosts, whereas lacking the membrane iron transport system, do have an operative accumulation system.
Because the concentration of intravesicular iron at equilibrium is severalfold higher than the extravesicular concentration, the question arose as to whether the observed kinetics were a reflection of the transport process or a subsequent process related to the drainage of the transported iron. Since the overall uptake rate should be determined by the ratelimiting step, the introduction of an artificial iron carrier should increase the observed rate only if the transport rate was limiting. The observed increase in the uptake rate caused by A23187 is therefore an indication that the observed rates are a reflection of the transport rate.
In studying putative transport processes, a fundamental problem is to differentiate between actual transport of the species and binding to the external surface. The effect of osmolarity in both the influx and the efflux of iron, as well as the effect of external iron on the efflux of previously internalized "Fez+, indicate that we were determining a true transport process. Moreover, the doubling of the efflux rate constant in the presence of extravesicular iron indicates that the iron binding sites of the transporters are evenly oriented towards the extravesicular and the intravesicular environment (DevCs and Krupka, 1979;Dev6s and Boyd, 1989). An unexpected observation was that the ghosts accumulated iron severalfold over the external concentration. It is possible that the driving force for accumulation reside in iron binding sites present in the internal face of the membrane, although more experiments are needed to investigate this problem.
Accumulation of iron by the ghost implies the neutralization of the positive charges carried into the ghosts by Fez+. The capacity of C1-and NO: to support maximal accumulation was similar to that found by Wright et al. (1988) in perfused liver and may reflect the capacity of these anions to neutralize the positive charges of the transported iron. Fez+ transport was competed by several 2+ transition metals, being Cu2+ the most effective. The observed selectivity in competition, Cu2' > Zn2+ > Co2+ > Mn2+ > Cd2+, is somehow different to that reported by Morgan (1988) for reticulocytes, but similar to that reported by Kaplan et al. (1991) for HeLa cells when assayed in saline buffer. Since in this study metabolic poisoning can be discarded, it is likely that the above sequence better expresses the possible selectivity of the Fez+ transport system.
In summary, we have determined that sealed right-side out reticulocyte ghosts transport and accumulate iron offered as Fe2+-ascorbate. The overall characteristics of the system make reticulocyte sealed ghosts an excellent model in determining the basic mechanisms of membrane iron transport.