On the Ca2+ Dependence of Non-transferrin-bound Iron Uptake in PC12 Cells*

Non-transferrin-bound iron (NTBI) uptake has been reported to follow two pathways, Ca2+-dependent and Ca2+-independent (Wright, T. L., Brissot, P., Ma, W. L., and Weisiger, R. A. (1986) J. Biol. Chem. 261, 10909–10914; Sturrock, A., Alexander, J., Lamb, J., Craven, C. M., and Kaplan, J. (1990) J. Biol. Chem. 265, 3139–3145). Studies reporting the two pathways have ignored the weak interactions of Ca2+ with the chelator nitrilotriacetate (NTA) and the reducing agent ascorbate. These studies used a constant ratio of total Fe2+ to NTA with and without Ca2+. We observed Ca2+ activation of NTBI uptake in PC12 cells with the characteristics reported for other cells upon using 1 mm ascorbate and a constant ratio of total Fe2+ to NTA with or without Ca2+. However, Ca2+ did not affect NTBI uptake in solutions without NTA. We then determined conditional stability constants for NTA binding to Ca2+ and Fe2+ by potentiometry under conditions of NTBI uptake experiments (pH, ionic strength, temperature, ascorbate, total Fe2+, and total Ca2+concentrations). In solutions based on these constants and taking Ca2+ chelation into account, Ca2+ did not affect NTBI uptake over a range of free Fe2+concentrations. Thus, the Ca2+ activation of NTBI uptake observed using the constant total Fe2+ to NTA ratio was because of Ca2+-NTA chelation rather than an activation of the NTBI transporter itself. It is suggested that the previously reported Ca2+ dependence of NTBI uptake be re-evaluated.

Mn 2ϩ , and Cu 2ϩ or by chelation with diethylenetriaminepentaacetate. Many cell types have also been reported to show an extracellular Ca 2ϩ -dependent high affinity NTBI uptake that does not distinguish between Fe 3ϩ or Fe 2ϩ (8,(12)(13)(14). The importance of Ca 2ϩ -dependent high affinity NTBI uptake versus Ca 2ϩ -independent uptake was presented (8,(12)(13)(14). In these experiments, iron was reduced to Fe 2ϩ using ascorbate, and the Fe 2ϩ was chelated using nitrilotriacetate (NTA). Interactions between Ca 2ϩ and NTA were apparently considered too weak to be important, because a constant iron to NTA ratio was used in the presence or absence of Ca 2ϩ .
The rat adrenergic neural tumor pheochromocytoma cell line, PC12, is used extensively as a neuronal model. Nerve growth factor (NGF)-treated PC12 cells exhibit sympathetic neuron-like properties characterized by neurite outgrowth and electrical excitability. They express cholinergic receptors, Na ϩ channels, N-type voltage-operated Ca 2ϩ channels, and neuronal nitric-oxide synthase (15)(16)(17). They also exhibit high levels of expression of sarcolemmal Ca 2ϩ pump and low levels of the organellar pump (18). Upon stimulation with acetylcholine, PC12 cells can secrete dopamine, thus making them a model of dopaminergic neurons. Here, we report that in these cells Ca 2ϩ produces an apparent activation of NTBI uptake when reported protocols are used. We demonstrate that this activation is an artifact caused by Ca 2ϩ binding to NTA and thus indirectly affecting Fe 2ϩ -NTA chelation rather than an activation of the transporter itself.

EXPERIMENTAL PROCEDURES
PC12 Cell Cultures-PC12 cells in passages 12 to 30 were cultured in bovine Achilles tendon collagen-coated plates in RPM1 1640 medium supplemented with penicillin and streptomycin (5000 units/5000 g), 10% fetal calf serum, and 5% calf serum as described previously (18). At each passage, the cells were split 1 to 4. The cells were treated with 40 ng/ml 2.5 S NGF, which was purified as described previously (19). The medium with NGF was replenished 2, 5, and 7 days after the initial plating. The cells were harvested in Ca 2ϩ -free physiological saline solution (PSS) after another 2 days. PSS contained 10 mM HEPES, 140 mM NaCl, 5 mM KCl, and 5 mM MgCl 2 at pH 7.4 (pH at 37°C).
Non-transferrin-bound Iron Uptake Experiments-NTBI uptake was carried out using the PC12 cells in suspension with shaking. Initially, we used a constant iron to NTA ratio (12,13,20). Typically, the uptake was for 2 min in a 300-l reaction mixture at 37°C in an ascorbate (1 mM)-PSS solution containing 0.43 M total iron including 55 FeCl 3 (1-2 Ci/ml, specific activity 13 Ci/g), 28 M NTA, and PC12 cells containing 0.4 -0.8 mg of protein. 200 l of the sample was filtered under suction through a 0.45-m nitrocellulose filter, which was presoaked in 1% bovine serum albumin and 100 mM KCl. The filter was washed 3 times with 5 ml of a chilled solution containing 30 mM imidazole-HCl, pH 6.8, 0.5 mM EGTA, and 250 mM sucrose. The amount of radioactivity remaining on the filter was determined by scintillation counting. Blank values obtained without any cells were subtracted before any further computations. A number of specific variations made in this protocol will be as described under "Results." Determination of Conditional Stability Constants-The stability con-stants for Ca 2ϩ and Fe 2ϩ binding to NTA were determined under the conditions of the NTBI uptake experiments using a CdS/AgS Cd 2ϩ ion selective electrode and a Model 290A pH/ISE meter (Orion Instruments) as described previously (21,22). The Cd 2ϩ ion selective CdS/AgS electrode was prepared and polished using a polishing pad and 0.5-m silica polishing powder immediately before each titration. Because chloride ions interfere with these measurements, nitrate was substituted for chloride. ] was determined using the standard curve. The data were analyzed using a nonlinear binding model and competition between Cd 2ϩ and Ca 2ϩ or Fe 2ϩ for binding to NTA. In this model, K a ϭ M⅐L n /(M⅐L n ), where M is the concentration of M 2ϩ , L n is the complex of n NTA molecules with M 2ϩ , and L is the concentration of free NTA. The K a values determined were designated as the conditional stability constants (K a(cond) ). This analysis was performed using the software package Matlab 5.2, release 10 (Mathworks Inc.). The formation constants and stoichiometry coefficients were optimized using a simplex minimizing of the mass balance equation.
Statistical Analysis-Values given throughout are the means Ϯ S.E. of n replicates. Student's t test was used where applicable, and p Ͻ 0.05 was considered statistically significant. All of the experiments were repeated at least three times.

RESULTS
Iron Uptake at Constant Iron to NTA Ratio-We followed the literature and examined NTBI uptake initially using fixed concentrations of iron and NTA in ascorbate-PSS with and without 2 mM Ca 2ϩ (12,13,20). The experiment clearly shows a higher rate of NTBI uptake in the presence of 2 mM CaCl 2 than in its absence (Fig. 1A). Subsequently, the uptake was determined for only 2 min. In 16 such experiments, 2 mM CaCl 2 increased the NTBI uptake by 151 Ϯ 23%. Activation of the NTBI uptake was [Ca 2ϩ ]-dependent (Fig. 1B). The [Ca 2ϩ ] required to produce half-maximum activation was 0.5-1 mM. Consistent with the processes being associated with an activation energy, the Ca 2ϩ -activated and Ca 2ϩ -independent components of the NTBI uptake were greater at 37 than at 4°C (Fig. 1C). We also examined whether Ca 2ϩ activation occurred for NTBI uptake using Fe 3ϩ (FeCl 3 plus NTA in ascorbate-free PSS). In 5 experiments, there was an activation of 251 Ϯ 43% by Ca 2ϩ . These properties are consistent with those reported in the literature for the two components of NTBI uptake, one being Ca 2ϩ -dependent and the other Ca 2ϩ -independent.
NTBI Uptake without NTA-Using literature values for the stability constants for various interactions of Fe 2ϩ , Mg 2ϩ , Ca 2ϩ , and H ϩ with NTA and ascorbate (23) and correcting these constants for the temperature and ionic strength, 0.43 M total iron and 28 M NTA gave different values for free [Fe 2ϩ ] with 0 and 2 mM CaCl 2 (4.8 and 9.5 nM, respectively). Therefore, we conducted an NTBI uptake experiment in ascorbate-PSS without NTA (Fig. 2). It is noted that this experiment is limited in that high concentrations of Fe 2ϩ cannot be sustained at this pH without chelation. In the range of Fe 2ϩ concentrations used, 2 mM Ca 2ϩ did not activate the uptake. Because Ca 2ϩ activated NTBI uptake when NTA was used ( Fig. 1) but not without it (Fig. 2), Ca 2ϩ -NTA interactions were explored further.
NTBI Uptake Using NTA Concentrations Based on Literature Values-Based on pH, ionic strength, and temperature corrected literature values of the stability constants for Fe 2ϩ , Mg 2ϩ , and Ca 2ϩ binding to NTA and ascorbate, 0.43 M total Fe 2ϩ in ascorbate-PSS with 0 mM Ca 2ϩ and 28 M NTA gave an identical free Fe 2ϩ concentration (4.8 nM) as using 2 mM Ca 2ϩ and 74 M NTA. Using these concentrations of NTA gave a lower value for the NTBI uptake with 2 than with 0 mM Ca 2ϩ (Table I). However, the computations for the experiment presented in Table I were based on 16 stability constants. Each of these had been determined at different metal ion concentrations, temperatures, and ionic strengths as reported in the literature and corrected to our experimental conditions. A dis-FIG. 1. Calcium activation of NTBI uptake at a constant ratio of iron to NTA. A, NGF-treated PC12 cells were incubated at a concentration of 0.33 mg protein/ml in a solution containing 0.43 M total iron and 28 M NTA at 37°C in the presence of 0 or 2 mM CaCl 2 in ascorbate-PSS. Aliquots were filtered after 2, 4, 6, 8, and 10 min. Ca 2ϩ -dependent uptake is the difference between uptake at 2 and 0 mM CaCl 2 . All of the values are the means Ϯ S.E. of 6 replicates. The experiment was conducted 5 times, and a Ca 2ϩ -dependent iron uptake was observed in all experiments. B, cells were incubated for 2 min at 37°C in solutions containing iron and NTA, as described in A, and the specified concentrations of CaCl 2 in ascorbate-PSS. All of the values are the means Ϯ S.E. of 6 replicates. The experiment was conducted 3 times, and a similar Ca 2ϩ concentration dependence was observed. C, uptake was carried out at 4 or 37°C for 2 min in the uptake solutions used in A. All of the values are the means Ϯ S.E. of 6 replicates. The experiment was conducted twice, and a similar temperature dependence was observed. crepancy in any one of them could give erroneous values. Therefore, we set out to determine stability constants under conditions of the NTBI uptake experiment.
Determination of Conditional Stability Constants-We determined the various conditional stability constants based on competition between Cd 2ϩ and other cations to bind NTA under conditions of the NTBI experiment (in the presence of ascorbate at the pH, ionic strength, temperature, and range of total [Fe 2ϩ ] and [Ca 2ϩ ]). The Cd 2ϩ -NTA binding constants were determined first on each day of the experiment. These data fit best with the model in which 0.94 Ϯ 0.01 molecule of NTA bound Cd 2ϩ with a K a(cond) of (3.32 Ϯ 0.09) ϫ 10 7 M Ϫ1 . In the next set of titrations, we used 2 or 5 mM Ca 2ϩ to compete with Cd 2ϩ for binding to NTA (Fig. 3A). These experiments fit a model in which NTA binds Ca 2ϩ with a stoichiometry of 0.907 NTA/Ca 2ϩ and K a(cond) 2.2 ϫ 10 4 M Ϫ1 . We then used 2 and 5 M FeSO 4 in the absence of added Ca 2ϩ to obtain a stoichiometry of 1.04 NTA/Fe 2ϩ and K a(cond) ϭ 1.9 ϫ 10 8 M Ϫ1 (Fig. 3B). Based on the binding constants for NTA to Ca 2ϩ and Fe 2ϩ , we computed the K a(cond) value for NTA binding to Fe 2ϩ in the presence of 2 mM Ca 2ϩ as 2.29 ϫ 10 6 M Ϫ1 . We then carried out the binding experiment using 2 and 5 M FeSO 4 in the presence of 2 mM Ca 2ϩ . This experiment gave K a(cond) ϭ 2.34 ϫ 10 6 M Ϫ1 , which was consistent with the computed value (Fig. 3C). Fig. 4A compares NTBI uptake in 0 and 2 mM Ca 2ϩ in ascorbate-PSS using a fixed total iron concentration and NTA concentrations based on the conditional stability constants to clamp free Fe 2ϩ concentration at 0.1 M. Ca 2ϩ did not produce an activation. In three experiments, activation of iron uptake in 2 min by 2 mM Ca 2ϩ was 9 Ϯ 8%. This experiment indicated that Ca 2ϩ chelation to NTA was the explanation for the 151 Ϯ 23% activation we had observed in the initial protocol using a constant amount of NTA with and without Ca 2ϩ . To determine whether the lack of activation in Fig. 4A was because of the free [Fe 2ϩ ] chosen for the experiment, we conducted an experiment at various concentrations of free Fe 2ϩ , which we prepared using a fixed concentration of NTA and a series of concentrations of total Fe 2ϩ based on the conditional stability constants and correcting for NTA binding to Ca 2ϩ . 2 There was no effect of Ca 2ϩ on NTBI (Fig. 4B). DISCUSSION The results presented here demonstrate that the Ca 2ϩ activation of NTBI uptake observed using the established literature protocol is an artifact caused by Ca 2ϩ -NTA interaction. The "Discussion" will focus on the validity of the methods used, the need for determining conditional stability constants, possible implications of these observations to the field of NTBI uptake in general, and the biological relevance of these findings.

NTBI Uptake Using [NTA] Based on Conditional Stability Constants-
NTBI transport has been reported in many cell types, and in some cases, shown to be activated by Ca 2ϩ with an EC 50 value of 0.5-1 mM (8,12,13); in these studies, iron was maintained as Fe 2ϩ using ascorbate and chelated using NTA. In several studies, a constant iron to NTA ratio was used with and without Ca 2ϩ , and in others it is not clear how the required concentrations of NTA were determined (8,(12)(13)(14). Using the established constant ratio method, PC12 cells showed NTBI uptake activation by Ca 2ϩ with the characteristics previously reported. However, this protocol is based on the binding of Ca 2ϩ to NTA to be too weak to affect the Fe 2ϩ -NTA buffering system. This issue has not been adequately addressed in the literature concerning NTBI uptake. Our computations using the previously reported absolute stability constants showed that 2 mM Ca 2ϩ bound 65% of the NTA in our initial experiments, which led to a doubling of the free Fe 2ϩ concentration. However, the matter is even more complex because Fe 2ϩ -NTA binding stoichiometry is not 1:1 (24). Furthermore, it is not known how this stoichiometry depends on metal ion concentration, pH, temperature, and ionic strength. Sixteen such constants were used in these  computations. Consequently, the validity of the classical methods for correcting standard constants to experimental conditions is questionable. We resolved this problem by determining conditional stability constants. These values were determined under the conditions of the NTBI uptake experiments. The results obtained using the conditional stability constants were in agreement with the experiments without using NTA in that NTBI uptake is not activated by Ca 2ϩ . The results using the established protocols, however, are flawed. Based on the conditional stability constants using 0.43 M total Fe 2ϩ and 28 M NTA in the experiments in Fig. 1, the free Fe 2 concentration would be higher in the presence of 2 mM Ca 2ϩ (52 nM) than at 0 mM (28 nM). Computations using the conditional stability constants (Table I) show that Ca 2ϩ decreased the NTBI uptake because the NTA concentration used resulted in lower free Fe 2ϩ in the presence of Ca 2ϩ (20 nM) than in its absence (28 nM). Thus, all four experiments are consistent with the conclusion that Ca 2ϩ binds NTA, indirectly affecting the free Fe 2ϩ concentration, but it has no direct effect on the NTBI transporter. Several studies have used citrate instead of NTA to chelate Fe 2ϩ (1), but the chelation problem highlighted in this study may also apply to citrate. These results suggest a need to re-evaluate the Ca 2ϩ -activated pathway for NTBI uptake.