Monensin Mediates a Rapid and Selective Transport of Pb 2 (cid:1) POSSIBLE APPLICATION OF MONENSIN FOR THE TREATMENT OF Pb 2 (cid:1) INTOXICATION*

The carboxylic acid ionophore monensin, known as an electroneutral Na (cid:1) ionophore, an anticoccidial agent, and a growth-promoting feed additive in agriculture, is shown to be highly efficient as an ionophore for Pb 2 (cid:1) and to be highly selective for Pb 2 (cid:1) compared with other divalent cations. Monensin transports Pb 2 (cid:1) by an elec-troneutral mechanism in which the complex PbMonOH is the transporting species. Electrogenic transport via the species PbMon (cid:1) may also be possible. Monensin catalyzed Pb 2 (cid:1) transport is little affected by Ca 2 (cid:1) , Mg 2 (cid:1) , or K (cid:1) concentrations that are encountered in living systems. Na (cid:1) is inhibitory, but its effectiveness at 100 m M does not exceed (cid:2) 50%. The poor activity of monensin as an ionophore for divalent cations other than Pb 2 (cid:1) is consistent with the pattern of complex formation constants observed in the mixed solvent 80% methanol/water. This pattern also explains why Ca 2 (cid:1) , Mg 2 (cid:1) , and K (cid:1)

Ionophores are lipophilic chelating agents that transport cations across phospholipid bilayer membranes, such as the plasma and subcellular membranes of cells. The naturally occurring compounds (about 100) are antibiotics produced primarily by soil bacteria of the Streptomyces genus (1). They are distinguished from pore-forming antibiotics such as gramicidin by the involvement of a discrete complex in the transport mechanism. Thus, true ionophores can be highly selective for particular cations (2,3). The known compounds are generally divided into two groups: the so-called electrogenic and the electroneutral ionophores (4,5). Electrogenic ionophores are typified by the well known compound valinomycin (6). They are neutral molecules that form cation complexes that carry a net positive charge, such that transbilayer charge movements accompany transport catalyzed by this class. Accordingly, the rate of transport is influenced by membrane electrical potential, which also determines the transmembrane distribution of the transported cation at equilibrium. Compounds such as monensin, nigericin, and A23187 ( Fig. 1) typify electroneutral ionophores, also called carboxylic acid or polyether ionophores. The anionic forms of these compounds complex the cation, which is exchanged for H ϩ , or another cation, without net charge movement. Transport catalyzed by this class is therefore influenced by transmembrane pH gradients, as is the equilibrium distribution of cations.
Among the carboxylic acid ionophores, it has been common to consider individual compounds as able to complex/transport a small group of cations having the same charge, with transport occurring through a single species by a mechanism that is purely electrogenic or electroneutral. Thus, monensin and nigericin were known as ionophores for Na ϩ and K ϩ , respectively, with transport occurring via a 1:1 complex and electroneutral mechanism; A23187 was known as a Ca 2ϩ ionophore transporting via a 1:2 complex and an electroneutral mechanism, and so forth. In work conducted over a period of time, we showed that the model depicted in Fig. 2 is a more accurate representation of the factors that establish the transport properties of a carboxylic acid ionophore (7)(8)(9)(10)(11)(12)(13)(14)(15)(16). The equilibrium shown at the top of Fig. 2 represents formation of 1:1 complexes and emphasizes that carboxylic acid ionophores react with a broad range of cations (n ϭ 1-3), not just those with a particular charge as often assumed. The 1:1 complexes then react to form higher order species by competing equilibria I-III, which involve free ionophore (A Ϫ ), hydroxide ion (OH Ϫ ), and other anions (X Ϫ ), respectively. Complexes of differing stoichiometry therefore arise, which include mixed species containing OH Ϫ and X Ϫ . Within this scheme, the mode and overall rate of transport will reflect the distribution of ionophore between these various complexes and their respective transmembrane diffusion constants. As a corollary of this interpretation, for a given ionophore and set of conditions, transport selectivity will also depend on how these factors vary with properties of the cation. The latter include size, charge, coordination and hydration number, preferred donor atom geometry, Lewis acidity, and ease of hydrolysis.
In considering the above model, the characteristics of the donor atoms in typical ionophores, and the metal ion complexation properties of analogous synthetic polyethers (17), we realized that some of the antibiotic compounds might be efficient ionophores for Pb 2ϩ and for other cations with biological toxicity.
Subsequently, we showed that ionomycin transports Pb 2ϩ rapidly and with selectivity over Ca 2ϩ near 10 3 when both cations are present simultaneously. Ionomycin was also found to affect an efficient transport of Pb 2ϩ into cultured cells as well as to facilitate the depletion of Pb 2ϩ when the cells had been previously loaded. We indicated that ionomycin should be considered primarily as an ionophore for Pb 2ϩ , rather than Ca 2ϩ , and suggested that its Pb 2ϩ -transporting activity might be adapted to improve existing treatments for Pb 2ϩ intoxication (15).
In the present report, we extend the investigation of ionophore-mediated Pb 2ϩ transport by demonstrating that monensin is also effective as a Pb 2ϩ ionophore and is more selective in that regard than is ionomycin. We also show that monensin promotes the excretion of Pb 2ϩ from rats when the cation has been previously provided in drinking water and the ionophore is administered in feed. Aspects of these data have been presented in abstract form (18).

EXPERIMENTAL PROCEDURES
Reagents and Solvents-High purity nitric acid (trace metal; Fisher) and perchloric acid (double-distilled; GFS Chemicals) were obtained from commercial sources. Synthetic 1-palmitoyl-2-oleoyl-sn-glycerophosphatidylcholine (POPC) 1 was obtained from Avanti Polar Lipids, Inc. Purity was confirmed by thin layer chromatography before use. For the transport studies, monensin and ionomycin were obtained from Calbiochem and used without further purification. Ionomycin stock solutions were prepared in ethanol and standardized spectrophotometrically using an extinction coefficient of 13,560 at 278 nm. In the case of monensin, standardization was gravimetric or by titration with Me 4 NOH. Quin-2 (K ϩ salt) from Sigma was purified by passage over Chelex 100 resin (100 -200-mesh) in the Cs ϩ form as described previously (19), or in the Na ϩ form when Na ϩ -loaded vesicles were employed. The nitrate and chloride salts of divalent cations were the ultrapure grade from Alfa Products. Stock solutions were standardized by titration with a primary standard EDTA solution (20) or by atomic absorption spectroscopy using certified solutions (Fisher).
For solution chemical studies, a mixed solvent of 80% (w/w) methanol in water was prepared gravimetrically, using distilled deionized water and reagent grade methanol (Fisher) that had been freshly distilled. The Et 4 NClO 4 that was used to maintain ionic strength in this solvent was prepared by reaction of Et 4 NOH (Aldrich) with 70% perchloric acid (distilled; GSH Chemicals). The salt obtained was recrystallized four times from water. Solvent containing Et 4 NClO 4 and H ϩ buffering compounds was further deionized by passage over Chelex 100. For this purpose, the resin was in the Et 4 N ϩ form, which was prepared as previously described (8).
Preparation of Phospholipid Vesicles-The preparation of freezethaw-extruded POPC vesicles loaded with Quin-2 has also been described previously (21,22). Briefly, 300 mg of POPC in chloroform was dried by rotation under a nitrogen stream to produce a film on the wall of a 25 ϫ 150-mm culture tube. Residual solvent was removed under high vacuum (4 h), and the film was subsequently hydrated in 6 ml of a solution containing 6.6 mM purified Quin-2 and 10.0 mM Hepes buffer adjusted to pH 7.00 with Chelex-treated CsOH or NaOH (19), depending on the internal composition required. The mixture was vortexed, and the resulting multilamellar vesicles were frozen in a dry ice-acetone bath, thawed in lukewarm water, and vortexed again. The freeze-thaw and vortexing procedures were repeated two additional times, after which the vesicles were extruded three times through two stacked 100-nm polycarbonate membrane filters. This step was followed by six additional freeze-thaw cycles coupled with additional extrusions. The resulting preparations were applied to Sephadex G-50 minicolumns (23) to remove extravesicular Quin-2. These columns were eluted by low speed centrifugation and had previously been equilibrated with a solution containing 10 mM Hepes buffer, pH 7.00. A single pass over such columns effectively removes the external Quin-2 (19,21,22).
FIG. 2. Equilibria between an ionized carboxylic acid ionophore (A ؊1 ), a metal cation of charge n؉ (M n؉ ), and selected anions (A ؊ , X ؊ , and OH ؊ ). Following formation of a 1:1 complex between the ionophore and a cation (MA n Ϫ 1), additional species can arise that may contribute to transport. In I, a 1:2 complex is formed by reaction with a second ionophore molecule. In II and III, mixed complexes arise by reaction with OH Ϫ or another anion (X Ϫ ), respectively. Quin-2 and Cs ϩ /Na ϩ were determined for each preparation by the methods described previously (19,25). Briefly, entrapped Quin-2 is determined by spectrophotometric titration with standard CaCl 2 following dispersion of the vesicles in deoxycholate. The entrapped monovalent cation is determined by atomic absorption spectroscopy, following replacement of the external medium with one containing a different cation, and dispersion of the vesicles in 0.1 N HCl. When of interest, buffer entrapment is determined from the other values by calculation, using the Henderson-Hasselbach equation, the buffer pK a , and the internal pH. When buffer entrapment is to be determined, the vesicles also contain the fluorescent pH indicator 2Ј,7Јbis(2-carboxyethyl)-5(6)-carboxyfluorescein, so that the internal pH can be ascertained. The internal pH and solute concentrations differ from those of the vesicle formation medium because of a freeze-thawdriven solute-concentrating effect that operates during preparation of the vesicles (21,22).
Pb 2ϩ Buffers and the Determination of Transport-A buffer system was used to control the concentration of Pb 2ϩ available for transport into the vesicles. 15 or 5 mM citrate was employed to buffer the concentration of this cation, whereas 10 mM each of Hepes and Mes were present to buffer H ϩ . Seventeen equilibria involving citrate 3Ϫ , H ϩ , Pb 2ϩ , and OH Ϫ were accounted for when calculating the free Pb 2ϩ concentration. When Ca 2ϩ or Mg 2ϩ was to be buffered simultaneously, five additional equilibria involving those cations were also considered. The respective equilibrium constants were taken from literature sources (26,27), and, when necessary, the Davies equation (28) was used to correct these to an ionic strength of 100 mM. The species distribution program COMICS (29) was used to solve the applicable sets of simultaneous equations at experimental conditions of interest and to allow the generation of standard curves. Examples of the latter were shown previously (15).
The transport of Pb 2ϩ and other divalent cations into Quin-2-loaded vesicles was determined by monitoring formation of the Quin-2-cation complexes spectroscopically. Unless otherwise indicated, vesicles containing Quin-2 were present at a nominal POPC concentration of 1.0 mM in a medium that also contained 50 mM CsCl and 10 mM each of Hepes and Mes. The medium pH ranged from 6.00 to 9.50 and was adjusted with CsOH that had been passed over Chelex 100 columns to remove contaminating divalent cations (19). In some cases, valinomycin (0.5 M) and carbonyl cyanide m-chlorophenylhydrazone (CCCP) (5 M) were also present to maintain internal pH at the external value and to dissipate any transmembrane electrical potential that might otherwise arise (25). Specific concentrations of ionophores, divalent cations, and pH values are given in the figure legends. Reactions were started by the addition of the ionophore, following an initial 2-min preincubation to allow equilibration of transmembrane pH.
The formation of Quin-2-cation complexes was followed continuously by difference absorbance spectroscopy, using an Aminco DW2a spectrophotometer operated in the dual wavelength mode. An Oriel 59800 band pass filter was used between the cuvette and the beam scamblerphotomultiplier assembly to prevent detection of the fluorescent light emitted by Quin-2. The sample wavelength used for all cations was 264 nm. The reference wavelengths were at an isosbestic point in the Quin-2/Quin-2-cation complex difference spectrum of interest. These wavelengths vary slightly from cation to cation, as previously described (13). Data were collected on disk using Unkel Scope software (Unkel Software, Inc., Lexington, MA).
To determine initial transport rates, an early portion of the progress curves was fit to Equation 1 using standard nonlinear least-squares methods.
In this expression, A T and A 0 are the observed and the initial absorbance values, respectively, B is the initial rate in units of absorbance/s, C is a correction factor for nonlinearity, and t is time in seconds. The values presented are in units of M external cation transported into the vesicles per second. B values obtained from Equation 1 were converted to the latter unit by referring to a standard curve for the cation of interest that was generated by titrating the vesicles in the presence of excess ionophore or after they had been lysed with 0.33% (w/v) of Cs ϩ -deoxycholate (13,19). Transport selectivities are expressed as S values defined by Equation 2.
S ϭ Initial rate of Pb 2ϩ transport/Initial rate of Ca 2ϩ transport (Eq. 2) When determining S, the concentration of the Pb 2ϩ or Ca 2ϩ was 20 M, and all other conditions were held constant. All transport data were obtained at 25.0°C.
Potentiometric Titrations and the Determination of pH in Aqueous Methanol-The protonation constants and complex formation constants of monensin were measured by potentiometric methods in the mixed solvent 80% (w/w) methanol/water. For these studies, Na ϩ monensin was purified by column chromatography on silica gel, using ethyl acetate as the eluent. The Na ϩ salt was converted to the acid form by repeated back-extraction of a CHCl 3 solution with 1.0 M HCl. The CHCl 3 solution was then washed three times with distilled water, the solvent was removed, and the product was dried under vacuum. Analysis by flame atomic emission showed that less than 0.05% Na ϩ (w/w) remained.
The pH meter-electrode system was calibrated in aqueous solution using standard buffers (Gram-Pac; Fisher); then the electrodes were equilibrated in 80% methanol/water for at least 2 h prior to measurement. The operational pH* scales developed by de Ligny et al. (30,31) and Gelsema et al. (32,33) were utilized to determine the value of pH*. The term pH* is defined as Ϫlog a H *, where a H * is the activity of H ϩ in the mixed solvent. Accordingly, the term pH* when used in reference to a specific methanol/water mixture has the same meaning as the term pH when used in reference to an aqueous solution (see Ref. 34 and references therein).
All titrations were carried out at 25.0°C using a thermostatted cell and 20 mM Me 4 NOH as the titrant. A nitrogen or argon atmosphere was maintained to minimize contamination by CO 2 . The Me 4 NOH was standardized using KH 2 PO 4 and checked for carbonate content (35). Typical titrations consisted of 50 -100 pairs of pH* versus ml Me 4 NOH readings. The titration data were analyzed using the computer programs PKAS for the protonation constants (36) and BEST for the metal ion complexation constants (37).
Treatment of Experimental Animals-Male rats were utilized when investigating the effects of monensin on the pathophysiology of Pb 2ϩ . They were housed in AALAC-approved animal facilities at the College of Medicine, Ohio State University. A 12-h light-dark cycle, single housing in plastic cages, or in metabolic cages, and conditions of constant temperature and humidity were employed. One week was allowed for acclimation before an experimental protocol began. During this period, a standard laboratory chow was employed, whereas the AIN-93 M diet containing 0.5% calcium was employed thereafter. During the administration of Pb 2ϩ and/or monensin, water and feed were provided ad libitum, and records of consumption were maintained, together with periodic measurements of body weight.
At the beginning of an experimental protocol, the rats weighed 245-255 g. They were divided randomly into groups of six or eight, and the administration of Pb 2ϩ was begun. It was provided at 100 ppm in the drinking water and was in the form of Pb(acetate) 2 . The water was rendered slightly acidic with acetic acid to prevent the precipitation of PbCO 2 (38). When utilized, monensin was provided in the feed, also at 100 ppm.
Determination of Pb in Biological Samples-Blood samples were taken periodically from the tail artery, and blood Pb levels were determined by electrothermal atomic absorption spectroscopy against certified standards (39). For the determination of Pb in urine and feces, total outputs were collected over 2-day periods from each rat individually. The urine samples were neutralized by the careful addition of nitric acid while the samples were stirred, and the pH was monitored with an electrode. Samples were then extracted with methyl isobutyl ketone in the presence of excess ammonium pyrrolidine dithiocarbamate, which quantitatively extracts Pb 2ϩ into the organic layer and concentrates it compared with the original concentration in urine (40). The Pb content was thereafter determined by flame atomic absorption spectroscopy. For the determination of Pb in feces, weighed samples comprising ϳ1.0 g were dispersed in 10 ml of concentrated nitric acid and allowed to stand at room temperature overnight. They were then concentrated to ϳ1.5 ml by heating on a hot plate and diluted to a total volume of 10.0 ml with water, and Pb was determined by flame atomic absorption spectroscopy (41).
At the completion of experimental protocols, the rats were sacrificed by the injection of excess Nembutal and perfused briefly with Hepes-buffered 0.9% NaCl, via the left ventricle, to remove blood from the organs. Organs and tissues of interest were then removed and stored at Ϫ20°C. For Pb analysis, the samples were thawed, and weighed portions were digested with 9 volumes of nitric/perchloric acid (3:1). After reducing the volume to 1 ml by heating, the samples were diluted to a total volume of 10.0 ml with a solution containing 140 mM NH 4 OH, 29 mM NH 4 (H 2 PO 4 ), and 3 mM EDTA. For samples containing relatively high levels of Pb, such as bone and kidney, an initial dilution with water preceded that step. Following all of the above steps, the Pb content was determined by electrothermal atomic absorption spectroscopy (42). The resultant values and other parameters pertaining to rats were examined by univariant analysis of variance to determine whether the differences observed were statistically significant. Fig. 3 compares the efficiency of ionomycin and monensin as ionophores for Pb 2ϩ and contrasts their selectivity for Pb 2ϩ compared with other divalent cations. Under the conditions employed, the two compounds are similarly effective from the perspective of rate, whereas monensin is much more selective. Selectivity for Pb 2ϩ over Ca 2ϩ , for example, as defined by Equation 2, is ϳ100 for ionomycin and ϳ3400 for monensin. 2 The stoichiometry of the complex between ionomycin and Pb 2ϩ that is responsible for transport is 1:1, cation/ionophore, based upon plots of log rate versus log ionomycin or log Pb 2ϩ concentration, which both display slopes of 1.0 (15). By the same criteria, monensin-mediated Pb 2ϩ transport also occurs through formation of a 1:1 complex (Fig. 4), although the plot of log rate versus log Pb 2ϩ concentration progressively deviates from a slope of 1 as the free Pb 2ϩ concentration rises above 1 M. This negative deviation suggests that the steady state fraction of monensin that is located at the external membrane interface approaches saturation with Pb 2ϩ at concentrations above that value.

Monensin-mediated Pb 2ϩ Transport-
Ionomycin is dibasic, because both the carboxylic acid function and the enolized ␤ diketone moiety can be ionized ( Fig. 1) (43). Accordingly, ionomycin can form uncharged complexes with divalent cations, presumably including Pb 2ϩ , and can exchange these for 2H ϩ in an electroneutral manner (19,25). In contrast, monensin is monobasic (Fig. 1) (44) and would form a 1:1 complex with Pb 2ϩ having a net charge of ϩ1. Thus, the 1:1 complex between Pb 2ϩ and monensin might result in electrogenic Pb 2ϩ transport, or it might associate with an anion as depicted in Fig. 2, to produce transport through a charge neutral mechanism. To examine the mode of Pb 2ϩ transport, we initially determined complex stability constants of potential transporting species. Potentiometric titration methods were employed using 80% methanol/water as the solvent. This mixed solvent provides an effective polarity similar to that experienced by ionophores at a POPC membrane interface (7,12,45,46). Analogous complexation equilibria for other cations were also examined to provide insight into the basis of the high selectivity that is seen in Fig. 3B. Fig. 5A shows examples of the primary data obtained, whereas Fig. 5B reports the complex stability constants. It is seen that the complex PbMon ϩ is relatively stable, displaying a log K value of 7.25. The uncharged complex PbMonOH that is formed upon reaction of PbMon ϩ with OH Ϫ is nearly 4 orders of magnitude more stable than the 1:2 complex formed from PbMon ϩ and a second molecule of the ionized ionophore. When these stability constants and the protonation constant of monensin are used to generate a species distribution diagram (29) (Fig. 5C), it is seen that significant levels of both PbMon ϩ and PbMonOH are expected at the membrane interface during Pb 2ϩ transport, within the pH range of 6.5-7.5. In contrast, the species Pb(Mon) 2 is negligible (Ͻ0.003% of total monensin) throughout the broad range of pH considered (not shown). Thus, the complex stability data indicate that monensin might transport Pb 2ϩ by a mixed mode (electrogenic and electroneutral transport occurring simultaneously) and that the fraction transported by the neutral process would probably occur via the species PbMonOH.
To further examine these possibilities, the effect of external pH on the rate of Pb 2ϩ transport was examined, using Na ϩcontaining vesicles with valinomycin and CCCP excluded. These conditions limit transport to the fraction occurring by neutral mechanisms, because no provision is made to collapse . After a 2-min preincubation, 0.10 M ionophore was added to initiate transport (designated as 0 s in the figure). The sequence of cations shown on the right corresponds to their relative rate of transport as observed directly for A or on an expanded scale for B. The y axis unit refers to cation concentration in the external medium. When this parameter reaches 20 M, all of the cation that was originally present in the external medium has been transported into the vesicles and sequestered by Quin-2. the membrane potential arising from electrogenic transport, which is very effective at limiting the process (e.g. Refs. 14 and 16). In addition, the use of Na ϩ rather than Cs ϩ -containing vesicles allows the ionophore to exchange Pb 2ϩ for Na ϩ , ensuring that the rate of Pb 2ϩ transport does not become limited by the protonation of monensin at the internal membrane interface. As seen in Fig. 6A, monensin remains active as an ionophore for Pb 2ϩ under these conditions, demonstrating that the electroneutral mode is indeed active.
Regarding the species PbMonOH, the data in Fig. 6B show that increasing external pH enhances the rate of transport, consistent with a model involving dissociation of a single proton from hydrated PbMon ϩ to form PbMonOH. This finding and a comparison of the transport data with the pH dependence for PbMonOH formation shown in Fig. 5C indicate that PbMonOH is indeed the major transporting species. In Fig. 6, the half-maximal rate was seen at pH 7.8 (pOH 6.2) or at pPbOH 6.9 if it is considered that that species is the one which is actually transported.
Competitive Relationships between Pb 2ϩ and Other Cations-Given the potential use of ionophores to manipulate Pb 2ϩ in living systems, we considered the possibility that the relatively high concentrations of other cations found in blood and intracellular compartments might substantially prevent monensin from transporting Pb 2ϩ in vivo. Specifically, the rates of Pb 2ϩ transport were compared when the free Pb 2ϩ concentration was buffered at 1.0 M alone and when 1.0 mM free Ca 2ϩ or Mg 2ϩ was also present. Only negligible differences were seen (Fig. 7, A and B), as might be expected based upon the complex stability constants shown in Fig. 5B. Potential interference by K ϩ and Na ϩ was also examined. The actions of K ϩ were again negligible when it was present at 5.0 or 100 mM, approximately the levels of this cation in blood and cytoplasm, respectively (data not shown). The same concentrations of Na ϩ did have modest inhibitory effects (Fig. 7C); however, these were smaller than expected based on the K ML values listed in Fig. 5B. In other words, the ratio K Pb /K Na is ϳ300, whereas at a ratio of free Na ϩ /Pb 2ϩ of 10 5 (i.e. when Na ϩ was present at 100 mM and Pb 2ϩ at 1.0 M), the rate of Pb 2ϩ transport is reduced by only a factor of Ͻ2.
To further examine the effectiveness of Na ϩ as an inhibitor of monensin-mediated Pb 2ϩ transport, the Na ϩ concentration was varied from 0 to 100 mM while another solute was varied in the opposite direction, so as to maintain the external ionic strength and/or the osmotic pressure at constant values. As seen in Fig. 8, Na ϩ is more effective when used to replace Cs ϩ , compared with K ϩ or tetraethylammonium cation, and it is ineffective when used to replace mannitol. These data presumably reflect modest differences in the stability of complexes between monensin and monovalent cations (47)(48)(49) together with effects of ionic strength on complex stability. However, of greater importance, they further indicate that concentrations of Na ϩ found in living systems have little effect on the efficiency of monensin as a Pb 2ϩ ionophore.
Monensin Promotes the Excretion of Pb 2ϩ in Rats-Given that the above data are consistent with the notion that monensin might alter the dynamics of Pb 2ϩ in whole organisms, we determined the effect of monensin on the accumulation and disposition of Pb in rats. In one experiment, the ionophore and Pb 2ϩ were administered simultaneously, at 100 FIG. 4. Pb 2؉ transport dependence on the concentrations of Pb 2؉ and monensin. Data were obtained as described in the legend to Fig. 3 except that the concentrations of Pb 2ϩ and monensin were varied. In addition, the external free Pb 2ϩ concentration was established by a 15 mM citrate-based buffer system, as described under "Experimental Procedures." A, the monensin concentration was 0.10 M, and the concentration of free Pb 2ϩ was varied from 10 nM (bottommost curve) to 5.63 M (topmost curve). B, the free Pb 2ϩ concentration was 1.0 M, and the concentration of monensin was varied from 1.17 nM (bottommost curve) to 2.40 M (topmost curve). C, the data from A and B are shown as log initial rate versus log Pb 2ϩ (E) or log monensin (q) concentration. Initial rates were obtained from the individual progress curves as described under "Experimental Procedures." ppm in feed and 100 ppm in drinking water, respectively, or Pb 2ϩ was administered without the ionophore. Blood samples were taken at weekly intervals over a 28-day period, and thereafter the rats were sacrificed to allow the determination of Pb in organs and tissues. Throughout the entire experimental period, the average concentration of Pb in blood was lower by ϳ25% in the rats that had received monensin, and this difference was significant at p ϭ 0.05 (Table I and data not shown). Monensin also reduced the accumulation of Pb in several organs, with a particularly large effect seen in heart, where it was reduced by 66% (Table I). The reduced values seen in brain, muscle, and bone were also significant.
In a second experiment, the rats were loaded with Pb 2ϩ over a 21-day period and in the absence of monensin. Thereafter, they were divided into groups and given drinking water that did not contain Pb 2ϩ . One group was given monensin at 100 ppm in feed, whereas another was given the same feed without monensin. After an additional 21 days, the rats were sacrificed, and lead in organs and tissues was determined. This time there  6. Effect of external pH on ionophore-mediated Pb 2؉ transport. Vesicles were prepared as described under "Experimental Procedures" except that Na ϩ rather than Cs ϩ was employed as the counter ion to Quin-2. The external medium contained 10 mM each of Mes, Hepes, and Ches, to provide for H ϩ buffering over a broad range of pH, and 5 mM citrate was present to buffer free Pb 2ϩ . K ϩ was used as the external counter ion for citrate and the H ϩ buffers. The external pH was set at a value of interest, and the concentration of Pb(NO 3 ) 2 required to produce a free Pb 2ϩ concentration of 0.10 M was added to the cuvette. Pb 2ϩ transport was initiated by the addition of 0.1 M monensin. A, individual progress curves covering the pH region of 6.5-9.0. The earliest portions are omitted, because the initial equilibration of H ϩ and monovalent cation gradients perturbs the UV-visible spectrum of Quin-2. Accordingly, rates pertaining at 140 s were determined rather than the initial rates. B, rates obtained from A at the pH indicated. Values are the mean of duplicate determinations. They were fit to the Henderson-Hasselbalch equation to obtain the solid line.
was no difference in blood lead, either at the time of sacrifice (Table II) or at earlier sampling times (data not shown). However, the rats given monensin had lower levels of lead in brain, kidney, liver, and bone. The greatest effect was seen in kidney (a 55% reduction), with the effect on liver being of similar magnitude. Consistent with the transport data shown in Fig. 3 and the complex stability constants shown in Fig. 5B, the reduced levels of Pb in organs were obtained without reducing the levels of zinc or copper (Table III). Thus, the actions of monensin on lead are not accompanied by perturbation of these trace elements having biological roles.
During this second experiment, the rats were housed in metabolic cages so that urine and feces could be collected to determine the fate of Pb released from the organs. Amounts excreted in urine were low and were unaffected by the presence or absence of monensin (Fig. 9). Determining excretion via the feces is more problematic, because the gastrointestinal tract in rats rejects most Pb that is ingested, resulting in high levels of Pb in feces during the loading period (Refs. 50 and 51 and data not shown). When Pb 2ϩ is withdrawn, feces Pb remains high for several days while the contents of the gastrointestinal tract turn over. These large values of excreted Pb obscure smaller differences between groups, such as differences that might be produced by monensin. Nevertheless, at later times, a statistically significant effect of monensin on Pb in feces was observed. In other words, during the second and third week of monensin administration, the treated rats lost lead at a rate of ϳ120 nmol/day, whereas without monensin the rate was about half that value. As further considered below, these data suggest that the lead mobilized from organs by monensin is ultimately excreted via the feces.

DISCUSSION
Monensin was among the first group of carboxylic acid ionophores to be discovered, a group that also includes dianemycin, nigericin, compound X-206, and lasalocid A (reviewed in Ref. 5). Soon after these compounds were reported to be potent anticoccidial agents (52) and to derive this activity through a direct interaction with monovalent cations, leading to altered cation transport (53). Among the monovalent alkali metal cations, monensin was subsequently shown to be most effective as an ionophore for Na ϩ , and it has long been utilized as a research tool in that regard (reviewed in Refs. 2 and 54). It has also been used as a feed additive in agriculture, because of the anticoccidial activity and because it promotes the growth of several animal species that are used in that industry (55,56). FIG. 7. Actions of physiological cations as inhibitors of Pb 2؉ transport. Vesicles were prepared as described under "Experimental Procedures" using Cs ϩ as the counter ion to Quin-2. The external medium contained 5 mM citrate (Cs ϩ ) plus Pb(NO 3 ) 2 sufficient to produce a free Pb 2ϩ concentration of 1.0 M (except for traces labeled ϪPb 2ϩ ), and 5 mM each of Mes and Hepes (both Cs ϩ ), pH 7.0. As indicated, the medium also contained CaCl 2 sufficient to produce a free Ca 2ϩ concentration of 1.0 mM (A), MgCl 2 sufficient to produce a free Mg 2ϩ concentration of 1.0 mM (B), or the indicated concentration of NaCl (C). Transport was initiated by the addition of 0.1 M monensin.
FIG. 8. Inhibition of Pb 2؉ transport by Na ؉ . Vesicles were prepared as described under "Experimental Procedures," using Cs ϩ as the counter ion to Quin-2. The external medium contained 5 mM citrate plus Pb(NO 3 ) 2 sufficient to produce a free Pb 2ϩ concentration of 1.0 M, and 5 mM each of Mes and Hepes, pH 7.0. At NaCl ϭ 0 the medium also contained 100 mM CsCl (q), 100 mM KCl (E), 100 mM tetraethylammonium percholorate (TEAP; OE), or 200 mM mannitol (OE). Cs ϩ , K ϩ , or tetraethylammonium cation was used as the counter ion for the external buffers, so as to match the cation that was added at 100 mM. In the case of mannitol, Na ϩ was used. As the NaCl concentration was increased, the concentration of the external cation was decreased by an equal value or was decreased by twice that value in the case of mannitol. Monensin was used at 0.1 M.
Relatively recent studies have shown that monensin also forms complexes with several divalent cations and that they are generally more stable than those formed with monovalent cations when compared in nonaqueous solvents such as methanol or diethylformamide (47,48,(57)(58)(59). More limited data have furthermore shown that monensin displays some capacity to extract divalent cations from a bulk aqueous to a bulk organic phase (57) and to convey these cations through three phase bulk solvent systems (60) and polyvinyl chloride membranes in ion-selective electrodes (61,62). Thus, previous studies have suggested that monensin might not be strictly an ionophore for monovalent cations, but they did not reveal that it is actually a highly selective and efficient ionophore for Pb 2ϩ when phospholipid membranes are employed. That activity is newly discovered and is shown in the present report (Figs. 3 and 4).
The stoichiometry of the complex that transports Pb 2ϩ is clearly 1:1, ionophore/cation, based upon the relationships between log rate and log monensin or log Pb 2ϩ concentration, which are both linear over a wide range and display slopes of 1.0 (Fig. 4). As illustrated in Fig. 10, the 1:1 complex PbMon ϩ may then be active as a transporting species and would function by an electrogenic mechanism. This type of mechanism may have contributed to the transport seen in Figs. 3 and 4, because valinomycin and CCCP were present during those experiments. Valinomycin and CCCP, acting together, are fully effective at collapsing an induced membrane potential in the present vesicle system (14,16), which is required for sustained electrogenic transport.
No provision was made to collapse membrane potential during the experiments reported in Figs. 6 -8, so only electroneutral transport was contributing. The transporting species under those conditions is PbMonOH, as shown by the relationship between rate and external pH (Fig. 6) and by comparing that relationship to the distribution of Pb/monensin species that arises when the pH is varied (Fig. 5C). PbMonOH can form by the pathways shown by Reaction 1 or 2, as further illustrated in Fig. 10, and both of these pathways can be considered to involve the acid ionization of a hydrating water molecule associated with Pb 2ϩ .
The present data do not provide insight into which of the pathways may be the most significant, but that may not be a point of central interest, assuming that all equilibria shown in Fig. 10 remain at or near equilibrium during monensin-catalyzed Pb 2ϩ transport.
Apart from the pathway forming PbMonOH, the extent of equilibration between reactants forming the transporting species is of interest when the origin of high selectivity for Pb 2ϩ transport is considered, compared with the transport of other divalent cations. If near equilibrium conditions are maintained, selectivity is expected to arise from differing complex stabilities, together with variation in the transmembrane diffusion constants of the various metal cationmonensin complexes. Several alkaline earth and first transition series divalent cations are poorly shielded when complexed by monensin (57), a factor that reduces transmembrane diffusion constants. However, the role of differential shielding in establishing the selectivity of monensin as an ionophore for Pb 2ϩ is uncertain, because shielding in Pb 2ϩmonensin complexes has not been investigated.
Regarding complex stabilities, the CaMon ϩ and the MgMon ϩ complexes are less stable than the PbMon ϩ complex by ϳ4 orders of magnitude ( Fig. 5B; Refs. 57 and 59), and in addition, no interaction of these complexes with OH Ϫ was detected during the potentiometric titrations. The latter finding was expected, given the hydrolysis properties of these cations (63), but more importantly, these considerations further explain why monensin has little or no activity as an ionophore for Ca 2ϩ (Fig.  3) and why neither Ca 2ϩ nor Mg 2ϩ is effective as an inhibitor of Pb 2ϩ transport (Fig. 7). Likewise, Zn 2ϩ is bound by monensin with low affinity compared with Pb 2ϩ , although the ZnMon ϩ and PbMon ϩ complexes bind OH Ϫ with similar affinities (Fig.  5A). From those considerations, one might expect monensin to have some activity as an ionophore for Zn 2ϩ , and that was observed in the vesicle system (Fig. 3). Thus, relative complex stabilities clearly contribute to establishing the transport selectivity sequence seen in Fig. 3.
On the other hand, Na ϩ is much less effective as an inhibitor of monensin-mediated Pb 2ϩ transport than would be expected on the basis of complex stability constants alone, were near equilibrium maintained between monensin, OH Ϫ , and the two competing cations (Figs. 7 and 8). Thus, kinetic constants of the cation complexation/decomplexation reactions, transmembrane diffusion constants, or a differing tendency of Pb 2ϩ and Na ϩ to accumulate at the membrane interface may contribute to the selectivity of monensin as an ionophore for Pb 2ϩ when TABLE I Monensin effects on Pb 2ϩ accumulation Two groups of six rats were given lead at 100 ppm in their drinking water, as further described under "Experimental Procedures." One group also received monensin at 100 ppm in feed, whereas the other group did not. After 28 days, the rats were sacrificed, and lead was determined in the indicated organs and tissues, as further described under "Experimental Procedures." "Muscle" and "bone" refer to the quadriceps muscle and the femur, respectively. All values are means Ϯ one S.D. value and are in units of nmol/g, wet weight, except for blood, which is in M. Control values (no lead, no monensin) are from the literature (69). For tissues marked with an asterisk, the lead plus monensin value was lower than the lead value at p Ͻ 0.05. For those marked # , the threshold was p Ͻ 0.10. For liver and kidney, the differences did not reach either threshold.  II Monensin effects on Pb 2ϩ clearance Two groups of eight rats were given Pb 2ϩ at 100 ppm in drinking water without the administration of monensin. After 21 days, lead was withdrawn, and the administration of monensin at 100 ppm in feed began in one of the groups, whereas normal feed was given to the other group. After an additional 21 days, the rats were sacrificed, and lead was determined in the indicated organs and tissues, as further described under "Experimental Procedures." "Muscle" and "bone" refer to the quadriceps muscle and the femur, respectively. All values are means Ϯ one S.D. value and are in units of nmol/g, wet weight, except for blood, which is in M. Control values (no lead, no monensin) are from the literature (69). As in Table II, for tissues marked with an asterisk, the plus monensin value was lower than the no monensin value at p Ͻ 0.05. For those marked # , the threshold was p Ͻ 0.10. For heart and muscle, the differences did not reach either threshold.  (15), monensin might be useful in the treatment of lead intoxication; however, there are no studies that have tested that possibility in animals. Accordingly, we determined whether monensin alters the accumulation or distribution of lead when the two agents are given simultaneously and whether monensin affects the disposition of lead that was accumulated previously. The level of Pb 2ϩ employed and the method of administration are typical of those used when investigating Pb toxicity in rats (64 -67). The level of monensin is the same as that usually given to chickens as an agricultural practice (56) and is below the toxic threshold for monensin in rats (68,69). 3 Monensin was administered as a component of feed, because that is the practice in agriculture and because activity upon oral administration is considered to be advantageous for agents used to treat lead intoxication.
The data show that in both types of experiment monensin reduced the level of lead in several organs and tissues (Tables   I and II). The design of studies that investigate the efficiency of agents used to treat lead intoxication varies widely, so it is difficult to compare the effectiveness of monensin with that of the traditional agents (e.g. EDTA and dimercaptosuccinate) in a succinct manner. Nevertheless, monensin appears to be similarly effective to these better investigated compounds (e.g. compare with Refs. 70 -72), although the dose employed here, the treatment interval, etc. have not been optimized. In addition, monensin reduces Pb in bone and brain, which is not true for the traditional compounds under some circumstances (64,73), and does not deplete the tissues of zinc and copper, which can occur as an unwanted side effect with some of the traditional compounds (74 -76). Thus, monensin may indeed be useful for promoting the elimination of lead from animals.
At this early stage, it is appropriate to consider the possibility than monensin acts differently than the traditional compounds as regards the mechanism by which lead excretion occurs. The traditional compounds are hydrophilic cation chelators that circulate in blood and form Pb 2ϩ complexes having stability constants of ϳ10 15 and higher. They are not  9. Effect of monensin on Pb excreted in urine. Two groups of eight rats were given 100 ppm Pb 2ϩ in drinking water, for 21 days, as further described under "Experimental Procedures." The rats were housed in metabolic cages, allowing total urine output by each rat to be collected over 2-day intervals. Lead was determined by atomic absorption spectroscopy and expressed as nmol of lead excreted per rat per day. Up to and including day 21, the values plotted are means of those obtained for all 16 rats. On day 22, Pb 2ϩ was withdrawn, and one group was shifted to a diet containing 100 ppm monensin, while the second was maintained without monensin. E, lead excreted by rats receiving monensin; q, lead excreted by rats not receiving monensin.
FIG. 10. Potential mechanisms of monensin catalyzed Pb 2؉ transport. The curved lines together represent a phospholipid bilayer, whereas the heavy arrows indicate transmembrane diffusion of a leadmonensin complex. The various species illustrated are assumed to be present near the membrane-bulk aqueous phase interface. membrane-permeant and are excreted together with Pb 2ϩ , typically via the kidney. The lead-monensin complex is much less stable (K ϭ 10 7.25 in 80% methanol/water) and has little solubility in an aqueous environment. Moreover, the complex easily crosses a phospholipid membrane, whereas Pb excretion occurs primarily via the colon. Thus, monensin may act primarily by transporting Pb out of cells, where it eventually returns to the lumen of the gastrointestinal tract and is excreted by a normal mechanism, possibly involving the enterohepatic circulation. Since monensin is highly active as an ionophore for Na ϩ as well as for Pb 2ϩ , its presence will couple the gradients of these two cations across cell membranes, providing a driving force for Pb transport out of the cell. Overall, this potential mechanism suggests that the application of monensin, together with a hydrophilic chelator, might be particularly effective at promoting the excretion of Pb from animals. That possibility is currently under investigation.