Uptake of substrate and electric charge was measured simultaneously in voltage-clamped Xenopus laevis oocytes expressing rat organic cation transporter 2 (rOCT2). At 0 mV, saturating substrate concentrations induced uptake of more positive elementary charges than monovalent organic cations, with charge-to-substrate ratios of 1.5 for guanidinium+, 3.5 for tetraethylammonium+, and 4.0 for 1-methyl-4-phenylpyridinium+. At negative holding potentials, the charge-to-substrate ratios decreased toward unity. At 0 mV, charge-to-substrate ratios higher than unity were observed at different extracellular pH and after replacement of extracellular Na+, K+, Ca2+, Mg2+, and/or Cl−. Charge-to-substrate ratios were not influenced by intracellular succinate2− or glutarate2−. The effects of membrane potential and ion substitution strongly suggest that the surplus of transported positive charge is not generated by passive ion permeabilities. Rather, we hypothetize that small cations are taken up together with organic cation substrates whereas the outward reorientation of the empty transporter is electroneutral. Nonselective cotransport of small cations was supported by the three-dimensional structures of rOCT2 in its inward-facing and outward-facing conformations, which we determined by homology modeling based on known corresponding structures of H+-lactose permease of E. coli, and by functional analysis of OCT mutants. In our model, the innermost cavity of the outward-open binding cleft is negatively charged by Glu448 and Asp475, whereas the inward-open innermost cavity is electroneutral, containing Asp379, Asp475, Lys215, and Arg440. Substitution of Glu448 by glutamine reduced the charge-to-TEA+ ratio at 0 mV to unity. The observed charge excess associated with organic cation uptake into depolarized cells may contribute to tubular damage in renal failure.
- organic cation uptake
- charge translocation
- membrane potential
organic cation transporters (OCTs) of the SLC22 family translocate a wide range of structurally diverse molecules (17, 18). They participate in small intestinal uptake, hepatic excretion and/or renal excretion of various endogenous and exogenous compounds, including choline, monoamine neurotransmitters, coenzymes, and drugs or xenobiotics. The cloning of rat OCT1 (rOCT1) (12) and the subsequent identification of other members of the SLC22 family, including OCT2, which is mainly expressed in the kidney, and OCT3, which has a broad tissue distribution including brain, has led to a much better understanding of these transporters. Particular progress has been made in elucidating their functional characteristics, tissue localization, regulation by hormones and second messengers, generation of knockout mice, and the identification of genetic polymorphisms (17). In addition, attempts have been made to elucidate the functional mechanism of these transporters. It has been shown first that OCTs are able to mediate sodium-independent electrogenic transport of organic cations in both directions across the plasma membrane (3, 17); second, that a nontransported competitive inhibitor can inhibit rat OCT2 (rOCT2) from either side of the plasma membrane, suggesting inwardly and outwardly facing orientations of the substrate binding site (31); third, that rOCT1 contains high- and low-affinity binding sites for individual organic cations (11); and fourth, that rOCT2 mediates electrogenic uptake of the inorganic cation Cs+ with high selectivity over Na+ and K+ (24). Mutagenesis identified seven amino acids in rOCT1 that are involved in substrate binding (9, 10, 21). In a three-dimensional model of rOCT1 based on the structure of the inward-faced conformation of LacY permease (1) these amino acids could be located within a large inward-open cleft (21). The data suggest that rOCT1 contains a large binding region with overlapping interaction sites for cations within the cleft. During transport rOCT1 may switch between conformations in which the cleft is alternatingly open to the extracellular and intracellular cell side (11), similar to the alternating access mechanism proposed for LacY permease and the glyceraldehydephosphate transporter (1, 14, 15, 19, 27).
In the present work, we tested whether rOCT2 functions as a simple uniporter for monovalent organic cations. This mechanism would imply a strict coupling ratio of unity between substrates and positive elementary charges. We therefore determined the charge-to-substrate ratio from simultaneous measurements of organic cation uptake and electric current in individual Xenopus laevis oocytes expressing rOCT2. Holding potential, organic cations, and ionic conditions were varied. Surprisingly, we observed organic cation-induced uptake of surplus positive charges at 0-mV membrane potential with a charge-to-substrate ratio strongly depending on the structure of transported organic cation. Removal of a negative charge in the innermost cavity of the modeled outward-open conformation of the substrate binding cleft abolished the surplus of positive charges during transport of TEA at 0 mV. The data suggest that small cations are translocated into cells together with organic cation substrates whereas the outward reorientation of the empty transporter is electroneutral.
MATERIALS AND METHODS
[14C]guanidinium (GUA+; 1.9 TBq/mmol), [14C]tetraethylammonium+ (TEA+; 1.9 TBq/mmol), and [3H]-1-methyl-4-phenylpyridinium+ (MPP+; 3.1 TBq/mmol) were obtained from Biotrend (Köln, Germany). [14C]succinate2− (2.0 TBq/mmol) was purchased from Hartmann Analytic (Braunschweig, Germany). Other chemicals were obtained as described earlier (2, 31).
Expression of rOCT2 in X. laevis oocytes.
Animal husbandry, partial ovarectomy, defolliculation, synthesis and injection of cRNAs, and incubation of oocytes were carried out according to standard protocols described previously (2). Animal handling was performed according to German laws and approved by the University of Würzburg and the government of Unterfranken. The exchange of Glu448 in rOCT2 by glutamine was performed as described (9). The cRNAs or rOCT2 wild-type and the rOCT2(Glu448Gln) mutant were transcribed with a commercial kit (T7 mMESSAGE mMACHINE, Ambion, Huntingdon, UK) from the vector pRSSP (4) in which the coding sequence of rOCT2 (2) (accession no. X98334) is flanked by the 3′-untranslated region of the X. laevis β-globin gene. Heterologous protein expression was allowed to proceed for 2–4 days at 16°C.
Solutions for measurements in oocytes.
As a standard solution, we used a modified ND96 buffer (100 mM NaCl, 5 mM HEPES, 2 mM KCl, 1.8 mM CaCl2 or calcium gluconate, and 1 mM MgCl2 or magnesium gluconate, yielding a pH of ∼7.4). Where necessary, pH was adjusted to 7.4 by acetic acid. The substrates TEA+ and GUA+ were added as chloride salts and at a standard concentration of 1 mM. The substrate MPP+ was added at a standard concentration of 200 μM as chloride salt. The dependency of the substrate-induced currents and isotope fluxes on extracellular ions was investigated in several substitution experiments, including the substitution of Cl− with gluconate− or lactate−, substitution of Na+ with lysine+ or with K+, omission of K+ (2 mM KCl vs. 2 mM NaCl), omission of Ca2+ or Mg2+, reducing the total ion concentration by substitution of NaCl with mannitol, addition of 10 mM BaCl2 to decrease the membrane potential for tracer uptake experiments in nonclamped oocytes, and variation of pH between 5.1 and 8.1. The dependency of charge-to-substrate stoichiometry on intracellular succinate2− was investigated by preincubation for 2 or 24 h in standard solution supplemented with 50 mM succinate−2 (replacing Cl−). Dependency of charge-to-substrate stoichiometry on intracellular glutarate−2 was tested by injecting 50 nl of 100 mM glutarate2− shortly before the measurements were started. Glutarate−2 was dissolved in modified ND96 buffer in which sodium was replaced by potassium (K buffer), and the pH was adjusted with KOH. Assuming an internal aqueous distribution volume of 400 nl/oocyte (33) the internal glutarate−2 concentration was increased by ∼12.5 mM.
Measurements of substrate-induced current in oocytes.
Two-electrode voltage-clamp experiments were carried out using a feedback amplifier (TEC-05 from NPI Electronic, Tamm, Germany), which was controlled by software (PULSE and X-CHART from Heka Electronics, Lambrecht, Germany) via an AD/DA converter (ITC-16 from Instrutech, Port Washington, NY). The custom-built oocyte chamber had a small volume of ∼300 μl and could be perfused via gravity-fed lines. The ground electrode was a silver/silver chloride wire. As an external voltage electrode, a commercial liquid reference half-cell (DRIREF-2 from WPI, Berlin, Germany) was used to avoid offset potentials in our ion substitution experiments, particularly in the absence of Cl−. The absence of offset potentials was confirmed in preliminary experiments (data not shown).
Measurement of isotope uptake under voltage-clamp conditions in oocytes.
Substrate uptake was determined by incubating the voltage-clamped oocytes inside the recording chamber with saturating concentrations of [14C]TEA+ (1 mM), [14C]GUA+ (1 mM), or [3H]MPP+ (200 μM) under various ionic conditions and holding potentials (3). These substrate solutions were prepared by spiking the “cold” solutions of the respective chloride salts with traces of [14C]TEA bromide, [14C]GUA chloride, or [3H]MPP iodide to a final specific activity of ∼2 dpm/pmol. Uptake was allowed to proceed for 4 min and was then terminated by superfusion of the oocytes with substrate-free solution; this solution was supplemented with 20 μM of the rOCT2-inhibitor tetrabutylammonium+ (TBuA+) to prevent substrate efflux (“stop solution”). Oocytes were removed from the recording chamber and washed six times in ice-cold stop solution, subsequently dissolved in 100 μl of 5% SDS, mixed with 1 ml of scintillation fluid, and counted in a scintillation counter. Preliminary experiments had shown a considerable overshoot of current upon substrate withdrawal in the absence of TBuA+, likely representing an outward current carried by substrate flowing out from the cell. With TBuA+, this outward current was quickly blocked, and membrane current returned to the true baseline level (see e.g., Fig. 1A), which is crucial for the correct determination of substrate-induced charge flow. The 4-min incubation time was used to collect sufficient radioactively labeled substrates in the oocytes. Experiments were discarded if current drift was apparent (i.e., membrane current did not return during washout to the same level as before substrate application).
Measurement of [14C]TEA+ uptake and [14C]succinate−2 efflux in nonclamped oocytes.
For uptake measurements, oocytes were incubated for 15 min at room temperature with 1 mM TEA+ (tracer plus nonlabeled compound), washed with ice-cold modified ND96 buffer containing 20 μM TBuA+, and radioactivity in the oocytes was measured as described (2). For efflux measurements of succinate−2, oocytes were preloaded with [14C]succinate2− by incubating them for 3 h at room temperature with modified ND96 buffer containing 50 mM total succinate2− replacing Cl− (tracer plus nonlabeled succinate−2). After a washing with ice-cold K buffer, five oocytes were transferred into tubes containing 100 μl of K buffer (room temperature). After various time intervals, 10-μl samples were removed and analyzed for radioactivity.
Whole-cell patch measurements in HEK293 cells expressing rOCT2.
Human embryonic kidney (HEK) 293 cells were stably transfected with the organic cation transporter rOCT2 (2, 5) and studied in the whole-cell voltage-clamp and current-clamp mode as described (20). Recordings were performed at room temperature using a bath solution consisting either of 150 mM NaCl, 3.5 mM KCl, 2.0 mM CaCl2, 1 mM MgCl2, 10 mM glucose, 10 mM HEPES, pH 7.4, or of a depolarizing high-potassium solution in which 140 mM NaCl was replaced by KCl. Effects of TEA+ were investigated using bath solutions in which 1 mM TEA+ was replaced for 1 mM Na+ or 1 mM K+. Patch pipettes were pulled from borosilicate glass capillaries (GB-150F-8P, Science Products, Hofheim, Germany) and heat-polished to give input resistances of 3–5 MΩ. The pipette recording solution contained 95 mM potassium gluconate, 30 mM KCl, 4.8 mM Na2HPO4, 1.2 mM NaH2PO4, 1 mM MgCl2, 5 mM d-glucose, 1 mM EGTA, 1 mM Na2ATP, and 5 mM HEPES (pH 7.3). Currents and membrane potentials in the absence and presence of 1 mM TEA+ were recorded with an EPC9 patch-clamp amplifier (Heka Electronics) and low-pass filtered at 1–2 kHz. Stimulation and data acquisition were controlled by the Pulse/Pulsefit software package (Heka Electronics) on a Macintosh computer, and data analysis was performed with Igor software (WaveMetrics, Lake Oswego, NY).
Modeling of the inward- and outward-facing conformation of rOCT2.
A model of the inward-facing conformation of rOCT2 (UniProtKB/TrEMBL entry Q9R0W2_RAT) was obtained by homology modeling using the crystal structure of LacY permease of Escherichia coli (PDB entry 1PV6) (1) and our previously reported homology model of rOCT1 (21) as templates. Two initial tertiary structure templates of rOCT2 were generated: one starting from the LacY permease (21) and the other from the rOCT1 homology coordinate set exchanging the amino acids which are different between rOCT2 and rOCT1. Both methods revealed similar structures. Refinement was performed with the structure derived from the rOCT1 homology coordinate. Close van der Waals contacts between noncovalently linked atoms were removed using the XBUILD tool of Quanta2005 software (Accelrys, San Diego, CA). Further refinement was performed by stepwise energy minimization runs of 500 steps utilizing the CharmM22 force field without electrostatic terms and the Adopted Newton Raphson minimization algorithm. In the first run, only side chains in locally restricted regions (5-Å spheres) with not fully resolved close contacts were minimized; all main chain atoms and side chains outside the minimization sphere were kept fixed. Then, a second minimization was performed with all side chains, but with main chain atoms kept fixed. Final energy minimization was performed with a harmonic positional restraint (force constant 100 kcal·mol−1·Å−2) on all atoms: minimization cycles were repeated with resetting of the harmonic positional restraints until the overall energy converged.
The outward-facing model was obtained from the final model of the inward-facing conformation by a rigid-body movement of the first six helices with respect to the last six helices as proposed Kaback and coworkers (1, 8, 15, 26) allowing intrahelix motions and drifting of individual domains as proposed by Holyake and Samson (13). The model of the inward-facing conformation was separated into two parts (the NH2-terminal half up to Ile283 representing the last residue in helix 6, and the COOH-terminal half starting with Gln343 representing the first residue in helix 7). To perform the rigid-body movements of the NH2- and COOH-terminal domains, the model was oriented such that the helices of both domains did not overlap in the plane. The NH2-terminal domain was then manually rotated around the z-axis until the open cleft at the inward-facing side was closed, e.g., due to close contacts between amino acids at the COOH terminus of helix 2 (Gly168) and at the NH2 terminus of helix 11 (Arg463) as well due to contacts between amino acids at the NH2-terminal start of helix 5 (Val237) and the COOH-terminal end of helix 8 (Ala391). The initial model exhibited only a small number of bad van der Waals contacts. Bad contacts between side chains were removed by side chain rotamer searches using Quanta2005 software (Accelrys). For further refinement allowing intrahelix motions (13), all helices were separated from interconnecting loops by breaking a peptide bond at the very NH2 and COOH terminus of the individual helices. The helical structure was maintained by using a set of strict distance restraints (force constants 250 kcal·mol−1·Å−2) which mimics the H-bond pattern in the individual helices. Helix packing was subsequently optimized by running alternating protocols of energy minimization (100 steps Adopted Raphson Newton algorithm, Quanta2005) and short in vacuo molecular dynamic simulations (10 ps at 300 K) employing only geometrical energy terms. At this point the “disconnected” loops were kept fixed. After 10 rounds of this packing optimization, the peptide bonds between the transmembrane helices and the interconnecting loops were closed and the loop structures were refined by a similar protocol but with restraining of the helices with a strong harmonic positional restraint (force constant 250 kcal·mol−1·Å−2). The final model of the outward-facing rOCT2 displays good backbone and side chain geometries, with 96% of the residues in the most favored or additionally allowed regions of the Ramachandran plot according to PROCHECK analysis. Residues (six) that occupy phi/psi backbone torsion angles in disallowed positions are located in loop regions. Like the model of the inward-open conformation, the model of the outward-open conformation does not include residues of loop regions for which the LacY permease could not provide coordinates. These regions comprise the NH2-terminal Met1 to Gln18, the long interhelical loops between α1 and α2 (residues Pro53 to His145) and α6 and α7 (residues Pro284 to Pro342), and the COOH terminus starting from residue Glu530.
Electrostatic potential map calculations of the inward- and outward-facing conformation of rOCT2.
Both model structures were protonated using the Quanta2005 Protein Design tool assuming regular pKa values for all amino acid residues. Potential map calculations were performed by two different programs, GRASP version 1.2.5 using the simplified charge parameter file full-charge and the more detailed charge parameter file AMBER and APBS (Adaptive Poisson-Boltzmann Solver). For the latter software, the protonation and charge assignment were carried out using the PDB2PQR webserver version 1.2.1 using the AMBER and CHARMM charge force field. Electrostatic potential maps were calculated at two ionic strengths, 0 and 150 mM salt, and the dielectric constant ε was set to 2.0 for the protein interior and to 80 for the solvent. A grid spacing of 1 Å was used. Using different force fields and more complex charge assignments or different ionic strength did not alter the general characteristics of the overall potential map. With higher ionic strength, the potential field did not reach out from the protein as far as for low ionic strength, but distribution of the charges of the potential map was highly similar for a particular coordinate set. For display, potential maps were loaded into PyMOL mapped onto a solvent-accessible surface at −8 and +8 kT/e. A contour map showing the characteristics of the electrostatic potential surrounding rOCT2 in both conformations was also calculated.
Computations and statistics.
Substrate-induced uptake of electric charge into the oocyte was obtained by integration of substrate-induced current (I) over time, and expressed in moles of positive elementary charges e+ (via division of charge uptake by Faraday's constant). To take into account any potential efflux of the substrate during the washout phase following the uptake phase, the integration was carried out from substrate addition to removal of the oocyte from the measuring chamber. Uptake of the substrate was computed from the measured radioactivity of the homogenized oocyte and the specific activity of the respective substrate. Herein, specific activity was determined from the known substrate concentration in the uptake solution together with the measured radioactivity of that solution. The dimensionless proportion between the number of positive elementary charges (e+) taken up by the oocyte and the number of substrate molecules taken up was computed by dividing charge uptake by substrate uptake. Following mathematical conventions, we here use pipe symbols (|) to express the cardinality, or absolute number, of elementary charges as |e+|, to distinguish this dimensionless number from the particular electric charge e+ (in Coulombs), and similarly to express the absolute number of substrate molecules as |S+|. The charge-to-substrate ratio is thus written |e+|:|S+| in general, or |e+|:|TEA+|, |e+|:|GUA+|, or |e+|:|MPP+| for the respective particular substrates. This “charge-to-substrate ratio” is numerically equivalent to a “current-to-flux ratio” that is computed analogously as the ratio of mean substrate-induced current Î (in moles of elementary positive charges per second) to mean net flux Φ̂ (i.e., ÎTEA/Φ̂TEA, etc.). To estimate the intracellular concentration of succinate2− from tracer uptake measurements, an aqueous distribution volume of 0.4 μl/oocyte was assumed (33). The software package GraphPad Prism Ver. 4.1 (GraphPad Software, San Diego, CA) was used to perform linear and nonlinear regressions and to test hypotheses (t-test against fixed value, unpaired or paired t-test between 2 groups, ANOVA test with post hoc Tukey comparison among more than 2 groups). The data are presented as means ± SD.
Charge flow through rOCT2 exceeds substrate flow.
We determined and compared the flows of organic cations and of electrical charge in individual X. laevis oocytes. In oocytes expressing rOCT2, the three classic substrates GUA+ (1 mM), TEA+ (1 mM), and MPP+ (0.2 mM) induced currents (typically between −20 and −200 nA) and uptake of radiolabeled substrate (typically between 20 and 400 pmol in 4 min) as observed previously (2). In water-injected oocytes tested at three different membrane potentials, on the other hand, these compounds did not induce significant current (not shown; see dotted line in Fig. 1A for GUA+) or uptake (at −100, −50, and 0 mV, respectively, for GUA+: 3.2 ± 1.7, 2.9 ± 1.4, and 1.8 ± 1.9 pmol; TEA+: 0.2 ± 0.3, 0.1 ± 0.4, and 0.9 ± 0.4 pmol, MPP+: 3.5 ± 1.6, 4.2 ± 0.9, and 4.9 ± 1.2 pmol; n = 5–6 each). The continuous trace in Fig. 1A shows the current response to superfusion of rOCT2-expressing oocytes clamped to 0 mV with 1 mM [14C]GUA+, followed by washout with stop solution. GUA+ induced an inward current, followed by a partial decay during the subsequent 4-min superfusion period. Upon substrate removal, the membrane current exhibited a transient overshoot and finally returned to its baseline. The overshoot, which is due to the efflux of substrate, was inhibited by 20 μM rOCT2 inhibitor TBuA+ (31) reaching the measuring chamber after a short delay (due to a small priming volume of inhibitor-free solution in the perfusion line). Subsequently, the oocyte was removed and the number of substrate molecules taken up by the oocyte was determined via scintillation counting of [14C]GUA+. In Fig. 1B, the findings from several such experiments in different oocytes are represented as a plot of charge uptake vs. GUA+ uptake at 0 mV. A straight line fitted to the data points had a slope of 1.5 ± 0.04. This value represents the charge-to-substrate ratio for GUA+ (|e+|:|GUA+|). The same result was obtained when the charge-to-substrate ratio was computed in an alternative way as the mean from the charge-to-substrate ratios of individual oocytes (1.5 ± 0.08, n = 19). Importantly, this value is significantly different from unity (P < 0.001 in t-test against a hypothetical value of 1.0). However, a value of unity would be expected if the uptake of guanidinium ions through rOCT2 had occurred via simple uniport. The finding of a ratio |e+|:|GUA+| greater than unity suggests that transport of GUA+ is associated with the translocation of another ion.
Charge-to-substrate ratio in rOCT2 is voltage dependent and substrate specific.
In this way, we have determined the charge uptake and substrate uptake for all three substrates (GUA+, TEA+, MPP+) at three different holding potentials (−100, −50, 0 mV) (Fig. 2, A and B). For any substrate (S+), the ratio |e+|:|S+| was strongly voltage dependent, in the sense of smaller ratios at more negative membrane potentials. Moreover, we found that the charge-to-substrate ratios also varied between different substrates at a given holding potential. Note that the charge-to-substrate ratios for TEA+ and MPP+ at 0 mV exceeded unity to a greater extent than for GUA+ (|e+|:|TEA+| = 3.5 ± 0.4; |e+|:|MPP+| = 4.0 ± 0.4). With GUA+ and TEA+, the excess of positive charge associated with substrate transport disappeared at −100 mV, and the ratio |e+|:|S+| became indistinguishable from 1.0 (|e+|:|GUA+| = 1.0 ± 0.17, |e+|:|TEA+| = 1.02 ± 0.27). Note 1) that the |e+|:|S+|ratio larger than one at 0 mV can be explained by a movement of cations together with the substrate or by a movement of anions in the opposite direction, and 2) that the observed potential dependence excludes the membrane potential as the main driving force for the uptake of nonsubstrate cations as well as for the efflux of anions.1
In Fig. 2C, the uptake of GUA+(1 mM), TEA+(1 mM), and MPP+(0.2 mM) into oocytes expressing rOCT2 at −100, −50, and 0 mV is compared. Uptake of these substrates was voltage dependent, with a significantly higher uptake at −100 or −50 mV compared with 0 mV. For example, TEA+ uptake at −100 mV was ∼2.4 times higher compared with 0 mV. When we measured the voltage dependence of currents in rOCT2-expressing oocytes induced by superfusion with 1 mM TEA+ (Fig. 2D), we observed only a very small current increase with increasingly negative membrane potential (0.039 ± 0.036 nA/mV; n = 3). Thus the current induced by TEA+ at −100 mV was only ∼12% higher compared with the current induced at 0 mV. This is consistent with the higher |e+|:|TEA+| ratio obtained at 0 mV compared with 100 mV (Fig. 2B). Since there is large scatter between expression of transport activity in individual oocytes, simultaneous measurement of tracer uptake and charge uptake allows a much more accurate determination of charge-to-substrate ratios than a comparison of currents and tracer uptake measured in different oocytes.
Charge-to-substrate ratio in rOCT2 is independent of extracellular pH.
The finding of an excess positive charge associated with uptake of organic cations by rOCT2 implies a simultaneous translocation of one or more other ions, prompting the question as to their chemical identity. One candidate ion was extracellular H+, whose role in ion transport by OCTs is not entirely clear (6, 7). To test for putative effects of H+ on the charge-to-substrate ratio of rOCT2, we measured |e+|:|GUA+| over a range of pH values in which H+ concentration ([H+]) was varied 1,000-fold (Fig. 3). Herein, a holding potential of 0 mV was applied because the excess charge was highest at this voltage. These measurements confirmed the existence of an excess charge, but did not show a significant effect of [H+] on this phenomenon. In particular, we did not observe an increase in |e+|:|GUA+| at lower pH values, as one would expect under the assumption that the observed access charge transfer is due to protons. Together, these experiments suggest that H+ ions do not contribute significantly to charge flow through rOCT2 under the employed experimental conditions (100 mM extracellular Na+). Incidentally, these results also showed that extracellular H+ did not inhibit transport by rOCT2 under strictly voltage-clamped conditions, in contrast to findings in unclamped X. laevis oocytes (28).
Charge-to-substrate ratios higher than one were observed in presence of different extracellular nonsubstrate cations.
Other possible extracellular cations that might underlie the excess charge flow associated with organic cation transport by rOCT2 are Na+ and K+. To test for a putative effect of extracellular Na+, we determined the uptake of GUA+ and the concomitant uptake of charge with and without Na+ (100 mM Na+ vs. 100 mM lysine+) at a holding potential of 0 mV (Fig. 4A). The results confirmed the previously observed high value of |e+|:|GUA+| and showed that Na+ replacement had no effect on |e+|:|GUA+|. Similarly, the removal of 2 mM K+ from the bath solution at 0 mV had no effect on the apparent ratio |e+|:|GUA+| (Fig. 4B).
We also investigated effects of extracellular cations on charge-to-substrate stoichiometry of TEA+ uptake by rOCT2 at 0 mV. After replacement of Na+ in the bath by lysine+, |e+|:|TEA+|was significantly decreased from 3.5 ± 0.98 to 2.5 ± 0.37 (Fig. 5, compare first two columns). This decrease is at variance to |e+|:|GUA+|; however, similar to GUA+, |e+|:|TEA+| was still significantly higher than unity. Replacing extracellular Na+ by K+, decreased |e+|:|TEA+|more strongly than replacing Na+ by lysine+ (1.7 ± 0.26) (Fig. 5, compare columns 1 and 3). Noteworthy, also this value was significantly higher than unity (P < 0.001). The data indicate that a surplus of positive charge entering the oocytes during TEA+ uptake is also observed in the absence of extracellular Na+ and of inwardly directed nonsubstrate cation gradients. The finding that |e+|:|TEA+| ratios higher than unity are observed in the absence of inwardly directed gradients of Na+ or lysine+ excludes that leak permeabilities for monovalent cations are responsible for the surplus of a translocated positive charge. At K+ equilibrium and −50 mV (absence of Na+), an |e+|:|TEA+| ratio close to unity (1.06 ± 0.15) was observed (Fig. 5, column 4) indicating a similar potential dependence as in the presence of an inwardly directed Na+ gradient. Removing Ca2+ or Mg2+ from the standard bath solution (containing 100 mM Na+ and 2 mM K+) had no effect on the |e+|:|TEA+| ratios measured at 0 mV (Fig. 5, compare column 1 with columns 5 and 6). This indicates that the surplus of positive charge translocated with TEA+ at 0 mV is not due to translocation of Ca2+ or Mg2+.
Since the surplus of translocated positive charge at 0 mV cannot be explained by passive permeability for inorganic cations, we raised the hypothesis that it may be due to nonselective translocation of nonsubstrate cations together with organic cation substrates. An obvious experiment to test nonselective translocation is to measure the charge-to-substrate ratio in the absence of extracellular nonsubstrate cations, keeping the osmolarity constant. Unfortunately, such an experiment could not be performed because the oocytes became unstable when the NaCl concentration in the bath was reduced below 20 mM. We largely decreased the total extracellular ion concentration by superfusing rOCT2-expressing oocytes with 5 mM HEPES, pH 7.4, containing 24 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, and 152 mM mannitol (isosmotic replacement of 76 mM NaCl). The oocytes were clamped to 0 mV, and the |e+|:|TEA+| ratio was measured (Fig. 5, column 7). The decrease in extracellular NaCl led to a significant reduction of the charge-to-TEA+ ratio from 3.4 ± 0.3 to 2.0 ± 0.3. When extracellular Ca2+ was removed in addition, no further significant change of the charge-to-TEA+ ratio was observed (1.6 ± 0.3) (Fig. 5, column 8). The data are consistent with our hypothesis that nonsubstrate inorganic cations are translocated nonspecifically together with organic cation substrates.
Charge-to-substrate ratios higher than one are independent of intracellular dicarboxylates.
In principle, a charge-to-substrate ratio greater than unity could be due to intracellular anions moving outward. In fact, the SLC22 family comprises transporters such as organic anion transporters of the OAT subfamily that do couple the uptake of organic anions to the efflux of cytoplasmic succinate−2, α-ketoglutarate, or glutarate2− (22, 23, 25, 29). Considerable functional and structural similarities exist between OCTs and OATs, and the OAT-ligands para-aminohippurate, α-ketoglutarate, and probenecid are low-affinity inhibitors of rOCT1 and rOCT2 (2). We therefore studied the effect of intracellular succinate−2 and glutarate−2 on the charge-to-substrate ratio of rOCT2. At variance to glutarate−2 X. laevis oocytes exhibit a relatively rapid uptake of succinate−2. We preloaded oocytes with succinate2− by preincubation for 2 or 24 h in 50 mM succinate2− (modified ND96 buffer in which 50 mM Cl− was replaced by 50 mM succinate−2). In parallel experiments, we added tracer amounts of [14C]succinate−2 and determined the intracellular concentrations of succinate−2 (see materials and methods). After 2-h incubation with 50 mM succinate−2, intracellular succinate2− concentrations of 9.4 ± 0.9 or 10.3 ± 0.9 mM (n = 13 each) were estimated in noninjected control oocytes and in rOCT2-expressing oocytes, respectively. After 24-h incubation, equilibrium between intracellular and extracellular succinate−2 was obtained. Measuring the ratios of |e+|:|GUA+| and |e+|:|TEA+| at 0 mV, the values obtained in succinate−2-treated oocytes were not different from those measured in control oocytes that were incubated without succinate−2 (Fig. 6).
To determine the effect of intracellular glutarate−2, we injected 50 nl of K buffer containing 10 mM potassium glutarate per oocyte and measured |e+|:|TEA+| at 0 mV. Assuming an internal aqueous distribution volume of 400 nl/oocyte (33), the intracellular concentration of glutarate2− was increased by ∼12.5 mM. A charge-to-substrate stoichiometry of 3.1 ± 0.5 (n = 6) was obtained, which was similar to the values obtained in the absence of intracellular dicarboxylates (Fig. 6).
We also investigated whether efflux of [14C]succinate−2 from rOCT2-expressing oocytes at low membrane potential (replacement of extracellular Na+ by K+) is stimulated by 1 mM extracellular TEA+. Noninjected oocytes (control) and oocytes expressing rOCT2 were preloaded with succinate2− for 2 h at room temperature in 50 mM succinate2− containing [14C]succinate2− as tracer. After washing oocytes at 0°C with modified N96 buffer in which Na+ was replaced by K+ (K buffer), oocytes were incubated at room temperature in K buffer or K buffer containing 1 mM TEA+, and the efflux of succinate2− was measured after different times of incubation. The efflux of succinate2− was linear for at least 4 min under all experimental conditions (data not shown). After 4-min incubation of oocytes in the absence and presence of extracellular TEA+, similar efflux rates were obtained (in nmol·oocyte−1·min−1: rOCT2 with TEA+ 2.7 ± 0.3, rOCT2 without TEA+ 2.3 ± 0.7, control with TEA+ 2.7 ± 0.2, control without TEA+ 3.3 ± 0.3; n = 3 each). These results indicate that succinate2− efflux is not stimulated by TEA+ uptake through rOCT2. Taken together, the data indicate that the observed charge-to-substrate ratios greater than unity are not due to countertransport of dicarboxylates.
Replacing extracellular Cl− by gluconate− or lactate− does not increase charge-to-substrate ratios.
In oocytes expressing rOCT2 that are superfused with modified N96 buffer (containing 107.6 mM Cl−) and clamped to 0 mV, a putative influx of Cl− during organic cation transport should decrease the charge-to-substrate ratio whereas export of intracellular Cl− could account for the observed excess uptake of positive charge. We replaced Cl− in the bath by gluconate− or lactate− (107.6 mM) and measured the |e+|:|GUA+| and/or |e+|:|TEA+|ratios at 0 mV (Fig. 7, open bars). This replacement reverts the direction of the Cl− concentration gradient from inward to outward and generates additional inward gradients for gluconate− or lactate−, respectively. Replacement of Cl− by gluconate− reduced the |e+|:|GUA+| and |e+|:|TEA+| ratios significantly (Fig. 7). However, an increase in charge-to-substrate ratios was expected if the ratios larger than unity measured in the presence of 107.6 mM extracellular Cl− were caused by Cl− export. The observed decrease in charge-to-substrate ratio after replacement of Cl− by gluconate− may be caused by a selective increase in a passive gluconate− permeability during organic cation uptake or by cotranslocation of gluconate− into oocytes. To test for the first possibility, we also determined the charge-to-TEA+ ratio after replacement of extracellular Cl− by gluconate− at −50 mV (Fig. 7). At −50 mV, |e+|:|TEA+| was significantly smaller compared with 0 mV (1.12 ± 0.03 vs. 2.4 ± 0.6), rather than larger as would be expected for a passive gluconate− permeability. We also measured the charge-to-TEA+ ratio at 0 mV after replacement of extracellular Cl− by lactate− (Fig. 7). After replacement of Cl− with this anion, the |e+|:|TEA+| ratio was not changed significantly (3.0 ± 0.2). The data indicate that charge-to-substrate ratios larger than unity at 0 mV are not due to substrate stimulated efflux of Cl−. The decreased charge-to-substrate ratio after replacement of Cl− by gluconate− may be explained by cotranslocation of gluconate− in addition to nonsubstrate cations.
Excess current is dependent on translocation of organic cations.
To test whether permeability changes in rOCT2 can be induced by binding of organic cations, we studied the effect of the nontransported competitive inhibitor TBuA+ (2, 3, 31) on membrane current in X. laevis oocytes expressing rOCT2. When oocytes expressing rOCT2 clamped at −100, −50, or 0 mV were superfused with 1 mM GUA+, inward currents of 99 ± 39 nA (−100 mV), 117 ± 64 nA (−50 mV), and 43 ± 18 nA (0 mV; n = 10–14 for each) were obtained. In contrast, no significant inward currents were detected when the oocytes clamped were superfused with 200 μM TBuA+ (−100 mV: 0.5 ± 0.9 nA, −50 mV: −0.9 ± 1.5 nA, 0 mV: −0.8 ± 1.9 nA; n = 10–14 each). The data indicate that TBuA+ does not inhibit passive permeabilities mediated by rOCT2 in the absence of substrate. They also show that binding of a nontransported organic cation inhibitor to the substrate site is not sufficient to induce permeability for inorganic ions.
TEA+ uptake under depolarized condition is not stimulated by inwardly directed gradients of Na+ or lysine+.
The data obtained so far suggest that at 0 mV small cations like Na+, K+, and lysine+ are transported together with organic cation. We tested whether Na+ or lysine+ is able to stimulate uptake of [14C]TEA+ into depolarized oocytes. Uptake of 1 mM TEA (traced with [14C]TEA+) was measured in the presence of 100 mM KCl and 10 mM BaCl2 plus either 1) 200 mM mannitol, 2) 100 mM NaCl, 3) 100 mM lysine chloride, or 4)190 mM mannitol plus 5 mM TBuA chloride (Fig. 8). In rOCT2-expressing oocytes, 5 mM TBuA+ inhibited TEA+ uptake to the same level as TEA+ uptake measured in noninjected control oocytes. In rOCT2-expressing oocytes, the uptake rate of 1 mM TEA+ measured in the presence of 200 mM mannitol was 136 ± 55 pmol·oocyte−1·15 min−1 (means ± SD of 3 independent experiments). Replacement of 200 mM mannitol by 100 mM NaCl or 100 mM lysine chloride did not change TEA+ uptake. The data suggest that small cations that are cotranslocated with TEA+ do not contribute to the driving force for TEA+ uptake.
Models indicating charge distribution of the inward- and outward-faced conformations of rOCT2.
The conformation of rOCT2 with an inward-open cleft was modeled using the crystal structure of LacY permase of E. coli (1) and our previously reported homology model of rOCT1 (21). A model of rOCT2 with an outward-facing cleft was obtained from the inward-facing conformation by applying the alternating access mechanism proposed for the glyceraldehydephosphate transporter and LacY permease (1, 14, 15, 19, 27). For modeling, intrahelix motions and drifting of individual domains were allowed (13). The resulting rOCT2 structures were highly similar to those obtained previously for rOCT1 (11, 21). This was expected since most of the amino acid differences between rOCT2 and rOCT1 are conservative. The above-described observation that a surplus of positive charge is translocated at 0 mV during transport of organic cation substrates may be explained by low-affinity binding of small extracellular cations to negatively charged amino acid residues within an innermost cavity of the outward-facing binding cleft of rOCT2, uptake together with an organic cation substrate, and subsequent release from the inward-facing conformation of the binding cleft together with the organic cation substrate. This implies that no net positive charge is translocated back during reorientation of the empty transporter. To determine whether the charge distribution within both conformations of the cleft is consistent with such a mechanism, we determined the electrostatic potential of our rOCT2 model in both conformations using two different software packages (GRASP and ABPS) that employ different force fields to describe the charge assignments. With both programs, similar results were obtained (data not shown). The potential maps (Fig. 9A) indicate that the outward-open cleft exhibits a negative potential that protrudes significantly into the extracellular space. In contrast, the inward-open conformation of the cleft exhibits a neutral or even slightly positive charge potential (Fig. 9A). The outer rim of the outward-open cleft comprises seven negatively charged residues (6 Asp, 1 Glu) and two positively charged Arg residues (Fig. 9B, left). Importantly, the innermost cavity of the outward-open cleft (Fig. 9, B, middle, and D, left and middle) is lined by two negatively charged amino acids (Asp475 and Glu448). Asp475 is conserved in the three OCT subtypes and has been shown to be critically involved in translocation of organic cations (10). Glu448 is conserved in OCT2 and OCT3 from different species whereas OCT1 from different species contain glutamine in this position. The outer rim of the inward-facing cleft comprises eight negatively charged residues (3 Asp and 5 Glu) and 14 positively charged amino acids (9 Arg and 5 Lys) (Fig. 9C, left). The inward-open innermost cavity is about neutral (Fig. 9, C, middle, and D, right). It is lined by two positively charged amino acids (Lys215 and Arg440) and two negatively charged amino acids (Asp379 and Asp475). These four amino acids are conserved in all OCT subtypes. Within the inward-open cleft of rOCT2, Glu448 obtains a peripheral position outside the innermost cavity (Fig. 9D, right). The models suggest that the change between the outward-facing and the inward-facing conformation of rOCT2 or vice versa results in a dramatic change of the electrostatic potential map (Fig. 9A). This may explain the recently described membrane potential-dependent conformational change of rOCT1 in the absence of organic cation substrates (11). Together with the above-described experimental data, the models support the hypothesis that small nonsubstrate cations bind to the innermost cavity of the outward-open cleft, are translocated together with organic cation substrates, and are subsequently released to the intracellular space whereas no net positive charge is translocated during reorientation of the empty transporter (nonselective-cotransport-of-small-cations hypothesis).
Replacement of glutamate 448 in rOCT2 by glutamine abolishes the surplus of positive charge translocated with TEA+.
Measuring |e+|:|TEA+| ratios in oocytes expressing rat OCT1 (rOCT1), we observed less pronounced potential-dependent differences in stoichiometry with a smaller surplus of translocated positive charge at 0 mV (100 mV: 0.94 ± 0.12, n = 3; −50 mV: 1.3 ± 0.10, n = 10; 0 mV: 1.8 ± 0.13, n = 3). According to the nonselective-cotransport-of-small-cations hypothesis, the difference between rOCT1 and rOCT2 may be explained by different charge distributions within the innermost parts of the outward-open cleft compared with the inward-open cleft. Different sizes of the outward- and inward-open innermost cavities of rOCT1 vs. rOCT2 may contribute.
To test our hypothesis, we replaced Glu448 in rOCT2 by glutamine and measured |e+|:|TEA+| at 0 mV. We obtained a reduced |e+|:|TEA+| ratio as predicted (1.27 ± 0.38, mean ± SD, 8 oocytes from 3 batches). This ratio is significantly lower compared with rOCT2 wild-type (P < 0.001) and to rOCT1 wild-type (P < 0.05). The data support the nonselective-cotransport-of-small-cations hypothesis. They suggest high-quality modeling of the innermost cavities of the inward-open and outward-open binding clefts.
Organic cation uptake by rOCT2 induces higher depolarization in depolarized cells vs. nondepolarized cells.
The observed translocation of a surplus positive charge during organic cation transport may have pathophysiological consequences. For example, when the membrane potential of proximal tubular cells is decreased during renal failure, transport of organic cation substrates by OCT2 may depoarize the cells further and thus aggravate renal failure. To test the effect of rOCT2-mediated TEA+ uptake on the membrane potential in eukaryotic epithelial cells, we performed whole-cell patch measurements in human embryonic (HEK) 293 cells that were stably expressing rOCT2. In the presence of 150 mM extracellular sodium, membrane potentials between −30.7 and −53.7 mV (44.8 ± 7.8 mV, n = 7) were measured. Superfusion with 1 mM TEA+ lead to currents between 53 and 132 pA (78.4 ± 34.1 pA) and to depolarizations between 1.8 and 4.7 mV (2.93 ± 0.99 mV). When 140 mM extracellular sodium was replaced by potassium, the cells were depolarized to values between +2.3 and −15.4 mV. Under this condition, the membrane potential was decreased by 4.1–11.9 mV (5.6 ± 2.8 mV) when the cells were superfused with 1 mM TEA+ (Fig. 10). TEA+-induced decreases in the membrane potential in the depolarized cells were significantly larger than those in polarized cells (P = 0.015, paired Student's t-test). The magnitude of the OCT2-mediated further depolarization in previously depolarized cells suggests that organic cations may strongly effect cellular functions in various pathophysiological contexts, such as ischemic acute renal failure.
Parallel measurements of charge and substrate uptake by the organic cation transporter rOCT2 in X. laevis oocytes revealed a surplus of translocated positive charges vs. positively charged substrate molecules; the magnitude of that surplus varied with the particular organic cation substrate and decreased with more negative membrane potential.
The observed potential dependence of the charge-to-substrate ratio and the effects after replacement of inorganic ions in the bath cannot be explained by opening of a passive permeability during organic cation transport without making highly improbable assumptions. However, they can be explained by a nonselective-cotransport-of-small-cations hypothesis suggesting that small cations are taken up into cells together with organic cation substrates whereas outward reorientation of the empty transporter occurs without transfer of positive charge. This hypothesis is supported by the results of our modeling and mutagenesis studies. The innermost cavity of the modeled outward-open substrate binding cleft of rOCT2 is negatively charged whereas the modeled inward-open innermost cavity is electroneutal. When glutamate 448 within the outward-open innermost cavity was replaced by glutamine, the surplus of positive charge translocation during organic cation transport was abolished.
Most experiments in the present study were carried out using X, laevis oocytes because they allow us to combine isotope uptake and voltage clamping in single cells overexpressing a transporter of interest. On the other hand, X. laevis oocytes also have a complex biology and various experimental pitfalls in store; the latter ones need to be ruled out before interpretation of the results in terms of transport mechanism. One concern are endogenous ion conductances of the oocytes that may contribute to the apparent charge-to-substrate ratio of heterologously expressed rOCT2 (32).
Several findings in this study provide strong evidence against that possibility. 1) Permeation of GUA+, TEA+, or MPP+ through endogenous ion channels or transporters was ruled out by the observed lack of uptake in water-injected oocytes. 2) Inhibition of endogenous ion transport pathways for other ions by GUA+, TEA+, or MPP+ was ruled out by the lack of substrate-induced currents in water-injected oocytes. 3) Endogenous conductances, presumably being independent of the expression level of rOCT2, would produce a constant charge offset in the plot of |e+| vs. substrate; no such offset was observed, and a straight line through the origin could be convincingly fitted to the data (e.g., Figs. 1B, 2A, 4A).
Another potential concern is that the experimental procedures to determine the charge-to-substrate ratios might not be sufficiently accurate. Indeed, these measurements depend critically on the high quality of current recordings and radioisotope uptake measurements. For instance, current recordings could be disturbed by current drift, and radiolytic or chemical decay of the radiolabeled substrates could result in deviation of the actual specific activity from the supplier's specifications. After including of an rOCT2 inhibitor in the washout solution, current drift due to efflux of organic cations via rOCT2 could be blocked. The remaining small drift observed in some experiments was likely of a random nature and is thus unable to account for our systematic, statistically significant effects of voltage, substrate, or ionic conditions. Problems with accuracy due to decay of radioactive compounds appear unlikely considering that the experiments were highly reproducible, with consistent results obtained using different batches of radioisotope over a course of several years, and that systematic variation of the ratio |e+|:|S+| with voltage was found in experiments carried out on the same day using the same batch of oocytes and radioisotope.
Translocation of small ions during transport of organic cation substrates.
One central result of this study is that the number of organic cations transported by rOCT2 at 0 mV is not matched by an identical number of positive elementary charges. Rather, there is a variable mismatch depending on the nature of the transported substrate. It is obvious that the excess of translocated positive charge must be carried by small ions. The nontransported inhibitor TBuA+ was not able to induce currents at 0 mV, indicating that the transport of organic cation substrate is required for translocation of small ions. We cannot exclude that rOCT2 promotes some permeability for small ions in the absence of organic cation substrates that cannot be blocked by inhibitors; however, such type of permeability is supposed to be small and does not influence the interpretation of the present results. So far we were not able to detect significant changes of the membrane potentials due to expression of organic cation transporters in the absence of organic cation substrates (data not shown).
The observed variability of charge translocation during transport of different organic cations indicates that rOCT2 operates neither as an organic cation cotransporter nor as an electrogenic organic anion antiporter with a fixed stoichiometry. In addition, antiport of organic cations with intracellular dicarboxylates similar to antiport of dicarboxylate and organic anions described for organic anion transporters (22, 23, 25, 29) was excluded by showing that rOCT2 did not mediate efflux of succinate2− and that the charge-to-cation stoichiometry of rOCT2 was not changed when the oocytes were preloaded with succinate2− or glutarate2−.
Other than by cotransport or antiport with fixed stoichiometry, the observed positive charge excess during organic cation uptake at 0 mV could, in principle, be due to two mechanisms or to a combination of both. First, a poorly selective conductance for small cations and/or anions could be opened during the transport cycle allowing electrodiffusion of small cations and/or anions. Second, small ions may be trapped within the innermost cavity of the substrate binding cleft of rOCT2 and translocated together with the transported organic cations or during reorientation of the transporter during the transport cycle. We cannot exclude that a nonselective conductance associated with organic cation substrate transport contributes to the observed charge-to-substrate ratios. However, the observed potential dependence of the charge-to-substrate ratio cannot be explained by this mechanism without making highly improbable assumptions. At 0 mV we observed a higher surplus of translocated positive charge compared with −50 mV. Opening of a passive conductance for small ions during organic cation transport should change the charge-to-substrate ratio in the opposite direction because more negative potential would increase both, influx of small cations and passive efflux of small anions. It is possible that the opening of a passive permeability through rOCT2 during the transport cycle is favored by the transporter conformation at 0 mV; however, the effects observed after ion replacement experiments do not allow deriving a presumed ion selectivity of such permeability that could explain the observed surplus of positive charge uptake at 0 mV. This interpretation is also at odds with the effect of the Glu448Gln mutation (see below). For these reasons, we rather propose that the excess of charge flow during organic cation uptake at 0 mV is due to nonselective low-affinity binding of small cations within the innermost cavity of the binding cleft and their subsequent translocation together with organic cations (nonselective-cotransport-of-small-cations hypothesis). Small cations may be attracted nonselectively by positively charged amino acid residues; however, van der Waals forces and hydrophobic interactions may be also effective.
Modeling of rOCT2 structural states predicts the functional consequences of a point mutation and supports the nonselective-cotransport-of-small-cations hypothesis.
We modeled the inward-open conformation of rOCT2 according to the crystal structure of the LacY permease, a member of the MFS superfamily like the OCTs, as previously described for rOCT1 (21). Convincing experimental evidence has been provided that the LacY permease operates by an alternating access mechanism and obtains an inward-facing conformation, a conformation with occluded substrate cavity, and an outward-facing conformation (1, 15, 19, 27). Since previous data suggest a similar mechanism for OCTs (11, 31), we modeled the outward-open conformations of rOCT1 (11) and rOCT2 (this study) using the helix rearrangement that has been proposed for the outward-facing conformation of the LacY permease. Due to the high conservation of amino acids between rOCT1 and rOCT2, the modeled structures of both subtypes are very similar. Since model structures include various uncertainties, it is remarkable that main features of the modeled innermost cavities of the inward-open and outward-open binding clefts could be confirmed by mutagenesis. In the case of rOCT1, mutations of four amino acids lining the outward-open as well as the inward-open innermost cavity (Phe160, Trp218, Arg440, Leu447 corresponding to Tyr447 in rOCT2; Fig. 9D) altered the affinities for the inhibition of TEA+ uptake by corticosterone that was either shortly applied from the extracellular side or was only accessible from the intracellular side (Volk C, Gorboulev V, Klotzsch A, Müller TD, and Koepsell H, unpublished observations). By mutating these four amino acids, the affinity for MPP+ was also altered. These data strongly suggest that OCTs contain an innermost substrate binding cavity that is alternatingly exposed to the extracellular and intracellular cell side during the transport cycle. The modeled charge distributions of the outward-open and inward-open conformations of rOCT2 shown in the present study support the interpretation of the observed charge-to-cation ratios higher than one at 0 mV (nonselective-cotransport-of-small-cations hypothesis). They predict that the innermost cavity of the outward-facing substrate binding cleft of rOCT2 is negatively charged containing two negatively charged amino acids (Glu448, Asp475), whereas the innermost cavity of the inward-facing cleft is about electroneutral containing two negatively charged (Asp379, Asp475) and two positively charged amino acids (Lys215 and Arg440). Note that Glu448 does not contribute to the inward-open innermost cavitiy. Small cations may be trapped nonselectively within the outward-open innermost cavity. After binding of an organic cation substrate the outward-facing innermost binding cavity may undergo structural changes during which it becomes closed to the extracellular side, may pass an occluded state, and may be opened to the intracellular side. During these transitions, the surface charge distribution of the innermost cavity may be changed and organic cation substrate may be released together with the trapped small cations. During reorientation of the empty transporter equal amounts of small cations and anions or a surplus of small anions trapped within the inward-open innermost cavity may be translocated to the extracellular site. Importantly we observed that removal of one negative charge from the outward-open innermost cavity (Glu448Gln mutation) abolished the surplus of positive charge that was translocated together with TEA+ at 0 mV. Nonselective cotransport of small cations proposed for rOCT2 differs to classic cotransport mechanisms as described for Na+-d-glucose cotransporters or Na+ neurotransmitter cotransporters. First, cotransport of small cations by rOCT2 has a low selectivity for small cations since it does not appear to differentiate between Na+ and K+. Second, cotransport by rOCT2 exhibits a variable stoichiometry between transported organic cation substrates and cotransported small cations. Third, cotransport by rOCT2 is not obligatory since it has been only observed under depolarized conditions.
The reasons for the observed potential and substrate dependence of charge-to-substrate ratio are poorly understood. Concerning the effect of increased membrane potential, it is possible that the weakening of cation attraction due to reduced electrostatic potential around the negatively charged amino acids at the rim of the outward-open cleft is more effective than the increase in cation attraction due to the increased membrane potential. Another explanation may be potential dependent structural differences of the outward-open and inward-open conformation of rOCT2 that may reflect the sizes and/or distribution of charge within the outward-facing and inward-facing innermost cavities. The different charge-to-substrate ratios observed for different organic cation substrates may be due to a different occupation of the innermost cavities, possibly resulting in different shielding of charges. Another possibility is that different organic cation substrates that are supposed to occupy high-affinity cation binding sites in peripheral parts of the binding cleft without being directly involved in translocation (11, 21) induce differential allosteric effects on the innermost cavities of the binding clefts.
Physiological and pathophysiological implications.
The above-described data may have physiological and pathophysiological implications. A charge-to-organic cation ratio close to one at physiological membrane potentials is important for the contribution of OCTs to secretion of organic cations in the liver and kidney. Since the membrane potential is essential for various cell functions and provides part of the driving force for organic cation uptake by OCTs, a too strong depolarization of hepatocytes or renal proximal tubular cells during organic cation uptake has to be avoided. The relatively small depolarization during organic cation uptake at normal membrane potential may be compensated by increasing open probability of K+ channels and (Na++K+)-ATPase activity. Pathological conditions such as ischemia or hyperkalemia tend to depolarize proximal tubular epithelial cells. Under depolarized conditions, where organic cation uptake is associated with a considerable excess current, high concentrations of organic cation substrates will tend to depolarize the cells even further, potentially triggering a vicious cycle that may aggravate renal dysfunction. Employing the renal cell line HEK293 that was stably transfected with rOCT2, we demonstrated that organic cation uptake by rOCT2 lead to a more pronounced decrease in membrane potential in depolarized cells vs. polarized cells. Notwithstanding 1) that there may be species differences between rOCT2 and hOCT2, 2) that the expression level of OCT2 in dysfunctional kidneys may be different than in transfected HEK293 cells, and 3) that mechanisms which counteract membrane potential changes may be different, the data demonstrate that cation-induced depolarization during renal dysfunction may have pathophysiological relevance.
The charge-to-substrate ratio greater than unity at 0 mV may be specifically relevant for neurons expressing OCT2 and OCT3 (4). Human OCT2 (hOCT2) mediates uptake of acetylcholine and serotonin with a Michaelis-Menten constant of 100–300 μM (17) and may thus contribute to the presynaptic reuptake of these neurotransmitters. The surplus of electric charge flowing through OCT2 may increase and/or prolong depolarization at axosomatic cholinergic or serotonergic synapses. OCT3 participates in reuptake of epinephrine, norepinephrine, and histamine and is involved in modulation of motor activity, behavior, and regulation of salt uptake (16, 30). It is presently unknown whether this OCT subtype also exhibits voltage-dependent variation of charge-to-substrate ratio, but such properties would be equally important for the role of OCT3 in brain function.
This work was supported by the Deutsche Forschungsgemeinschaft Grant SFB487/A4.
↵* B. M. Schmitt, D. Gorbunov, and P. Schlachtbauer contributed equally to this work.
↵1 With MPP+, the ratio |e+|:|S+| at −100 mV was smaller than 1.0 (Fig. 2B, |e+|:|MPP+| = 0.37 ± 0.014). This observation points to movement of anions together with this specific substrate, and/or of cations in the opposite direction at −100 mV. In the present study, we did not investigate this MPP+-specific effect further.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society