D-Enantiomers Take a Close Look at the Functioning of a Cardiac Cationic L-Amino Acid Transporter

Abstract

Cationic amino acid transporters are highly selective for L-enantiomers such as L-arginine (L-Arg). Because of this stereoselectivity, little is known about the interaction of these transporters with D-isomers. To study whether these compounds can provide information on the molecular mechanism of transport, inward currents activated by L-Arg with low apparent affinity were measured in whole-cell voltage-clamped cardiomyocytes as a function of extracellular L-Arg and D-Arg concentrations. D-Arg inhibited L-Arg currents in a membrane-potential (V_M)-dependent competitive manner, indicating the presence of D-Arg binding sites in the carrier. Analysis of these steady-state currents showed that L- and D-Arg binding reactions dissipate a similar small fraction of the membrane electric field. Since D-Arg is not transported, these results suggest that enantiomer recognition occurs at conformational transitions that initiate amino acid translocation. The V_M dependence of maximal current levels suggests that inward currents arise from the slow outward movement of negative charges in the unliganded transporter. Translocation of the L-Arg-bound complex, on the other hand, appears to be electroneutral. D-Arg-dependent transient charge movements, also detected in these cells, displayed a V_M-dependent charge distribution and kinetics that are consistent with amino acid binding in an ion well in a shallow, water-filled extracellular binding pocket.

Introduction

Cardiac muscle cells depend entirely on transport from the circulation to ensure adequate intracellular levels of cationic amino acids. In particular, arginine-activated currents measured in these cells represent import of this amino acid through a low-affinity process that is consistent with the activity of CAT-2A (1), a member of the system-y⁺ family of cationic amino acid transporters (2,3). This low-affinity transport activity accounts for at least 50% of the total incorporated arginine at normal plasma levels of this amino acid, the other half being transported by the high-affinity/low-capacity CAT-1 (4). Among its many roles in physiology, arginine is the substrate for the biosynthesis of nitric oxide, a signaling molecule with major regulatory effects on the cardiovascular system (5,6).

System-y⁺ carriers, which include the high-affinity CAT-1, CAT-2B, and CAT-3, as well as the low-affinity CAT-2A, display high selectivity for the L-enantiomers of arginine (L-Arg), lysine (L-Lys), and ornithine (L-Orn). These amino acids are transported in a Na⁺-independent manner down their electrochemical potential gradient and, within a given isoform, with similar apparent affinities (reviewed in Devés and Boyd (7)). In addition, radiolabeled cationic L-amino acid uptake mediated by system-y⁺ transporters is accelerated by the presence of unlabeled substrate at the other side of the membrane, a signature feature called trans-stimulation (8) that indicates that the kinetics of binding-site translocation are faster in the presence of substrate (reviewed in Devés and Boyd (7)). The high-affinity isoforms display up to 10-fold trans-stimulation, whereas CAT-2A has been reported to be somewhat irresponsive to stimulation by trans-substrate (2). In contrast, our recent work has shown that low-affinity L-Lys uptake is trans-stimulated approximately threefold in cardiac sarcolemmal vesicles (4). Besides these general characteristics, the molecular mechanisms of amino acid transport by members of this family remain largely unknown.

The stereoselectivity of these carriers and the biological significance of L-amino acids have drawn attention away from D-enantiomers. Nonetheless, it is pertinent to the transport mechanism to learn whether these proteins bind D-isomers and, if so, how they distinguish between compounds that are three-dimensional mirror images. To the best of our knowledge, only two publications have dealt with D-Arg effects on members of system y⁺, yielding overall inconclusive results. In one case, D-Arg was reported to be a competitive inhibitor of L-Arg uptake, with a K_i 10-fold larger than the K_m for L-Arg activation of transport (8). The other status publication literally reports “…the failure of D-Arg to inhibit competitively the transport of its stereoisomer…” at concentrations up to 3 mM (9).

In light of this scarce and contradictory information, we decided to investigate the effect of D-Arg on L-Arg-activated currents in whole-cell voltage-clamped ventricular cardiomyocytes, with the aim to clarify whether the D-enantiomer is a competitive inhibitor of L-Arg transport and to use this compound as a tool to gain insights into how these carrier proteins work.

Portions of this study have been previously published in abstract form (10).

Methods

Adult male rats were injected intraperitoneally with Nembutal, 100 mg kg^–1, in accordance with Institutional Animal Care and Use Committee guidelines, and their hearts were removed under complete anesthesia. Single ventricular myocytes were enzymatically isolated using published methods (11).

Steady-state voltage-clamp experiments

Freshly isolated ventricular cardiomyocytes were placed in a superfusion chamber on the stage of an inverted microscope and superfused at 36 ± 1°C with a Tyrode solution containing (in mM) 145 NaCl, 5 KCl, 10 dextrose, 2 CaCl₂, 1 MgCl₂, and 10 Hepes/NaOH, pH 7.40 (23°C). Myocytes were voltage-clamped in the whole-cell configuration with low-resistance (1.0–1.5 MΩ) patch electrodes back-filled with an intracellular salt solution containing (in mM) 110 potassium aspartate, 20 TEACl, 4 MgCl₂, 0.7 ATP-Mg²⁺ salt, 10 EGTA/Tris, 5 glucose, and 10 Hepes/KOH, pH 7.30 (23°C). Voltage-clamp experiments were performed with a low-noise Axopatch 200B patch-clamp amplifier using a Digidata 1400 and pCLAMP 10 software for data acquisition and analysis (Molecular Devices, Sunnyvale, CA).

After establishing a Giga-Ohm seal, the superfusion solution was switched to a Na⁺- and K⁺-free solution containing (in mM) 145 tetramethylammonium chloride (TMACl), 2.3 MgCl₂, 0.2 CdCl₂, 5.5 dextrose, and 10 Hepes/Tris, pH 7.40 (23°C). L-Arg- and D-Arg-containing solutions were prepared on the day of the experiment by equimolar substitution of TMA. Extracellular TMA and Cd²⁺, as well as intracellular TEA, were added to block contaminating ionic currents, as previously described to isolate Na,K-pump current (12). Cells were exposed to these blockers for 5 min before further manipulations.

Voltage-clamp protocol

Step changes of 100 ms in V_M were produced from a holding potential of −40 mV to various V_M in the range −100 to +40 mV, in 10-mV increments at 2 Hz. These V_M jumps were applied before and during superfusion with L-Arg-containing solution, and again after L-Arg removal, to obtain the respective current-V_M relationships. L-Arg currents were defined as the difference between current levels measured in the absence and presence of the amino acid.

Pre-steady-state voltage-clamp experiments

To measure transient charge movements, protocols were followed as previously outlined (12). Experiments were performed at 19–21°C (to slow the kinetics of current relaxation) with Na⁺- and K⁺-free superfusion solutions containing (in mM) 145 TMACl, 2.3 MgCl₂, 0.2 CdCl₂, 5.5 dextrose, and 10 Hepes/Tris, pH 7.4 (20°C). The hydrochloride salt of D-arginine was added to this solution as indicated. The patch electrode solution contained (in mM) 110 cesium sulfamate, 20 TEACl, 4 MgCl₂, 0.7 ATP-Mg²⁺ salt, 10 EGTA/Tris, 5 dextrose, and 10 Hepes/CsOH, pH 7.2 (20°C).

Voltage-clamp protocol

In all experiments, the holding potential was −40 mV. When needed, voltage steps 100 ms long from −160 to +100 mV were elicited in 20-mV increments at 2 Hz. Each protocol was performed in the presence and absence of D-Arg so that D-Arg-sensitive difference currents could be obtained. Corrections for junction potentials were not carried out.

Data analysis

Current traces were sampled at 10 kHz and 25 Hz, and low-pass filtered at 3 kHz and 6 Hz for pre-steady-state and steady-state measurements, respectively. Linear cell capacitance was calculated with Clampex using 5-mV depolarizing pulses, and current analysis was performed with Clampfit. Clampex and Clampfit routines are included in pClamp 10 (Molecular Devices). Data are displayed as the mean ± SE for the indicated number of experiments. Statistical significance was determined using the Student's t-test (P < 0.05). Curve-fitting was performed with nonlinear least-squares routines included in SigmaPlot v10.0 (Systat Software, Chicago, IL) using statistical weights proportional to (SE)^–2.

Results

Effect of D-Arg on L-Arg-activated currents

To study the effect of D-enantiomers on cationic L-amino acid transport, L-Arg currents were measured in myocytes voltage-clamped at −40 mV with patch electrodes filled with a high-K⁺ intracellular salt solution and superfused with a Na⁺- and K⁺-free, 145-mM TMA-containing solution. Under these conditions, increasing concentrations of extracellular L-Arg (L-Arg_o) elicited progressively larger currents, as shown in Fig. 1 A for 1, 5, and 10 mM L-Arg_o. The V_M dependence of current was studied in the range −100 to +40 mV by applying a voltage-clamp protocol (described in Methods) after L-Arg-activated currents reached a steady state at −40 mV. Baseline curves obtained 30 s before L-Arg application and 2–3 min after withdrawal of the amino acid were averaged and subtracted from those obtained in the presence of L-Arg. The corresponding difference current (I_M)-V_M relationships collected from 12 myocytes are shown in Fig. 1 B for the concentrations of L-Arg_o used in Fig. 1 A. Current levels increased with more negative V_M at all L-Arg_o concentrations tested. This V_M dependence was still apparent at L-Arg concentrations as high as 50 mM (not shown), suggesting the existence of additional electrogenic reaction steps not involved in L-Arg binding/release (see also Peluffo (1)).

Figure 1.

Effect of D-Arg on extracellular L-Arg-activated currents in whole-cell voltage-clamped cardiomyocytes. (A) Inward currents were elicited by increasing concentrations of L-Arg at a holding potential of −40 mV in the absence or presence of increasing concentrations of D-Arg. (B) I_M-V_M relationships obtained by applying voltage jumps to currents measured in the presence of 1 (●), 5 (○), and 10 (▾) mM L-Arg_o. Symbols represent average difference current values calculated for the last 70 ms of the applied voltage pulses, defined as steady-state current. (C) I_M-V_M relationships for 10 mM L-Arg-activated currents obtained in the presence of 0 (○), 10 (●), and 30 (▵) mM D-Arg_o. A curve obtained with 10 mM D-Arg_o and no added L-Arg is also included (▾) to show that the D-enantiomer is poorly (if at all) transported. Symbols represent the mean ± SE of three to five experiments for each condition. Curves in B and C were calculated with Eq. 1 and the best-fit parameters obtained by simultaneously fitting this equation to the entire data set (see text for details).

When myocytes were superfused with a 10 mM L-Arg_o-containing solution, increasing concentrations of D-Arg_o produced a concomitant decrease in current density levels (Fig. 1 A). Inspection of I_M-V_M relationships obtained from 14 cells treated with 0, 10, and 30 mM D-Arg shows that the inhibitory effect of the D-enantiomer was more pronounced at hyperpolarizing potentials (Fig. 1 C). As an example, 10 and 30 mM D-Arg inhibited L-Arg currents by 48 and 77% at −100 mV, respectively, and by 26 and 53% at −10 mV, respectively. For comparison, application of 10 mM D-Arg_o in the absence of L-Arg_o elicited negligible inward currents at all V_Ms tested (Fig. 1 C, solid triangles). Altogether, these results indicate that L-Arg currents are blocked by D-Arg in a V_M-dependent manner.

Inhibition type

To determine whether D-Arg inhibition was competitive with L-Arg activation of current, the experiments described above (see Fig. 1) were repeated with L-Arg concentrations in the range 1–50 mM. Whenever possible, the effects of 10 and 30 mM D-Arg, as well as control-current levels, were tested in the same myocyte. The concentration dependence of D-Arg effects on L-Arg-activated current at selected V_M is summarized in Fig. 2. In a preliminary analysis of these results, data at each D-Arg concentration and V_M were fit with single decreasing hyperbolic functions. The L-Arg concentration required for half-maximal activation of current (K_0.5) and maximal current density levels (I_max) are presented as functions of V_M and D-Arg concentration in Fig. 3. L-Arg K_0.5 values increased both with depolarization and with the concentration of D-Arg (Fig. 3 A). For example, K_0.5 at zero V_M was found to have values of 10.9 ± 0.9, 20.6 ± 2.6, and 41.7 ± 2.7 mM at 0, 10, and 30 mM D-Arg, respectively. I_max increased (in absolute value) with hyperpolarization, showing no signs of saturation at up to −100 mV (Fig. 3 B). It is important to note that I_max values were not statistically significantly different at any of the D-Arg concentrations tested. This behavior of K_0.5 and I_max as a function of D-Arg is, by definition, that of a competitive inhibitor. Therefore, considering the results in Fig. 1 C, D-Arg appears to be a V_M-dependent competitive blocker of L-Arg transport. In addition, since I_max is defined for L-Arg concentrations approaching infinity, its V_M dependence (Fig. 3 B) indicates that steps other than L-Arg binding move a substantial amount of charge across the membrane electric field.

Figure 2.

L-Arg_o-concentration dependence of current at V_M = −100 mV (A), −40 mV (B), and +20 mV (C) for 0 (●), 10 (○), and 30 mM D-Arg (▾). Ordinate scaling in B and C was kept proportional to that in A to highlight the dramatic decrease in current levels with depolarization at all L-Arg concentrations tested. Curves were generated with Eq. 1 and the best-fit parameters obtained by fitting this equation to the entire dataset.

Figure 3.

Effect of D-Arg and V_M on kinetic parameters of L-Arg transport. (A) L-Arg K_0.5-V_M relationships in the presence of 0 (●), 10 (○), and 30 mM D-Arg (▾). Symbols were obtained by fitting single hyperbolic functions to the data at each D-Arg concentration and V_M. Curves were calculated at each D-Arg using the best-fit parameters from Eq. 1 on the expression for K_0.5,

$$K_{0.5}=K_{0.5}^0\exp \left({\lambda }_{\text{L}-\text{Arg}}U\right)\left[1+\frac{[\text{D}-\text{Arg}]}{K_{\text{i}}^0}\exp \left(-{\lambda }_{\text{D}-\text{Arg}}U\right)\right],$$

that is included in the denominator of Eq. 1. (B) I_max-V_M relationships for 0 (●), 10 (○), and 30 mM D-Arg (▿). Symbols represent I_max values obtained by fitting hyperbolic functions to the data at each D-Arg concentration and V_M. The curve, which represents the numerator of Eq. 1, was calculated using the best-fit values of I⁰_max and λ_turnover.

Pseudo-three-state reaction scheme

The effect of D-Arg on L-Arg transport was analyzed according to the reaction scheme (Scheme 1)where T_o is the outward-facing conformation of the cationic amino acid transporter, T_oD is the D-Arg-bound complex, and T_oL is the L-Arg-bound complex. One of the hemiloops represents extracellular L-Arg binding; the other lumps together L-Arg translocation and release, as well as reorientation of the protein to the outward-facing conformation. Solution of this Scheme in terms of electrogenic L-Arg transport yielded the expression

$$I_{\text{M}}=\frac{I_{\max }^0\exp (-{\lambda }_{\text{turnover}}U)}{1+\left[\frac{K_{0.5}^0}{[\text{L}-\text{Arg}]}\exp \left({\lambda }_{\text{L}-\text{Arg}}U\right)\right]\left[1+\frac{[\text{D}-\text{Arg}]}{K_{\text{i}}^0}\exp (-{\lambda }_{\text{D}-\text{Arg}}U)\right]},$$ (1)

where superscripts “0” indicate that the parameter is defined at V_M = 0; λ values are dielectric coefficients that represent the portion of the membrane electric field dissipated by the given reaction; K_0.5 is the concentration of L-Arg that produces half-maximal activation of transport; K_i is the concentration of D-Arg that produces half-maximal inhibition of transport; and U is the dimensionless transmembrane voltage expressed as zFV_M/RT. Consistent with results presented in Figs. 1–3, Eq. 1 is the expression for L-Arg current with a V_M-dependent turnover rate in the presence of a V_M-dependent competitive inhibitor.

Scheme 1.

Equation 1 was simultaneously fit in three variables to the entire dataset, yielding the six best-fit parameter values shown in Table 1. Notice that λ_turnover > λ_L-Arg = λ_D-Arg, and that K⁰_0.5 ≅ K⁰_i. To visualize whether Scheme 1 accounts for all experimental features of these currents, Eq. 1 was solved with these parameter values and plotted in Figs. 1, B and C, 2, and 3 (solid lines). Scheme 1 properly describes the V_M dependence of L-Arg currents at different L-Arg (Fig. 1 B) and D-Arg concentrations (Fig. 1 C). The reaction scheme also accounts well for the L-Arg and D-Arg concentration dependence of current at any given V_M (Fig. 2). Accordingly, the values of L-Arg K_0.5 and I_max, calculated by individually fitting hyperbolic functions at each D-Arg concentration and V_M, are reasonably well described by the pseudo-three-state model (Fig. 3). Taken together, these results demonstrate that D-Arg is a V_M-dependent competitive inhibitor of L-Arg transport in cardiac myocytes. Furthermore, D-Arg and L-Arg dissipate a similar small fraction of the membrane electric field during binding reactions.

Table 1.

Parameter values describing the pseudo-three-state model

I0max (pA pF–1)	–1.06 ± 0.06
λturnover	0.41 ± 0.02
K00.5 (mM)	10.1 ± 0.8
λL-Arg	0.14 ± 0.05
K0i (mM)	8.3 ± 0.5
λD-Arg	0.13 ± 0.05

The pseudo-three-state model is depicted in Scheme 1. Parameter values were obtained by simultaneously fitting the entire dataset with Eq. 1.

D-Arg-dependent charge movements

Based on the findings above, it is anticipated that to block L-Arg transport in a V_M-dependent manner, D-Arg will move its net positive charge within the membrane electric field, thus producing D-Arg-dependent charge movements. To test this prediction, ventricular cardiomyocytes were whole-cell voltage-clamped at a holding potential of −40 mV with wide-tipped patch electrodes backfilled with a 115 mM Cs⁺-containing solution and superfused with a 145 mM TMA-containing, Na⁺- and K⁺-free external solution at 20°C. Voltage steps (see Methods) were applied before, during, and after exposure to 10 mM D-Arg, yielding current traces such as those shown superimposed in Fig. 4 A, before (left) and after (right) 2 min in the presence of the amino acid. The concentration of D-Arg was selected to be close to the K⁰_i value determined above because nonsaturation of the binding site with ligand is a prerequisite for the detection of charge movements. A detail of traces obtained with voltage steps to −100 and +60 mV in the absence and presence of D-Arg is shown in Fig. 4 B. The surface enclosed by both curves represents the amount of charge that was moved at each V_M. Upon subtraction, D-Arg-sensitive difference currents were obtained. Fig. 4 C shows selected transient currents elicited by voltage pulses from −40 mV to −120, −60, 0, and +60 mV (“on” currents), and back to −40 mV from −60, 0, and +60 mV (“off” currents). The off-current trace corresponding to −120 mV was omitted because of current contamination at V_M <−100 mV. The quantity of charge moved at the onset of voltage pulses was calculated to be −4.5, −1.2, +1.5, and +3.1 fC pF⁻¹, respectively. Likewise, current relaxation followed single-exponential kinetics (see superimposed fitting for “on” currents), with apparent rate-constant (k_tot) values of 1100, 828, 941, and 1055 s⁻¹, respectively. In all cases, currents decayed at rates that were at least fourfold slower than charging of linear membrane capacitance (typical clamp time constants were 210–250 μs). Voltage-clamp stability was tested by comparing current traces obtained before D-Arg application and 3 min after washout of the amino acid (total elapsed time, 5.5 min). Superimposed difference currents yielded residual charge amounts that were 5–9 % of the D-Arg-sensitive charge in Fig. 4 C at equivalent voltage pulses (Fig. 4 D). Therefore, changes in clamp stability were negligible over the duration of these experiments and, thus, transient charge movements are genuinely associated with D-Arg binding/unbinding reactions.

Figure 4.

D-Arg-dependent transient charge movements. (A) Superimposed current traces elicited by voltage pulses from −40 mV to voltages in the range −160 to +100 mV were measured in a whole-cell voltage-clamped cardiac myocyte superfused with D-Arg-free (left) or 10 mM D-Arg-containing solution (right). (B) Current traces from A in the absence and presence of D-Arg are displayed superimposed at two selected V_M to show the area contained within both sets of curves. (C) Superimposed D-Arg-sensitive difference currents at selected V_M. The solid curves through the decaying phase of current traces are best-fit exponential functions. (D) Difference currents determined by subtracting membrane currents recorded before D-Arg application from those obtained 3 min after removal of the amino acid. Superimposed currents are shown for voltage-clamp steps to −120, −60, 0, and +60 mV.

Experiments were also performed with 5 mM D-Arg to study the concentration dependence of charge movements (see Fig. S1 in the Supporting Material). The salient feature of these transient currents is that depolarizing V_M greatly increased current relaxation rates. In the example shown, k_tot values went from 490 to 880 s⁻¹ by changing V_M from −20 to +60 mV.

V_M dependence of steady-state charge distribution

The quantity of charge moved (ΔQ), calculated as the time integral of transient currents produced by voltage pulses from −160 to +100 mV (“on” charge) and after returning to the holding potential (“off” charge), was characterized as a function of V_M. Fig. 5 A shows ΔQ-V_M relationships for 10 mM D-Arg-sensitive “on” and “off” charge. Values of ΔQ were found to saturate at large negative and positive potentials, which suggests that these charge movements involve redistribution of a finite number of charged particles in the membrane. The nonlinearity of the steady-state charge distribution enabled analysis with the Boltzmann function (Fig. 5 A, solid lines):

$$\Delta Q=Q_{\min }+Q_{\text{tot}}/\left\{1+\exp \left[{\text{z}}_{\text{q}}F\left(V_{1/2}-V_{\text{M}}\right)/\text{RT}\right]\right\},$$ (2)

where Q_tot = Q_max − Q_min is the total quantity of mobile charge, z_q represents the apparent valence obtained from steady-state charge distribution measurements, and V_1/2 is the midpoint potential, i.e., the V_M at which the concentrations of D-Arg-bound and -free transporter are equal. Best-fit parameters for “on” and “off” charge are reported in the figure legend (Fig. 5). The absolute values of Q_tot were not statistically significantly different for Q_on and Q_off, indicating that the same amount of charge was reversibly moved in each direction. The midpoint potential, V_1/2, was found to be slightly more negative for “on “than for “off” charge. Finally, the apparent valence of the mobile charges was similar and slightly <1 in both cases.

Figure 5.

V_M dependence and kinetics of D-Arg-dependent charge movement. (A) V_M dependence. Ten millimolar D-Arg-sensitive charge calculated during voltage pulses to the indicated V_M (ΔQ_on) and after returning to the holding potential (ΔQ_off) are represented by solid circles and open triangles, respectively (n = 5). Curves through the data represent fitting of Eq. 2 with the best-fit parameter values:

	ΔQon (10 mM)	ΔQoff (10 mM)	ΔQon (5 mM)
Qtot (fC pF−1)	9.6 ± 1.3	−9.3 ± 1.5	10.2 ± 0.8
V1/2 (mV)	−29.8 ± 3.7	−20.1 ± 3.8	−79.1 ± 4.8
zq	0.74 ± 0.19	0.76 ± 0.18	0.70 ± 0.13
Qmin (fC pF−1)	−4.3 ± 0.6	3.8 ± 0.9	−7.0 ± 1.5

(B) ΔQ_on-V_M curves comparing data at 5 (○, n = 3) and 10 mM D-Arg (●). Data were normalized as follows, using the parameter values listed above:

$$\left(\Delta Q-Q_{\min }\right)/Q_{\text{tot}}={\left\{1+\exp \left[z_{\text{q}}F\left(V_{1/2}-V_{\text{M}}\right)/\text{RT}\right]\right\}}^{-1}$$

(C) Kinetics. The decaying portions of D-Arg-sensitive charge movements were fit with exponential functions (see Fig. 4C) and the resulting rate constants (k_tot) were plotted against V_M for 5 (○) and 10 mM D-Arg (●). Data points were then simultaneously fit with Eq. 3 to yield the solid curves and the best-fit parameter values:

kf0 (s−1 mM–1)	76.3 ± 6.2
δ	0.19 ± 0.06
zk	0.50 ± 0.15
kr0 (s−1)	214.6 ± 51.1

Charge-V_M relationships obtained with 5 mM D-Arg were also analyzed with Eq. 2, and best-fit parameters for Q_on are reported in Fig. 5. The value of Q_tot was not statistically significantly different from that obtained with 10 mM D-Arg. In a similar way, z_q remained unchanged at both D-Arg concentrations. On the other hand, the midpoint potential shifted ∼50 mV when the concentration of D-Arg was doubled. This shift in V_1/2 is shown in Fig. 5 B for normalized ΔQ-V_M relationships. Altogether, this behavior is consistent with D-Arg binding in an ion well.

The V_M-dependent kinetics of current relaxation

To study the effect of voltage on the kinetics of current decay, k_tot values were obtained at all V_M tested by fitting exponential functions to D-Arg-sensitive current traces. The results, plotted as k_tot-V_M relationships for 5 and 10 mM D-Arg-dependent charge movements (Fig. 5 C), show a U-shaped dependence on V_M. Data analysis was performed by simultaneous fitting in two variables of Eq. 3, which considers k_tot as the sum of V_M-dependent forward and reverse rate constants that describe binding reactions:

$$k_{\text{tot}}\left(V_{\text{M}};\left[\text{D}-\text{Arg}\right]\right)=k_{\text{f}}^0\left[\text{D}-\text{Arg}\right]\exp \left\{-\varphi {\text{z}}_{\text{k}}F{V}_{\text{M}}/\text{RT}\right\}+k_{\text{r}}^0\exp \left\{(1-\varphi ){\text{z}}_{\text{k}}F{V}_{\text{M}}/\text{RT}\right\},$$ (3)

where k_f⁰ and k_r⁰ are the second-order forward and first-order reverse rate constants at zero V_M; ϕ is a symmetry factor (0 ≤ ϕ ≤ 1) that accounts for the effect of V_M on the forward and reverse reaction steps, and z_k represents the apparent valence of the mobile charges as determined by kinetic measurements. Best-fit parameters, listed in the figure legend, show three distinct features. First, the value of k_f⁰ (76,300 ± 6160 s^–1 M^–1 at V_M = 0) corresponds to a process far slower than diffusion. Second, a ϕ = 0.19 indicates that the V_M dependence lies largely on the reverse reaction (D-Arg release). Finally, the apparent valence of the mobile charges, although lower on average, was not significantly different from z_q, the value obtained from steady-state charge distribution.

In conclusion, these results confirm the prediction that as a V_M-dependent blocker of L-Arg transport, D-Arg binding reactions move charge within the membrane electric field.

Discussion

The D-enantiomer of arginine was found to be a V_M-dependent competitive inhibitor of low-affinity cationic L-amino acid transport in cardiac myocytes and, as such, was used as a tool to gain insights into how these carrier proteins work. The result λ_L-Arg = λ_D-Arg indicates that both stereoisomers bind to sites located at the same electrical distance within the transporter. Together with the comparable values found for K⁰_0.5 and K⁰_i, and the fact that D-Arg is not transported, this result suggests that D- and L-Arg compete equally efficiently for the same binding site and, thus, enantiomer recognition might occur downstream at conformational transitions associated with amino acid translocation. Furthermore, these conformational transitions are probably what rate-limit the kinetics of D-Arg-dependent charge movements (Fig. 5 C). The ability of the binding site to accommodate both stereoisomers is reminiscent of the induced-fit model proposed for enzyme kinetics (13), which postulates that enzymes are rather flexible structures with active sites continually being reshaped by their interactions with the substrate.

Extracellular binding reactions dissipate ∼15% of the membrane electric field, i.e., are weakly V_M-dependent, so that the main electrogenic step must involve either amino acid translocation, or release to the opposite side, or reorientation of the unliganded carrier to the extracellular side. Such a prediction correlates well with the observed V_M dependence of I_max (λ_turnover = 0.41), a parameter defined at [L-Arg] → ∞ and, thus, independent of extracellular L-Arg binding reactions. Moreover, the shape of the I_max-V_M relationship is consistent with a single charge-translocation event that is also the rate-limiting step in the reaction cycle (14). Considering a four-state alternating model for L-Arg transport (Scheme 2), reasonable candidates for this charge-moving slow step include translocation of the loaded or unliganded transporter.

Scheme 2.

In this model, T represents a transporter molecule that can exist in extracellular- (T_o) and intracellular-facing conformations (T_i). T_oL and T_iL are the corresponding conformations for the L-Arg-bound complex. Rate constants k₁, k₂, k₃, and k₄ describe the L-Arg (L_o) inward transport process. Rate constants k₋₂, k_–1, and k₋₄ describe outward translocation of the L-Arg-bound complex, L_o release, and the inward reorientation of the unliganded transporter, respectively, under zero-trans conditions ([L_i] = 0). The steady-state expression for current generated by L-Arg import can be derived using the King and Altman algorithm (15) as

$$I_{\text{M}}^{\text{o}\to \text{i}}=-zF{\left[T\right]}_{\text{T}}\frac{k_1{k}_2{k}_3{k}_4\left[{\text{L}}_{\text{o}}\right]}{\left[k_2\left(k_3+k_4\right)+k_4\left(k_3+k_{-2}\right)\right]k_1\left[{\text{L}}_{\text{o}}\right]+\left[k_{-1}\left(k_3+k_{-2}\right)+k_2{k}_3\right]\left(k_4+k_{-4}\right)},$$ (4)

where [T]_T is the total concentration of cationic amino acid transporters in the membrane. Accordingly, the expression for I_max is

$$I_{\max }^{\text{o}\to \text{i}}=\frac{-zF{\left[T\right]}_{\text{T}}}{\frac{1}{k_4}+\frac{1}{k_3}+\frac{1}{k_2}+\frac{k_{-2}}{k_2{k}_3}}.$$ (5)

Equation 5 can be expressed as an explicit function of V_M:

$$I_{\max }^{\text{o}\to \text{i}}=\frac{-zF{\left[T\right]}_{\text{T}}}{\frac{\exp \left(\eta U\right)}{k_4^{\text{0}}}+\frac{\exp \left(\gamma U\right)}{k_3^{\text{0}}}+\frac{\exp \left(\beta U/2\right)}{k_2^{\text{0}}}+\frac{k_{-2}^{\text{0}}\exp \left[\left(\beta +\gamma \right)U\right]}{k_2^{\text{0}}{k}_3^{\text{0}}}},$$ (6)

where k_j⁰ indicates rate constants defined at zero V_M, z is the valence of the charged particles crossing the membrane, and β, γ, and η are dielectric coefficients for the partial reactions described by the respective rate constants. The explicit expression for k₄ as a function of V_M is given here as an example: k₄ = k₄⁰ exp(−ηU).

Simulations with Eq. 6 that assign the V_M-dependent rate-limiting step to either translocation of the unloaded carrier (k₄, η > 0, β = 0) or the L-Arg-bound complex (k₂ and k₋₂, β > 0, η = 0) yield indistinguishable I_max-V_M relationships (Fig. S2). However, the phenomenon of trans-stimulation (8), a distinctive feature displayed by this family of transporters in which unlabeled trans-substrate increases severalfold the rate of cationic amino acid uptake, indicates that the reaction T_oL ↔ T_iL is indeed fast (2,4,7). Therefore, translocation of the unliganded transporter T_i → T_o (described by k₄) appears to be the slow V_M-dependent reaction. In this case, the outward movement of a negative charge must accompany reorientation of the unoccupied transporter to be consistent with the observed current polarity and V_M dependence of I_max. This scenario suggests that a negatively charged residue located at or near the cationic amino acid binding site interacts with the distal positive charge of these compounds upon binding, thus leading to the electroneutral translocation of the L-Arg-bound complex (β = 0) and the electrogenic outward reorientation of the unliganded transporter (η > 0).

Based on the proposed topology of the human CAT-2A isoform (16), which displays a >90% homology with its murine counterparts (17), we identified three negatively charged residues located in the extracellular-facing portion of transmembrane segments (TMs) 4, 9, and 10. Two of these residues, Asp¹⁶⁶ (TM4) and Asp⁴¹¹ (TM10) are conserved in all four CAT isoforms (Fig. 6). Asp⁴⁰⁵ (TM9), on the other hand, is replaced by a positively charged lysine in CAT-3 and, thus, this residue is unlikely to participate in stabilizing cationic amino acid binding unless the transport mechanism is substantially different in this isoform. Therefore, Asp¹⁶⁶ and Asp⁴¹¹ are potential candidates to supply the negative charge at the cationic amino acid binding pocket. Nonetheless, small shifts in TM assignments may bring into consideration additional amino acids, such as the highly conserved Glu¹⁶³ (Fig. 6) located outside the membrane electric field in this model (16).

Figure 6.

Amino acid sequence comparison of human CAT-1, -2A, -2B, and -3. Sequences of putative transmembrane segments (TMs) 4, 9, and 10 (according to Habermeier et al. (16)) were aligned using BLAST. Numbers above the sequences refer to the positions of four negatively-charged residues within or near the TMs. Gray boxes indicate highly conserved amino acid residues located in extracellular-facing portions of TMs 4 and 10. The white box indicates a residue that is not conserved in CAT-3. Amino acid residue 163 is not located within the membrane electric field with this arrangement of TMs (16).

The dielectric coefficients in Eq. 6 are bound by the constraint

$$\beta +\gamma +\eta =\nu -\alpha ,$$ (7)

where ν is the number of elementary charges moved per cycle from one side to the opposite side of the membrane and α is the dielectric coefficient for extracellular Arg binding (k₁) and release (k₋₁). To estimate η, we used the notion that the electrogenicity of I_max must reflect that of the V_M-dependent rate-limiting step and, thus, this dielectric coefficient (which affects k₄) was assigned the value 0.41 (λ_turnover). As discussed above, this implies that translocation of the L-Arg-bound complex is electroneutral, i.e., β = 0. In addition, the value found for λ_L-Arg (0.14) was chosen as a reasonable approximation for α, even though the expression for K⁰_0.5 in Eq. 1 is an algebraic combination of all rate constants that describe Scheme 1. In terms of the actual value of ν, it is important to notice that these organic compounds do not have fixed charges. Instead, the net charge of cationic amino acids is set by the pH of the bathing solution. Multiple equilibria analysis of L-Arg intermediates with different degrees of dissociation indicates that the form ArgH₂⁺ accounts for practically 98% of total Arg at pH 7.4, the remaining 2% corresponding to the uncharged ArgH (see Supporting Material). Thus, using ν = 0.98, Eq. 7 yields γ = 0.43. This result predicts that L-Arg will sense almost half of the membrane electric field during intracellular release. However, since this reaction is anticipated to be fast (probably diffusion-limited), its electrogenicity will not be easily detected.

The value of k₄, which likely describes the charge-moving rate-limiting step, will become larger than k₂ at sufficiently negative V_M values (Eq. 6). In such a case, it is anticipated that the I_M-V_M relationship will display a plateau associated with the electroneutral T_oL ↔ T_iL transition. However, it is not known whether these voltages fall within the experimentally achievable range. In the case of the high-affinity CAT-1 heterologously expressed in Xenopus oocytes, currents do not saturate at V_M as negative as −180 mV (18).

Transient charge movements

Transient currents were measured in the presence of extracellular D-Arg, indicating that this compound binds to the carrier at extracellular-facing sites located within the membrane electric field. This finding was consistent with, and provided further support for, steady-state measurements of D-Arg as a V_M-dependent competitive blocker of L-Arg transport. Boltzmann analysis of the quantity of mobile charge as a function of V_M showed that a total (Q_tot) of ∼10 fC pF⁻¹ were moved by D-Arg carrying roughly one net positive charge (z_q = 0.75 ± 0.20). Knowing that Q_tot = zF[T]_T, a value of ∼600 μm⁻² was estimated for the density of low-affinity cationic amino acid transporters on the rat cardiomyocyte plasma membrane.

The shift in midpoint potential produced by doubling the concentration of extracellular D-Arg was consistent with binding of this amino acid in an ion well. In fact, the ion well depth (δ) can be calculated from this shift in V_1/2 as follows (for a derivation of Eq. 8, see Peluffo (19)):

$$\Delta V_{1/2}=\frac{\text{RT}}{\delta F}\ln \frac{{\left[\text{D}-\text{Arg}\right]}_2}{{\left[\text{D}-\text{Arg}\right]}_1}.$$ (8)

Solution of Eq. 8 at 293.15 K yields δ = 0.36, a value indicative of the small portion of the membrane electric field sensed by mobile charges during binding reactions.

The common assumption of an apparent molecularity of n = 1 for the charge-moving process, knowing that the valence of the charged species has an upper-limit value of ν = 1 with our experimental conditions (see above), leads to an inconsistency between the value of δ and the measured apparent valence, defined as z_q = nνδ (20). Indeed, the actual stoichiometry for these transporters has not been determined. The assumption that one L-amino acid molecule is bound and transported per cycle is largely based on the hyperbolic dependence of uptake (4,7) and current ((1) and Fig. 2 of this work) curves with the concentration of substrate. However, a similar behavior is expected from binding of two (or more) substrate molecules to identical, noninteracting sites in the carrier. Therefore, a simple way to reconcile the values of δ and z_q is to assume that n = 2. In the case of L-Arg, transport of two molecules of substrate per cycle may also help to explain in part the high capacity of this low-affinity transporter.

Kinetic analysis showed that charge movements are rate-limited by a conformational change, which follows fast D-Arg binding reactions and is likely associated with TM rearrangements that prepare L-Arg translocation. The apparent rate constant for current relaxation was found to be a mild U-shaped function of V_M, with the V_M dependence residing largely in the reverse (release) reaction. This asymmetry, together with the small portion of the membrane electric field dissipated by D- and L-Arg binding reactions and the value found for δ, suggests that the transporter might have a shallow, water-filled extracellular binding pocket.

Although pure binding reactions were not measured, an apparent equilibrium constant for the dissociation of D-Arg, K_d = 2.8 ± 0.7 mM, was estimated as k_r⁰/k_f⁰. This K_d value is not expected to match that of K⁰_i because of the different temperatures used for transient charge movement and steady-state current measurements, and the fact that K⁰_i is also affected by the rate of reorientation of the unoccupied transporter.

In summary, the combination of pre-steady-state and steady-state measurements using the D-enantiomer of arginine yielded new quantitative information on the V_M-dependent mechanism of L-Arg transport, as well as some testable predictions on the structure-function of a low-affinity cationic amino acid transporter.