Electrophysiological characterization of a recombinant human Na⁺-coupled nucleoside transporter (hCNT1) produced in Xenopus oocytes

Methods

Heterologous expression in Xenopus oocytes

hCNT1 cDNA in pGEM-T (Ritzel et al. 1997; Loewen et al. 1999) or pGEM-HE (Ritzel et al. 2001) was linearized, respectively, with Not1 or Nhe1 and transcribed with T3 or T7 polymerase using the mMESSAGE mMACHINE^TM. (Ambion, Austin, TX, USA) transcription system. Stage V–VI oocytes were isolated by collagenase treatment of ovarian lobes from female Xenopus laevis (Biological Sciences Vivarium, University of Alberta) that had been anaesthetized by immersion in 0.2% tricaine methanesulphonate (pH 7.4; Sigma, Oakville, ON, Canada). Frogs were humanely killed following final collection of oocytes in compliance with guidelines approved by the Canadian Council on Animal Care. Defolliculated oocytes were injected with 10 nl of water ± 10 ng of capped hCNT1 RNA transcript and incubated for 4 days at 18°C in modified Barth's solution (changed daily). The enhanced Xenopus expression vector pGEM-HE (Liman et al. 1992) produced greater hCNT1 functional activity than pGEM-T and was used in most experiments.

Steady-state electrophysiological studies

Oocyte membrane currents were measured using a GeneClamp 500B oocyte clamp (Axon Instruments, Inc., Foster City, CA, USA) in the two-electrode, voltage-clamp mode that was interfaced to an IBM compatible computer via a Digidata 1200A/D converter and controlled by pCLAMP software (Version 9.0, Axon Instruments, Inc.). Current signals were filtered at 20 Hz (four-pole Bessel filter) and sampled at a sampling interval of 20 ms. For data presentation, the signals were further filtered at 0.5 Hz by the pCLAMP program suite. Microelectrodes were filled with 3 m KCl and had resistances in the range 0.5–2.5 MΩ. All experiments were performed at room temperature (20°C) and oocytes were discarded if the membrane potential was unstable or more positive than −30 mV. Unless otherwise indicated, the membrane potential was clamped at a holding potential (V_h) of −50 mV and nucleoside was added at a concentration of 100 μm. The transport medium contained (mm): NaCl, 100; KCl, 2; CaCl₂, 1; MgCl₂, 1; Hepes, 10 (pH 7.5). In some experiments, Na⁺ was replaced by equimolar choline.

Current–voltage (I–V) curves were determined from differences between steady-state currents generated in the presence and absence of permeant during 175 ms voltage pulses to potentials between −90 and +60 mV (10 mV increments). For I–V relations, the voltage rise time of the clamp was adjusted by use of an oscilloscope such that it varied between 200 and 500 μs. Currents were filtered at 2 kHz (four-pole Bessel filter) and sampled at a rate of 200 μs point⁻¹ (corresponding to a sampling frequency of 5 kHz). The ability of H⁺ to drive nucleoside transport was tested by replacing Na⁺ in the transport medium with choline and varying the pH between 7.5 and 5.5 (10 mm MES (2-[N-morpholino]ethanesulphonic acid) was used in place of Hepes in solutions with pH values ≤ 6.5). Exposure of oocytes to acid pH was kept to intervals of < 2 min to minimize toxicity. For studies of phloridzin inhibition, currents were measured in the same oocyte before and after a 10 min preincubation with inhibitor (the time required for onset of maximum inhibition). Phloridzin remained present during uridine perfusion.

Current values are presented as means ±s.e.m. of 4 or more oocytes. Each experiment was repeated at least twice on oocytes from different frogs. Uridine kinetic parameters (apparent affinity, K_m^uridine; maximal current, I^uridine_max) were determined by current measurements at different uridine concentrations and analysed by least squares fits to the Michaelis-Menten eqn I =I^uridine_max[U]/(K_m^uridine+[U]), where I is the permeant-induced current and [U] represents the uridine concentration (SigmaPlot Version 4, Jandel Scientific Software, San Rafael, CA, USA). Kinetic parameters for other permeants were determined in similar fashion. The results from Na⁺ activation experiments were fitted to the Hill equation, I=I_max[Na⁺]ⁿ/(K_m^Nan+[Na⁺]ⁿ), where n is the Hill coefficient, K^Na_m is the half-saturation constant for Na⁺ activation, I is the uridine-induced steady-state current, and I_max is the predicted current maximum (SigmaPlot Version 4).

Radioisotope flux studies (phloridzin and β-DFP-5M inhibition)

Initial rates of hCNT1-mediated transport of 10 μm¹⁴C-labelled uridine (1 μCi ml⁻¹, Amersham Pharmacia Biotech, Canada) were measured at room temperature (1 min flux) as previously described (Huang et al. 1993; Ritzel et al. 1997). Values are presented as means ±s.e.m. of 10–12 oocytes, and each experiment was repeated at least twice on different batches of cells.

Cation–nucleoside coupling ratios

hCNT1 Na⁺–nucleoside stoichiometry was determined by the simultaneous measurement of Na⁺ current and [¹⁴C]uridine influx under voltage-clamp conditions in the same oocyte (Chen et al. 1998; Loewen et al. 2003). Coupling ratios (±s.e.) were calculated from slopes of least-squares fits of uridine-dependent charge versus uridine accumulation for seven or more oocytes.

Presteady-state currents

Presteady-state currents were measured using a voltage step protocol. The membrane voltage was stepped using 250 ms voltage pulses from the holding potential (V_h) of −50 mV to a series of test potentials (V_t) ranging from −170 to +130 mV (20 mV increments). In experiments to determine the turnover rate of the transporter, membrane voltage was stepped from V_h of −50 mV to V_t from −170 to +150 mV in 40 mV increments to ensure maximal charge displacement while reducing the number of voltage pulses to which the oocyte was subjected. The maximal steady-state inward Na⁺ current (I_max) was measured at V_h of −50 mV with a saturating concentration of uridine (100 μm). Currents were filtered and sampled as described for I–V relationships. For data presentation, the current at each test potential was averaged from five sweeps. If necessary, signals were further filtered at 750 Hz (pCLAMP 9.0). Presteady-state currents due to hCNT1 were fitted using the Chebyshev method with two exponential functions (pCLAMP 9.0). Since the capacitive transients were longer than 1–2 ms, amplitudes were extrapolated to 1 ms after the onset of the step. Current–time integrals were calculated using these extrapolated amplitude values. Curve fits were considered successful only if the correlation coefficient (r) was 0.95 or higher. Charge movements (Q) obtained from the current-time integral of the curve fits were plotted against voltage and fitted to the Boltzmann function: where the total charge Q_T=Q_dep–Q_hyp (Q_dep and Q_hyp representing Q at depolarizing and hyperpolarizing limits, respectively), z_d is the product of the valency of the charge (z) and the apparent fraction of the field (δ) sensed by that charge, V_t is the membrane voltage during the pulse, V_0.5 is the membrane voltage at which half of the total charge has moved, F is Faraday's constant, R is the gas constant and T is the absolute temperature (Hazama et al. 1997). Mean values of V_0.5 and z_d (±s.e.m.) were determined from individual Boltzmann fits to data from three to six separate experiments in different oocytes.

Chemicals

Nucleosides, nucleoside analogues and phloridzin were purchased from Sigma (Oakville, ON, Canada). β-DFP-5M 1-(2-deoxy-β-d-ribofuranosyl-2,4-difluoro-5-methylbenzene) and β-DFP-5I (1-(2-deoxy-β-d-ribofuranosyl)-2,4-difluoro-5-iodobenzene) were synthesized as previously described (Wang et al. 2001).

Results

General characteristics

Measured in Xenopus oocytes at V_h=−50 mV and using choline as Na⁺ substitute, transport of uridine by recombinant hCNT1 was electrogenic, Na⁺ dependent and H⁺ independent (shown in Fig. S1 in Supplementary material, available online). In contrast to reports for h/rCNT1 by other investigators (Dresser et al. 2000; Lostao et al. 2000), addition of permeant to hCNT1-producing oocytes in the absence of Na⁺ did not generate any detectable inward current (Fig. S1, Supplementary material). This agrees with previous radiotracer uptake studies that found only very small amounts of nucleoside uptake in the absence of Na⁺ (Huang et al. 1994; Ritzel et al. 1997). This minor component of  Na⁺-independent transport had the characteristics of ‘slippage’ (i.e. uncoupled nucleoside transport) and would not be expected to be electrogenic. hCNT1 steady-state currents were voltage dependent and increased at more negative potentials (Fig. S2, Supplementary material). Currents approached zero, but did not reverse polarity at potentials up to +60 mV. No steady-state currents were observed in control water-injected oocytes.

Transport of physiological nucleosides

hCNT1 selectivity for pyrimidine nucleosides has been demonstrated previously using conventional radioisotope flux measurements (Ritzel et al. 1997). It has also been found that hCNT1 and rCNT1 mediate low, but significant radioisotope fluxes of adenosine, but not of inosine or guanosine (Huang et al. 1994; Fang et al. 1996; Yao et al. 1996b; Ritzel et al. 1997; Loewen et al. 1999). These results were confirmed and extended in Fig. 1 using electrophysiological techniques. hCNT1 currents elicited by application of test permeants in Na⁺-containing transport medium were: uridine, thymidine, cytidine (100 μm) > adenosine (100 μm and 1 mm); guanosine and inosine (100 μm and 1 mm) were without effect. The nucleobases of uridine (uracil) and inosine (hypoxanthine) (100 μm and 1 mm) were also not transported. No currents were observed in control water-injected oocytes (data not shown). Therefore, hCNT1 is specific for pyrimidine nucleosides and adenosine. In agreement with radiotracer uptake measurements (Ritzel et al. 1997), adenosine elicited larger currents than 2′-deoxyadenosine (Fig. 2B).

In radioisotope flux studies, adenosine is transported by rCNT1 with a similar apparent K_m to uridine (∼30 μm), but with a much lower V_max due to slow conversion of the CNT1–adenosine complex from outward-facing to inward-facing conformations (Yao et al. 1996b). Competition experiments were undertaken with hCNT1 to determine if the same kinetic behaviour could be demonstrated electrophysiologically. As shown in Fig. S3 (Supplementary material) the current produced by a saturating concentration of uridine was substantially higher than that produced in the same oocyte by simultaneous perfusion of both uridine and adenosine.

Transport of nucleoside analogues

We also used electrophysiology to determine transportability of 100 μm and 1 mm concentrations of a panel of clinically important antiviral and anticancer nucleoside drugs and other nucleoside analogues (Fig. 2). Large to moderate inward currents were elicited by application of 2′-deoxyuridine, 2′,3′-dideoxyuridine, 5-fluorouridine, 5-fluoro-2′-deoxyuridine, zalcitabine (ddC, 2′,3′-dideoxycytidine) and zidovudine (AZT, 3′-azido-3′-deoxythymidine) (Fig. 2A). Smaller, but significant inward currents were also observed for cladribine (2-chloro-2′-deoxyadenosine) (Fig. 2B). Cytarabine (1-β-D-arabinofuranosylcytosine) and tubercidin (7-deazaadenosine) generated small inward currents only at the higher permeant concentration of 1 mm, while didanosine (ddI, 2′, 3′-dideoxyinosine) was without effect (Fig. 2B). As illustrated for zidovudine in Fig. S4 (Supplementary material), currents were reversible and abolished in Na⁺-free medium. No currents were observed in control water-injected oocytes.

Transport of nucleoside mimics

The novel thymidine mimetics β-DFP-5M and β-DFP-5I (Fig. 3A and B, respectively), in which the pyrimidine base was replaced by a substituted aromatic ring, were similarly tested as potential hCNT1 permeants. Both compounds induced reversible, Na⁺-dependent inward currents in hCNT1-producing oocytes, but not in control water-injected oocytes (Fig. 3C and D). β-DFP-5M inhibited hCNT1-mediated ¹⁴C-labelled uridine influx with an IC₅₀ value (±s.e.) of 0.56 ± 0.06 mm (r= 0.99) (Fig. S5, Supplementary material).

Na⁺ and uridine steady-state kinetics: order of binding

When the dependence of hCNT1-mediated Na⁺ currents on uridine concentration (0–1000 μm) was examined at three different extracellular Na⁺ concentrations (5, 25 and 100 mm), saturable inward current responses that were consistent with simple Michaelis-Menten kinetics were observed (Fig. 4A). At 5, 25 and 100 mm external Na⁺, the apparent affinity for uridine increased as [Na⁺]_out increased, with no significant change in the maximal current, yielding apparent K^uridine_m values (±s.e.) of 139 ± 10, 80 ± 7 and 32 ± 5 μm, respectively, with I^uridine_max values (±s.e.) of 54 ± 1, 55 ± 2 and 54 ± 2 nA, respectively. The corresponding dependence of hCNT1-mediated Na⁺ currents on the external concentration of Na⁺ (0–100 mm) was examined at two different concentrations of extracellular uridine (25 and 100 μm) (Fig. 4B). The Na⁺ concentration dependence of the steady-state transport current also conformed to simple Michaelis-Menten kinetics. Both the apparent affinity of the transporter for Na⁺ and the maximal current increased when the external concentration of uridine increased. At 25 and 100 μm uridine, apparent K_m^Na values (±s.e.) were 12 ± 2 and 3 ± 1 mm, respectively, with I_max^Na values (±s.e.) of 38 ± 2 and 64 ± 3 nA, respectively.

Together, the data in Fig. 4A and B indicate a sequential ordered binding mechanism in which Na⁺ binds to the transporter first, increasing its affinity for the nucleoside, which then binds second (Jauch & Lauger, 1986; Stein, 1990; Klamo et al. 1996; Mackenzie et al. 1996b). Transport of nucleoside and ion is simultaneous because decreasing the concentration of either Na⁺ or uridine decreased the apparent affinity of the other (Eskandari et al. 1997). A sequential ordered binding mechanism is consistent with studies of native cit and cif transport activity in bovine renal brush-border membrane vesicles showing that the apparent affinity for nucleoside increased as the external Na⁺ concentration was raised (Williams & Jarvis, 1991). The predicted hCNT1 Na⁺–nucleoside coupling ratio was 1: 1, since fitting the 25 and 100 μm uridine current data of Fig. 4B to the Hill equation yielded Hill coefficients (±s.e.) of 0.90 ± 0.12 and 0.79 ± 0.06, respectively.

Na⁺ and uridine steady-state kinetics: voltage dependence

We also used steady-state kinetics to investigate the mechanism behind hCNT1 voltage dependence. The apparent affinities for Na⁺ (K_m^Na) and uridine (K^uridine_m) and corresponding I_max values were measured at four different holding potentials (V_h=−10, −30, −50 and −70 mV) (curves not shown). K_m^Na was determined at an external uridine concentration of 100 μm, while K^uridine_m was determined at both 10 and 100 mm external Na⁺. K^uridine_m was unaffected by membrane potential at 100 mm Na⁺, but was voltage dependent at 10 mm Na⁺, decreasing from 84 to 44 μm as the membrane potential was increased from −10 to −70 mV. At high negative membrane potentials therefore K^uridine_m (10 mm Na⁺) approached that observed at 100 mm external Na⁺, indicating that the voltage dependence of K^uridine_m is the result of voltage dependence of Na⁺ binding (Birnir et al. 1991; Parent et al. 1992a). Consistent with this conclusion, we found that K_m^Na was voltage sensitive, decreasing from 5 to 1 mm as the membrane potential was varied from −10 to −70 mV. I^uridine_max and I_max^Na also showed voltage dependence, their magnitudes increasing as the membrane potential was made more negative. Membrane potential therefore influences both ion-binding and carrier translocation (Birnir et al. 1991; Parent et al. 1992a).

Na⁺ and uridine steady-state kinetics: phloridzin inhibition

Figure S6 (Supplementary material) demonstrates phloridzin inhibition of uridine currents in hCNT1-producing oocytes. Inhibition was partial (∼80%), even at high phloridzin concentrations, and the IC₅₀ value for inhibition of the phloridzin-sensitive component of current was 0.21 ± 0.05 mm (r= 0.98). A similar inhibition profile was obtained for ¹⁴C-labelled uridine influx (IC₅₀ of 0.35 ± 0.12 mm; r= 0.99) (Fig. S6, Supplementary material). In kinetic experiments, phloridzin (5 mm) reduced both I^uridine_max and I_max^Na, but had opposite effects on K^uridine_m and K_m^Na (Fig. 5). K^uridine_m and I^uridine_max values (±s.e.) (100 mm NaCl) were 22 ± 3 μm and 151 ± 5 nA, respectively, in the absence of phloridzin, and 131 ± 27 μm and 73 ± 9 nA, respectfully, in the presence of phloridzin (Fig. 5A). Corresponding K_m^Na and I_max^Na values (±s.e.) (100 μm uridine) were 3.0 ± 0.3 mm and 95 ± 2 nA, respectively, in the absence of phloridzin, and 0.8 ± 0.2 mm and 22 ± 1 nA, respectively, in the presence of phloridzin (Fig. 5B). The Hill coefficient (±s.e.) for the control data in Fig. 5B was 1.1 ± 0.1.

Nucleoside analogue steady-state kinetics

Nucleoside analogues from Fig. 2A exhibiting robust steady-state currents were analysed kinetically as shown in Fig. S7 (Supplementary material). Apparent K_m and I_max (100 mm NaCl) values derived from these data are compared to uridine in Table 1. Relative affinities were in the order 5-fluoro-2′-deoxyuridine, 5-fluorouridine > uridine, 2′-deoxyuridine ≫ zidovudine, with calculated I_max: K_m ratios (a measure of transport efficiency) highest for uridine and 2′-deoxyuridine. The hCNT1 zidovudine apparent K_m of 0.45 mm is in good agreement with values determined for rCNT1 transport of zidovudine and zalcitabine by radioisotope flux studies (Yao et al. 1996a).

Cation–nucleoside coupling ratio

The stoichiometry of Na⁺–uridine cotransport was determined in individual hCNT1-producing oocytes by simultaneously measuring uridine-evoked hCNT1 current and [¹⁴C]uridine uptake under voltage-clamp conditions (Fig. 6). Figure 6A is a representative uridine-dependent current recording (200 μm[¹⁴C]uridine, 100 mm NaCl) in an hCNT1-producing oocyte clamped at −50 mV. Current reached an initial maximal value and then progressively decreased, a phenomenon that has also been observed for other cotransporters and is thought to result from (i) decreased ion concentrations in the immediate vicinity of the extracellular membrane, and (ii) trans-inhibition of transport activity by the accumulation of intracellular permeants and/or ions (Chen et al. 1998; Mackenzie et al. 1998; Chen et al. 1999). Results for groups of seven to nine different oocytes at holding potentials of −30, −50 and −90 mV yielded linear plots of charge (pmol) versus uptake (pmol), the slopes of which were independent of voltage and equal to the Na⁺–nucleoside coupling ratio (Fig. 6B–D). At V_h=−30 mV, the linear correlation between uridine-dependent charge and uridine accumulation gave a stoichiometry (±s.e.) of 0.92 ± 0.15 (Fig. 6B), compared to 0.89 ± 0.02 at −50 mV (Fig. 6C) and 0.90 ± 0.09 at −90 mV (Fig. 6D).

Presteady-state currents of hCNT1

Unless otherwise specified, presteady-state experiments were performed in the absence of nucleoside to eliminate steady-state inward currents of hCNT1 and to isolate partial reactions of the transport cycle. Oocytes were voltage clamped at a holding potential (V_h) of −50 mV, and presteady-state currents were activated by voltage steps to a series of test potentials (V_t). Figure S8 (Supplementary material) shows representative total current recordings in an hCNT1-producing oocyte bathed in 100 mm Na⁺-containing transport medium. Current relaxations, which persisted for tens of milliseconds after the time required to charge the membrane capacitance, were apparent in both the ON response, when V_h was stepped to V_t, and in the OFF response, when V_t was returned to V_h. These relaxations were also observed in hCNT1-producing oocytes in the absence of external Na⁺, but were not seen in control water-injected oocytes (Fig. S8, Supplementary material). In the presence of external Na⁺, the charge movement at the onset of the voltage pulse (Q_ON) was found to be equal and opposite to that at the return to the prepulse potential (Q_OFF), demonstrating conservation of charge during ON and OFF voltage steps (Fig. S9, Supplementary material). Figure 7A shows Q_OFF, normalized to Q_T, in a representative hCNT1-producing oocyte plotted as a function of voltage (25 mm NaCl). The Q–V relation obeyed a Boltzmann function, reversing at V_h and approaching saturation with both hyperpolarization and depolarization. The experiment was repeated in five different oocytes and at three additional Na⁺ concentrations (10, 50 and 100 mm NaCl). Mean values of z_d (±s.e.m.) from individual Boltzmann fits were unaffected by Na⁺ concentration (−0.47 ± 0.04, −0.51 ± 0.03, −0.50 ± 0.04 and −0.50 ± 0.02 at 10, 25, 50 and 100 mm external NaCl, respectively), and similar to those that can be calculated from the data of Larráyoz et al. (2004), while estimates of V_0.5, plotted versus the log of Na⁺ concentration, shifted towards more negative potentials as the concentration of Na⁺ was reduced (Fig. 7B). The fitted line corresponded to a shift (±s.e.) of 41 ± 1 mV for an e-fold change in Na⁺ concentration, and was converted to the effective fraction of the electric field (δ) sensed by Na⁺ using the relationship δ=kT/(e_o× 41 mV), where k is the Boltzmann constant and e_o is the elementary charge (Mager et al. 1993). The value of δ was 61 ± 1%, and implies binding of sodium to site(s) that traverse ∼61% of the membrane electric field. The valency of the moveable charge (z), calculated from the relationship z_d=δz, was −0.81 ± 0.03, consistent with the transporter having one net negative charge. The effects of uridine (0–100 μm) on hCNT1 presteady-state currents and on Q_T (calculated for the OFF response) were also examined (Fig. S10, Supplementary material). External uridine increased the hCNT1 steady-state uridine-induced current (100 mm NaCl), but reduced presteady-state currents and Q_T. At 25 μm uridine, a concentration close to the uridine apparent K^uridine_m, Q_T was decreased by ∼50%. Adenosine also has the ability to inhibit hCNT1 presteady-state currents (Larráyoz et al. 2004).

We also used the Boltzmann parameters to estimate the turnover rate (also known as turnover number) of hCNT1 and the number of transporter molecules present in the oocyte plasma membrane (Loo et al. 1993; Panayotova-Heiermann et al. 1995; Wadiche et al. 1995). Linear regression analysis of the steady-state transport current (100 mm NaCl) at a saturating concentration of uridine (100 μm) at −50 mV (I_max) versusQ_T (calculated for the OFF response) in oocytes with differing levels of hCNT1 expression yielded a straight line with a slope (±s.e.) of 17.2 ± 4.4 s⁻¹, corresponding to a charge transfer rate (Φ) of 8.6 ± 2.2 s⁻¹ (slope ×z_d= 17.2 × 0.50) (Fig. 8) (Wadiche et al. 1995). Since the turnover rate of the transporter is given by Φ/v (Wadiche et al. 1995), where v is the number of fundamental charges translocated per molecule of uridine and equals the Na⁺–uridine coupling ratio (−50 mV) of 0.89 ± 0.02 (Fig. 6C), the number of uridine molecules transported per hCNT1 protein per second was 9.6 ± 2.5 s⁻¹. The numbers of recombinant hCNT1 transporters expressed in the oocyte plasma membrane (N), determined from Fig. 8 and the equation Q_T=Ne_oz_d (Wright et al. 1994; Klamo et al. 1996; Eskandari et al. 1997), were in the range (5.4–8.5) × 10¹⁰ per oocyte, with a mean value (±s.e.m.) of (6.8 ± 0.2) × 10¹⁰.

Discussion

Na⁺-dependent hCNT1 (Ritzel et al. 1997), the prototypic human member of the CNT family of nucleoside transport proteins, is responsible for the concentrative cellular uptake of both physiological nucleosides and clinically important anticancer and antiviral nucleoside drugs. In immunolocalization studies, the rat orthologue of hCNT1 (rCNT1) is expressed predominantly in the brush-border membranes of the polarized epithelial cells of jejunum and renal cortical tubules, and in the bile canalicular membranes of liver parenchymal cells (Hamilton et al. 2001). In the present study, we have used the two-electrode voltage clamp in combination with heterologous expression in Xenopus oocytes to undertake steady-state and presteady-state electrophysiological studies of recombinant hCNT1.

Transport of nucleosides by hCNT1 was electrogenic and specific for pyrimidine nucleosides and adenosine. The latter nucleoside functions as a high-affinity low-capacity permeant, allowing it to act, in appropriate circumstances, as an hCNT1 inhibitor. Inosine, guanosine and nucleobases were not transported, even at high concentrations. Together with previous radioisotope flux studies and our parallel electrophysiological studies of rCNT1 (data not shown), the present findings contradict reports that adenosine is not transported by either hCNT1 or rCNT1 (Dresser et al. 2000; Larráyoz et al. 2004). Consistent with a physiological role of hCNT1 in renal handling of nucleosides, larger currents were elicited by adenosine (which is reabsorbed) than with 2′-deoxyadenosine (which is excreted).

Radioisotope flux studies have provided evidence that hCNT1 and rCNT1 also transport various nucleoside analogues, including clinically important nucleoside drugs with antineoplastic and/or antiviral activities (Huang et al. 1994; Fang et al. 1996; Yao et al. 1996a,b; Ritzel et al. 1997; Mackey et al. 1998; Yao et al. 2001). In the present study, inward currents were observed with the anticancer drugs 5-fluorouridine and 5-fluoro-2′-deoxyuridine, and with the antiviral drugs zidovudine and zalcitabine. Both fluorinated compounds were well tolerated (K_m values ∼15 μm). Unlike adenosine and 2′-deoxyadenosine, lack of the C(2′)-OH in 5-fluoro-2′-deoxyuridine compared to 5-fluorouridine had no discernable effect on transport, a finding confirmed by kinetic comparisons between the two parent compounds 2′-deoxyuridine and uridine. Similarly, gemcitabine, an anticancer analogue of 2′-deoxycytidine, is also a good hCNT1 permeant (apparent K_m∼25 μm) (Mackey et al. 1999). In contrast, absence of the C(3′)-OH in zidovudine (and zalcitabine) resulted in a > 10-fold decrease in transportability. The zidovudine K_m of 0.45 mm exceeds therapeutic levels of zidovudine in plasma, but is consistent with a role of hCNT1 in intestinal absorption of the drug during oral administration. Smaller inward currents were observed with the anticancer nucleoside drugs cladribine (an analogue of adenosine) and cytarabine (an analogue of cytidine). One millimolar cytarabine was required to produce a detectable inward current, suggesting that it functions as a low-affinity hCNT1 permeant, a conclusion supported by radiotracer flux studies in transfected mammalian cells, where 0.5 mm cytarabine caused only partial inhibition of uridine transport activity (Graham et al. 2000).

Two novel pyrimidine nucleoside mimics (β-DFP-5M and β-DFP-5I) (Wang et al. 2001) also functioned as low-affinity hCNT1 permeants. These compounds demonstrate that the pyrimidine ring is not required for translocation by hCNT1. The ability of the aromatic ring of β-DFP-5M and β-DFP-5I to functionally substitute for the pyrimidine moiety of nucleosides indicates that π–π interactions corresponding to those documented for trypanosomal ENT proteins (de Koning & Jarvis, 1999) may also be important in CNT–permeant interactions. While inhibitor-sensitivity assays have revealed potential hydrogen bonds formed between hCNT1 and uridine C(3′)-OH, C(5′)-OH and N(3)-H (Zhang et al. 2003), the present results showing inward currents for 2′,3′-dideoxyuridine, zidovudine, zalcitabine, β-DFP-5M and β-DFP-5I demonstrate that C(3′)-OH and N(3)–H interactions are not obligatory for transport.

Na⁺-dependent cotransporters are found mostly in animal cells, whereas H⁺-dependent cotransporters are widely distributed in plants, bacteria and animals. A number of cotransporters utilize more than one cation. For example, the Na⁺–glucose cotransporters SGLT1 and SGLT2 are able to couple sugar transport to the electrochemical gradients of Na⁺, Li⁺ or H⁺ (Hirayama et al. 1994; Mackenzie et al. 1996b), and the bacterial melibiose transporter utilizes Na⁺ or H⁺ to drive melibiose transport (Tsuchiya & Wilson, 1978; Bassilana et al. 1987). In contrast, hCNT1 did not demonstrate detectable nucleoside transport when Na⁺ was replaced with H⁺, a behaviour that is different from hCNT3 and mCNT3 which are able to use the electrochemcial gradient of either Na⁺ or H⁺ to accumulate nucleosides within cells (KM Smith, SK Loewen, E Karpinski & JD Young, unpublished observations). CNTs from C. albicans (CaCNT), C. elegans (CeCNT3) and E. coli (NupC) function exclusively as H⁺-dependent nucleoside cotransporters. Some protozoan and plant ENT family members differ from their mammalian counterparts and are also H⁺-coupled (Mohlmann et al. 2001; Stein et al. 2003). Analysis of hCNT1 steady-state kinetics revealed that Na⁺ binds to the transporter first, followed by nucleoside.

Na⁺–nucleoside coupling ratios for members of the CNT family have previously been determined indirectly from Hill-type analyses of relationships between nucleoside fluxes and Na⁺ concentration. For example, Na⁺–nucleoside coupling ratios of 1: 1 have been proposed for recombinant rCNT1 transport of both adenosine and uridine (Yao et al. 1996b), and similar ratios have been found in studies of Na⁺-dependent cit and cif nucleoside transport in renal brush-border membrane vesicles (Lee et al. 1988; Williams & Jarvis, 1991). Since Hill analysis of Na⁺ activation curves does not determine the number of Na⁺ ions that actually enter the cell as a result of transport activity (Weiss, 1997), we utilized simultaneous measurement of hCNT1-specific currents and radioactive nucleoside uptake from individual oocytes under voltage-clamp conditions to determine this parameter directly. When charge was converted to picomoles, the ratio of charge to nucleoside uptake for hCNT1 yielded a stoichiometry of 1: 1 that was independent of membrane potential. Therefore, both direct and indirect methods agree on a Na⁺–nucleoside coupling ratio of 1: 1. These results differ from those of Larráyoz et al. (2004) who incorrectly reported a Na⁺–nucleoside stoichiometry of 2: 1. Examination of Fig. 6A of that paper reveals an apparent scaling error. A 1: 1 Na⁺–nucleoside stoichiometry for hCNT1 contrasts with parallel studies of hCNT3, where the coupling ratio approached 2: 1 as the membrane was hyperpolarized (KM Smith, SK Loewen, E Karpinski & JD Young, unpublished observations). In this respect, CNTs resemble some other transporter families. For example, SGLT transporters have Na⁺–glucose coupling ratios of either 1: 1 (SGLT2) or 2: 1 (SGLT1/3) (Chen et al. 1995; Mackenzie et al. 1996b, 1998; Diez-Sampedro et al. 2001). The PEPT1 and PEPT2 proton-linked peptide transporters also have different H⁺–peptide coupling ratios of 1: 1 and 2: 1, respectively (Chen et al. 1999). While the stoichiometry of hCNT1 was independent of membrane potential, transport of uridine increased at more negative potentials, a finding consistent with earlier experiments in isolated rat hepatocytes (Gomez-Angelats et al. 1996).

Phloridzin is a potent inhibitor of SGLT1-3 (Lee et al. 1994; Mackenzie et al. 1996b; Hirayama et al. 2001) that has also been shown to reduce intestinal and renal Na⁺-dependent nucleoside transport activity (Lee et al. 1988; Huang et al. 1993). In the present study, phloridzin functioned as a partial hCNT1 inhibitor with an IC₅₀ of 0.2 mm, a value similar to that observed in parallel studies of hCNT3 (0.3 mm) (KM Smith, AML Ng, SYM Yao, E Karpinski & JD Young, unpublished observation). Thus, phloridzin appears to be a general CNT inhibitor. Phloridzin effects on hCNT1 uridine and Na⁺ steady-state kinetics were consistent with mixed non-competitive inhibition and uncompetitive inhibition, respectively (Dixon & Webb 1958; Wong, 1975), suggesting that phloridzin binds to hCNT1 after Na⁺ at a site possibly overlapping with, but not identical to, that occupied by uridine. Similarly, phloridzin binding to SGLT1 is Na⁺-dependent (Vick et al. 1973; Lin & Hahn, 1983; Parent et al. 1992a,b) and competitive with glucose (Wielert-Badt et al. 2000; Hirayama et al. 2001; Novakova et al. 2001).

In the absence of nucleoside, and in response to step-wise changes in membrane potential, oocytes producing hCNT1 exhibited slow current relaxations (presteady-state currents) in the presence and absence of Na⁺ similar to those observed for several other Na⁺- or H⁺-coupled cotransporters produced in Xenopus oocytes (Parent et al. 1992a,b; Loo et al. 1993; Mager et al. 1993; Chen et al. 1996; Klamo et al. 1996; Mackenzie et al. 1996a,b; Eskandari et al. 1997; Hazama et al. 1997; Chen et al. 1999). hCNT1 current–time integrals obeyed a Boltzmann function and were used to provide quantitative estimates of the fraction of the membrane field sensed by Na⁺ (61%), the valency of the movable charge (−0.81), and the average number of transporters present in the oocyte plasma membrane (6.8 × 10¹⁰ per cell). The first of these parameters reflects the location of the Na⁺ binding site within the hCNT1 translocation cleft. A valency of −0.81 is consistent with the determined Na⁺–nucleoside coupling ratio of 1: 1, while the estimate of hCNT1 membrane abundance allows determination, for the first time, of the turnover rate (turnover number) of a member of the CNT protein family. The calculated hCNT1 turnover rate of 9.6 uridine molecules transported per hCNT1 protein per second at −50 mV is similar to that of other cotransporters such as GAT1 (Mager et al. 1993) and SGLT1 (Panayotova-Heiermann et al. 1994), but is much lower than the mammalian ENT uridine transporter turnover rate of 10⁴ s⁻¹ determined from NBMPR binding studies (Young & Jarvis, 1983; Cass, 1995). Table 1 lists turnover rates for other hCNT1 permeants characterized in the present study.

In conclusion, the present studies provide important new mechanistic insights into hCNT1 transport of both physiological nucleosides, including adenosine, and anticancer and antiviral nucleoside drugs. This information will guide the development of detailed kinetic models of CNT-mediated Na⁺–nucleoside cotransport, and provides a functional framework to interpret CNT mutagenesis studies. Turnover rates can be combined with immunohistochemical patterns of protein expression to predict in situ hCNT1 fluxes of nucleosides and nucleoside drugs in normal and clinical human samples.

Supplementary material

The online version of this paper can be accessed at:

DOI: 10.1113/jphysiol.2004.068189

http://jp.physoc.org/cgi/content/full/jphysiol.2004.068189/DC1 and contains 10 supplementary figures, Figs S1–S10, examining steady-state and presteady-state currents of hCNT1.

This material can also be found at:

http://www.blackwellpublishing.com/products/journals/suppmat/tjp/tjp374/tjp374sm.htm