Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Nucleoside transporters mediate the transport of nucleosides in many cell types and organs. Studies using both cell culture models and tissue and plasma membrane preparations have identified two classes of nucleoside transporters in mammalian cells: equilibrative and concentrative (Griffith and Jarvis, 1996). Equilibrative nucleoside transporters transport nucleosides down their concentration gradient, whereas concentrative nucleoside transporters couple uphill substrate transport to downhill sodium transport. The concentrative transporters are further categorized based on their interactions with endogenous purine and pyrimidine nucleosides; they are purine-, pyrimidine-, or broadly-selective. During the past decade, a number of nucleoside transporter genes have been isolated from mammalian, bacterial, and parasitic species (Wang et al., 1997a; Baldwin et al., 1999). In addition to their established role in the cellular uptake of endogenous nucleosides, these transporters play a role in the transport of nucleoside analogs as well (Pastor-Anglada et al., 1998). Numerous nucleoside analogs have been found to be clinically effective in the treatment of cancer and viral infections. For example, ribavirin and zidovudine exhibit potent antiviral activity (McCarthy et al., 1995; Sperling, 1998). Likewise, the purine nucleoside 2-chlorodeoxyadenosine has proven useful for treatment of indolent lymphoproliferative disorders, especially hairy-cell leukemia (Piro et al., 1990; Saven et al., 1998). Many of the nucleoside analogs currently used clinically can cause toxic responses in some patients; responses which are thought to arise from their lack of selectivity for infected or cancerous cells. Furthermore, many nucleoside analogs are poorly available after oral administration. Therefore, the design and synthesis of new nucleoside analogs remains a very active area of research, with the goal of discovering more selective drugs with optimal oral bioavailability.

Because many nucleoside analogs are hydrophilic and diffuse very slowly across cellular membranes, nucleoside transporters play a primary role in their cellular uptake and release. Understanding the substrate selectivity of these transporters is critical for the development of new therapeutic nucleoside analogs with optimal pharmacokinetic properties, including high oral bioavailability and selective distribution. Indeed, there are a growing number of examples of synthetic analogs that are targeted to plasma membrane transporters to increase their bioavailability and to enhance their cell-specific delivery (Veyhl et al., 1998; Tsuji, 1999). The assays currently used to investigate the substrate selectivity of nucleoside transporters either require radiolabeled compounds or rely on inhibition studies, which are indirect, because an inhibitor may not actually be transported.

In this study, we investigated the substrate selectivity of the concentrative, nucleoside transporter, rCNT1. rCNT1 was cloned from a rat jejunal cDNA library and was found to mediate the sodium-dependent uptake of uridine and thymidine in Xenopus laevis oocytes injected with rCNT1 cRNA (Huang et al., 1994). rCNT1 is expressed in the brain, intestine, liver, and kidney and may therefore play important roles in the distribution of therapeutic nucleoside analogs in these tissues (Huang et al., 1994; Anderson et al., 1996; Felipe et al., 1998). Understanding the structural features of nucleoside analogs required for transport by rCNT1 is important in the rational design of therapeutic agents with optimal pharmacokinetic properties. In this study, we investigated the interaction of 19 structurally diverse nucleosides and nucleoside analogs with rCNT1. In particular, we used electrophysiology methods (Birnir et al., 1991; Lostao et al., 1994) to directly determine the ability of rCNT1 to translocate these compounds. Structural features of nucleoside analogs important in substrate recognition and translocation by rCNT1 are discussed. This report demonstrates the electrogenic nature of nucleoside transport via rCNT1 and describes key positions on the nucleosides that affect the substrate selectivity of the transporter.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

In Vitro Transcription and Oocyte Preparation. rCNT1 cDNA was amplified by reverse transcription-polymerase chain reaction from rat intestine using gene-specific primers designed from the published sequence (Huang et al., 1994). The cDNA was then subcloned into the pOX vector (Jegla and Salkoff, 1997). Capped cRNA was transcribed in vitro with T3 polymerase (mCAP RNA Capping Kit; Stratagene) from linearized rCNT1-pOX plasmids. Oocytes were harvested from oocyte positive Xenopus laevis (Nasco, Fort Atkinson, WI) and were dissected and treated with collagenase D (Boehringer-Mannheim Biochemicals, Indianapolis, IN) in a calcium-free ORII solution (2 mg/ml), as previously described (Giacomini et al., 1994). Healthy stage V and VI oocytes were injected with 50 nl of capped rCNT1 cRNA (1 µg/µl). Oocytes were maintained at 18°C in modified Barth's medium.

Radiotracer Transport Assays. Transport of ³H-uridine (11 µM) (37.8 Ci/mmol; Moravek, Brea, CA) in oocytes was measured 3 to 4 days after cRNA injection, as described previously (Wang et al., 1997b). The composition of the sodium buffer was: 100 mM NaCl, 2 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES/Tris, pH 7.4. In the sodium-free buffer, NaCl was replaced with choline chloride. For the inhibition study (Table 1), unlabeled nucleosides and nucleoside analogs (Sigma Chemical, St. Louis, MO or Aldrich, Milwaukee, WI) were added to the reaction solutions as needed.

Electrophysiology. Three to eight days after injection, two-electrode voltage clamp recordings were performed at room temperature (GeneClamp 500B; Axon Instruments, Foster City, CA) (Parent et al., 1992). Oocytes were voltage clamped at 50 mV and were perfused with sodium buffer before substrate perfusions (30-60 s). The substrates were removed by washing in sodium-free choline chloride buffer. Recordings were obtained in a recording chamber at flow rates of ~3 ml/min; the functional volume of the recording chamber is ~25 µl. The steady-state substrate-induced currents were taken as the difference between the average current measured in the presence of substrate and the holding current recorded in sodium buffer before substrate perfusion. (When data were expressed as the difference between the peak current in the presence of substrate and the holding current similar results were obtained.) Voltage pulse protocols were applied from a holding potential of 50 mV using 100-ms voltage pulse steps ranging from 150 to +50 mV in 20 mV increments, with 1-s interpulse intervals. Signals were filtered using a 4-pole low-pass Bessel filter set at a 100 Hz cutoff before sampling at 2 kHz. To obtain current-voltage curves, currents were averaged during the last 25 msec of the 100-ms voltage pulses and then plotted as a function of V_m. Uninjected oocytes were used as controls, undergoing the same treatments as transcript-injected oocytes.

Data Analysis. Values are expressed as mean ± S.D. or mean ± S.E.M., as indicated in the figure legends. The inhibition experiments were performed at least twice on different batches of oocytes. For the electrophysiology experiments, the number of oocytes (n) assayed for each experimental condition are reported in the figure legends and were from more than one donor frog, unless otherwise noted.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Sodium, Voltage, and Concentration Dependence of Uridine-Induced Currents. Nucleoside transport mediated by rCNT1 is sodium-dependent and chloride-independent (Huang et al., 1994). Uridine (50 µM) induced an inward current in sodium uptake buffer in rCNT1-expressing oocytes; 500 µM adenosine reduced the uridine-induced current almost entirely (Fig. 2A). This current is largely sodium-dependent (Fig. 2B). However, very small uridine-induced currents were measured in sodium-free choline chloride buffer in several oocytes. The uridine-induced currents are voltage-dependent and increased ~5-fold upon hyperpolarization of V_m from 10 to 150 mV (Fig. 2C). This suggests that transport of substrate by rCNT1 is voltage-dependent. Uridine displayed saturation kinetics (Fig. 3) with an estimated K_0.5 of 21 ± 2 µM.

Selectivity of rCNT1 for Endogenous Nucleosides. Previous radiotracer uptake and inhibition studies have shown that rCNT1 is a pyrimidine-selective nucleoside transporter; it can also translocate adenosine at slow rates (Huang et al., 1994; Fang et al., 1996). It has also been suggested that the sodium/nucleoside coupling stoichiometry of rCNT1 is 1:1 (Yao et al., 1996). Based on this information, one would predict that 1) uridine, thymidine, and cytidine should induce sodium currents in rCNT1-expressing oocytes; 2) guanosine and inosine should not induce currents; and 3) adenosine may or may not induce currents due to its slow uptake rate. Figure 4 shows that thymidine, uridine, and cytidine induced currents in rCNT1-expressing oocytes, whereas guanosine, inosine, and adenosine did not.

Selectivity of rCNT1 for Nucleosides Analogs. The selectivity of rCNT1 for nucleoside analogs (Fig. 1) was investigated by substrate inhibition and two-electrode voltage clamp studies. [³H]Uridine uptake mediated by rCNT1 was potently inhibited by uridine (U),¹ cytidine (C), thymidine (T), adenosine (A), 4-thiouridine (4TU), 2'-deoxyuridine (UdR), 5-fluoro-2'-dideoxyuridine (FUdR), 5-chloro-2'-dideoxyuridine (ClUdR), 5-bromo-2'-deoxyuridine (BrUdR), 5-iodo-2'-deoxyuridine (IUdR), 3-deazauridine (3DAU), and 5-azacytidine (5AC), and was moderately inhibited by 5-hydroxymethyl-2'-deoxyuridine (HMUdR), 5-fluoro-5'-deoxyuridine (5'dFU), and ribavirin (R). rCNT1 activity was weakly inhibited by inosine (I), guanosine (G), 6-azacytidine (6AC), and 6-azauridine (6AU) (Table 1).

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Structures of the nucleosides and nucleoside analogs used in this study. Compound names are given in Table 1.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

To reach their cellular targets, therapeutic nucleosides must pass a number of membrane barriers, such as the intestinal epithelium and the blood-brain barrier. Membrane transporters play an important role in the cellular uptake and release of these agents (Wang et al., 1997a; Pastor-Anglada et al., 1998). The sodium-coupled nucleoside cotransporter studied in this investigation, rCNT1, is expressed in the intestine, kidney, liver, and brain and is likely to play a primary role in the intestinal uptake of nucleoside analogs (and thereby influence rate and extent of absorption) and in their central nervous system distribution (and thereby influence drug targeting) (Huang et al., 1994; Anderson et al., 1996; Felipe et al., 1998). Knowledge of the substrate selectivity of rCNT1 will greatly enhance our understanding of tissue-specific distribution of therapeutic nucleoside analogs and may facilitate the design of selectively targeted, orally available nucleoside analogs. Our knowledge of the selectivity of rCNT1 has been greatly limited by the lack of robust assays to investigate the structural features of nucleoside analogs essential for translocation by rCNT1. In this report, we used the two-electrode voltage clamp technique to investigate the substrate selectivity and electrogenic properties of rCNT1. Electrophysiology methods can be used to differentiate between transporter substrates and inhibitors (Lostao et al., 1994; Mackey et al., 1999).

rCNT1 has been classified as a pyrimidine-selective nucleoside transporter (Huang et al., 1994). Consistent with this classification, we found that pyrimidine nucleosides induced currents in rCNT1-expressing oocytes, whereas purine nucleosides did not. In addition, ribavirin, an antiviral agent that is structurally related to guanosine (Fig. 1), did not induce a current. This is in agreement with a report that showed that ribavirin is a substrate of the purine-selective transport system in human intestine (Patil et al., 1998). Adenosine, a potent inhibitor of rCNT1 activity, has been shown to be transported by rCNT1 at very slow rates (Fang et al., 1996). At the rCNT1 expression levels achieved in this study, one would predict adenosine-induced currents of <2.4 nA, which are too small to be accurately measured with our two-electrode voltage clamp system. Adenosine did, however, block uridine-induced currents.

Nucleoside transport mediated by rCNT1 is voltage-dependent and also largely dependent on cotransport with sodium. However, in the absence of sodium, small uridine-induced currents were observed, which may represent proton- or choline-coupled uridine currents. It has been shown that other cations, in the absence of sodium, can drive substrate transport for several sodium-coupled cotransporters (Hirayama et al., 1994; Pajor et al., 1998). The voltage dependence of the uridine-induced currents suggests that substrate movement via rCNT1 across the cell membrane is sensitive to membrane potential in native cells. rCNT1 modulation by membrane potential may be important in neural tissue, where rCNT1 transcripts have been detected (Anderson et al., 1996).

The substrate-induced currents of the analogs were found to depend on both the position and the nature of their modifications. Uridine, deoxyuridine, and cytidine analogs with modifications at the 3-, 4-, or 5-positions were all transported by rCNT1. Uridine and cytidine analogs with modifications at the 6-position were not transported. More specifically, six deoxyuridine analogs containing a halogen atom or methoxy functional group at the 5-position were all substrates of rCNT1. Replacement of carbon with nitrogen at the 5-position of cytidine slightly affected the inhibition potency and I_substrate/I_uridine value of 5AC compared with cytidine, whereas replacement of carbon with nitrogen at the 6-position of cytidine dramatically reduced the inhibition potency and I_substrate/I_uridine value of 6AC. Similar results were found for 6AU. The seemingly small change in moving the nitrogen from the 5-position to the 6-position of azapyrimidines has been shown to dramatically alter their conformations, which may explain why the 6AC and 6AU are not substrates of rCNT1 (Singh and Hodgson, 1974a,b). Two uridine analogs, 4TU and 3DAU, were studied. The oxygen at the 4-position is replaced with sulfur in 4TU, resulting in the loss of a good hydrogen-bond acceptor. The nitrogen at the 3-position is replaced with a carbon in 3DAU, resulting in an altered nucleoside structure. Despite these changes, both 4TU and 3DAU were found to be substrates of rCNT1. Finally, 5'dFU was a substrate, which demonstrates that a 5'-hydroxyl group on the sugar is not essential for transport by rCNT1.

Most of the compounds that inhibited rCNT1 activity were substrates of the transporter, but there are notable exceptions. For example, ribavirin and 5'dFU both inhibited [³H]uridine uptake by ~50%; however, while the I_substrate/I_uridine value for 5'dFU was 1.6 (the highest of the 19 compounds studied), ribavirin did not induce a current. Likewise, adenosine, a potent inhibitor of rCNT1 activity, failed to induce a current (Fig. 2A). These results suggest that inhibition data are not good predictors of the substrate selectivity of rCNT1.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Electrogenicity of rCNT1 (A); sodium dependence of uridine-induced currents (B); and representative steady-state current-voltage relationship of the current induced by uridine (500 µM) in a single oocyte expressing rCNT1 (C). A, representative tracing from a single rCNT1-expressing oocyte voltage-clamped at 50 mV. The oocyte was perfused with the solutions indicated in the upper bars; striped bar indicates sodium buffer. Addition of 50 µM uridine induced an inward current. Adenosine (500 µM) abolished the uridine-induced inward current. B, representative tracing from a single rCNT1-expressing oocyte voltage-clamped at 50 mV. Striped bar indicates sodium buffer, open bar indicates sodium-free (choline) buffer, and dark bars indicate uridine applications. For clarity, the current tracings are not shown between washes. Similar results were obtained in other rCNT1-expressing oocytes (n = 4). Uridine-induced currents were not observed in control (uninjected) oocytes (n = 2). C, voltage pulse protocols were applied from a holding potential of 50 mV using 100-ms pulse steps ranging from 150 to +50 mV in 20 mV increments, with 1-s interpulse intervals. Net uridine-induced currents were obtained by subtracting the current generated in the absence and presence of uridine. Similar results were observed from replica experiments (n = 3).

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Concentration dependence of uridine-induced sodium currents. A, representative recordings from a single rCNT1-expressing oocyte exposed to increasing uridine concentrations (for clarity, only four of the nine concentrations are shown). Between applications, the substrate was removed by washing in sodium-free buffer. After washing, the original sodium baseline current was restored before application of a new substrate solution. B, data from a single experiment using one rCNT1-expressing oocyte was fit using the Michaelis-Menten equation (K0.5 = 30 ± 3 µM; Imax = 153 ± 3 nA). Similar K0.5 values were determined from replica experiments (n = 7) (average ± S.E.M.: 21 ± 2 µM). The K0.5 is comparable to values obtained by tracer uptake measurements in oocytes (i.e., 37 and 21 µM) (Huang et al., 1994; Fang et al., 1996). Uridine-induced currents were not detected in control (uninjected) oocytes (n = 2) (data not shown).

View larger version (11K): [in this window] [in a new window]   Fig. 4.   Representative sodium currents in the presence of different endogenous nucleosides (the concentrations used and substrate names are listed in Table 1) (A); and nucleoside-induced currents under voltage clamp conditions in rCNT1-expressing oocytes (B). A, an rCNT1-expressing oocyte was voltage-clamped at 50 mV in sodium buffer. Applications (30 s) of nucleoside are indicated by the dark bars above the traces. Between applications, the substrates were removed by washing in sodium-free buffer. After washing, the original sodium baseline current was restored. B, the currents (like the ones shown in A) were individually normalized to the uridine-induced currents within one set of recordings. Data represent mean ± S.D. (n = 9). The nucleosides did not induce currents in control (uninjected) oocytes (n = 3) (data not shown).

In summary, we have used an electrophysiological assay to investigate the substrate selectivity of rCNT1. rCNT1 was able to transport pyrimidine nucleoside analogs modified at the 3-, 4-, or 5-position of the base and an analog lacking the 5'-hydroxyl group of the sugar. rCNT1 was not able to transport analogs with modifications at the 6-position. Substrate-induced currents in rCNT1-expressing oocytes are substrate-, sodium-, concentration-, and voltage-dependent. The two-electrode voltage clamp assay could be further utilized to screen for rCNT1 substrates and inhibitors.