Differential Kinetics of Transport of 2′,3′-Dideoxyinosine and Adenosine via Concentrative Na⁺ Nucleoside Transporter CNT2 Cloned from Rat Blood-Brain Barrier

Experimental Procedures

Materials.

The mMessage mMachine in vitro transcription kit was obtained from Ambion (Austin, TX). [2,8-³H]adenosine (38.6 Ci/mmol) and [¹⁴C]sucrose (0.6 Ci/mmol) were purchased from PerkinElmer Life Science Products (Boston, MA). [2′,3′-³H(N)] Dideoxyinosine (34.9 Ci/mmol) was purchased from Moravek Biochemicals (Brea, CA). [methyl-³H]Thymidine (86.0 Ci/mmol) was purchased from Amersham Pharmacia Biotech (Arlington Heights, IL). Unlabeled DDI, 2′,3′-dideoxyadenosine (DDA), adenosine, and other molecular biology grade reagents were obtained from Sigma Chemical (St. Louis, MO). NotI was obtained from New England Biolabs (Beverly, MA). Female oocyte-positive Xenopus frogs (50–70 g) were purchased from Nasco (Fort Atkinson, WI).

In Vitro Transcription and Expression in Oocytes.

Rat CNT2 RNA was synthesized with the full-length A-11 CNT2 clone in the expression vector pSPORT, which was isolated from rat BBB cDNA library, as described previously (Li et al., 2001). Clone A-11 is 2905 nucleotides in length, with a 252 nucleotide 5′-untranslated region, a 1977 nucleotide coding region, a stop codon, and a 629 nucleotide 3′-untranslated region followed by a 44-mer poly A tail (Li et al., 2001). The pSPORT-A-11 vector was linearized with NotI and incubated with T7 polymerase for the production of 5′-capped cloned RNA (cRNA) using the mMessage mMachine kit. Capped cRNA was characterized by denaturing gel electrophoresis and ethidium bromide staining. Frog oocytes were isolated as described previously (Boado et al., 1999) and injected with 50 nl of water or cRNA solution (20 ng/oocyte) using a nanoliter injector (World Precision Instruments, Sarasota, FL). Oocytes were kept in Barth's/gentamycin solution for 3 days at 18°C to allow for expression of the mRNA within the oocyte.

Transport Assays.

The uptake, also called clearance, of labeled DDI or adenosine by oocytes injected with either water or CNT2 RNA was measured as previously described (Li et al., 2001). Seven to eight healthy oocytes were incubated in 100 μl of sodium buffer (0.1 M NaCl, 2 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 10 mM Hepes, pH 7.5) containing 2 μCi³H-nucleoside, and 0.08 μCi [¹⁴C]sucrose at 22°C for 5 min. The reaction was stopped with 3 ml of ice-cold choline buffer (0.1 M choline chloride, 2 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 10 mM Hepes, pH 7.4), followed by three additional washes with the same buffer. Oocytes were individually dissolved in 0.5 ml of 1 M NaOH for 30 min at 60°C. The radioactivity was measured in a liquid scintillation counter. The volume of distribution (VD), microliter per oocyte, for either the labeled nucleoside or the labeled sucrose was calculated as follows: VD = (dpm per oocyte)/(dpm per μl of medium). The VD for [¹⁴C]sucrose was subtracted from the total VD for ³H-nucleoside. The clearance is the VD divided by the time of incubation. The flux ratio is the VD in RNA-injected oocytes divided by the VD in water-injected oocytes. Substrate influx into the RNA-injected oocytes was linear for at least 30 min. Each transport saturation study was performed on one to three different oocyte isolates, and a total of 10 frog oocyte isolations was performed for this study. A water-injected oocyte control uptake measurement was performed for each study, and this nonspecific uptake by water injected oocytes was 0.2, 1.4, and 7% of the uptake by cRNA-injected oocytes for adenosine, DDI, and thymidine, respectively.

Parameter Estimation.

TheK_m andV_max of DDI, thymidine, or adenosine uptake, and the constant of nonsaturable transportK_D, were determined by fitting the clearance (Cl) data to the Michaelis-Menten equation Cl = [V_max/(K_m+ S)] + K_D using a nonlinear regression analysis (P3R of the BMDP Statistical Software package; UCLA Biomedical Computing Facility, Los Angeles, CA), where S is the medium substrate concentration. Clearance is the ratio of influx/S, and the clearance plot is formally equivalent to a Michaelis-Menten plot. The adenosine self-inhibition plot was performed in the presence of a constant concentration, 100 μM, of unlabeled DDI, and in this case the apparent K_m(K_m^app) was computed, where K_m^app=K_m(1 + [I]/K_I), andK_m is the absolute adenosineK_m for CNT2, [I] is the concentration of DDI, and K_I is the concentration of DDI that yields half-inhibition of adenosine transport. The uptake of a tracer concentration, 0.5 μM, of [³H]DDI by the oocytes was measured in the presence of increasing concentrations of either unlabeled adenosine or DDA. The [³H]DDI clearance was fit to the following equation: Cl = {(V_max/K_m)[K_I/(K_I+ I)] + K_D}, whereV_max,K_m, andK_D are parameters of DDI transport,K_I is the adenosine or DDA half-inhibition constant, and I is the concentration of unlabeled adenosine or DDA, and the [³H]DDI is present at tracer concentrations. The parameters of sodium dependence were determined from the Hill equation, which isv/V_max = Naⁿ/(Kⁿ + Naⁿ), where v is the transport flux,V_max is the flux at maximal sodium, K is the sodium K_0.5, Na is the sodium concentration, and n is the Hill coefficient, which is the number of sodium ions that are cotransported with each adenosine molecule.

Statistical significance was determined with analysis of variance and ap < 0.05 was considered statistically significant.

Results

The uptake of [³H]DDI by frog oocytes injected with the cloned A-11 CNT2 RNA was 16-fold higher than the uptake of DDI by oocytes injected with water (Fig.1A). DDI transport into the oocytes expressing CNT2 was completely suppressed by 2 mM concentrations of purine nucleosides (adenosine, guanosine, inosine), and the pyrimidine nucleosides (thymidine, uridine), but was marginally inhibited by 2 mM concentrations of AZT, DDA, or DDI, and was not inhibited by dideoxycytidine (DDC) (Fig. 1A). In contrast, the flux ratio of adenosine transport into oocytes expressing the A-11 CNT2 clone was 470 (Fig. 1B), or 29-fold higher than the flux ratio for DDI (Fig. 1A). Whereas adenosine transport was strongly self-inhibited, there was only minor cross-competition of adenosine transport by 2 mM DDI or thymidine (Fig. 1B).

DDI transport into frog oocytes expressing the A-11 CNT2 clone was sodium-dependent (Fig. 2). The half saturation constant (K_0.5) was 8.7 ± 0.9 mM sodium and the Hill coefficient (n) was 1.00 ± 0.10 at a [DDI] of 0.5 μM. These values are similar to the K_0.5 and Hill coefficient for adenosine transport via the A-11 CNT2 clone (Table1).

The clearance (uptake) of [³H]DDI by oocytes expressing the A-11 CNT2 clone was self-inhibited by increasing concentrations of DDI (Fig. 3). Approximately 50% of the DDI transport was nonsaturable. The data in Fig. 3 were fit to the Michaelis-Menten equation (underExperimental Procedures) to yield a DDIK_m of 29.2 ± 8.3 μM, a DDIV_max of 0.40 ± 0.11 pmol/oocyte/min, and a nonsaturable transport constant (K_D) of 15.7 ± 0.6 nl/oocyte/min (Fig. 3; Table 1). The [(V_max/K_m) + K_D] determined from the transport parameters for DDI is 16 nl/oocyte/min, which is 29-fold less than the calculated [(V_max/K_m) + K_D] for adenosine, which is 468 nl/oocyte/min (Table 1).

The clearance of [³H]DDI by the oocyte expressing the A-11-CNT2 clone was cross-inhibited by increasing concentrations of adenosine, and a kinetic analysis of these data yielded an adenosineK_I of 14.8 ± 1.6 μM and a DDIK_D of 0.66 ± 0.47 nl/oocyte/min (Fig. 4). The clearance of [³H]DDI by oocytes expressing the A-11 CNT2 clone was also cross-inhibited by increasing concentrations of DDA and a kinetic analysis of these data yielded a DDAK_I of 8.7 ± 2.2 μM and a DDIK_D of 13.9 ± 0.6 nl/oocyte/min (Fig. 5).

A comparison of the adenosine and DDI kinetic parameters indicated theK_m value for both ligands was comparable, but the V_max for DDI transport via the A-11 CNT2 clone was 27-fold lower than theV_max of adenosine (Table 1). To further examine the mechanism of the interaction of adenosine and DDI with the A-11 CNT2 transporter, the self-inhibition of uptake of [³H]adenosine into oocytes expressing the A-11 CNT2 clone was examined in the presence of a fixed concentration (100 μM) of DDI. In the presence of 100 μM DDI, the adenosineK_m^app was 73.3 ± 12.2 μM with a V_max of 19.2 ± 2.8 pmol/oocyte/min (Fig. 6).

The inhibition of [³H]DDI transport by thymidine (Fig. 1A) suggested thymidine is transported by rat CNT2, and a thymidine saturation curve is shown in Fig.7. The parameters of CNT2 transport of thymidine are K_m = 130 ± 44 μM, V_max = 1.7 ± 0.5 pmol/oocyte/min, and K_D = 2.7 ± 0.5 nl/oocyte/min (Fig. 7). The clearance of a tracer concentration of [³H]thymidine by the CNT2-injected oocytes was 15 nl/oocyte/min (Fig. 7), and this was 16-fold greater than the uptake of [³H]thymidine in water-injected oocytes. Thymidine transport into oocytes injected with the A-11 cRNA was sodium-dependent, and the sodium saturation curve (not shown) for [³H]thymidine yielded a sodiumK_0.5 = 8.4 ± 0.6 mM (Table 1) with a Hill coefficient (n) = 1.1 ± 0.1 in the presence of 0.2 μM thymidine. Adenosine inhibited the uptake of [³H]thymidine by the oocytes injected with A-11 cRNA; the curve of adenosine inhibition of [³H]thymidine uptake (not shown) was similar to the adenosine inhibition of [³H]DDI uptake (Fig. 4). A kinetic analysis of the inhibition of [³H]thymidine transport by unlabeled adenosine yielded an adenosine K_I of 27.8 ± 11.0 μM, a [³H]thymidineV_max/K_mratio of 13.2 ± 1.2 nl/oocyte/min, and a thymidineK_D not significantly different from zero. This adenosine K_I is not significantly different from the K_m of adenosine transport via CNT2 (Table 1). Unlabeled thymidine inhibited the uptake of [³H]DDI by the oocytes injected with A-11 cRNA. The curve of thymidine inhibition of [³H]DDI uptake is shown in Fig.8, and a kinetic analysis (underExperimental Procedures) of this inhibition data yielded an thymidine K_I of 347 ± 113 μM, a [³H]DDIV_max/K_mratio of 17.9 ± 1.7 nl/oocyte/min, and a DDIK_D of 0.87 ± 1.79 nl/oocyte/min, which was not significantly different from zero. This thymidineK_I is not significantly different from the K_m of thymidine transport via CNT2 (Table 1).

Discussion

The results of these studies are consistent with the following conclusions. First, DDI, adenosine, and thymidine are transported via the A-11 clone of rat CNT2 that was isolated from a rat BBB cDNA library (Fig. 1; Table 1) in a sodium-dependent mechanism (Fig. 2; Table 1). Second, DDI and adenosine have comparableK_m values, but theV_max of adenosine transport is 27-fold greater than the V_max of DDI transport (Fig. 3; Table 1). DDI, but not adenosine, has a significant nonsaturable component of transport, represented by theK_D value (Fig. 3; Table 1). However, this nonsaturable transport is not free diffusion; the DDIK_D determined with either saturating adenosine or thymidine concentrations is low and near zero (Figs. 4 and8), and comparable with the low K_Dvalue for adenosine (Table 1). Fourth, the nonsaturable component of DDI transport is not inhibited by high concentrations of DDA (Fig. 5), although the K_I of DDA (Fig. 5) is comparable with the K_m of adenosine transport via CNT2 (Table 1). Fifth, the total inhibitory effect of DDI on adenosine transport is modest (Fig. 1) owing to the reduced DDIV_max, but kinetic evidence supports a mechanism of competitive inhibition of adenosine and DDI transport via rat CNT2 (Fig. 6). Sixth, thymidine is transported via CNT2 with reduced affinity (high K_m) and reduced capacity (low V_max) compared with adenosine (Table 1); thymidine strongly suppresses DDI transport (Figs.1 and 8), but is a weak inhibitor of adenosine transport (Fig. 1).

The transport of DDI via the CNT2 cloned from the rat BBB (rCNT2) confirms previous findings showing DDI transport via CNT2 cloned for human small intestine (hCNT2) (Ritzel et al., 1998). However, the DDI flux into oocytes expressing the A-11 rCNT2 clone, 0.25 pmol/oocyte/min at [DDI] = 10 μM (Fig. 3), is 50-fold greater than DDI flux on the hCNT2, which is 0.005 pmol/oocyte/min at 10 μM DDI (Ritzel et al., 1998). The high rate of transport of DDI via the A-11 rCNT2 clone parallels similar findings with adenosine transport via the A-11 CNT2 clone. Adenosine transport via the rCNT2 transporter expressed from clone A-11 RNA is characterized by aV_max value (Li et al., 2001) that is 50-fold higher than the V_max values of adenosine transport via rCNT2 or hCNT2 isoforms cloned from peripheral rat or human tissues (Che et al., 1995; Yao et al., 1996; Ritzel et al., 1998). There is a high degree of homology between rCNT2 and hCNT2 with a 81% conservation of amino acids. The molecular mechanism causing the high activity of the A-11 rCNT2 clone is unclear, but may be related to a high level of expression of the transporter in frog oocytes, owing to the very long poly A tail (44 nucleotides) of the A-11 rCNT2 clone. The long poly A tail may either stabilize the mRNA or enhance the efficiency of translation leading to increase transporter protein in oocytes. This high transporter activity of the A-11 clone enables a detailed kinetic analysis of substrate transport via the rCNT2, such as described here for DDI. Owing to the reducedV_max of DDI transport on rCNT2, it would be difficult to analyze the kinetics of transport into oocytes expressing the cloned CNT2 if the activity of the transporter was reduced up to 50-fold.

The transport of DDI via rCNT2 is much slower than the flux of adenosine on this transporter. The oocyte volume of distribution at tracer concentrations of [³H]DDI, 0.14 μl/oocyte, is 7-fold less than theV_D value of [³H]adenosine at tracer concentrations (Fig.1). The reduced V_D value for DDI is due in part to the 27-fold lower V_maxof DDI transport on rCNT2, compared with theV_max of adenosine transport via this carrier (Table 1). The DDI V_D is not reduced in proportion to the DDI V_maxbecause approximately 50% of DDI transport is contributed by the nonsaturable component (K_D), whereas the K_D value for adenosine transport is nearly zero (Table 1). The reducedV_max of DDI via the BBB CNT2 explains, in part, the very low rate of DDI penetration through the BBB in vivo (Anderson et al., 1990).

The molecular mechanism underlying the significantK_D value for DDI does not appear to be free diffusion. The removal of the 2′ and 3′ hydroxyl groups would be expected to increase the lipid solubility of DDI, compared with the lipid solubility of adenosine. However, if free diffusion were responsible for the DDI K_D, then the nonsaturable component of transport would not be subject to either 1) inhibition by high concentrations of adenosine (Fig. 4) or thymidine (Fig. 8), or 2) inhibition by removal of sodium (Fig. 2). The [³H]DDI clearance is reduced to a very low value at high adenosine concentrations (Fig. 4); the DDIK_D value at saturating adenosine concentrations, 0.66 ± 0.47 nl/oocyte/min (Fig. 4), is not significantly different from the K_Dvalue reported previously for adenosine (Table 1). In contrast, in the presence of high concentrations of either unlabeled DDI (Fig. 3) or DDA (Fig. 5), the [³H]DDI clearance is not reduced below a value of 14 to 16 nl/oocyte/min, which is equal to theK_D value of DDI transport (Table 1). The nonsaturable component of DDI transport is also sodium-dependent. Whereas the K_D of DDI transport represents more than 50% of the total DDI transport (Fig. 3), the complete removal of sodium from the media effectively eliminates any DDI transport via the rCNT2 (Fig. 2). Free diffusion of DDI owing to lipid solubility of the molecule would not be sodium-dependent.

The two principal differences in the kinetic parameters of DDI and adenosine transport via the A-11 rCNT2 are the 27-fold lowerV_max for DDI relative to adenosine, and the >16-fold increase in the K_Dvalue for DDI relative to adenosine (Table 1). These findings are consistent with a model of asymmetric CNT2 transport sites within the membrane: a high V_max site favored by adenosine and a low V_max site favored by DDI. Further evidence for this model is the finding that thymidine is also transported by CNT2 (Fig. 7) and that thymidine is a strong inhibitor of DDI transport (Fig. 8), but a weak inhibitor of adenosine transport (Fig. 1). Adenosine is transported by both sites, because adenosine strongly inhibits both DDI transport (Fig. 4) and thymidine transport (under Results). DDI may be bound by the highV_max site, but not transported. The inhibition of the principal adenosine transport site by DDI is shown in Fig. 6. The K_m^app of adenosine transport via rCNT2 is 73.3 ± 12.2 μM in the presence of 100 μM DDI (Fig. 6). In the absence of competitive inhibitors, theK_m of adenosine transport via the A-11 rCNT2 clone is 23.1 ± 3.7 μM (Table 1). Therefore, theK_m^app/K_mratio for adenosine is 73.3/23.1 = 3.2 at a [DDI] = 100 μM. If the mechanism of inhibition of adenosine transport by DDI was competitive inhibition, then theK_m^app/K_mratio should equal the [I]/K_I ratio, where [I] = the DDI concentration (100 μM) and theK_I is equal to theK_m of DDI transport, 29.2 ± 8.3 μM (Table 1). Given [I] = 100 μM andK_I = 29.2 μM, the I/K_I ratio is 100/29.2 = 3.4, which is not different from the adenosineK_m^app/K_mratio. These findings indicate DDI inhibits at the principal adenosine transporter site.

These kinetic results are consistent with a CNT2 transporter model of asymmetric transport sites within the membrane, which is comprised of a high V_max site, used primarily by adenosine and not by DDI or thymidine, and a lowV_max site, used by DDI and thymidine. These kinetic properties of the transporter may be indicative of complex topology of the rCNT2 transporter protein within the membrane. Other sodium cotransporters form homodimeric structures within the membrane (Kilic and Rudnick, 2000), and electron microscopy provides evidence for asymmetric structures of transport monomers within the functional oligomeric unit (Tate et al., 2001). The structural asymmetry of the transporter subunits enables an alternating two-site mechanism (van Veen et al., 2000).

In summary, the present study demonstrates that DDI is transported via the rat CNT2 cloned from the BBB (Li et al., 2001). However, the influx of DDI into brain via transport on the BBB adenosine carrier does not lead to the accumulation in brain of DDI in pharmacologically significant amounts. There is decreased influx of DDI from blood to brain, owing to the reduced V_max of the BBB CNT2 for DDI (Table 1), and there is increased efflux from brain to blood of DDI (Takasawa et al., 1997). These combined influx and efflux mechanisms explain the reduced brain penetration of circulating DDI (Anderson et al., 1990; Morgan et al., 1992).