- Top of page
- Materials and methods
In membrane preparations, CFT, a phenyltropane cocaine analog, and dopamine (DA) interact with the recombinant human dopamine transporter (hDAT) in Na+-free medium. Na+ markedly increased the transporter's affinity for CFT, but had little or no effect on DA potency for inhibiting CFT binding. Raising [Na+] from 20 to 155 mm reduced Li+-induced increase in DA Ki, but not CFT Kd. The presence of 155 mm Na+ enhanced the tolerance to low pH of CFT Kd but not DA Ki. Leucine substitution for tryptophan 84 (W84L) in transmembrane domain (TM) 1 or asparagine substitution for aspartate 313 (D313N) in TM 6 did not or only modestly enhance the affinity of Na+-independent CFT binding, and retained the near normal ability of DA, Li+, K+, or H+ to inhibit this binding. However, the mutations significantly enhanced the Na+ stimulation of CFT binding as well as the Na+ antagonism against Li+ and H+ inhibition of CFT binding. In contrast, the mutations neither changed the Na+-insensitive feature of DA Ki nor enhanced the Na+ protection of DA Ki against Li+'s inhibitory effect, though they caused Na+ protection of DA Ki against H+'s inhibitory action. These results are consistent with the existence of binding conformations for DA that are distinguishable from those for CFT, and with a differential association of cation interactions with DA and CFT binding. The mutations likely alter Na+-bound state(s) of hDAT, preferentially strengthening the positive allosteric coupling between Na+ and CFT binding, and reducing the impact of Li+ or H+ on the CFT binding.
The involvement of monovalent alkali cations and protons in interactions of the human dopamine (DA) transporter (hDAT) with substrates and cocaine analogs has been documented extensively. Like other members of the Na+/Cl–-dependent neurotransmitter transporter family, the hDAT not only requires Na+ for uphill transport of substrates, but also generates a transport-associated current carried by Na+ as well as a leak current that can be carried by multiple ions such as K+, Li+, H+, and Na+ (Sonders et al. 1997). Despite their similarities in carrying the leak current that can be blocked by both substrates and inhibitors of the transporter, Na+ mainly promotes cocaine analog binding and substrate transport, but K+, Li+, and H+ inhibit these processes (Shank et al. 1987; Amejdki-Chab et al. 1992; Xu and Reith 1996; Chen et al. 1997, 1999; Berfield et al. 1999; Li and Reith 1999). Interactions of Na+ with other cations at the transporter level have been observed in several members of the Na+/Cl–-dependent transporter family. Thus, Na+ prevents K+ inhibition of cocaine analog and DA binding in rat and human DATs (Chen et al. 1997; Li and Reith 1999, 2000), and antagonizes the effect of Li+ on chemical inactivation and current in serotonin transporters (Ni et al. 2001). Additionally, in serotonin and GABA transporters, the mutation-induced change in Na+ sensitivity can be accompanied by changes in pH sensitivity (Barker et al. 1999; Forlani et al. 2001). It has been demonstrated that the Na+ binding at the transporter protein can bring about conformational changes (Mager et al. 1996; Loo et al. 1998; Maehrel et al. 1998; Li et al. 2000). Recently, antagonism between Na+ and Li+ at the serotonin transporter has been suggested to result from different conformational changes induced by the two cations (Petersen and DeFelice 1999; Kamdar et al. 2001; Ni et al. 2001).
Kinetic analysis of DA transport (McElvain and Schenk 1992; Chen et al. 1999; Earles and Schenk 1999) and mathematical modeling of cocaine analog and substrate binding (Chen et al. 1997; Li and Reith 1999, 2000) have provided evidence implicating different Na+ requirements between substrates and cocaine analogs. In our mutagenesis studies on negatively charged residues and dipole tryptophan residues that are highly conserved in the Na+/Cl–-dependent neurotransmitter transporter family, mutations often differentially modified cation-sensitivity of cocaine analog binding and DA uptake (Chen et al. 2001a). Moreover, some of these mutations also shifted the relative selectivity of the transporter towards either cocaine analogs or DA. We speculated that the binding of cocaine analogs and DA involve non-identical conformations differentially coupled to cation-binding sites. Thus, a shift in ligand selectivity by a mutation could be caused by alterations in cationic interactions that are critical in directing the binding preference of the hDAT. The present study was initiated to address this hypothesis by assessing the effect of individual cations Na+, K+, Li+, and H+, as well as the impact of Na+ interactions with Li+ or H+, on the binding of a phenyltropane analog of cocaine, 2β-carbomethoxy-3β-(4fluorophenyl)tropane (CFT), and DA. DA binding was assessed indirectly through inhibition of [3H]CFT binding. We sought to distinguish cationic interactions with CFT binding from those with DA binding at wild-type as well as mutated hDATs. Two mutated hDATs were used in this study: leucine substitution for tryptophan 84 (W84L) in transmembrane domain (TM) 1 or asparagine substitution for aspartate 313 (D313N) in TM 6. Both residues are highly conserved among the member of Na+/Cl–-dependent neurotransmitter transporter family. Their sequence conservation could reflect their involvement in features common for the Na+/Cl–-dependent transporters, such as ion regulation. Moreover, both mutations resulted in not only changes in cationic sensitivity (Chen et al. 2001a) but also a shift in the relative selectivity towards cocaine analogs (Chen et al. 2001b). Knowledge regarding unique cationic susceptibility for cocaine analog and substrate binding could open the possibility to develop anticocaine medications aimed at the cation-modulating site on the hDAT.
- Top of page
- Materials and methods
The current study was performed to address the possibility of CFT and DA binding to different hDAT conformations that are differentially impacted by Na+ and by other cations the binding of which can be affected by Na+. For this purpose, cationic interactions at the hDAT were explored in parallel. When interpreting cation interactions from their effects on ligand binding at the transporter, one faces multiple possibilities. First, Na+ and Li+ (or H+) can be thought to interact with hDAT mutually exclusively either via competing for the same site or inducing opposite conformational changes. In the present study, mutation at W84 or D313 specifically enhanced Na+ sensitivity (Fig. 2) but retained near normal Li+, K+, and H+ sensitivity in the absence of Na+ (Tables 3 and 4, Fig. 5a), suggesting that the sites at which Na+ exerts a stimulatory effect differ from those at which the inhibitory cations Li+, K+, or H+ act. Second, if Na+ and Li+ (or Na+ and H+) bind to different DAT conformations and the binding of each fully discourage the binding of the other (extreme negative allosterism resulting in mutually exclusivity), Na+ would prevent Li+ (or H+) from inhibiting not only CFT but also DA binding, even though the binding of DA itself was not sensitive to Na+. However, in wild-type hDAT, Na+ failed to completely overcome Li+ (or H+) inhibition of CFT binding and did not antagonize H+ inhibition of DA binding at all. Moreover, the inhibition by K+, Li+, and H+ of CFT binding was actually stronger at 5 mm Na+ than in Na+-free medium (Tables 3 and 4, Fig. 6). These findings do not support exclusive interactions between Na+ and Li+ (or H+). A third possibility that can be contemplated is that Na+ and Li+ (or H+) bind to hDAT simultaneously and do not interact with each other, but both interact with a third party on the hDAT. In this case, the apparent antagonism by Na+ of Li+ (or H+) inhibition as judged from [3H]CFT binding may be primarily due to the high affinity of CFT binding induced by Na+, which offsets the cation inhibition. However, Na+-induced reduction in Li+ inhibition remained even after correcting for CFT affinity (Table 3) or even when the Ki for DA was insensitive to [Na+] (Table 5). Additionally, Na+ antagonism against H+ inhibition was also observed when the control affinity for CFT was comparable between the two Na+ concentrations tested (Fig. 8). More likely, Na+ and Li+ (or H+) bind to the hDAT simultaneously, but the Na+ binding induces conformational changes that alter the interactions of Li+ (or H+) with the hDAT.
As previously reported (Reith and Coffey 1993; Corera et al. 2000), high concentrations of Na+ modestly reduced the CFT binding to a level below its peak stimulatory effect (Fig. 2b). Intriguingly, W84L mutation only shifted the rising phase while D313N mutation shifted both the rising and the declining phases of the Na+ curve to the left (Fig. 2). These observations raise the possibility that the dual effects of Na+ may represent different hDAT states triggered by Na+ binding at different sites of the transporter. Moreover, when comparing the inhibitory effect of Li+, K+, and H+ on CFT binding in the absence and presence of Na+, we found that Na+ influenced their effects also in a biphasic fashion, enhancing them at low concentrations and reducing them at moderate to high concentrations (Tables 3 and 4, Fig. 6). Such a complexity may reflect the contribution of more than one Na+-bound hDAT state to the mix of conformations existing at Na+ concentrations varying from 0 to 155 mm, causing different susceptibility to the inhibitory cations. All together, the results do not seem compatible with the idea that there is one hDAT conformation in the absence of Na+, another one in the presence of Na+, with [Na+] simply determining the proportion of transporters in each of those two states, and with Na+ counteracting the inhibitory action of Li+, K+, and H+ at the Na+-bound state. Rather, the results for membrane preparations under the current conditions suggest the existence of at least one more Na+ bound state where Na+ in fact enhances the inhibitory effect of other cations; this was not previously measured in other studies because the buffers commonly used do not take into account the 0–10 mm Na+ concentration range. Our experiments cannot answer the question of whether the states correspond to DAT free of Na+, DAT bound with one Na+ ion, and DAT bound with two Na+ ions (as during transport of DA). As observed for the allosteric effect of scorpion toxin on batrachotoxin binding to sodium channels (Catterall et al. 1985), if the transition between different hDAT conformations is very fast, one would actually measure a single time-averaged affinity for CFT depending on Na+ concentration. Therefore, the effect of the inhibitory cations would be expected to reflect their preference for a given hDAT state as well as the manner they interact with Na+ at that state.
In the present study, several lines of evidence from wild-type and mutated hDATs support the existence of binding conformations for DA that are distinguishable from those for CFT. Whereas Na+ markedly increased the transporter's affinity for CFT, it had no effect on the potency for DA to inhibit CFT binding. The observations that Na+ and CFT reciprocally enhanced each other's interaction with hDAT (Fig. 3a) suggest that Na+ and CFT bind to different sites and binding of each drive the protein towards a conformation favored by the other (coupling by positive allosterism). The lack of Na+ dependence of DA Ki observed here is consistent with the poor fits found for models assuming low DA binding to membrane preparations in the absence of Na+ (Li and Reith 2000), and with the observed lack of Na+ stimulation of DA binding with sulfate as the covarying anion (Li et al. 2002). The mutagenesis data further highlight the differences in Na+ dependence between CFT and DA binding. Thus, W84 or D313 mutation significantly enhanced the Na+ stimulation of CFT binding as well as the high-affinity CFT binding, but failed to modify the Na+-insensitive feature of DA Ki (Fig. 4). We conclude that CFT prefers a Na+-bound hDAT conformation(s), whereas DA can bind to Na+-free and Na+-bound hDAT conformations equally well, at least in membrane preparations. As a carrier transporting its substrate bidirectionally, the hDAT is expected to have multiple conformational transitions during the transport cycle. Thus, DA would have the ability to interact with multiple hDAT states where the DA binding sites face either high extracellular [Na+] or low intracellular [Na+]. In this context, membrane preparations may provide an opportunity to study conformational changes that cannot be readily detected in intact cells due to the restricted sidedness and cationic accessibility. However, more work is needed to characterize the role of Na+ in the binding of cocaine analogs and DA in functional plasma membranes of intact cells.
For CFT binding at wild-type hDAT, modest antagonism by Na+ of Li+ inhibition was observed when raising [Na+] from 5 mm to 20 mm (Table 3), but this effect was not further enhanced by raising [Na+] from 20 mm to 155 mm (Table 5). By contrast, Na+-induced reduction in Li+ inhibition of DA binding was observed when [Na+] was increased from 20 mm to 155 mm (Table 5). It can be considered that the proportion of various Na+-bound states of hDAT differs between moderate and saturating [Na+]. Further, on different and even on the same Na+-bound hDATs, the residues contributing to formation of CFT and DA binding sites are not identical. Thus, Li+ may have less of an effect on the formation of CFT binding sites on an intermediate hDAT state induced by modest [Na+] than on an hDAT state held by saturating [Na+]; the opposite could apply to the formation of DA binding sites. The findings that W84L or D313N mutation reduced Li+ inhibition of CFT binding at moderate but not high [Na+] as well as the observed lack of effect of Na+ antagonism against Li+ inhibition of DA binding (Table 5) lend further support to the above-mentioned ideas.
The discrepancy between CFT and DA binding was also evident when the effect of Na+–H+ interactions were examined in wild-type hDAT. At low [Na+], CFT Kd was much more sensitive to H+ than DA Ki; further, raising [Na+] attenuated H+ inhibition on CFT Kd but not DA Ki (Fig. 8a,d). Protonation of some critical residues may alter the charge-carrying status of the transporter, which not only directly impedes ligand binding, but also affects allosteric coupling between Na+ and CFT binding. Indeed, H+ lowered the plateau in the Na+ stimulation curve for CFT binding (Fig. 7), suggesting a less efficient Na+–CFT coupling. This could explain why CFT binding is more vulnerable to H+ than DA binding at low [Na+] and why increasing [Na+] cannot completely overcome the H+ inhibition of CFT Kd. In favor of a less efficient Na+–CFT coupling resulting from H+ action, the CFT binding affinity for W84L and D313N mutants that allowed higher Na+ plateau in stimulating CFT binding at low pH (Fig. 7b,c) was significantly tolerant to H+ inhibition (Fig. 8b,c).
In addition to cations inducing conformational changes in DAT, such as Na+ promoting a state preferentially binding cocaine-like compounds over DA, one also needs to consider the impact of ligand binding itself. For example, the thermodynamic analysis of Bonnet et al. (1990) indicated more drastic global conformational changes in the DAT upon binding of blockers than substrates, and Do-Régo et al. 1998) reported an early step in blocker binding that represented a conformational change. In addition, Ferrer and Javitch (1998) demonstrated that cocaine enhanced accessibility of C90 in hDAT to charged methanethiosulfonate (MTS) reagents, whereas benztropine was found to be inert in this respect as opposed to other blockers and DA itself (Reith et al. 2001). Another substrate, m-tyramine was reported to enhance accessibility to MTS of C342 in hDAT-expressing cells, requiring cocaine-sensitive translocation (Chen et al. 2000). Thus, cocaine analogs and DA may well induce different conformational changes in DAT, and it is very possible that these distinct conformational states are differentially sensitive to cations such as Li+ and H+. In this view, in the presence of saturating [Na+], rather than having one Na+-bound DAT state that binds CFT or DA with Li+ or H+ impacting binding of the two ligands differentially through non-identical intermediaries, one would have a state for the CFT–DAT complex that is different from the state of the DA–DAT complex.
The smaller EC50 value for Na+ stimulation of CFT binding at W84L and D313N could indicate an enhanced Na+-binding to Na+-preferring conformations. However, a smaller EC50 value could also result from kinetic changes in equilibrium binding assays as a consequence of enhanced coupling by positive allosterism between Na+ and CFT binding. Notably, the W84L mutation caused, in addition to a smaller EC50 value, a higher plateau in the Na+ stimulation curve for CFT affinity (Fig. 4), which indicates a stronger positive allosteric coupling effect. Further, for both mutants, the wild-type like Na+ antagonism against Li+ inhibition of DA binding (Table 5) as well as the greater H+-induced shift in Na+'s EC50 (Fig. 7) argue that the mutations neither enhanced the binding of Na+ nor reduced the binding of Li+ (or H+) in the presence of Na+. We consider it more likely that the mutations alter a Na+-bound form(s) of the hDAT. This alteration preferentially strengthens the positive allosteric coupling between Na+ binding and CFT binding, resulting in a shift in the binding selectivity of the mutants towards CFT over DA. Such an enhanced Na+–CFT coupling may also make CFT binding less sensitive to Li+ action (at modest [Na+]) and to the uncoupling effect of H+. In addition, there could be, even when [Na+] is saturating, multiple Na+-bound states with different affinities for CFT, giving an average plateau Kd; in this case, the mutations could cause increased frequency (and/or life-span) of a Na+-bound hDAT state(s) with higher affinity for CFT. Such changes would also explain the increased plateau in the Na+ curve for CFT affinity for W84L.
Although the mutations generally do not seem to affect DA binding in membrane preparations, they converted the H+ inhibition of DA Ki from being Na+-insensitive to Na+-sensitive (Fig. 8). This change could be accounted for by a Na+-bound form of the transporter less mobile in adopting H+-induced modification in conformational state or/and charge-carrying status that perturbs DA binding. We have also found that the DA uptake is damaged by the mutation of W84 and D313 (Chen et al. 2001b). How the altered cationic interactions occurring at recognition steps influence the transport cycle is currently under investigation. Because of the high conservation of W84 and D313 among the members of Na+/Cl–-dependent neurotransmitter transporter family, the present findings may apply to other transporters. Indeed, mutation of a tryptophan residue equivalent to W84 of the hDAT in the GABA transporter GAT1 (W68L) enhanced the charge movement linked to Na+ binding (Mager et al. 1996), favoring a role of this residue common for the Na+/Cl–-dependent transporter family.