A conformationally sensitive residue on the cytoplasmic surface of serotonin transporter.

Serotonin transporter (SERT) contains a single reactive external cysteine residue at position 109 (Chen, J. G., Liu-Chen, S., and Rudnick, G. (1997) Biochemistry 36, 1479-1486) and seven predicted cytoplasmic cysteines. A mutant of rat SERT (X8C) in which those eight cysteine residues were replaced by other amino acids retained approximately 32% of wild type transport activity and approximately 56% of wild type binding activity. In contrast to wild-type SERT or the C109A mutant, X8C was resistant to inhibition of high affinity cocaine analog binding by the cysteine reagent 2-(aminoethyl)methanethiosulfonate hydrobromide (MTSEA) in membrane preparations from transfected cells. Each predicted cytoplasmic cysteine residue was reintroduced, one at a time, into the X8C template. Reintroduction of Cys-357, located in the third intracellular loop, restored MTSEA sensitivity similar to that of C109A. Replacement of only Cys-109 and Cys-357 was sufficient to prevent MTSEA sensitivity. Thus, Cys-357 was the sole cytoplasmic determinant of MTSEA sensitivity in SERT. Both serotonin and cocaine protected SERT from inactivation by MTSEA at Cys-357. This protection was apparently mediated through a conformational change following ligand binding. Although both ligands bind in the absence of Na(+) and at 4 degrees C, their ability to protect Cys-357 required Na(+) and was prevented at 4 degrees C. The accessibility of Cys-357 to MTSEA inactivation was increased by monovalent cations. The K(+) ion, which is believed to serve as a countertransport substrate for SERT, was the most effective ion for increasing Cys-357 reactivity.

Serotonin transporter (SERT) 1 is a member of a large family of homologous integral membrane proteins (1)(2)(3)(4). These transporters take up extracellular substrate in a process that is coupled to the transmembrane movement of Na ϩ , Cl Ϫ , and, in some cases, K ϩ (5). In SERT, serotonin (5-HT) reuptake into neurons and peripheral cells such as platelets is believed to occur through cotransport with Na ϩ and Cl Ϫ and countertrans-port with K ϩ (5). A widely studied aspect of these proteins is their role in the removal of neurotransmitter after its release into the synaptic cleft of neurons, by which they regulate synaptic activity. The role of SERT in behavior is demonstrated by the action of SERT inhibitors, which are clinically effective as antidepressants (6). SERT also interacts with psychostimulants, some of which, such as cocaine, are inhibitors (7), while others, such as amphetamine derivatives, are alternative substrates (8). Members of this family include transporters for dopamine, norepinephrine, glycine, ␥-aminobutyric acid, proline, creatine, and betaine (9 -21). SERT is most closely related to transporters for the catecholamines dopamine and norepinephrine (DAT and NET, respectively) (1,22). These biogenic amine transporters stand out as a distinct subfamily. They are all inhibited by cocaine and share many structural and functional properties.
Hydropathy analysis of the cDNA sequence coding for SERT (23)(24)(25) predicted 12 ␣-helical transmembrane domains connected by six extracellular and five cytoplasmic loops with cytoplasmic NH 2 and COOH termini (9,10). Previous work from this laboratory established that cysteine and lysine residues in each of the predicted external loops reacted with impermeant reagents added to the extracellular medium (26). However, this technique did not establish the topology of any intracellular domains. In mutants with no reactive residues in external loops, cysteine and lysine reagents reacted with SERT only when the cell membrane was permeabilized with detergent, providing a means to identify intracellular residues.
A single endogenous cysteine residue at position 109 (Cys-109) in SERT is responsible for inactivation by extracellular application of the cysteine reagents 2-(aminoethyl)methanethiosulfonate hydrobromide (MTSEA) and [2-(trimethylammonium)ethyl]methanethiosulfonate (MTSET) (27). This residue is predicted to lie in the first extracellular loop. A mutant of SERT in which Cys-109 was replaced by alanine (C109A) was insensitive to external cysteine reagents. However, in membrane preparations of cells expressing C109A, we observed modest inactivation of binding activity by these reagents, suggesting that one or more internal cysteine residues were modified, leading to inactivation (28).
Work on related transporters has identified internal cysteines as determinants of sensitivity to cysteine reagents. In DAT, two predicted intracellular cysteines, Cys-135 in the first internal loop (IL1) and Cys-342 in IL3 (equivalent to Cys-155 and Cys-357 in SERT, respectively), were shown to confer sensitivity to cysteine reagents (29). In ␥-aminobutyric acid transporter (GAT-1), a predicted internal cysteine at position 399 was shown to be the major site of inactivation by cysteine reagents (30). There is no cysteine in SERT at the position corresponding to ␥-aminobutyric acid transporter Cys-399.
In the present work, we undertook identification of the cysteine residues on the internal surface of rat SERT that were responsible for inactivation of binding in membrane preparations. Our results suggest that a reactive cysteine in the third intracellular loop is sensitive to conformational changes that result from ion and ligand binding.

MATERIALS AND METHODS
Mutagenesis-Mutant transporters were generated by site-directed mutagenesis of the C109A mutant of rat SERT, which contains sequences encoding a c-myc epitope tag at the N terminus and a FLAG epitope tag at the C terminus (27,31). The mutated regions were excised by digestion with appropriate restriction enzymes and subcloned back into the original plasmid. All mutations were confirmed by DNA sequencing.
Expression-Confluent HeLa cells were infected with recombinant vTF-7 vaccinia virus and then transfected with plasmid bearing SERT mutant cDNA under control of the T7 promoter as described previously (32). Transfected cells were incubated for 14 -20 h at 37°C and then used for the determination of transport and binding activities.
Transport Assays-Transfected HeLa cells in 24-well culture plates were washed twice with 500 l of phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 , and 1.4 mM KH 2 PO 4 , pH 7.3) containing 0.1 mM CaCl 2 and 1 mM MgCl 2 (PBS/CM). To the washed cells, 250 l of PBS/CM containing 4.9 nM 5-HT (PerkinElmer Life Sciences; catalog no. NET-498) was added, and the incubation was continued for 10 min at room temperature when each well was washed three times by aspiration with ice-cold phosphate-buffered saline. The cells were lysed in 500 l of 1% SDS for 30 -60 min, transferred into scintillation vials, and counted in 3 ml of Optifluor scintillant (Packard Instrument Co.).
Membrane Preparation and Binding Assays-HeLa cells grown in 75-cm 2 cell culture flasks were transfected with SERT cDNA as described above. The cells were rinsed once with 10 mM HEPES buffer (adjusted to pH 8.0 with NaOH) and scraped into 5 ml of homogenization buffer (10 mM HEPES, pH 8.0, containing 1:500 (v/v) protease inhibitor mixture (Sigma; catalog no. P8340) and 100 M phenylmethanesulfonyl fluoride). The cells were then disrupted by homogenization in 20 ml of homogenization buffer on ice, using a Polytron homogenizer (Brinkman Inc., New York) at a setting of 7 for 20 s. The homogenization was repeated after 1 min on ice. The membranes were collected by centrifugation at 48,000 ϫ g for 20 min at 4°C. Each preparation from a single 75-cm 2 flask was resuspended in 1 ml of homogenization buffer and stored as 0.1-ml samples at Ϫ80°C.
To determine binding activity, the high affinity cocaine analog, 2-␤carbomethoxy-3-␤-(4-[ 125 I]iodophenyl)tropane (␤-CIT) was used. Membrane suspensions were thawed on ice and diluted with 1 ml of binding buffer (10 mM HEPES, pH 8.0, with NaOH, 150 mM NaCl, 0.1 mM CaCl 2 , and 1 mM MgCl 2 ). 100 l of the diluted suspension was added per well in 96-well filtration plates (Multiscreen Filtration System; Millipore Corp.). The membranes were washed twice by filtration with 200 l of binding buffer, and then binding was initiated by the addition of 200 l of binding buffer containing 10,000 cpm of ␤-CIT (RTI-55; PerkinElmer Life Sciences, catalog no. NEX272). For determination of MTSEA sensitivity, the membranes were incubated with MTSEA for 15 min following initial washing of the membranes. Then MTSEA was removed by washing the membranes three times with binding buffer. The membranes were incubated with ␤-CIT for 1.5 h at room temperature with gentle agitation, and then the reaction was terminated by washing the membranes three times with 200 l of binding buffer. The filters from each individual well were removed and placed in scintillation vials containing 3 ml of Optifluor scintillation fluid (Packard Instrument Co.). The filters were allowed to soak for 2 h and were then counted.
To measure the effect of cations on the action of MTSEA, the membranes were washed in binding buffer with all Na ϩ replaced by the given replacement ion, and MTSEA was diluted into that buffer. After the MTSEA incubation, the membranes were washed in normal, Na ϩcontaining binding buffer in which binding was measured.
For the protection assays, ligands (cocaine or serotonin) were added to the washed membranes and incubated for 10 min. MTSEA was subsequently added to the membranes for 15 min, and the membranes were then washed with binding buffer five times to remove unbound MTSEA and ligand. Binding was then measured as described above.
Reactivation Experiments-Following MTSEA incubation of membrane suspensions, either 12 mM free cysteine, 10 mM 2-mercaptoethanol, or 10 mM dithiothreitol were added to the suspension for 180, 90, or 15 min, respectively. At the end of the incubation, the membranes were washed three times with binding buffer, and binding was measured as described above.
Data Analysis-Nonlinear regression fits of experimental and calculated data were performed with Origin (Microcal Software, Northampton, MA), which used the Marquardt-Levenberg nonlinear least squares curve fitting algorithm. Each figure shows a representative experiment that was performed at least twice. The statistical analysis given was from multiple experiments. Unless indicated otherwise, data with error bars represented the mean Ϯ S.D. for four samples from two separate experiments. The asterisks indicate significance at the p Ͻ 0.05 level in the paired Student's t test.

RESULTS
The C109A mutant of SERT was shown to be insensitive to externally applied methanethiosulfonate (MTS) reagents (26,27). However, in membrane preparations from cells expressing C109A, ␤-CIT binding was inactivated by MTS reagents (28). In these membrane preparations, MTS reagents have access to the cytoplasmic surface of the plasma membrane, suggesting that one or more internal cysteines were responsible for the inactivation of binding in C109A.
To identify the internal cysteine residues responsible for this inactivation, a mutant of SERT was prepared (X8C) with eight predicted intra-and extracellular cysteines (at positions 15, 21, 109, 147, 155, 357, 522, and 622) replaced by other amino acids. Cys-357, which was highly conserved within the NaCl-coupled transporter gene family, was changed to isoleucine, which was found in the corresponding position in the proline transporter. Cys-522 was changed to serine as found in the corresponding positions in NET and DAT. Cys-147 was changed to alanine as found in NET and DAT. Cys-155 was highly conserved but was changed to alanine, which was one of the few amino acids that was found at that position when it was not cysteine. The other predicted internal cysteine residues were changed to alanine, because there was little sequence conservation at that position within the gene family. X8C, which retained significant transport and binding activity, was further modified by restoring each of the endogenous predicted internal cysteines in the X8C background. These constructs, as well as one with only Cys-109 and Cys-357 replaced (C109A/C357I) were all found to be active for transport and binding ( Table I).
All of these mutants retained significant transport and binding activities (Table I). X8C, the mutant with all predicted intracellular cysteines replaced, had about the lowest transport activity (31.9 Ϯ 3.8% of C109A activity). Reintroduction of Cys-155 and Cys-357 into X8C resulted in the greatest recovery TABLE I Transport and binding activities of SERT mutants Transport activity was assayed in HeLa cells following expression of the mutants as described under "Materials and Methods." Binding activity was assayed using a membrane preparation from HeLa cells expressing each construct as described under "Materials and Methods." Data are means Ϯ S.D. from six measurements in three separate experiments expressed as the percentage of transport or binding activity relative to SERT C109A. X8C is SERT with the following substitutions: C15A, C21A, C109A, C147A, C155A, C357I, C522S, and C622A. In constructs such as X8C-357C, one or more of the endogenous cysteine residues was restored in the X8C background. of transport activity (44 and 40%, respectively). The binding activities of the SERT mutants were significantly greater than the transport activities when expressed as a percentage of C109A activity. X8C had 56% of C109A binding activity, and again, reintroduction of Cys-155 resulted in the greatest binding activity of all single cysteine replacement mutants (70% of C109A). However, reintroduction of Cys-357 into X8C did not substantially increase binding activity. As described previously (28), 5-HT transport by cells expressing SERT C109A was insensitive to MTSEA. We chose to use this reagent in studying putative cytoplasmic cysteine residues, because it is more reactive than MTSET (33). MTSEA is also more permeant across lipid bilayers than MTSET (34), but for this study it was seen as an advantage that the reagent would have unrestricted access to the cytoplasmic face of the transporter. Fig. 1A shows that intact cells expressing Cys-109 or the X8C mutant were resistant to a 15-min treatment with 1.5 mM MTSEA. In membrane preparations, where both faces of the transporter are exposed to the reaction medium, binding to C109A was inhibited markedly by the same treatment, but X8C was resistant (Fig. 1B). It follows that one or more of the predicted internal cysteines replaced in X8C was responsible for the sensitivity of SERT to MTSEA. We examined mutants of X8C, each with one or two of the original cysteines reintroduced, to determine their sensitivity to MTSEA. Only X8C-357C was sensitive to MTSEA (Fig. 1B), suggesting that the presence of Cys-357 is sufficient to confer sensitivity to SERT. In intact cells, transport by X8C-357C was insensitive to the same treatment, suggesting that Cys-357 was located on the cytoplasmic face of SERT (Fig. 1A). To test whether Cys-357 was the only intracellular residue responsible for MTSEA inactivation, we generated a mutant of C109A in which Cys-357 was replaced with isoleucine (C109A/C357I). This mutant was insensitive to MTSEA both in intact cells (Fig. 1A) and in membranes (Fig. 1B), demonstrating that Cys-357 was the only residue contributing to the MTSEA sensitivity in membranes from cells expressing C109A.
The sensitivity of C109A, X8C, C109A/C357I, and X8C-357C to MTSEA inactivation is shown in Fig. 2. Essentially no inhibition was seen with X8C and C109A/C357I. However, C109A and X8C-357C were both markedly inactivated by concentrations of MTSEA above 0.1 mM. It is clear from these results that all of the sensitivity of C109A is accounted for by the cysteine at position 357. From the concentration of MTSEA required for half-maximal inactivation in 15 min, we estimated the pseudofirst order rate of inactivation to be 185 Ϯ 15 min Ϫ1 ⅐M Ϫ1 .
To evaluate the possibility that modification of Cys-357 by MTSEA inactivates binding by directly occluding the binding site, we examined the effect of incubating membrane preparations from cells expressing C109A with 5-HT or cocaine during the treatment with MTSEA. The results, shown in Fig. 3, demonstrate that these ligands markedly decreased the extent of inactivation. In the absence of ligand, 1.5 mM MTSEA inactivated over half of the binding activity in a 15-min incubation. When the incubation was performed in the presence of increasing concentrations of 5-HT or cocaine, the amount of residual activity progressively increased to over 75% of the control activity. In this experiment, the membranes were washed free of MTSEA, 5-HT, and cocaine prior to measuring binding with ␤-CIT, and the amount of inhibition is plotted as a percentage of controls treated similarly but without MTSEA. It is noteworthy that protection was not complete. We estimate that even in the presence of saturating concentrations of 5-HT or cocaine, MTSEA would have inactivated 17 Ϯ 2 or 21 Ϯ 2%, respectively. Although not shown, almost identical results were obtained with X8C-357C.
Protection by ligand binding might occur by direct steric blockade of Cys-357, or alternatively, binding might induce a conformational change that reduces the reactivity of Cys-357. The results shown in Fig. 4 suggest that the latter possibility is more likely. In this experiment, membranes from cells expressing C109A or X8C-357C were incubated with MTSEA in the presence or absence of Na ϩ and at room temperature or reduced temperature. At 4°C, the reaction rate for MTSEA inactivation is reduced, and to compensate, a higher MTSEA concentration was used, leading to greater inactivation. However, in each case, and for both C109A and X8C-357C, protection by 5-HT and cocaine was blocked by low temperature or Na ϩ removal. A clear increase in activity was observed for both constructs when 5-HT or cocaine was present during MTSEA treatment in Na ϩ at 25°C. However, at 4°C or in the absence of Na ϩ , no such increase was observed (Fig. 4). Previous experiments demonstrated that 5-HT and cocaine bind at 4°C and in the absence of Na ϩ (35, 36). Thus, bound 5-HT and cocaine altered the reactivity of Cys-357 only when Na ϩ was present and the temperature was permissive.
Since the reactivity of Cys-357 was sensitive to conformational changes induced by substrate and inhibitor binding, it was of interest to determine the effects of monovalent cations, some of which are involved in the proposed reaction cycle of SERT. Fig. 5 shows the remaining activity after a 15-min treatment of C109A and X8C-357C with 25 M MTSEA in media where all of the Na ϩ (150 mM) was replaced with the indicated cations. For both constructs, the most activity (least inactivation) was observed in the absence of alkali cations (NMDG). Although each of the alkali cations increased the extent of inactivation, they did so to various extents. Na ϩ had the least effect, and K ϩ had the greatest, with intermediate amounts of inactivation in the presence of Li ϩ , Cs ϩ , or Rb ϩ (Fig. 5). Incubation of membranes in these monovalent cation solutions in the absence of MTSEA had no effect on their binding activity subsequently measured in Na ϩ -containing binding buffer (data not shown).
Reversibility of Inactivation-Following inactivation by 1.5 mM MTSEA in membrane preparations, the C109A and X8C-357C mutants could not be reactivated by incubation with 12 mM free cysteine or 10 mM dithiothreitol for up to 90 min. Reactivation could not be observed even when Na ϩ was replaced with NMDG, K ϩ , Li ϩ , Cs ϩ , or Rb ϩ , or when 10 M 5-HT or 10 M cocaine was present during the incubation (data not shown). Similar treatments have been shown to reactivate some SERT mutants after inactivation with MTSET (36). DISCUSSION Previous results (28) uncovered a modest sensitivity of SERT C109A to MTSET inactivation of ␤-CIT binding activity. The inactivation was observed in membranes from cells expressing C109A, although transport in intact cells expressing the same SERT mutant was insensitive to externally added MTSET (27). The work presented here demonstrates that the sensitivity of SERT C109A membranes to MTS reagents is due to a single reactive residue, Cys-357, on the cytoplasmic face of SERT. When this cysteine was present, either in C109A or X8C-357C, binding was inactivated by MTSEA treatment of membranes ( Figs. 1 and 2). In contrast, when this residue was converted to Also shown are residual activity in the presence and absence of ligands when the inactivation was carried out at low temperature (Na ϩ , 4°C) with 5 mM MTSEA for 15 min or in the absence of Na ϩ (NMDG ϩ ). The left three sets of columns represent data from X8C-357C, and the right sets are from C109A. The asterisks indicate significant differences between samples containing only MTSEA and those also containing 5-HT or cocaine. buffer where all Na ϩ was replaced by NMDG ϩ , K ϩ , Li ϩ , Cs ϩ , or Rb ϩ . After incubation, the membranes were washed three times with Na ϩ -containing binding buffer, and ␤-CIT binding was subsequently measured. The asterisks indicate significant differences between inactivation in Na ϩ and in other conditions. isoleucine in C109A/C357I or X8C, binding was resistant to inactivation by MTSEA (Figs. 1 and 2). The same treatment of intact cells expressing these mutants did not inhibit transport (Fig. 1). This lack of sensitivity suggests that Cys-357 is accessible only from the cytoplasmic face of the membrane, which is more accessible in membrane preparations than in intact cells. Although MTSEA is known to cross biological membranes (34), here it apparently behaves like an impermeant reagent because the limited rate of MTSEA influx is less than the rate at which it reacts with competing intracellular thiols, such as glutathione. For the same reasons, we observed no labeling of intracellular SERT cysteine residues in a previous study with N-biotinylaminoethyl methanethiosulfonate (MTSEA-biotin) unless cells were permeabilized with digitonin (26).
Cys-357 was found to be the only residue conferring sensitivity to SERT. This contrasts to the case of DAT, where the corresponding cysteine residue, Cys-342, was one of two whose modification by MTS reagents led to inactivation of binding (29). The other sensitive residue in DAT was Cys-135 in IL1. Although that cysteine residue is conserved in SERT as Cys-155, SERT constructs containing Cys-155 but not Cys-357 were resistant to inactivation by MTSEA (mutants C109A-C357I and X8C-155C in Fig. 1). Aside from Cys-357, there are six endogenous cysteine residues predicted to lie on the cytoplasmic face of SERT at positions 15, 21, 147, 155, 522, and 622. Apparently, none of these cysteines are sites of inactivation by MTSEA. They may be inaccessible to the reagent, despite their predicted location, or if they do react, their modification by MTSEA apparently does not affect ␤-CIT binding to the transporter. Experiments are currently under way using biotinylation to evaluate these possibilities and to define further the internal topology of SERT.
At sufficient concentrations (above 2 mM, data not shown) MTSEA modification of Cys-357 leads to complete inactivation of ␤-CIT binding. The fact that both 5-HT and cocaine protect against the inactivation at concentrations close to their K D values for binding to SERT (Fig. 3) raises the possibility that Cys-357 is located in proximity to the ␤-CIT binding site, that its modification sterically interferes with binding, and that occupation of the site by 5-HT or cocaine blocks access of MTSEA to Cys-357. However, other evidence suggests that the binding site is formed from transmembrane domains (28,(37)(38)(39), and there are additional compelling reasons to reject this conclusion. Protection by 5-HT and cocaine against MTSEA inactivation does not occur in the absence of Na ϩ or at low temperature (Fig. 4). We know from previous studies that these conditions do not prevent binding (35,36). Therefore, it is not binding per se that prevents MTSEA modification of Cys-357, but rather a process, almost certainly a conformational change, that follows binding, requires Na ϩ , and is blocked at low temperature. Somehow, the Na ϩ -dependent changes that follow 5-HT or cocaine binding alter the accessibility of Cys-357 and possibly other neighboring residues in IL3. A likely consequence of this coupling between the binding site and IL3 is that modification of IL3 at Cys-357 distorts the binding site and thereby prevents high affinity ␤-CIT binding.
The rate at which Cys-357 reacts with MTSEA is lower than previously observed for some other residues in SERT. The pseudo-first order rate constant for modification of Cys-357 was 185 Ϯ 15 min Ϫ1 ⅐M Ϫ1 from the data presented in Fig. 2. By comparison, modification of SERT I179C by MTSET occurred with a rate of 782 min Ϫ1 ⅐M Ϫ1 (28). Furthermore, the rate of Cys-357 modification is affected by ligand and ion binding as discussed above. Finally, the modification rate was accelerated in the presence of alkali cations, particularly K ϩ (Fig. 5). These results strongly suggest that the accessibility of Cys-357 to MTSEA is limited and varies with the conformation of the transporter.
Consistent with the limited accessibility of Cys-357 is the observation that the MTSEA-modified transporter was not reactivated with free cysteine or dithiothreitol. The reaction with MTSEA generates a disulfide that should react with free sulfhydryl compounds to regenerate an unmodified cysteine residue, as we have observed with other SERT mutants (36). The lack of reactivation suggests that, after modification, the disulfide was inaccessible, either because the added 2-aminoethanethiol moiety blocked an already restricted access pathway to Cys-357 or because the modification induced a conformational change that occluded Cys-357. To evaluate whether the disulfide was inaccessible, we are currently using biotinylating MTS reagents to test the reversibility of labeling.
Thus, Cys-357 and, by extension, IL3 are in a flexible part of the transporter that is sensitive to the conformational changes that follow ligand and ion binding. It is linked to the substrate and inhibitor binding site as evidenced by the effect of ligand binding on Cys-357 accessibility and by the inhibition of binding after modification of Cys-357. Conformational changes are expected to be key steps in the transport cycle (5,22), and the variable accessibility of residues like Cys-357 and Ile-179 (28,36) suggests that they participate in those changes. It is therefore of special interest that K ϩ increased Cys-357 reactivity. Cytoplasmic K ϩ is thought to be countertransported by SERT in a step that follows 5-HT release to the cytoplasm (40,41). The effect of K ϩ on Cys-357 modification may result from conformational changes that follow K ϩ binding as SERT undergoes the countertransport step of its transport cycle.