The Role of External Loop Regions in Serotonin Transport
1999; Elsevier BV; Volume: 274; Issue: 51 Linguagem: Inglês
10.1074/jbc.274.51.36058
ISSN1083-351X
AutoresYoel Smicun, Scott Campbell, Marisa A. Chen, Howard H. Gu, Gary Rudnick,
Tópico(s)Analytical Chemistry and Chromatography
ResumoChimeric transporters were constructed in which the predicted external loops of the serotonin transporter (SERT) were replaced one at a time with a corresponding sequence from the norepinephrine transporter (NET). All of the chimeric transporters were expressed at levels equal to or greater than those of wild type SERT, but the transport and binding activity of the mutants varied greatly. In particular, mutants in which the NET sequence replaced external loops 4 or 6 of SERT had transport activity 5% or less than that of wild type, and the loop 5 replacement was essentially inactive. In some of these mutants, binding of a high affinity cocaine analog was less affected than transport, suggesting that the mutation had less effect on the initial binding steps in transport than on subsequent conformational changes. The more severely affected mutants also displayed an altered response to Na+. In contrast to the dramatic reduction in transport and binding, the specificity of ligand binding was essentially unchanged. Chimeric transporters did not gain affinity for dopamine, a NET substrate, or desipramine, an inhibitor, at the expense of affinity for serotonin or paroxetine, a selective SERT inhibitor. The results suggest that external loops are not the primary determinants of substrate and inhibitor binding sites. However, they are not merely passive structures connecting transmembrane segments but rather active elements responsible for maintaining the stability and conformational flexibility of the transporter. Chimeric transporters were constructed in which the predicted external loops of the serotonin transporter (SERT) were replaced one at a time with a corresponding sequence from the norepinephrine transporter (NET). All of the chimeric transporters were expressed at levels equal to or greater than those of wild type SERT, but the transport and binding activity of the mutants varied greatly. In particular, mutants in which the NET sequence replaced external loops 4 or 6 of SERT had transport activity 5% or less than that of wild type, and the loop 5 replacement was essentially inactive. In some of these mutants, binding of a high affinity cocaine analog was less affected than transport, suggesting that the mutation had less effect on the initial binding steps in transport than on subsequent conformational changes. The more severely affected mutants also displayed an altered response to Na+. In contrast to the dramatic reduction in transport and binding, the specificity of ligand binding was essentially unchanged. Chimeric transporters did not gain affinity for dopamine, a NET substrate, or desipramine, an inhibitor, at the expense of affinity for serotonin or paroxetine, a selective SERT inhibitor. The results suggest that external loops are not the primary determinants of substrate and inhibitor binding sites. However, they are not merely passive structures connecting transmembrane segments but rather active elements responsible for maintaining the stability and conformational flexibility of the transporter. the serotonin transporter 5-hydroxytryptamine (serotonin) transmembrane domain γ-aminobutyric acid the GABA transporter the glycine transporter internal loop external loop the norepinephrine transporter phosphate-buffered saline PBS containing 0.1 mm CaCl2 and 1 mm MgCl2 2β-carbomethoxy-3β-(4-[125I]iodophenyl)tropane [2-(trimethylammonium)ethyl]methanethiosulfonate The serotonin transporter (SERT)1 is responsible for the accumulation of serotonin (5-hydroxytryptamine, 5-HT) by neurons, platelets, and other cells (1Heisler S. Uvnas B. Acta Physiol. Scand. 1972; 86: 145-154Crossref PubMed Scopus (22) Google Scholar, 2Rudnick G. Holmsen H. Platelet Responses and Metabolism. II. CRC Press, Inc., Boca Raton, FL1986: 119-133Google Scholar, 3Greenberg B.D. McMahon F.J. Murphy D.L. Mol. Psychiatry. 1998; 3: 186-189Crossref PubMed Scopus (50) Google Scholar, 4Balkovetz D.F. Tiruppathi C. Leibach F.H. Mahesh V.B. Ganapathy V. J. Biol. Chem. 1989; 264: 2195-2198Abstract Full Text PDF PubMed Google Scholar). It is a member of a large family of carriers that couple the uphill movement of neurotransmitters and other metabolites with the downhill flux of Na+ and Cl− (5Rudnick G. Reith M. Neurotransmitter Transporters: Structure, Function, and Regulation. Humana Press Inc., Totowa, NJ1997: 73-100Crossref Google Scholar). The strong homology within this family suggests that the mechanism of transport is similar and that structural elements conserved within the family perform similar functions in each transporter. A prominent feature in each of the primary sequences is the presence of 12 regions rich in hydrophobic amino acids, linked by hydrophilic regions of variable length (6Blakely R. Berson H. Fremeau R. Caron M. Peek M. Prince H. Bradely C. Nature. 1991; 354: 66-70Crossref PubMed Scopus (687) Google Scholar, 7Hoffman B.J. Mezey E. Brownstein M.J. Science. 1991; 254: 579-580Crossref PubMed Scopus (508) Google Scholar). The hydrophobic regions were originally modeled as membrane-spanning domains linked by alternating external and internal loops (6Blakely R. Berson H. Fremeau R. Caron M. Peek M. Prince H. Bradely C. Nature. 1991; 354: 66-70Crossref PubMed Scopus (687) Google Scholar, 7Hoffman B.J. Mezey E. Brownstein M.J. Science. 1991; 254: 579-580Crossref PubMed Scopus (508) Google Scholar). One of these loops, which links putative transmembrane domains 3 and 4 (TM3 and TM4), is much larger than the rest and contains consensus sites forN-linked glycosylation (8Tate C. Blakely R. J. Biol. Chem. 1994; 269: 26303-26310Abstract Full Text PDF PubMed Google Scholar) and an intramolecular disulfide (9Chen J.G. Liu-Chen S. Rudnick G. Biochemistry. 1997; 36: 1479-1486Crossref PubMed Scopus (150) Google Scholar) in all family members. The original topological model for these NaCl-coupled transporters was challenged in studies using mutants of the γ-aminobutyric acid (GABA) and glycine transporters (GAT-1 and GLYT-1) (10Bennett E.R. Kanner B.I. J. Biol. Chem. 1997; 272: 1203-1210Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar, 11Olivares L. Aragon C. Gimenez C. Zafra F. J. Biol. Chem. 1997; 272: 1211-1217Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar). These studies suggested that the predicted first external loop (EL1) was not exposed on the cell exterior and that the first internal loop (IL1) was actually extracellular. One drawback of these studies was that the conclusions about EL1 and IL1 topology were drawn from the properties of inactive mutants. Subsequent studies with active mutants of SERT supported the original topology in which EL1 was external and IL1 was internal (12Chen J.G. Liu-Chen S. Rudnick G. J. Biol. Chem. 1998; 273: 12675-12681Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar). The mechanism usually considered to account for carrier-mediated transport depends on alternating access of bound substrate to the external and internal sides of the membrane (13Mitchell P. Res. Microbiol. 1990; 141: 286-289Crossref PubMed Scopus (34) Google Scholar). According to this model, the substrate must be bound in a way that allows it to be exposed to either the external medium or the cytoplasm. It has been proposed that transmembrane domains form the binding site and that conformational changes of the protein alternately expose this site to the two sides of the membrane. Evidence from cysteine-scanning mutagenesis of SERT TM3 supports the proposal that the binding site is formed by transmembrane domains. Two cysteine replacement mutants in TM3, I172C, and Y176C behave as if the inserted cysteine is located in close proximity to bound serotonin (14Chen J.G. Sachpatzidis A. Rudnick G. J. Biol. Chem. 1997; 272: 28321-28327Abstract Full Text Full Text PDF PubMed Scopus (167) Google Scholar). The function of external loop regions has not been established. In the GABA transporter subfamily, residues in EL5 were shown to participate in the ability of the transporter to discriminate between GABA and β-alanine as substrates (15Tamura S. Nelson H. Tamura A. Nelson N. J. Biol. Chem. 1995; 270: 28712-28715Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar). Residues in this loop were also implicated in the uncoupled ion flux mediated by SERT (16Cao Y. Li M. Mager S. Lester H.A. J. Neurosci. 1998; 18: 7739-7749Crossref PubMed Google Scholar). In EL2 of SERT, however, no change in substrate or inhibitor sensitivity was observed when a large portion was replaced with the corresponding sequence from the closely related norepinephrine transporter (NET), although the chimeric transporter was severely defective in transport activity (17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar). This study suggested that at least some external loop regions were involved in conformational changes required for transport and did not participate in substrate and inhibitor binding. In the present study, we have extended this approach by performing loop-scanning mutagenesis on the extracellular domain of SERT. Each predicted external loop was converted to the corresponding sequence in NET, which is the transporter closest to SERT in sequence. The chimeric mutants were tested for transport and binding activity, surface expression, and sensitivity to specific SERT and NET inhibitors. The results suggest that residues responsible for the selectivity of SERT and NET inhibitors are not located in external loops. However, many of the mutations exhibited markedly decreased intrinsic transport activity, suggesting that they participated in the conformational changes required for turnover. Transfected HeLa cells (one 75-cm2 flask) expressing SERT or SERT-NET chimeras were incubated overnight, washed once with 5 ml of binding buffer (BB, 300 mm NaCl containing 10 mm HEPES buffer adjusted with LiOH to pH 8.0) at 25 °C and then scraped off the flask into 1 ml of ice-cold BB with protease inhibitors (10 μg/ml each leupeptin and pepstatin A and 100 μm α-toluenesulfonyl fluoride) and transferred to a 12 × 75-mm polystyrene tube. The flask was rinsed with an additional 2 ml of ice-cold BB with protease inhibitors and combined with the first suspension. Clumps of cells were broken up by pipetting the suspension up and down. The suspension was then sonicated on ice using a Virtis Virsonic 50 at a power level of 50–60% for 20–30 s. Sonication was repeated once after the suspension had cooled. The membranes were collected by centrifugation at 44,000 ×g for 20 min at 4 °C and resuspended thoroughly in 0.5 ml of BB plus protease inhibitors. This membrane preparation was used fresh or stored at −80 °C until use. The SERT mutant Q111K (SNEL1′) was generated by site-directed mutagenesis (12Chen J.G. Liu-Chen S. Rudnick G. J. Biol. Chem. 1998; 273: 12675-12681Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar). Using this mutant, Ile-108 was then replaced with leucine using primer A (see below) to yield SNEL1. NET DNA was amplified with primers B and C (see below), which contained restriction sites for RsrII andPpuMI, respectively, and the product was digested with these enzymes. The digested fragment was then ligated into epitope-tagged wild type SERT (prSTag) that had the corresponding fragment removed by treatment with RsrII and PpuMI. Using site-directed mutagenesis, BsmI andPpuMI sites were introduced into NET at the positions corresponding to those same unique sites in SERT using the primers C and D, which include the KYSKYK sequence of NET. TheBsmI-PpuMI fragment was cut out and used to replace the corresponding fragment in prSTag. NET DNA was amplified using polymerase chain reaction with primers E and C, which contained restriction sites forBsmI and PpuMI, respectively. The resulting product was digested with these two enzymes and ligated into prSTag that had the corresponding fragment removed by treatment withBsmI and PpuMI. prSTag was cut with AlwNI, the 1877-base pair fragment was further digested with BglII, and theAlwNI-BglII fragment was retained along with the 3055-base pair AlwNI fragment. A BglII site was introduced into pNET by site-directed mutagenesis using the primer F, and a 60-base pair pNET fragment was created by digestion withBspMI and BglII. This NET fragment and the two SERT fragments were ligated with a BbsI-BspMI linker (sense, G; antisense, H) to generate SNEL3′. The SERT sequence WRV, near the border of EL3 with TM5, remained in this chimera. SNEL3′ was further mutated by site-directed mutagenesis with primer I to convert the WR sequence to the NET sequence SN to yield the construct SNEL3′′, and the Val was then replaced with an isoleucine residue by mutagenesis with primer J to yield the full SERT-NET EL3 chimera, SNEL3. Restriction sites for AvrII andBsu36I were introduced into the borders of EL4 in both SERT (primers K and L) and NET (primers M and O) by site-directed mutagenesis, and the AvrII-Bsu36I fragment from NET was used to replace the corresponding region of SERT. Since the SNEL4 chimera was shorter than SERT by 1 residue in EL4, an alanine residue was introduced into SNEL4 between Glu-400 and Gly-401 by site-directed mutagenesis using primer P. These chimeric transporters were created by site-directed mutagenesis using mutagenic primers Q and R, respectively, containing the entire NET loop sequence flanked by SERT sequences. Oligonucleotides: A 5′-GCGGTTTCCTTACCTATGCTACAAGAATGGCB 5′-CTCACGGACCGCCTGCCCTGGACCGACTGTGC 5′-TAGGTCCTGAATCCCGCTCGTCTCGTGD 5′-GCTGCATTCCAAGTACTCCAAGTACAAGTTCACGE 5′-CTGCATTCCTTCACGCCGGCAGCCGAGTTTTATGF 5′-GCCGCAACTCAGATCTTTTTTTCCTTGGGGGCTGGG 5′-GGGGTCAATGCCTACCTGCACATH 5′-GTCGATGTGCAGGTAGGCATTGAI 5′-CCTTCCTGGAGCTTCGAACGGGGTCAATGCCJ 5′-CCTTCCTGGAGCTTCGAACGGGATCAATGCCTACCK 5′-GTCATCTTCACGGTCCTAGGCTACATGGCGGL 5′-CCTCTTCATCACGTATCCTGAGGCAATAGCCAACM 5′-CCATCTTCTCCATCCTAGGTTACATGGCCCN 5′-CATCCTGTATCCTGAGGCCATTTCTACCCO 5′-GAGGATGTGGCAACAGAAGCAGGAGCTGGCCP 5′-CCTGCTCACACTGACCTCTGGAGGGATCTATGTACTGACTCTGCTAGACACGTTCGCCACGGGGCCAGCQ 5′-CTGATGAGCCCACCCCTGACATACGATGACTACATCTTTCCACCTTGGGCGATCGTCTTGGGC Mutations were introduced using the Chameleon and QuickChange kits (Stratagene, La Jolla, CA). For the modifications using QuickChange, each of the primers was used together with an antisense primer. In all cases, the mutated regions were subcloned into the original plasmid using flanking restriction sites and sequenced to ensure that no unwanted modifications were introduced. Cell surface expression of the transporters was determined using the membrane-impermeable biotinylation reagent, NHS-SS-biotin (Pierce) by a modification of the procedure of Gottardi et al. (18Gottardi C. Dunbar L. Caplan M. Am. J. Physiol. 1995; 37: F285-F295Crossref Google Scholar) as described previously (9Chen J.G. Liu-Chen S. Rudnick G. Biochemistry. 1997; 36: 1479-1486Crossref PubMed Scopus (150) Google Scholar). Briefly, cells were labeled with NHS-SS-biotin, the excess reagent was quenched, and the cells were solubilized. Cell surface proteins were isolated from the cell extract with immobilized streptavidin, and transporter was detected in the pool of surface proteins by gel electrophoresis and Western blotting using an anti-FLAG antibody (Sigma-Aldrich). Immunoblots were quantitated using an Alpha Innotech IS-1000. Nonlinear regression fits of experimental and calculated data were performed with Origin (Microcal Software, Northampton, MA), which uses the Marquardt-Levenberg nonlinear least squares curve-fitting algorithm. The statistical analysis was done with data from single experiments. All the experiments were repeated a total of 2–4 times and in all cases gave essentially the same results. Unless otherwise indicated, data with error bars represent the mean ± S.D. for duplicate samples. The expression system used has been described in detail elsewhere (17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar, 19Blakely R.D. Clark J.A. Rudnick G. Amara S.G. Anal. Biochem. 1991; 194: 302-308Crossref PubMed Scopus (151) Google Scholar). Briefly, the plasmid prSTag and all the mutants used here contain a promoter for T7 RNA polymerase upstream from the rat brain SERT cDNA. This promoter was used to express wild type and mutant SERTs in the vaccinia/T7 polymerase/HeLa cell system. HeLa cells were plated in 48-well plates at 50% confluency and allowed to grow overnight in Dulbecco's modified Eagle medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (Gemini Bio-Products, Inc.; Calabasas, CA), 100 units/ml penicillin G, and 100 μg/ml streptomycin sulfate (Life Technologies, Inc.). The next day they were infected with a vaccinia virus strain, VTF-7 (19Blakely R.D. Clark J.A. Rudnick G. Amara S.G. Anal. Biochem. 1991; 194: 302-308Crossref PubMed Scopus (151) Google Scholar), which makes T7 RNA polymerase (added to the cells in 40 μl/well of Dulbecco's modified Eagle medium without serum). After a 15-min incubation with the virus at 37 °C, the cells were transiently transfected with wild type and mutant plasmids (400 ng of plasmid DNA and 1.2 μl of Lipofectin (Life Technologies, Inc.) per well in 80 μl of Dulbecco's modified Eagle medium without serum). Wild type SERT, no plasmid (mock-transfected) controls, and mutants were each transfected in duplicate wells. [3H]Serotonin transport assays were carried out the next day between 19 and 24 h post-infection. The cells were washed with 250 μl of phosphate-buffered saline (PBS), 137 mm NaCl, 2.7 mm KCl, 4.3 mmNa2HPO4, and 1.4 mmKH2PO4, pH 7.3) containing 0.1 mmCaCl2 and 1 mm MgCl2 (PBSCM). Transport was measured by incubating the cells with 14.6 nm n-[1,2-3H]serotonin (NEN Life Science Products; NET-498) in 80 μl of PBSCM for 15 min at room temperature, an interval previously determined to include only the initial, linear phase of transport. For determination of V max, a range of 5-HT concentrations was generated by the addition of unlabeled 5-HT. Each well was washed very quickly three times with 320 μl of ice-cold PBS. The cells were lysed in 100 μl of 1% SDS, transferred to scintillation vials, and counted in 3-ml of Optifluor (Packard Instrument Co.). The transport activity of each mutant was tested at least three different times, in duplicate wells each time, and the results were averaged. Binding of the high affinity cocaine analog, 2β-carbomethoxy-3β-(4-[125I]iodophenyl)tropane (β-CIT), was measured in isolated membranes. The membrane suspension was thawed on ice, and a sample of 10 μl was diluted into 50 μl of binding buffer (300 mm NaCl, 10 mm HEPES, adjusted to pH 8.0 with LiOH) containing 0.07 nmβ-[125I]CIT unless indicated otherwise. For measurements of Na+-dependent β-CIT binding, the NaCl concentration was varied, and LiCl was added to maintain constant osmolarity. Binding was allowed to approach equilibrium by incubation for 1 h at room temperature. The samples then were diluted rapidly with 1 ml of binding buffer and filtered through a #32 glass fiber filter (Schleicher & Schuell) previously soaked in 0.3% polythyleneimine. The filters were washed twice with 1 ml of binding buffer, placed in the scintillation vials containing 3 ml of Optifluor (Packard), and counted after 3 h. To evaluate the involvement of external loop domains on the function of SERT, we generated a series of chimeric transporters in which residues in external loops were replaced with the corresponding residues from the norepinephrine transporter (NET). Fig. 1 shows the external loop regions and the residues (black circles) that are identical in SERT and NET. We had previously replaced a portion of the second external loop (EL2) with corresponding NET sequence (17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar). In the current work, we replaced the remainder of EL2 with NET sequence. NET EL2 is longer than SERT EL2 by six residues. We, therefore, also generated a chimera in which six residues of NET sequence were removed. Although the sequence similarity between SERT and NET is low in EL2, there is a notable lysine-rich sequence, KYSKYK, present in NET and not in SERT. The NET sequence in chimera SNEL2b begins just before (and includes) that lysine-rich sequence, and SNEL2b′ begins just after the KYSKYK and therefore has the same EL2 length as SERT (Fig. 1). Another EL2 chimera containing 29 residues of NET sequence replacing residues 194 through 222 of SERT (SNEL2a) (17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar) was adjusted by removing the sequence PKLLNG from the NET region. One other chimera was also adjusted so that the total loop length was the same as that of SERT. SNEL4′ contains one SERT residue (Ala-401) added back to the NET sequence to keep the loop length equal to that in SERT. Table I shows a list of chimeras and their locations within the primary sequence.Table ISERT-NET chimeric transportersChimeraNET sequenceRemarksSNEL1108–115Only Ile-108 and Gln-111SNEL1′Q111KSNEL2194–253Entire loopSNEL2b223–253Includes KYSKYKSNEL2b′223–247KYSKYK removedSNEL3303–331SNEL3′310–331SNEL3″303–308, 310–331SNEL4384–411SNEL4′384–412Ala-401 addedSNEL5488–494SNEL6561–574 Open table in a new tab Transport activity of the various chimeras transiently expressed in HeLa cells varied from close to that of wild type SERT to completely inactive. The sensitivity of transport activity varied with the position of the NET sequence replacement. Replacement of EL4, EL5, or EL6 led to complete loss of transport activity, but replacements in some other regions had a much smaller effect on transport (Table II). Significant amounts of activity were retained in chimeras where most or all of EL1, EL2, or EL3 was replaced with NET sequence. The location of the substitution was apparently more important than the divergence in sequence. For example, the replacements in EL3 and EL4 were similar in the number of residues replaced but very different in activity. In some cases, we were able to identify specific residues that led to activity loss. For example Ile-108 in EL1 and Try-306–Arg-307 in EL3 accounted for most of the activity loss in SNEL1 and SNEL3 (Table II).Table IITransport and binding activity of SERT-NET chimerasChimeraTransport activityBinding activityWild type100 ± 5100 ± 27SNEL144 ± 866 ± 13SNEL1′80 ± 10SNEL24 ± 239 ± 12SNEL2b6 ± 2.5SNEL2b′30 ± 4.4SNEL338 ± 1050 ± 12SNEL3′69 ± 17SNEL3″42 ± 5SNEL46 ± 32 ± 1SNEL4′1.1 ± 1SNEL51.25 ± 111 ± 6SNEL65.8 ± 483 ± 19 Open table in a new tab The length of the loop replacement also was not the critical factor determining the activity. In EL2 and EL4, the NET loop size differs from SERT. However, in both cases, size was not the only factor leading to activity loss. Although EL2 in SNEL2b′ lacked the KYSKYK sequence and was the same size as in SERT, that chimera had only about 30% wild type activity. This was significantly more than chimera SNEL2b, which contained the KYSKYK sequence, although removal of the NET sequence PKLLNG from SNEL2a (17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar) to make EL2 identical in length to SERT had no effect on activity (data not shown). Likewise, the inactivity of chimera SNEL4 does not result from the fact that EL4 is shorter in that mutant by one residue, since adding back the missing residue in SNEL4′ failed to restore activity. Finally, SNEL5 and SNEL6 lacked transport activity even though the loop length was unchanged. To test for the possibility that inactive mutants were misfolded, we prepared membranes from cells expressing the mutants and tested their ability to bind the high affinity cocaine analog β-CIT. Chimeric transporters with functional binding sites should bind β-CIT even if the protein is retained in intracellular pools or is prevented from catalyzing rapid transport. The results in Table II indicate that some of the chimeras, even those with very low transport activity, retained the ability to bind β-CIT with high affinity. In particular, chimera SNEL6, like SNEL2a (17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar), retained wild type binding activity, although it was not functional for transport. In SNEL4 and SNEL5, binding was so weak as to prevent accurate measurements of ligand affinity, but binding in the other chimeras was closer to wild type levels than was transport (Table II). Of all the loop replacements, the short sequence of seven residues in EL5 had the most dramatic effect on transporter function when replaced by the corresponding NET sequence, which differs only in three residues. To assess surface expression of the mutants, we treated cells expressing each chimeric transporter with NHS-SS-biotin, an impermeant reagent that reacts with exposed lysine residues and is known to label SERT from the cell exterior (12Chen J.G. Liu-Chen S. Rudnick G. J. Biol. Chem. 1998; 273: 12675-12681Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar). The labeled cells were solubilized, and biotinylated proteins were isolated using immobilized streptavidin as described (12Chen J.G. Liu-Chen S. Rudnick G. J. Biol. Chem. 1998; 273: 12675-12681Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar, 17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar). The abundance of SERT in the biotinylated cell surface fraction was determined by Western blotting, as shown in Fig. 2. The results clearly indicate that each of these chimeric transporters is delivered to the cell surface at least as efficiently as wild type SERT. The defects observed in transport by these mutants, therefore, are due to an intrinsic inability to catalyze efficient 5-HT translocation, rather than an inability of the protein to reach the plasma membrane. To determine whether the loss of transport or binding activity was due to altered ion coupling, we examined the Na+ ion dependence of these activities in chimeras SNEL1, SNEL2, SNEL3, and SNEL6. The Na+ K m for transport in SNEL1 and SNEL3 was moderately increased relative to the wild type prSTag from 17 ± 3 mm to 70 ± 19 and 58 ± 14 mm, respectively (Fig. 3 A). In SNEL2 and SNEL6, we determined the Na+ dependence of β-CIT binding, since these mutants were unable to transport. Fig. 3 B shows that in wild type SERT, as described previously (20Humphreys C.J. Wall S.C. Rudnick G. Biochemistry. 1994; 33: 9118-9125Crossref PubMed Scopus (55) Google Scholar), β-CIT binding does not absolutely require Na+. In SNEL2, however, no binding was observed when Li+ replaced Na+, and the effect of Na+ was not maximal even at 300 mm. Chimera SNEL6 was intermediate between these two. Both chimeras demonstrated a greater dependence on Na+(or inhibition by Li+) than the wild type and required higher [Na+] for saturation, as estimated by fitting the binding data. To determine if the substrate and inhibitor selectivity of the transporter is determined by external loops, we tested the ability of two substrates, 5-HT and dopamine, and two inhibitors, paroxetine and desipramine, to bind to the chimeric transporters. Dopamine is a good substrate for NET and a poor substrate for SERT, and 5-HT is similarly selective for SERT. Fig. 4 shows that dopamine was not an effective inhibitor of β-CIT binding in SERT or any of the chimeric constructs. 5-HT, however, displaced β-CIT from each chimera at approximately the same concentration that was effective in wild type SERT (Fig. 4). Paroxetine is one of the most selective inhibitors of SERT, having little affinity for NET, whereas desipramine binds tightly to NET and less well to SERT. Fig. 5 shows results in which the potency of desipramine and paroxetine for β-CIT displacement was tested. Table III lists theK I values of 5-HT, dopamine, paroxetine, and desipramine for inhibition of β-CIT binding to the SERT-NET chimeras. In addition, we have calculated an index of relative SERT versus NET affinity from the inhibitor dissociation constants. None of the chimeric mutants showed a major decrease in paroxetine affinity and a corresponding increase in desipramine affinity as would be expected if any external loop significantly contributed to the inhibitor binding site (Fig. 4). The SERT/NET affinity index did not dramatically decrease for any of the chimeras (Table III). Thus, it appears as if residues in the external loops of SERT do not contribute significantly to the selectivity of substrate and inhibitor binding.FIG. 5Displacement of β-CIT from SERT-NET chimeras by paroxetine and desipramine. Equilibrium binding measurements were performed as in Fig. 4 using the selective inhibitors paroxetine and desipramine to displace β-CIT.View Large Image Figure ViewerDownload (PPT)Table IIIAffinity of chimeras for 5-HT, paroxetine, and desipramineChimera5-HTParoxetineDesipramineSERT/NETμmnmnmWild type0.46 ± 0.020.76 ± 0.15226 ± 161SNEL10.49 ± 0.050.97 ± 0.17125 ± 50.43SNEL20.50 ± 0.051.02 ± 0.39261 ± 200.86SNEL30.33 ± 0.040.49 ± 0.13161 ± 201.10SNEL60.71 ± 0.060.72 ± 0.2793 ± 220.43 Open table in a new tab From an initial analysis of the SERT sequence, it was postulated that the polypeptide chain contained six extracellular hydrophilic loops (6Blakely R. Berson H. Fremeau R. Caron M. Peek M. Prince H. Bradely C. Nature. 1991; 354: 66-70Crossref PubMed Scopus (687) Google Scholar, 7Hoffman B.J. Mezey E. Brownstein M.J. Science. 1991; 254: 579-580Crossref PubMed Scopus (508) Google Scholar). Recent studies with impermeant reagents have demonstrated that residues in each of the predicted external loops are exposed to the cell exterior, largely validating the initial structural model (12Chen J.G. Liu-Chen S. Rudnick G. J. Biol. Chem. 1998; 273: 12675-12681Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar). According to the alternating access model for transport (13Mitchell P. Res. Microbiol. 1990; 141: 286-289Crossref PubMed Scopus (34) Google Scholar), the substrate binding site is likely to be made from the side chains of residues in transmembrane domains and not internal or external loops. The function of these loops, then, might be completely passive, serving only to connect transmembrane segments. However, the loops also might serve to coordinate the conformational changes required for transport or even participate in forming the substrate binding site. Previous attempts to understand neurotransmitter transporters using chimeras have had mixed results. Sufficient differences exist between NET and SERT so that chimeras containing transmembrane domains from both proteins were inactive (21Moore K.R. Blakely R. Biotechniques. 1994; 17: 130-136PubMed Google Scholar), although substitution of either the NH2- or COOH-terminal hydrophilic domain was tolerated well (22Blakely R. Moore K. Qian Y. Soc. Gen. Physiol. Ser. 1993; 48: 283-300PubMed Google Scholar). The higher extent of sequence identity between NET and the dopamine transporter allowed the formation of functional chimeras between these two proteins (23Buck K. Amara S. Soc. Neurosci. Abstr. 1993; 19: 40.12Google Scholar, 24Giros B. Wang Y. Suter S. McLeskey S. Pifl C. Caron M. J. Biol. Chem. 1994; 269: 15985-15988Abstract Full Text PDF PubMed Google Scholar, 25Buck K. Amara S. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 12584-12588Crossref PubMed Scopus (194) Google Scholar, 26Buck K.J. Amara S.G. Mol. Pharmacol. 1995; 48: 1030-1037PubMed Google Scholar). Although these studies suggested that certain regions of the sequence were responsible for substrate and inhibitor binding, it was not possible to distinguish between transmembrane and loop regions. In the GABA transporter subfamily, residues in EL5 were shown to contribute to substrate specificity (15Tamura S. Nelson H. Tamura A. Nelson N. J. Biol. Chem. 1995; 270: 28712-28715Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar), and residues in EL5 were also implicated in the uncoupled ion flux carried by SERT (16Cao Y. Li M. Mager S. Lester H.A. J. Neurosci. 1998; 18: 7739-7749Crossref PubMed Google Scholar). However, in EL2 of SERT, partial substitution with NET sequence did not affect substrate or inhibitor sensitivity (17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar). The present work extends the NET substitution study and represents the first attempt to scan the entire extracellular face of a neurotransmitter transporter. The results of this approach allow several conclusions about the role of extracellular loops in SERT function. 1) These loops are not merely passive elements connecting transmembrane domains. 2) The loops do not represent a major determinant of binding selectivity for substrates and inhibitors. 3) Certain external loops, especially EL4 and EL5, are extremely sensitive to modification, as if they are critical for proper assembly or stability of SERT. Although it is possible to generalize about some of the consequences of replacing SERT external loops with NET sequence, such as the retention of selectivity for SERT inhibitors and substrates, it is more difficult to make general conclusions about the role of these loops in transport. It is likely that many of these loops participate in conformational changes that occur during transport. Some of the loop replacements lead to dramatically lower transport activity but only moderately reduced binding, and this is not due to lower levels of cell surface expression. Such behavior would be expected if the initial steps in transport, namely binding, were less perturbed than subsequent steps involving conformational changes. At least part of the decreased activity observed with the chimeric mutants may be related to their ionic requirements. In particular, replacing Na+ with Li+ caused a loss of binding activity in SNEL2 and SNEL6 that was not observed in the wild type transporter. Perturbation of the transporter structure may lead to increased Na+ dependence or increased Li+sensitivity. In addition to the results with SNEL2 and SNEL6, the Na+ K m for transport was increased in SNEL1 and SNEL3. It is unlikely that any one of these loops serve as a specific Na+ binding site since chimeras from loops far apart in the primary sequence all have similar effects on the ionic dependence. Replacement of Na+ with Li+ also increases the reactivity of Cys-109 in EL1. This may represent a structural deformation that is tolerated better in the wild type transporter than in these loop chimeras. Alternatively, if some loop chimeras, such as SNEL2 and SNEL6, are limited in their ability to undergo conformational changes, then the exposure of Cys-109 in Li+ might also be limited. Future experiments will address this possibility. By default, if the substrate binding site is not formed by external loops, then transmembrane domains are most likely to contribute to the binding pocket. In previous work (14Chen J.G. Sachpatzidis A. Rudnick G. J. Biol. Chem. 1997; 272: 28321-28327Abstract Full Text Full Text PDF PubMed Scopus (167) Google Scholar), we demonstrated that two residues in TM3, Ile-172 and Tyr-176, are close to the binding site for 5-HT and cocaine. In cysteine replacement mutants, these positions, although predicted to be deep in the membrane interior, were sites of inactivation by the hydrophilic external reagent MTSET ([2-(trimethylammonium)ethyl]methanethiosulfonate). Moreover, the transporter was protected from inactivation by 5-HT and cocaine, and after modification by MTSET the transporter was unable to bind β-CIT (14Chen J.G. Sachpatzidis A. Rudnick G. J. Biol. Chem. 1997; 272: 28321-28327Abstract Full Text Full Text PDF PubMed Scopus (167) Google Scholar). These are characteristics expected for residues in a binding pocket. Taken together with the current work, these experiments suggest that the complimentary functions of binding and conformational changes in the transport cycle are subserved by the transmembrane and loop domains, respectively, of SERT. In the present work, we found that substitution of SERT loops EL4, EL5, or EL6 with NET sequence led to a dramatic loss in activity. Of these, perhaps the most interesting is EL5, where only three residues differ. Tamura et al. (15Tamura S. Nelson H. Tamura A. Nelson N. J. Biol. Chem. 1995; 270: 28712-28715Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar) showed that, in the GABA transporter family, conversion of EL5 residues from GAT-1 to GAT-3 sequence led to the ability to bind β-alanine as an inhibitor. They postulated that EL4, EL5, and EL6 formed a binding site for GABA and other substrates in the GABA transporter family. Previous work by Cao et al.(16Cao Y. Li M. Mager S. Lester H.A. J. Neurosci. 1998; 18: 7739-7749Crossref PubMed Google Scholar) on SERT EL5 residues indicated that the uncoupled ion flux carried by SERT was influenced by this loop and proposed that EL5 contributed to an external gate. The results of Tamura et al. (15Tamura S. Nelson H. Tamura A. Nelson N. J. Biol. Chem. 1995; 270: 28712-28715Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar) with GAT-1 and GAT-3 agree with the present results and those of Caoet al. (16Cao Y. Li M. Mager S. Lester H.A. J. Neurosci. 1998; 18: 7739-7749Crossref PubMed Google Scholar) with SERT, in that they all point to an important role for EL5. The mutations in GAT-1 EL5 led to a change in substrate selectivity (15Tamura S. Nelson H. Tamura A. Nelson N. J. Biol. Chem. 1995; 270: 28712-28715Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar), whereas changes in SERT EL5 resulted in a change in uncoupled flux (16Cao Y. Li M. Mager S. Lester H.A. J. Neurosci. 1998; 18: 7739-7749Crossref PubMed Google Scholar) or loss of activity with no detectable change in selectivity (this work). A more detailed investigation of this interesting region of the transporter seems warranted. It should be emphasized that the NET sequences used in these loop chimeras were fully functional in the background of NET. If these sequences behaved as interchangeable parts, they would be expected to function also in SERT. The fact that so many of the chimeras were defective in transport activity indicates that the loop sequences function by interaction with other parts of the protein. NET loop sequences apparently did not fit properly in the SERT background, leading to a defect in transport function. We tested the possibility that these required interactions occur within a single loop by combining the SNEL2a chimera (17Stephan M.M. Chen M.A. Penado K.M.Y. Rudnick G. Biochemistry. 1997; 36: 1322-1328Crossref PubMed Scopus (66) Google Scholar) with the SNEL2b chimera to generate a chimera (SNEL2) in which the entire EL2 consisted of NET sequence. Rather than increasing activity, SNEL2 activity was lower than that of either SNEL2a or SNEL2b alone (Table II). We conclude that interactions between the loops or between loops and transmembrane domains are important for the proper function of SERT. The authors thank Xiaohong Wu for expert technical assistance.
Referência(s)