The serotonin transporter (SERT) is a member of a highly homologous family of proteins responsible for the reuptake of biogenic amines from the synaptic cleft. We took advantage of native restriction sites in SERT to construct a chimeric transporter containing a small (34 amino acid) region of the norepinephrine transporter. The substituted region corresponds to about half of the largest extracellular loop. This chimera transports serotonin very slowly compared to wild type SERT. However, it binds serotonin and the cocaine analog 2beta-carbomethoxy-3beta-(4-[125I]iodophenyl)tropane with a high affinity indistinguishable from wild type. It has the same specificity as wild type SERT for the antidepressants paroxetine and desipramine. The low rate of transport does not appear to be due to poor expression, since the chimeric transporter is expressed at the membrane surface at close to wild type levels as measured by cell surface biotinylation. These observations lead us to conclude that, rather than playing a role in substrate or drug binding, this region of the large extracellular loop may be involved in the conformational changes associated with substrate translocation into the cell.

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 (SERT)1 is responsible for the accumulation of serotonin (5-hydroxytryptamine, 5-HT) by neurons, platelets, and other cells (1-4). 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 ϩ and Cl Ϫ (5). 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 (6, 7). The hydrophobic regions were originally modeled as membrane-spanning domains linked by alternating external and internal loops (6, 7). 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 for N-linked glycosylation (8) and an intramolecular disulfide (9) in all family members.The original topological model for these NaCl-coupled transporters was challenged in studies using mutants of the ␥-aminobutyric acid (GABA) and glycine transporters (GAT-1 and GLYT-1) (10, 11). 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 prope...