Fig 1.
Schematic representation of key steps in the physiological transport cycle of hSERT.
The outward-open conformation of hSERT bound to Na+ and Cl−is entered by a second Na+ ion and the endogenous substrate 5HT+. Once all these required components are bound, the extracellular gates can close, allowing the transporter to undergo progressive conformational changes through the outward-occluded, fully-occluded, and inward-occluded transporter states. Once the inward-open state is reached, one Na+ ion and 5HT+ are released into the cytosol. Upon binding of a K+ ion, the transporter resets via equivalent conformational changes, but in reverse order, finishing its cycle by release of K+ into the synaptic cleft.
Fig 2.
Transmembrane topology of hSERT comprising twelve membrane-spanning helices connected by intra- and extracellular loops (ILs, ELs).
The inverted-topology structural repeats are highlighted by orange (TM1-5) and blue (TM6-10) triangles. The “rocking bundle” proposed to facilitate substrate transport consists of TM1, TM2, TM6, and TM7.
Fig 3.
(A, B) The two template structures used were outward-open hSERT (A, PDB ID: 5I71) and outward-occluded LeuTAa (B, PDB ID: 3F48). The extracellular bundle components whose conformations define the occlusion of the extracellular pathway in NSS transporters are shown in green (hSERT) and brown (LeuTAa). C) To reposition the outward-open hSERT bundle segments according to their expected locations in the outward-occluded state, the former were individually superposed onto the corresponding elements of LeuT by structural alignment. D) The resultant modeling template consists of the unchanged segments (black) plus the now outward-occluded, repositioned bundle segments (cyan), both originate from hSERT itself.
Fig 4.
Plot of the local ProQM structural model quality scores for each residue in the X-ray structure (hSERT 5I71, cyan) and the 16 best ranked models (average in black, standard errors in gray). The ProQM plot shows a clear quality comparability between experimental and modeled structures and a striking consistency amongst the models.
Fig 5.
Conformations of 5HT in the most-favored clusters from induced-fit docking.
Selected docking pose of 5HT (blue) within the cluster in A) the outward-open structure (PDB ID 5I71, 30 poses) and B) the outward-occluded model (57 poses). Previously-identified protein-ligand interactions with D98, F341, and T439 are observed in both cases (blue dashes).
Fig 6.
Differences between outward-open (left) and outward-occluded (right) structures of hSERT.
The surface representation highlights the accessibility of the orthosteric binding site from the extracellular side. Specifically, the substrate molecule is visible from outside in the outward-open structure, but is covered in the outward-occluded structural model, where TM1b and TM6a bend over and close this same pathway from the synaptic cleft.
Fig 7.
Behavior of the extracellular gates over time during molecular dynamics simulations of hSERT.
The average minimum distances of the atoms between the R104-E493 and Y176-F335 gates are shown in 4 different trajectories for the outward-open (left) and outward-occluded states (right), with standard errors shown in the back (gray).
Fig 8.
Interactions between the protonated nitrogen of 5HT and hSERT during molecular dynamics simulations.
Aside from the salt-bridge with the deprotonated side chain of D98 (not shown), the amine coordinates the backbone oxygen atoms of Y95 (gray), F335 (orange), and S336 (blue). In the outward-open state (left), these distances increase over time, whereas they form a stable network in the outward-occluded state (right). Data are shown as normalized distributions over the combined simulation time.
Fig 9.
Distances between residues forming the intracellular gates during molecular dynamics simulations of hSERT.
Residues in the cytoplasmic ends of TM5, TM6, TM8, TM9, and the N-terminus were tracked in simulations of the outward-open (left) and outward-occluded (right) states. The normalized distributions of distances between these gate-forming residues reveal a higher number of hydrogen-bond breaking events in the outward-occluded models.
Fig 10.
Hydration of the pathways in simulations of hSERT.
A) The number of water molecules entering the central binding site, intra- and extracellular vestibules is shown as a normalized distribution over the total simulation time for the outward-open (brown) and outward-occluded states (green). B) Water occupancy within 20 Å of 5HT over the trajectory (1 μs simulation time) of one representative run for each state. The protein is shown as cylinders, highlighting specific helices colored according to Fig 2. The ligand (with C atoms in dark-gray) and sodium (orange) and chloride (green) ions are shown as spheres.
Fig 11.
Comparison of hSERT conformations: shown are the ibogaine-bound cryo-EM structures in outward-open (brown), outward-occluded (green), and inward-open (pink) states, alongside the outward-occluded, 5HT-bound model (blue). TM1b and TM6a are shown in color, whereas TM3 and TM8 of the hash region are in grey. Ligands bound in the orthosteric binding site are ibogaine (green sticks) in outward-occluded 6DZV, and 5HT in the outward-occluded model (blue sticks). A) Comparison of the cryo-EM structures in outward-open (6DZY), outward-occluded (6DZV), and inward-open (6DZZ) states. B) The outward-occluded model presented in this study (blue, from the final frame of MD replica 2), bound to 5HT, has bundle helix positions midway between the outward-occluded and inward-open, ibogaine-bound cryo-EM structures. All structures in both panels A and B were aligned using the backbone atoms in TM3, TM4, TM8, and TM9, i.e. the hash domain.