Modeling and dynamics of the inward-facing state of a Na+/Cl- dependent neurotransmitter transporter homologue

Abstract

The leucine transporter (LeuT) has recently commanded exceptional attention due mainly to two distinctions; it provides the only crystal structures available for a protein homologous to the pharmacologically relevant neurotransmitter: sodium symporters (NSS), and, it exhibits a hallmark 5-TM inverted repeat ("LeuT-fold"), a fold recently discovered to also exist in several secondary transporter families, underscoring its general role in transporter function. Constructing the transport cycle of "LeuT-fold" transporters requires detailed structural and dynamic descriptions of the outward-facing (OF) and inward-facing (IF) states, as well as the intermediate states. To this end, we have modeled the structurally unknown IF state of LeuT, based on the known crystal structures of the OF state of LeuT and the IF state of vSGLT, a "LeuT-fold" transporter. The detailed methodology developed for the study combines structure-based alignment, threading, targeted MD and equilibrium MD, and can be applied to other proteins. The resulting IF-state models maintain the secondary structural features of LeuT. Water penetration and solvent accessibility calculations show that TM1, TM3, TM6 and TM8 line the substrate binding/unbinding pathway with TM10 and its pseudosymmetric partner, TM5, participating in the extracellular and intracellular halves of the lumen, respectively. We report conformational hotspots where notable changes in interactions occur between the IF and OF states. We observe Na2 exiting the LeuT-substrate- complex in the IF state, mainly due to TM1 bending. Inducing a transition in only one of the two pseudosymmetric domains, while allowing the second to respond dynamically, is found to be sufficient to induce the formation of the IF state. We also propose that TM2 and TM7 may be facilitators of TM1 and TM6 motion. Thus, this study not only presents a novel modeling methodology applied to obtain the IF state of LeuT, but also describes structural elements involved in a possibly general transport mechanism in transporters adopting the "LeuT-fold".

Figure 1. The crystal structure of LeuT.

Views of the LeuT structure as would be seen along the plane of the lipid bilayer, oriented such that the upper side represents the extracellular half. Left: The pseudosymmetric 5+5 repeat is shown in color (green: first repeat, TM1 to TM5; pink: second repeat, TM6 to TM10) with the two broken helices in the core, TM1 and TM6, shown in ribbon representation. The rest of the protein, including TM11 and TM12 are in gray. Bound ions are in yellow, and the substrate, leucine, is colored by element name (H: white, C: gray, N: blue, O: red). TM10 is transparent for clarity. Center: A 180-rotated view of the structure is shown. TM5 is transparent for clarity. Right: The outer scaffolding helices, TM4, TM5, TM9, and TM10, are shown separated (above) from TM1, TM2, TM3, TM6, TM7, and TM8 (below) for clarity.

Figure 2. Schematic of the modeling methodology.

The combined approach adopted for generation of the inward-facing model is presented (see Methods). The location of the EC and IC halves of the lumen in the LeuT and vSGLT structures, respectively, are marked in brown. The initial phase of structural alignment and threading (gray), the pre-TMD phase (blue), and the final TMD phase (red) resulting in the two models discussed in this study, are illustrated. Dotted lines are used for structural input in addition to the starting LeuT structure.

Figure 3. The LeuT-vSGLT sequence alignment adopted for “pre-target” generation.

The sequences corresponding to TM1-TM10 along with intermediate loops (EL: extracellular loops, IL: intracellular loops) are indicated. The intracellular ends of the helices are marked as “IC”. Identical and similar residues between the two sequences are colored red, with the former in bold. LeuT residues that bind Na1 (blue triangles), Na2 (cyan triangles) and substrate (orange circles) are marked. Also marked are important residues (red stars), determined from this and previous studies, involved in gating and Na2 release.

Figure 4. RMSD variation in and .

Root mean square deviations (RMSDs) of C positions from the crystal structure (red, blue) and from the starting ( ns) structure (orange, cyan) during the 50 ns TMD+20 ns equilibration in monomer A (red, orange), compared to those in 70 ns free MD of monomer B (blue, cyan). The starting structure ( ns) of these simulations corresponds to a structure obtained after 10 ns of equilibration performed previously (see Methods). For both (left) and (right), the structures relax to a conformation distinctly different from the reference OF-occ structures.

Figure 5. Water accessibility of the IF state models.

A. Comparison of TM domains and water penetration in the initial structure input to TMD (left), and in Model1 and Model2 (right). Water molecules (white atoms, blue surface) in the lumen are shown. TM4, TM5, TM9, and TM10 are hidden for clarity. The radius profile of the lumen (center) shows EC narrowing and IC widening in both Model1 (orange curve, radii averaged over 1 ns) and Model2 (cyan curve, radii averaged over 1 ns) as compared to the initial structure (black curve), and remains occluded at the substrate binding site in all structures. These regions are highlighted with grey bands. B. Solvent accessible surface area (SASA) of EC (violet) and IC (green) lumen residues for TM1-TM10 in the initial structure, Model1, and Model2. C. Number of water molecules in the EC and IC halves of the lumen for the initial structure, Model1 and Model2.

Figure 6. Changes in contacts and conformation in and simulations.

A. Residue-residue differential contact maps (see Methods) for Model1 (left triangle) and Model2 (right triangle). Contacts broken (black dots) and formed (red dots) are shown. Residue ranges corresponding to TM1-TM10 are marked (blue). The intracellular ends are marked by dotted blue lines in the Model1 map. Regions showing notable contact breakage/formation excluding those beyond TM10 or residues on the diagonal are highlighted (green ovals). Similar trends are observable in and . In the center, EC (top) and IC (bottom) views of TM1-TM10 are shown with superimposed snapshots, taken before (transparent) and after (solid) the simulation. TM3 and TM8 C atoms were used for the superposition. Pseudosymmetric pairs of helices are colored the same, with darker colors for TM1-TM5, and lighter for TM6-TM10 i.e. TM1 and TM6 are red, TM2 and TM7 are blue, TM3 and TM8 are golden, TM4 and TM9 are violet, and, TM5 and TM10 are green. Closing of TM1, TM7, and TM10 in the EC half and opening of TM6, TM2, and TM5 in the IC half are clearly visible, corresponding to the differential contact maps. B. Side views of the less mobile TM3-TM4-TM5-TM8-TM9-TM10 scaffold (left) and the highly mobile TM1-TM2-TM6-TM7 bundle (right) are shown separately, in superimposed snapshots taken before (transparent) and after (solid) the simulation. The coloring scheme is as above. The position of the substrate (gray) and ions (yellow) before (transparent) and after (solid) the simulation are also shown. Dotted black lines highlight the motion in the helices.

Figure 7. Variation in salt bridge interactions in the EC and IC halves of the lumen.

Snapshots of salt bridges in the EC (top) and IC (bottom) half-lumens before (light) and after (dark) the simulation. The coloring scheme is as in Fig. 6. Variation of distance between salt bridging residues in the EC (top) and IC (bottom) half-lumens for (left) and (right) are shown. The distances are compared between monomer A, which undergoes TMD and monomer B, which undergoes free MD and serves as a control system.

Figure 8. Na2 release and water access to the substrate in the simulation.

A. The Na2 binding site before (left) and after (right) is shown, with the Na2 binding residues marked. E192, which binds Na2 (yellow) along its unbinding and release from the binding site is also shown. The substrate (Leu) is shown in gray. N21 and S256, which protect the substrate from water within a cavity below the substrate, are shown as a green surface. TM1 (red) and TM8 (golden), which bind Na2, TM6 (light red) which carries S256, and TM5 (green, transparent), which carries E192 are shown. Other parts of the protein are hidden for clarity. Water is shown as white molecules with a blue surface. B. The distance of Na2 from O atoms of binding residues is shown for the simulation. Na2 release occurs around ns. C. SASA of the substrate is shown for the simulation. The substrate SASA increases around the same time as Na2 release occurs, due to water access through the empty Na2 site.