Structure Modeling of the Norepinephrine Transporter

Abstract

The norepinephrine transporter (NET) is one of the monoamine transporters. Its X-ray crystal structure has not been obtained yet. Inhibitors of human NET (hNET) play a major role in the treatment of many central and peripheral nervous system diseases. In this study, we focused on the spatial structure of a NET constructed by homology modeling on Drosophila melanogaster dopamine transporter templates. We further examined molecular construction of primary binding pocket (S1) together with secondary binding site (S2) and extracellular loop 4 (EL4). The next stage involved docking of transporter inhibitors: Reboxetine, duloxetine, desipramine, and other commonly used drugs. The procedure revealed the molecular orientation of residues and disclosed ones that are the most important for ligand binding: Phenylalanine F72, aspartic acid D75, tyrosine Y152, and phenylalanine F317. Aspartic acid D75 plays a key role in recognition of the basic amino group present in monoamine transporter inhibitors and substrates. The study also presents a comparison of hNET models with other related proteins, which could provide new insights into their interaction with therapeutics and aid future development of novel bioactive compounds.

Keywords: homology modeling; ligand docking; norepinephrine transporter; reuptake inhibitors.

Figure 1

Representation of fragment of sequence alignment for core domains and extracellular loop EL4 of human NET, DAT, and SERT. The most important residues found at hNET models involved in binding ligands are marked in blue. First line depicted as red represented the alpha-helix secondary structure predicted by PROMALS3D [29].

Figure 2

Quality assessment of chosen models.

Figure 3

Generated three-dimensional hNET models built on (A) 4xpg dDAT template (QMEAN value −3.85) and (B) 4m48 dDAT template (QMEAN value −3.82). (C) Scheme of the monoamine transporter construction with indicated substrate binding site. Prokaryotic homolog is shown in black, differences found in eukaryotic equivalents in blue.

Figure 4

The hNET model presentation: (A) Superposition of two selected models. Comparison of the whole proteins. Model built on 4xpg template is shown as yellow, while on 4m48 as blue. Sodium ions are indicated in purple, chloride ions in green. (B) Surface of the model built on 4xpg template structure. By visualization of the area occupied by amino acid residues, an access gate to the internal part of the transporter could be easily detected. Surface of core domains and EL4 have been colored in red for contrast. Proposed entry pathway from extracellular site for NET inhibitors are indicated by white arrow. (C) Representation of model built on 4m48 structure both as surface and cartoon in blue. Core domains and EL4 are marked in black.

Figure 5

Structural features of hNET models which might be connected with specific interactions with ligands: (A) Alignment of core domains and EL4, (B) overview of nonconserved residues found between hNET and hDAT represented on received NET models: 4xpg and 4m48. Nonconserved residues are shown for: (C) EL4, (D) putative secondary binding site, (E) primary binding site. For 4xpg, sites are marked in gray, orange, and yellow in (C–E) respectively, while for 4m48 corresponding ones are shown as blue, green, and gray.

Figure 6

Representation of primary binding site at generated hNET models within 4 Å of the ligand. Residues are colored according to the models from which they originate, i.e., 4xpg model as yellow sticks, 4m48 model as blue. Sodium ions are shown as purple spheres, chloride ions as green. Identified ionic interactions are represented as yellow dashed lines.

Figure 7

Superposition of hNET model built on 4xpg template presented in yellow and LeuBAT 4mmd shown in magenta. Comparison of (A) core domains and extracellular loop 4, (B) binding pocket with subsite A and B, (C) amino acid residues in subsite C (shown as sticks) together with residues from subsites A&B (shown as lines).

Figure 8

Representation of ion binding sites from different perspectives in generated models: The 4xpg shown in yellow and 4m48 in blue within 4 Å of the ions. Bond length between residues and ion molecules represented as orange dashed lines measured for (A) Na⁺1, (B) Na⁺2, (C) Cl⁻ are shown as purple and green spheres: A/B and C.

Figure 9

Values of pK_i (negative logarithm of the dissociation constant of ligand—transporter complex) of selected monoamine transporter inhibitors; pK_i have been measured for human MATs. Exceptions are modafinil and compound X whose affinity has been appointed for Rattus norvegicus equivalent. All of the values were taken from the ChEMBL Database [46].

Figure 10

Structures of selected monoamine transporter inhibitors with a representation of pharmacological effects.

Figure 11

Binding mode of TCAs (A) nortriptyline, (B) desipramine in hNET on 4m48 template. Initial contacts with measured distances between inhibitors and transporter are shown as orange dashed lines. Amino acid residues involved in binding drug molecules have been represented as sticks. The remaining residues also important in recognition are shown as lines. Core domains are represented as a blue cartoon, the other transmembrane spanning domains are omitted for clarity.

Figure 12

Binding mode of (A) reboxetine, (B) modafinil, (C) venlafaxine in hNET on 4m48 template. Initial contacts with measured distances between inhibitors and transporter are shown as orange dashed lines. Amino acid residues involved in binding drug molecules are represented as sticks. The remaining residues also important in recognition are shown as lines. Core domains are represented as a blue cartoon, the other transmembrane spanning domains are omitted for clarity.

Figure 13

Binding mode of (A) bupropion in 4m48 model, (B) nocaine, (C) compound X in 4xpg hNET model. Amino acid residues involved in binding drug molecules are represented as sticks. The remaining residues also important in recognition are shown as lines. Core domains are represented as a blue cartoon, the other transmembrane spanning domains are omitted for clarity.

Figure 14

Comparison of the binding mode of (A) mazindol, (B) duloxetine at superimposed protein structures of hNET on 4xpg template and LeuBAT crystal structures: (A) 4mme, (B) 4mmd. Amino acid residues are colored according to the proteins from which they originate: 4xpg as yellow and 4mme, 4mmd as pink. Duloxetine and mazindol molecules docked to hNET are presented as light gray sticks, while inhibitors crystallized with LeuBAT are dark gray lines. The most important residues involved in binding drug molecules are presented as sticks. Measured distances between particular functional groups and docked compounds in 4xpg hNET are shown as orange dashed lines, while in 4mme and 4mmd they are omitted for clarity. Sodium ions are shown as purple spheres, chloride ions as green for hNET and LeuBAT alike.