henhaobio
Sign in

Journal of Chemical Information and ModelingPharmacology & ToxicologyPeer-reviewed journal article

MD simulations of Human Sigma1 Receptor Trimer Uncovers Cholesterol Dependent Stabilization and Ligand Specific Dynamics

Vittoria NannaiD, Costanza Paternoster, Alessio BartocciiD, Dritan SiliqiiD, Domenico AlbergaiD, Carmen AbateiD, Gianluca LattanziiD, Giuseppe Felice MangiatordiiD

PDFDOIPreprintData

Abstract

The sigma-1 receptor (S1R) is an endoplasmic reticulum transmembrane protein implicated in a wide range of physiological and pathological processes, including neurodegeneration, cancer, and pain modulation. Although X-ray crystallography has revealed S1R as a trimeric assembly with a distinctive triangular architecture, the dynamic behavior of this oligomeric state and its modulation by ligands and membrane composition remain poorly understood. In particular, agonists and antagonists have been experimentally proved to differentially regulate S1R oligomerization although the underlying molecular mechanisms are still obscure. Here, we present the first atomistic molecular dynamics study of trimeric S1R embedded in a physiologically relevant lipid environment. Using a total of 12 {micro}s of simulation time, we investigate the impact of membrane composition, with a specific focus on cholesterol, as well as the conformational response of S1R to pharmacologically distinct ligands: the agonist (+)-pentazocine and the antagonist haloperidol. Our simulations reveal how ligands can alter S1R interprotomer interaction through a mechanism involving the {beta}6-strand of the protein and in particular W136, data that correlate with experimentally observed differences in S1R oligomerization. These findings provide new molecular-level insights into S1R regulation and establish a framework for rationalizing the distinct functional outcomes induced by agonists and antagonists.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (8 Mar 2026) · CC BY 4.0

Across 12 µs of all-atom simulations, cholesterol stabilizes the sigma-1 receptor trimer, while haloperidol, unlike (+)-pentazocine, engages β6-strand W136 to tighten interprotomer coupling.

Crystallography shows the sigma-1 receptor (S1R) as a trimer, yet agonists and antagonists shift its oligomerization differently, and earlier simulations mostly used monomers in simplified, typically pure POPC, bilayers. This work simulates the full trimer in a cholesterol-rich bilayer that mimics the mitochondria-associated membrane (MAM) and proposes a residue-level mechanism for ligand-specific oligomeric effects.

Key results

  1. ShownThe S1R trimer stayed structurally stable in a cholesterol-containing, MAM-mimicking bilayer. In pure POPC, Cα RMSD was higher and fluctuated more. Bilayer thickening followed membrane composition, not ligand identity.

    Thickness was 41.10 ± 0.04 Å (Hal) and 41.19 ± 0.03 Å (PnT) in POPC/POPE/CHOL vs 38.29 ± 0.02 Å and 38.21 ± 0.02 Å in pure POPC. RMSD differences are described only qualitatively, from three 1 µs replicates per system. · Ligand-bound Sigma1 receptor trimers modulate…; Supporting Information

  2. SuggestedS1R pulls the bilayer into a local bowl-shaped pit above its core. This comes from the lipid imbalance between leaflets and from strong tilting of lipids packed against the apolar α4/α5 protein platform.

    Mean and Gaussian curvature were both positive (H>0, K>0) near S1R. Lipids within 6 Å of the protein tilted up to 90°, while bulk lipids sat mostly near 20°. The inner leaflet started with 400 lipids vs 450 in the outer leaflet. Crystal structures show ordered lipid density near α1. · Ligand-bound Sigma1 receptor trimers modulate…; Supporting Information; Figure legends

  3. ShownCholesterol gathered around S1R on its own, with no restraints applied. The longest-lasting contacts were at the transmembrane CARC-like motif and the α1/α4/α5 junction. Contacts with α5 were most frequent; contacts with α1 were rarer but longest-lived. The pattern was similar with either ligand.

    Persistence was measured as the longest unbroken cholesterol contact within 6 Å, averaged over monomers and three replicas. It peaked at CARC-like residues 7 to 14 and at the α1-α4-α5 junction, a site occupied by ordered lipids in the crystal structures. Hal and PnT complexes gave matching patterns. · Cholesterol-protein contacts; Cholesterol binds S1R in specific hotspots; Figure legends

  4. ShownThe two ligands contacted mostly the same residues. Haloperidol, whose more elongated structure penetrates deeper into the cavity, engaged the β6 strand through its chlorophenyl ring more persistently. That strand carries W136, which sits at the interface with the neighbouring protomer.

    Contacts within 3.5 Å differed mainly at residues 83-87 and 132-137. In the crystal structures, both ligands contact W89, Y103 and F107 and form the E172 salt bridge, and the two complexes differ by an RMSD of 0.439Å. W136 packs against F83, A92 and T109 of the adjacent monomer. Prior experiments found that replacing W136 with glycine impairs multimerization. · Ligand-protein contacts; W136 emerges as a key determinant of…; Figure legends

  5. SuggestedCorrelation network analysis found W136 to be the main hub in both complexes. Haloperidol gave a more cohesive network and coherent communication paths from E172 to W169 across the protomer interface. With (+)-pentazocine, the paths split apart at the interface.

    W136 centrality was markedly higher in S1R-Hal. S1R-PnT had 24 communities vs 14 in S1R-Hal (edges kept at |cij| ≥ 0.75). In S1R-Hal, all 100 computed paths ran through T109 and A110 to W169; in S1R-PnT the paths diverged after crossing between monomers. · Correlation network analysis (CNA); Ligands alter S1R interprotomer interaction and…

Methods, data and reproducibility

Unbiased all-atom MD (GROMACS 2023, CHARMM36m/CGenFF) of the human S1R trimer bound to haloperidol or (+)-pentazocine, in pure POPC versus POPC/POPE/cholesterol bilayers. Each system ran as three 1000 ns replicates and was analyzed for membrane properties, cholesterol contact persistence, ligand contacts and Bio3D correlation networks.

  • Crystal structure of the S1R homotrimer bound to haloperidol (S1R-Hal), used as the starting model · human · PDB 6DJZ
  • Crystal structure of the S1R homotrimer bound to (+)-pentazocine (S1R-PnT), used as the starting model · human · PDB 6DK1
  • S1R-progesterone structure, used only to show the steroid-binding regions (SBDLI, SBDLII, CARC-like motif) · PDB 8w4b
  • MD trajectories of four systems (Hal or PnT bound; pure POPC or POPC 56%/POPE 22%/CHOL 22%), approximately 370,000 atoms each · human (S1R) · n = twelve replicates of 1000 ns (12 µs total) · Zenodo 10.5281/zenodo.18787104 (described as raw data; contents not itemized in the text)
Methods & tools
GROMACS 2023, CHARMM36m force field, CHARMM General Force Field (CGenFF), Modeller, Protein Preparation Wizard (Schrödinger Suite 2024-1), CHARMM-GUI, Orientations of Proteins in Membranes (OPM) database, LipidDyn, MDAnalysis (MembraneCurvature), Bio3D v2.4 (R), TIP3P water model, In-house Python scripts (lipid tilt, contact analyses)
Reagents & environment
GROMACS 2023; CHARMM36m and CGenFF force fields; TIP3P water; Schrödinger Suite 2024-1 Protein Preparation Wizard; Modeller; CHARMM-GUI; Bio3D v2.4 (R); LipidDyn; MDAnalysis MembraneCurvature.
Reproducibility gaps
No code repository is linked. Lipid tilt used an in-house Python script and ligand contacts a custom MDAnalysis script; the code for the cholesterol-contact analysis is not described. Full GROMACS parameters, box sizes and membrane compositions appear only in Supplementary Tables S1 to S3.
Limitations (4)
  • The bilayers have an unavoidable lipid imbalance between leaflets (400 inner vs 450 outer), forced by keeping helices α4/α5 inserted in the membrane. The authors say this asymmetry partly causes the local thinning and pit-shaped curvature seen near S1R. · stated by the authors
  • Default neighbor-list and pressure-coupling settings are known to cause membrane instabilities, which the authors note could be amplified by their leaflet asymmetry, so these parameters were re-tuned following Kim et al. (Supplementary Table 3). · stated by the authors
  • The area-per-lipid values printed in the text (49.64 ± 0.04 Å^2 pure POPC vs 62.49 ± 0.02 Å^2 MAM-like, Hal) contradict the stated cholesterol-driven drop in APL and the Figure S3 legend. The labels appear to be swapped. · noted in this summary
  • The link from ligand to oligomerization is indirect. Each system has three 1 µs replicates, and two of them branch from replicate 1 (at 100 and 110 ns). No simulations of oligomer dissociation, free energies or W136 mutants were run. · noted in this summary

Read this if you study sigma-1 receptor pharmacology or oligomerization, or simulate membrane protein assemblies in cholesterol-rich bilayers and want residue-level hypotheses for how agonists and antagonists differ.

Citations
2
FWCI
4.10
Field percentile
top 6%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Similarity Map
SelectedSame fieldCross-field
THISSchmidt '18Zhemkov '20Zhao '22Yano '19Isu '24Salgado-Pol '26
75% semantic matchPharmacology & Toxicology
H. R. Schmidt, R. M. Betz, R. O. Dror+1 more·Nature Structural & Molecular Biology·Oct 5, 2018