The flexible stalk domain of sTREM2 modulates its interactions with brain-based phospholipids

  1. David Saeb
  2. Emma E Lietzke
  3. Daisy I Fuchs
  4. Emma C Aldrich
  5. Kimberley D Bruce
  6. Kayla G Sprenger  Is a corresponding author
  1. Department of Chemical and Biological Engineering, University of Colorado Boulder, United States
  2. Department of Endocrinology, Metabolism, and Diabetes, University of Colorado Anschutz Medical Campus, United States
4 figures and 1 additional file

Figures

Overview of structural domains in soluble form of TREM2 (sTREM2) and full-length Triggering Receptor Expressed on Myeloid Cells 2 (TREM2).

(A) Pre-molecular dynamics (MD) structure of sTREM2WT. (B) Full sequence of TREM2, indicating significant structural domains.

Figure 2 with 2 supplements
The stalk of sTREM2WT modulates the Ig-like domain via binding to Surface 1.

(A) Distance between complementarity-determining region (CDR) loop residues 45 and 70 over time for wild-type (WT) and R47H models of soluble form of TREM2 (sTREM2) and TREM2 (Ig), averaged across six replicates. Error bars represent the SEM. (B) Temporally averaged Cα root-mean-square fluctuation (RMSF) of the isolated Ig-like domain of WT and R47H sTREM2 compared to the Ig-like domain of TREM2, averaged across six replicates. Error bars represent the SEM. (C–D) Cα RMSD of (C) WT and (D) R47H models of sTREM2 and Triggering Receptor Expressed on Myeloid Cells 2 (TREM2) over time during the initial simulations, with corresponding snapshots shown for the starting and equilibrated structures in each case. The stalk domain of sTREM2 is shown in dark purple, and the CDR2 loop of TREM2 is shown in light purple. (E–F) Normalized fractional occupancy of residues in the Ig-like domain of (E) sTREM2WT and (F) sTREM2R47H by the stalk during the initial simulations.

Figure 2—figure supplement 1
Structural characterization of (s)TREM2 without ligands.

(A) Distance between complementarity-determining region (CDR) loop residues 45 and 70 over time for wild-type (WT) and R47H models of soluble form of TREM2 (sTREM2) and Triggering Receptor Expressed on Myeloid Cells 2 (TREM2) (Ig). (B) Percentage of simulation time that the WT (left) and R47H (right) sTREM2 stalk region remained within a converged root-mean-square deviation (RMSD) window (±2 Å over 50 ns) for each replicate, calculated using MDAnalysis, shown alongside a visualization of a representative structure from each simulation, color-coded by stalk position (pink for ‘Surface 1’ and blue for ‘Surface 2’). Normalized fractional occupancy of residues in the Ig-like domain of (C) sTREM2WT and (D) sTREM2R47H by the stalk, averaged across six replicates each.

Figure 2—figure supplement 2
Complementarity-determining region (CDR) distance, root-mean-square fluctuation (RMSF), and root-mean-square deviation (RMSD) for (s)TREM2 without ligands.

Analyses across molecular dynamics (MD) replicates for (A) IgWT, (B) IgR47H, (C) sTREM2WT, and (D) sTREM2R47H. Each row shows: (i) distance between CDR loop residues 45 and 70 versus simulation time, (ii) time-averaged Cα RMSF per residue, and (iii) Cα RMSD over time.

Figure 3 with 2 supplements
The stalk of sTREM2WT modulates phospholipid binding.

(A–D) Normalized fractional occupancy of residues in (A) IgWT by stearoyl-oleoyl-PS (SOPS), (B) sTREM2WT by SOPS, (C) IgWT by stearoyl-oleoyl-PC (SOPC), and (D) sTREM2WT by SOPC, averaged across seven, five, five, and six 150-ns simulations, respectively. (E–H) Visual Molecular Dynamics (VMD) snapshots of representative structures from the PL/(s)TREM2WT simulations, with phospholipids (PLs) and corresponding complex binding free energies shown in matching colors.

Figure 3—figure supplement 1
Root-mean-square deviation (RMSD) and root-mean-square fluctuation (RMSF) analysis for (s)TREM2WT with phospholipids (PLs).

Comparative RMSD and RMSF analyses across individual molecular dynamics (MD) trials for each wild-type (WT) (s)TREM2/PL system: (row A) SOPS/IgWT, (row B) SOPS/sTREM2WT, (row C) SOPC/IgWT, and (row D) SOPC/sTREM2WT. The subfigures in each row show: (i) Cα RMSD of the protein structure vs. simulation time, (ii) RMSD of the PL calculated relative to the protein vs. simulation time to capture broad PL movements, and (iii) temporally averaged Cα RMSF of protein residues.

Figure 3—figure supplement 2
Protein-phospholipid (PL) complex renderings and interaction entropy calculations.

Visual Molecular Dynamics (VMD)-rendered images from PL/(s) TREM2 simulations (A–H), with the PLs colored to match their binding interaction entropy contributions (-TΔS) (kJ/mol) in panel I. (I) Comparison of the interaction entropy contributions averaged across all wild-type (WT) and R47H models for each PL/protein system. Error bars represent the standard error of the mean. Black brackets with a single asterisk represent statistical significance with 0.01<p<0.05.

Figure 4 with 1 supplement
The R47H mutation decreases ligand discrimination but increases binding to the flexible stalk domain.

(A–D) Normalized fractional occupancy of residues in (A) IgR47H by stearoyl-oleoyl-PS (SOPS), (B) sTREM2R47H by SOPS, (C) IgR47H by stearoyl-oleoyl-PC (SOPC), and (D) sTREM2R47H by SOPC, averaged across four, seven, six, and eight 150-ns simulations, respectively. (E–H) Visual Molecular Dynamics (VMD) snapshots of representative structures from the PL/(s) TREM2R47H simulations, with phospholipids (PLs) and corresponding complex binding free energies shown in matching colors. (I) Comparison of the binding free energies averaged across all WT and R47H models for each PL/protein system. Error bars represent the standard error of the mean.

Figure 4—figure supplement 1
Root-mean-square deviation (RMSD) and root-mean-square fluctuation (RMSF) analysis for (s) TREM2R47H with phospholipids (PLs).

Comparative RMSD and RMSF analyses across individual molecular dynamics (MD) simulation trials for each R47H (s) TREM2/PL system: (row A) SOPS/sTREM2R47H, (row B) SOPS/IgR47H, (row C) SOPC/sTREM2R47H, and (row D) SOPC/IgR47H. The subfigures in each row show: (i) Cα RMSD of the protein structure vs. simulation time, (ii) RMSD of the PL calculated relative to the protein vs. simulation time, and (iii) temporally averaged Cα RMSF of protein residues.

Additional files

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. David Saeb
  2. Emma E Lietzke
  3. Daisy I Fuchs
  4. Emma C Aldrich
  5. Kimberley D Bruce
  6. Kayla G Sprenger
(2025)
The flexible stalk domain of sTREM2 modulates its interactions with brain-based phospholipids
eLife 13:RP102269.
https://doi.org/10.7554/eLife.102269.3