Lead
A team led by Yifan Cheng and David Julius reports (Nature, 25 March 2026) high-resolution structures and local energetic measurements that explain how the cold receptor TRPM8 senses temperature. Using cryo-EM of cell-derived membrane vesicles together with hydrogen–deuterium exchange mass spectrometry (HDX–MS), the authors capture two interconverting architectures—a canonical “fully swapped” and a newly described “semi-swapped” arrangement—and link those conformations to cold activation and menthol binding. They show menthol and low temperature bias a population toward the semi-swapped state, identify the outer pore and pore helix as principal energetic determinants of cold sensitivity, and present a cold-activated open structure stabilized by PIP2. The work combines structural snapshots with residue-level thermodynamics to map a mechanistic landscape for TRPM8 gating.
Key takeaways
- Cryo-EM in cell-derived vesicles yielded structures at ~3–3.5 Å and revealed both fully swapped and a new semi-swapped transmembrane architecture, with an S6 bend of ~52° in the semi-swapped form.
- Single-particle distributions shifted with menthol: semi-swapped particles rose from 24% (apo) to 37% (1 mM menthol), while fully swapped particles fell from 18% to 6% under the same conditions.
- HDX–MS identified bimodal exchange in the pore and TRP helices; one envelope exchanged >100-fold faster than the other, implying a local folding free-energy difference ≥3 kcal mol−1 between states.
- Menthol binding (1 mM) stabilizes the TRP helix by ~1.4 kcal mol−1 and repositions the TRP helix toward the VSLD, consistent with a ligand-driven population redistribution toward the semi-swapped/open pathway.
- Species comparisons show 92% sequence identity but distinct thermodynamic behavior: the human TRPM8 pore helix displays a pronounced enthalpic change across 30–22 °C (near the ~26 °C activation threshold), a feature weak or absent in the avian orthologue.
- A single outer-pore residue correlates with species differences: tyrosine in birds (Y905 in Pm) versus valine in mammals (V915 in Hs); swapping these residues shifts cold sensitivity and structural stability.
- A bona fide cold-open structure was determined at 4 °C and high pH (pH 9) for avian TRPM8 in GDN where F969 forms the lower gate and a cation-like density is visible in the permeation path; an endogenous PIP2 acyl chain (modelled as 18:0-20:4) occupies an S5–S6 cleft and stabilizes the upshifted S6 register.
Background
Temperature-activated TRP channels exhibit unusually steep thermosensitivity that has been attributed to changes in heat capacity (ΔCp), local solvation of hydrophobic residues, or partial unfolding transitions. Prior mutagenesis and structural work identified many residues that modulate thermosensitivity, suggesting gating is distributed across multiple domains rather than confined to a single thermostat. However, most high-resolution TRPM8 structures were obtained from detergent-solubilized protein and failed to capture a clear cold-open state or to resolve local energetic contributions across physiologically relevant temperature ranges.
To bridge structure and thermodynamics, the authors combined cryo-EM in native-like cell-derived membrane vesicles with HDX–MS on purified protein. The vesicle approach preserves native bilayer lipids and allowed visualization of conformations not resolved in detergent. HDX–MS reports residue or peptide-level folding free energies and can reveal temperature-dependent enthalpic contributions that are invisible to global calorimetry methods.
Main event
The investigators expressed an avian TRPM8 orthologue (Parus major, Pm) and human TRPM8 in HEK293 cells and prepared inside-out membrane vesicles without detergents. After an additional size-exclusion step to improve particle contrast, cryo-EM 2D class averages and 3D refinements produced maps at ~3–3.5 Å that showed cytoplasmic domains oriented outward from vesicles and well-resolved transmembrane features. In addition to previously seen closed/desensitized classes, they identified a distinct semi-swapped transmembrane arrangement in which each S6 helix remains paired with its cognate S1–S4 domain rather than domain-swapping with a neighbour.
Structurally, the semi-swapped state is characterized by a ~52° bend at the S6–TRP junction, reorientation of outer pore loops and a π→α helical register change near N958 that shifts the lower S6 register by one residue. The lower gate constriction changes: V966 forms the gate in the fully swapped π form, whereas L965 and F969 constrict in the semi-swapped α form; in the cold-open structure F969 translates and reorients to create a wider permeation path.
Particle counting and HDX–MS concur: menthol (1 mM) reduced the fraction of fully swapped particles and increased semi-swapped representation, and HDX spectra from pore and TRP helix peptides showed slow and fast exchanging populations with bimodal envelopes that interconvert on the minutes-to-hours timescale. The slower-exchanging population aligns with the semi-swapped configuration and is enriched by menthol, providing independent thermodynamic evidence for ligand-driven redistribution.
Detailed maps in the menthol-bound structures reveal a menthol-like density in the VSLD lower cavity. Docking places menthol hydroxyl near R832 (R842 in human), with a hydrophobic cavity that stabilizes the ligand and a ~15° repositioning of the TRP helix toward the VSLD. Differential HDX quantification shows the TRP helix is selectively stabilized by ~1.4 kcal mol−1 on menthol binding, consistent with menthol acting allosterically to bias gating.
Analysis & implications
Residue-level thermodynamics from van’t Hoff analysis reveal non-linear temperature dependence (ΔCp) in many human TRPM8 regions but not in the avian orthologue. The human pore helix in particular shows a marked decrease in standard folding enthalpy (ΔH°) across the 30–22 °C window that spans the channel’s activation threshold, indicating enthalpically driven stabilization on cooling. This provides a mechanistic basis for the heat-capacity model of TRP gating and identifies the pore helix as a primary energetic driver of cold activation in mammals.
Species-specific cold sensitivity thus arises from thermodynamic tuning rather than gross structural difference: despite 92% sequence identity and similar low-energy conformations, the human channel samples a broader energetic landscape with larger local ΔCp changes. The outer-pore interface (Y905 in birds versus V915 in mammals) kinetically traps avian TRPM8 in a closed semi-swapped basin; substituting residues or the entire pore loop shifts enthalpic profiles and function accordingly, validating the causal link between local energetics and physiological cold sensitivity.
The structures emphasize a central role for lipid cofactors: a long PIP2 acyl chain invades an S5–S6 hydrophobic cleft and stabilizes an upshifted S6 register that supports gate opening. The authors model the lipid as 18:0-20:4 PIP2, consistent with endogenous plasma-membrane species, and propose that temperature-dependent dynamics modulate the ease of PIP2 engagement. Functionally, this places phosphoinositides as active modulators that can bias the conformational ensemble toward states competent for opening.
Comparison & data
| Feature | Avian (Pm) | Human (Hs) |
|---|---|---|
| Sequence identity | ~92% overall (Pm vs Hs) | |
| Key outer-pore residue | Y905 (tyrosine) | V915 (valine) |
| Cold activation enthalpy (pore helix, 30–22 °C) | Weak ΔH° change | Pronounced ΔH° decrease |
| Menthol effect on particle distribution | semi-swapped 24%→37% | similar bias observed |
| Cold-open lower gate | F969 coordinates ion, pore radius ~2 Å | F969-based gate also observed |
This compact table highlights how small sequence differences map to measurable thermodynamic and functional differences; the open gate involves a conserved S6 α-helical register that can be stabilized either by ligand (menthol) or by PIP2 engagement in colder conditions.
Reactions & quotes
Authors presented the results as an integration of structure and thermodynamics; selected paraphrased remarks are given below with context.
“Combining native-membrane cryo-EM with HDX–MS lets us see both the structures and the energetic drivers behind cold sensitivity,” the team notes, arguing this dual approach is required to connect conformational states to physiological gating.
Choi, Lin, Cheng & Julius (authors)
External peer reviewers highlighted that the semi-swapped architecture and lipid engagement provide a plausible route by which local ΔCp changes in the pore helix can be transduced to gate opening.
Peer review summary (Nature)
Unconfirmed
- The central cation-like density observed at the open lower gate is interpreted as calcium in the presented maps but its chemical identity was not directly measured in situ.
- The assignment of the invading lipid as 18:0-20:4 PIP2 is modelled from density and length compatibility and is consistent with co-purifying PI species, but lipidomics confirmation in the same samples was not reported here.
- How the described ensemble behavior maps onto native sensory neurons and in vivo temperature detection remains to be demonstrated experimentally.
Bottom line
This study delivers a cohesive structural and energetic model for TRPM8 cold sensing: TRPM8 samples two tetrameric transmembrane architectures that interconvert, and cold or menthol biases the conformational ensemble toward an S6 α-helical configuration that, together with PIP2 engagement and pore-helix enthalpic changes, drives opening. The pore helix and outer-pore interface emerge as the principal thermodynamic loci that differentiate mammalian from avian cold sensitivity.
Practically, the results suggest routes to tune TRPM8 pharmacology and temperature sensitivity via ligands that stabilize the TRP helix, mutations in the outer pore, or by targeting the channel–lipid interface—avenues relevant to analgesia, thermosensory biology and bioengineering of temperature-sensitive probes. Future work should validate lipid identity, measure channel energetics in native neuronal membranes and explore whether similar outer-pore thermodynamic tuning governs other thermosensitive TRP channels.
Sources
- Choi K.Y. et al., Structural energetics of cold sensitivity, Nature (2026) — peer-reviewed journal article and primary dataset description (journal).
- PDB 9P91 / EMD-71395 — deposited semi-swapped avian TRPM8 structure (structural database).
- PDB 9P90 / EMD-71394 — deposited fully swapped avian TRPM8 structure (structural database).
- PRIDE / PXD064468 — mass-spectrometry proteomics deposition for HDX–MS (data repository).