Effects of a Serotonergic Psychedelic on the Lipid Bilayer

The experimental approach used reconstituted lipid systems and a set of complementary physical techniques to probe membrane partitioning and mechanics in the absence of proteins. Two primary lipid compositions were employed: a complex four-lipid mixture intended to mimic synaptic vesicle composition (POPC/POPE/POPS/cholesterol in a 3/5/2/5 molar ratio, referred to as PCPEPSChl 3525) and a phase-separating raft-like mixture (DOPC:egg sphingomyelin:cholesterol 2:2:1, referred to as DEC221). Partitioning of DOI into membranes was measured by a dialysis retention assay: small unilamellar vesicles (SUVs, lipid concentration 4 mM) were incubated with DOI (1.5 mM), dialysed against HEPES buffer (20 mM HEPES, pH 7.4, 150 mM NaCl) using a 100 kDa cutoff membrane for 8 h, and the DOI remaining associated with vesicles was quantified by UV absorption at 296 nm using an external calibration curve. A parallel measurement was performed for serotonin for comparison. Membrane order and effective chain length were assessed by solid-state 2H NMR on multilamellar vesicles (MLVs) composed of deuterated POPC within the PCPEPSChl 3525 mixture. Experiments were performed at 25 °C with DOI included at 0, 10 and 20 mol% relative to lipid; order parameters for individual carbons were extracted and mean quantities (<SCD>) and effective chain lengths computed using the mean torque model. Mechanical stiffness and pore-formation propensity were measured by atomic force microscopy (AFM) indentation on supported lipid bilayers (SLBs) of PCPEPSChl 3525 to obtain the rupture/indentation force (Fx). AFM height profiling was also used to characterise domain structure in DEC221 SLBs and to quantify domain area and perimeter for an estimate of interfacial surface tension. Membrane fluidity was probed using the polarity-sensitive dye Laurdan; emission spectra and the generalized polarization (GP) parameter were recorded for lipid vesicles (1 mM lipids, Laurdan 60 nM) in the presence and absence of DOI. Functional assays of vesicle association and fusion used two fluorescence-based techniques. Fluorescence correlation spectroscopy (FCS) tracked Nile red-labelled SUVs in solution to derive translational diffusion times and hydrodynamic radii (RH) as a function of DOI concentration; autocorrelation traces were fitted to a three-dimensional diffusion model containing two diffusing species (free dye and vesicle-bound dye). Total internal reflection fluorescence (TIRF) microscopy imaged Rh-PE-labelled vesicles interacting with an otherwise unlabeled SLB to measure binding kinetics and fusion. The investigators modelled vesicle–bilayer interaction as a two-step process (binding followed by either dissociation or irreversible fusion) and extracted kon (reported indirectly via collision-limited behaviour), koff, kfus and the net fusion rate Rfus at DOI concentrations up to 1.5 mM. Where serotonin comparisons are reported, matching experimental setups were used.

DOI partitions strongly into the model lipid membrane. The dialysis assay yielded a partition coefficient (mole-fraction units) of 2941 for DOI in PCPEPSChl 3525 (log Px = 3.46), compared with 713 for serotonin (log Px = 2.85). The authors note both partition coefficients exceed simple lipophilicity predictions and that DOI’s partitioning is only ~4× higher than serotonin’s, indicating additional specific interactions may influence membrane binding. Solid-state 2H NMR showed DOI decreases acyl-chain order in a concentration-dependent and approximately homogeneous manner along the sn-1 chain of POPC. The average chain order parameter <SCD> fell by ~11% with 10 mol% DOI and ~18% with 20 mol% DOI; by comparison, 10 mol% serotonin reduced <SCD> by only ~5%. Using the mean torque model, effective chain length (hydrophobic thickness) decreased by 0.55 Å and 0.92 Å at 10 and 20 mol% DOI respectively, versus a 0.25 Å decrease with 10 mol% serotonin. AFM indentation measurements indicated a marked reduction in the force required to rupture SLBs in the presence of DOI. The indentation force decreased by 38 ± 4% with 0.5 mM DOI; even 0.1 mM DOI produced a substantial effect. Serotonin showed negligible effect on stiffness at 5 mM and required ≥100 mM to produce comparable changes, implying DOI is roughly three orders of magnitude more potent than serotonin in facilitating pore formation. Laurdan measurements supported increased fluidity: GP decreased from 0.37 ± 0.01 to 0.32 ± 0.01 in the presence of 1.5 mM DOI, consistent with NMR and AFM results. In phase-separated DEC221 bilayers, DOI induced rapid reorganisation of ordered domains. Incubation with 0.5 mM DOI nucleated disordered domains within ordered regions and drove a time-dependent shrinkage of the ordered fraction from ~0.28 at t = 0 to ~0.12 at 120 min. The estimated interfacial surface tension between phases fell by 51.7 ± 6.7% with 0.5 mM DOI. At 0.1 mM DOI the ordered area and surface tension decreased modestly (surface tension −13.4 ± 1.8%). For comparison, ~4 mM serotonin reduced surface tension by 34.0 ± 2.1%, so 0.5 mM DOI produced a stronger effect at an eight-fold lower concentration. FCS measurements revealed DOI-induced vesicle association in solution: the hydrodynamic radius RH increased from 14.1 ± 0.2 nm to 62.6 ± 7.0 nm in the presence of 1.5 mM DOI. Higher concentrations could not be measured by FCS because very large particles formed. This degree of association at 1.5 mM DOI was comparable to that caused by 5 mM serotonin. TIRF-based single-vesicle assays showed DOI prolongs vesicle residence on the SLB and increases fusion probability. The dissociation rate koff decreased from 7.1 ± 0.3 s−1 to 1.1 ± 0.4 s−1 with 1.5 mM DOI, while the fusion-rate constant kfus increased from 0.13 ± 0.03 s−1 to 0.71 ± 0.1 s−1. The net fusion rate Rfus rose ~20-fold, from 0.027 ± 0.002 to 0.51 ± 0.06 vesicles s−1, at 1.5 mM DOI. By contrast, serotonin at 5 mM lowered koff but did not alter kfus and produced roughly a 10× increase in fusion rate. Comparing across techniques, the authors report that DOI is more potent than serotonin by factors ranging from ~2× to >200× depending on the property measured; partitioning alone (4× difference) cannot fully account for these potency differences.

Saha and colleagues interpret the data as demonstrating that DOI markedly perturbs lipid bilayer physical properties at concentrations substantially lower than those required for serotonin to produce similar effects. They argue these membrane-mediated changes—reduced chain order and thickness, lower indentation force, increased fluidity, domain reorganisation, and enhanced vesicle association and fusion—represent a receptor-independent mechanism by which lipophilic psychedelics could influence cellular processes that depend on large membrane rearrangements, such as endo- and exocytosis and synaptic vesicle trafficking. Compared with earlier work showing serotonin’s membrane effects, the authors emphasise DOI’s greater potency and suggest that specific interactions (for example hydrogen-bonding with lipid headgroups), orientation and location within the bilayer, as well as differences in lipophilicity, likely combine to determine the magnitude of perturbation for different compounds. They also stress that the various measurement modalities probe different spatial and temporal scales—ranging from atomic-scale order parameters (2H NMR) to mesoscopic functional processes (FCS, TIRF)—which helps explain why relative potencies vary across assays. The investigators acknowledge important limitations and uncertainties. All experiments were performed in protein-free model membranes, whereas biological fusion and vesicle trafficking in neurons are regulated by many proteins (for example SNAREs and SM family proteins); thus the in vitro fusion assay is a simplification. They note that concentrations at which strong membrane effects are seen (hundreds of micromolar to millimolar) exceed typical experimental animal dosing levels (the authors cite studies that administer ~10 μM DOI), so direct membrane effects in vivo may be limited at therapeutic doses. Nevertheless, the authors point out scenarios where local accumulation (for example within organelles), long-term exposure, or misuse could yield concentrations at which membrane perturbation is meaningful. They also raise the possibility that membrane-mediated pathways could provide a more distributed and longer-lasting influence across cell types that express few canonical receptor targets. Finally, the authors propose that appreciating membrane-mediated actions opens new avenues for drug discovery: compounds that are receptor-orthogonal but designed to modulate membrane properties could be explored as modulators of processes tied to membrane mechanics. They caution that neglecting such pathways may constrain understanding of the full spectrum of effects exerted by serotonergic psychedelics.

The study concludes that DOI, a serotonergic psychedelic, substantially alters multiple physical properties of model lipid bilayers—chain order, membrane thickness, stiffness, domain organisation, and propensity for vesicle association and fusion—at concentrations much lower than those required for serotonin to cause comparable changes. While recognising that in vitro concentrations differ from many in vivo dosing regimens, the authors suggest that membrane-mediated effects could contribute to the biological actions of lipophilic psychedelics, particularly for processes that demand significant membrane remodelling. They further note the implication that receptor-independent, membrane-acting molecules could represent a novel direction for drug discovery and that accounting for membrane-mediated pathways may broaden understanding of psychedelic effects on the brain.