Cyclic peptides cross cell membranes by sensing environmental polarity and switching between open and closed conformations. In nonpolar surroundings, the closed state buries polar groups, lowers desolvation energy, and becomes the main permeation state, crossing the lipid core through a four-step…
Cyclic peptides cross cell membranes by switching between two shapes in response to the local environment. In polar surroundings, such as water, a flexible cyclic peptide adopts an open conformation in which its hydrogen-bonding atoms are exposed to solvent. When the peptide reaches the nonpolar interior of a lipid bilayer, it closes into a conformation in which those polar atoms form intramolecular hydrogen bonds with one another, leaving a hydrophobic surface facing the lipid. This polarity-sensing behavior, called chameleonic switching , is the core mechanism behind cyclic peptide membrane permeability. It also explains why cyclosporine A, a natural cyclic peptide, works as an oral drug: its oral bioavailability is high, and it has been given as an oral formulation since 1983.
Oral peptide delivery still faces two core obstacles: stability against enzymatic degradation and the ability to cross cell membranes. Stability is a chemistry problem that backbone modification can address. Membrane crossing has proven harder. Flexible linear peptides cannot hide their backbone amides, and each amide pays a desolvation cost in the lipid core. Cyclic peptides can hide them, because the ring allows the backbone to fold back on itself. In water, the open conformation maximizes favorable contacts with solvent. In a membrane, the closed conformation trades those contacts for intramolecular hydrogen bonds, burying polar groups and displaying nonpolar groups on the surface. The switch reduces polar surface area and desolvation energy, helping balance water and lipid solubility and increasing cell permeability.
The closed state is usually considered the main permeability state, but that is a simplification. The open-closed equilibrium is dynamic, and which state dominates depends on sequence, molecular size, and hydrophobic surface area. Some cyclic peptides are permanently closed rigid rings, highly permeable but poorly soluble. Others stay open and never cross. The molecules most relevant to drug design sample both states and shift the population in response to where they are.
The most detailed mechanistic account of this behavior comes from a molecular dynamics study by Stephanie M. Linker and colleagues at ETH Zurich in Switzerland, published in 2023 in the Journal of Medicinal Chemistry 66 4 : 2773-2788. The authors simulated a flexible decacyclic peptide crossing a lipid membrane and resolved the process into four steps governed by the open-closed switch. The study does more than show that permeation happens; it identifies which side chains initiate contact, which orientation the peptide must adopt, and why an open-state peptide stalls halfway across.
The simulated crossing is passive permeation : no transporter, no energy input, only a peptide diffusing through a lipid bilayer. The pathway divides into four steps, each with its own conformational requirement.
| Step | What happens at the membrane | Conformational requirement |
|---|---|---|
| 1 | Specific amino acid side chains anchor the peptide to the membrane surface before insertion | Anchor residues depend on the starting conformation |
| 2 | The peptide inserts at the interface between polar heads and nonpolar tails, parallel to the membrane plane | Open state: orientation A or B, with A preferred; closed state: orientation B only |
| 3 | The open state converts to the closed state inside the membrane | Conversion required before onward diffusion |
| 4 | The closed peptide crosses the nonpolar core, anchors to the opposite leaflet, and flips | Closed state only |
Step 1 is anchoring. Before insertion, specific amino acid side chains reach into the membrane surface and establish the first contact. The simulation identifies these residues as molecular anchors . The anchor residues are not the same in the open and closed conformations, because the residues contact the membrane to different degrees in each. A residue buried in the interior of the closed conformation cannot serve as an anchor, however hydrophobic it is. Anchoring is therefore a conformation-dependent event, and the conformation a peptide adopts in solution, before it touches a membrane, helps determine whether it can initiate entry.
Step 2 is insertion. The peptide jumps into the interface between the polar head and nonpolar tail regions, where it lies parallel to the membrane plane. Two insertion orientations are possible, designated orientation A and orientation B . An open-state peptide can adopt either orientation but favors A. A closed, pre-folded peptide adopts only orientation B. The distinction matters because only orientation B favors the open-to-closed transition. A peptide that inserts in orientation A is positioned to get stuck.
Step 3 is the critical conversion. Inside the membrane, an open-state peptide finds it difficult to continue diffusing. Its hydrogen-bonding atoms are still exposed, and the nonpolar core imposes a desolvation penalty on them. The peptide must convert to the closed state before it can proceed. The simulation describes this open-to-closed transition as required for onward movement. The conversion changes the molecule's spatial structure and its interaction with the membrane, and therefore determines how it moves and how efficiently it diffuses.
Step 4 is crossing and exit. Only the closed-state peptide can diffuse through the nonpolar lipid core to the opposite polar/nonpolar interface. There the final event resembles the first: the peptide anchors to the polar heads of the lower leaflet and reverses its direction, flipping to complete the transmembrane journey. This reversal also reverses the hydrophobic moment. The molecule enters with one face leading and exits with the opposite face leading, after re-anchoring on the far side.
The four-step pathway is a sequence of distinct molecular events, not a single partition coefficient. A peptide must present exposed side chains to get in, then close to get across, then re-anchor to get out. Each of those events is a separate opportunity for design.
The simulation was run to improve peptide bioavailability, especially oral bioavailability, and to enable more orally deliverable peptide drugs. Five design principles follow from the mechanism. First, characterize the conformational behavior of a candidate before testing permeability. Because anchoring is conformation-dependent, the residues a peptide presents to a membrane depend on which conformation dominates in solution and at the membrane surface. The practical step is to measure or predict the dominant conformations of a cyclic peptide before membrane contact, and the exposure of side-chain residues in those conformations. Knowing which residues will be available to anchor is as important as measuring overall lipophilicity.
Second, design macrocyclic peptides with continuous hydrophobic surfaces toward orientation B. Orientation B is the only insertion mode for a closed, pre-folded peptide, and it is the orientation that favors the open-to-closed transition. A peptide that enters already closed is set up for a complete crossing. This argues for building a continuous hydrophobic surface into macrocyclic drug candidates rather than leaving the insertion orientation to chance.
Third, treat side-chain anchoring as a general permeability-enhancing strategy. The simulation's authors suggest the mechanism may generalize beyond peptides: adding a moderate aliphatic chain can increase the membrane permeability of other molecules. Appending a well-placed hydrophobic group is a standard medicinal chemistry move, and this mechanism gives it a molecular rationale. The anchor does not need to be part of a peptide backbone.
Fourth, balance water and lipid solubility. A peptide locked permanently closed maximizes permeability but may be too insoluble for formulation and too hydrophobic to survive in aqueous compartments. The chameleon strategy works because both states are accessible. Designs that cannot open, or cannot close, fail in one environment or the other.
Fifth, do not assign intrinsic permeability values to individual amino acids. The simulation reports that the influence of residues such as leucine, phenylalanine, and proline depends on position and neighboring residues. The same amino acid can promote anchoring in one sequence and bury itself in another. This is consistent with the wider peptide literature: a comparative review of linear and cyclic cancer-targeting peptides found that cyclization generally enhances binding and stability, but does not guarantee improved targeting, because some linear analogues exhibit better properties PMID 22917248 . Ring closure alone does not decide the outcome.
These principles come with a built-in tension. The open conformation exposes polar atoms to solvent, which is also what a target-binding interface requires; the closed conformation hides them, which is what a membrane crossing requires. A peptide drug often has to do both, bind a target in an aqueous compartment and then cross a membrane to reach it. The same hydrogen-bonding atoms cannot serve both jobs at once. Chameleonic behavior lets one backbone satisfy each environment in turn, but a peptide optimized for open-state binding may lose permeability, and a peptide optimized for closed-state permeation may lose affinity. A comparative review of cancer-targeting peptides found that cyclization generally improves binding and stability, yet some linear analogues outperform their cyclic counterparts on the properties that matter for targeting PMID 22917248 . The same caution applies to permeability.
Computational support for this design workflow is emerging. AutoDock CrankPep, extended to cyclic peptides, places the correct interactions among the top 10 solutions for 71% of a 38-complex dataset PMID 31505931 . Docking predicts how a peptide binds a protein, not how it crosses a bilayer, so it cannot substitute for permeability assays or membrane simulations. But it can help predict the conformations and exposed side chains that determine anchoring, which is exactly the information the four-step mechanism says to collect first.
The Linker simulation should be read alongside what the broader literature shows about cyclic and macrocyclic peptides. The first point is structural diversity. The space of macrocyclic peptides is far larger than any single molecular dynamics run can sample. A 2023 genomic study uncovered 11 novel P450-modified ribosomally synthesized peptides from four classes with distinct cross-linking patterns, markedly increasing the known structural diversity of these macrocycles PMID 38069901 . Aromatic cross-links and nonstandard residues constrain the backbone and shift the open-closed equilibrium. The four-step pathway was derived from one flexible decacyclic peptide, and it cannot be assumed to hold unchanged for heavily cross-linked macrocycles.
The second point is that environment sensitivity is not unique to lipid polarity. The cyclic octapeptide Lys2CP8 forms nanotubes in a pH-dependent manner and inhibits ice recrystallization at pH 3 but not at pH 11, likely because nanotube bundling blocks its ice-binding sites PMID 35045706 . This study concerns supramolecular assembly in water, not membrane permeation, so it is not direct evidence for the chameleon mechanism. It does demonstrate that a cyclic peptide's behavior can be gated sharply by its environment, and that a small change in conditions can switch a peptide between active and inactive states. Oral peptide designers must therefore consider the full journey of environments a molecule travels through, from stomach to intestinal lumen to enterocyte to blood.
The limits of the evidence need to be stated plainly. The four-step pathway comes from molecular dynamics simulation, not from direct experimental observation of a peptide crossing a membrane. The trajectory is a physical model constrained by force-field parameters. The simulation focuses on one specific flexible decacyclic peptide. The source reports no numerical oral bioavailability or permeability values; cyclosporine A is described only qualitatively as highly bioavailable. And the statement that the closed state is the main permeability state simplifies a dynamic equilibrium.
What would experimentally test the mechanism? The simulation makes specific, falsifiable predictions: a defined insertion orientation at the interface, a conformational conversion inside the bilayer, and re-anchoring at the far leaflet. Techniques that report the conformation of a peptide as a function of its depth within a membrane could test each prediction directly. None has been applied to this peptide. Standard permeability measurements, which report how much material crosses a barrier, cannot distinguish an open-state insert that stalls from a closed-state peptide that completes the journey.
There is also a gap between crossing a bilayer in silico and bioavailability in a human. Oral bioavailability depends on solubility in gastrointestinal fluids, resistance to proteases, efflux transporters that pump peptides back into the intestinal lumen, and first-pass metabolism. A simulation of passive diffusion through a model lipid bilayer captures only one step of that chain. The 1983 cyclosporine A formulation succeeded because the molecule's overall property set, not permeability alone, allowed oral dosing.
Four open questions follow from this mechanism. The first is prediction. Can a peptide's open-closed conformational behavior and membrane permeability be predicted from its amino acid sequence before experimental testing? The simulation shows that leucine, phenylalanine, and proline exert position-dependent effects, but it does not provide a rule for anticipating which residues will act as anchors in a given sequence. Building such a rule requires a larger set of simulations matched against permeability measurements across diverse sequences.
The second is orientation control. Why is orientation A preferred for open-state peptides, and which sequence features determine whether a peptide adopts orientation A or B at the interface? Only orientation B favors the open-to-closed transition, so a peptide that reliably inserts in orientation B would be the better permeation candidate. The structural determinants of orientation preference have not been identified.
The third is generalization. Can the side-chain anchoring principle be translated into broadly applicable medicinal chemistry rules for non-peptide oral drugs? The observation that a moderate aliphatic chain can increase the membrane permeability of other molecules is encouraging, but it is a single suggestion, not a validated rule. The position of the chain relative to hydrogen-bonding groups, and the required chain length, remain undefined.
The fourth is validation. How accurately do molecular dynamics simulations of passive membrane permeation predict actual oral bioavailability in vivo? The simulation captures one model bilayer and one peptide. Experimental permeability assays and pharmacokinetic studies are needed to calibrate its predictions. Until that calibration exists, the four-step pathway is a mechanism to test, not a result to assume.
The stated purpose of the simulation study is concrete: to improve peptide bioavailability, especially oral bioavailability, and to enable more orally deliverable peptide drugs. The mechanism gives medicinal chemists a target to engineer. A candidate oral peptide should present anchor residues in its solution-state conformation, insert at the interface in an orientation that permits closure, close completely in the hydrophobic core, and retain enough polarity to re-enter water on the far side. That is a demanding list, and it explains why oral peptides remain rare. It also turns a vague aspiration, making a peptide permeable, into a specific set of molecular events that can be simulated, measured, and designed for.
Related reading: Cyclic Hexapeptides and beta-Branched Residues for PPI Targeting, Reversible Double Linkers Reduce Amyloid Peptide Aggregation, Enzymatic Synthesis of Oligopeptides: Five Enzyme Families, Amyloid Fibrils in Therapeutics, Bionanomaterials, and Biophysics.