How to choose the right membrane protein mimetic for your research

What are the considerations when deciding upon which option is best to present a target membrane protein?

In this article Riccardo Cavalieri and Steven Harborne discuss the subject and present examples of the current state of play within the field.

Introduction

Membrane proteins present numerous technical challenges to experimental research. They are expressed at low levels, are often highly dynamic, and can become unstable when removed from the native lipid bilayer. The cell lipid environment, in fact, plays a central role in maintaining their structural integrity and functional dynamics, and its disruption can lead to misfolding, aggregation, or loss of activity. As a result, careful consideration of the membrane mimetic system to be chosen is required from the earliest stages of experimental design.

Choosing a membrane mimetic for membrane protein purification requires balancing biological relevance against experimental tractability. If a membrane protein cannot be successfully extracted from its native membrane and maintained in a soluble, functional state, other factors become largely irrelevant. The choice of membrane mimetic is thus often dictated by the necessity to enable efficient extraction of the target protein, and provide sufficient stability for purification, storage, and subsequent characterisation, rather than by considerations on preserving a “native–like” environment.

What is a membrane protein mimetic?

Numerous membrane mimetics with distinct physicochemical properties are available to scientists for the study of membrane proteins. They can be grouped in four broad categories: detergents, polymer lipid particles (PLP) such as styrene maleic acid lipid particles (SMALPs), protein-based nanodiscs, and liposomes.

• Detergents
  • Detergents are the most established and versatile approach, with extensive precedent in the literature; however, they strip away most native lipids and the harsher among them can destabilize some fragile membrane proteins.
• SMALPs
  • In contrast, polymer based nanodiscs such as SMALPs enable direct extraction of proteins from the membrane together with their surrounding native lipid environment, better preserving physiologically relevant interactions. Despite this advantage, purification can be more challenging, the resulting particles are often heterogeneous, and the methodology is less firmly established.
• Nanodiscs
  • Nanodiscs provide a controlled lipid bilayer environment with defined particle size, making them particularly attractive for structural studies such as cryo-electron microscopy (cryo-EM), although they require prior solubilization in detergent and optimization of lipid composition and reconstitution conditions.
• Liposomes
  • Finally, liposomes most closely resemble biological membranes. The possibility to create electrochemical gradients across the membrane makes them the gold standard for functional assays. They are also widely used in mechanistic studies and antibody generation. Nonetheless, their preparation is complex and requires that the sample has previously been purified in detergents. Liposomes can also be fragile, exhibiting limited stability during storage and freeze–thaw cycles before starting to leak.

Membrane mimetics are not just passive surfactants that keep the protein in solution, but active physical environments that reshape the conformational behaviour of membrane proteins. Rising evidence from cryo-EM structures, functional assays, biophysical measurements, computational studies, and comparative analyses across detergents, protein nanodiscs, polymers, and liposomes points to the fact that the choice of membrane mimetic fundamentally determines which regions of a protein’s conformational space are accessible or suppressed.

The current state of play

The first point of note is that retention of endogenous lipids does not equate to preservation of native membrane physics. Nanodiscs (polymer or MSP‑based) co‑extract or reconstitute lipids, yet systematically alter their physical properties: lipid packing, lateral pressure profiles, and bilayer elasticity. Whilst polymers that directly extract proteins from the membrane (e.g. SMA, DIBMA, etc) are often thought of as native-like environment, native lipids can coexist with non-native mechanical constraints, leading to conformational biases.

Secondly, stability (especially as measured by thermal shift assays) is often considered a good proxy for correct, physiologically relevant folding and sample quality. While a good stability of the sample is certainly an indicator of tractability for structural studies, recent data on the neurotensin receptor and ABC transporters ABCB10, ABCB25 and LmrA, show that detergents such as LMNG and GDN, as well as small MSP nanodiscs (especially the most used MSP1D1 based), increase stability, homogeneity, and cryo-EM tractability of the protein samples, but do so at the expenses of the conformational ensemble sampled. In both GPCRs and ABC transporters, this stabilisation correlates with reduced ligand efficacy, altered basal activity, and dampened regulatory responses.

Third, recent publications in the literature collectively show that single structures are insufficient descriptors of membrane protein function. High resolution cryo-EM structures frequently capture only the most stabilised state of the protein, and investigating the protein in multiple membrane mimetic environment might be necessary to gain a more complete understanding of its conformational flexibility. An example of this approach has been a paper by Hoffman and colleagues on the ABC transporter MsbA, whose structure was solved in 12 different membrane mimetics (detergents and nanodiscs).

So, what’s the best fit for different applications?

A distinction between structural resolution and functional or biophysical interrogation thus seems to be required when approaching a new membrane protein: no single mimetic optimally serves both objectives, and an exhaustive approach will often need to study the protein of interest in multiple environments.

For single particle cryo-EM applications, success appears to correlate with mimetics that impose strong constraints (high stability and rigidity, low population heterogeneity), such as rigid detergents (LMNG, GDN) or small nanodiscs. These environments are optimal to yield high resolution structures but bias the system towards deep, narrow free energy minima. For proteins with strong lipid dependence (GPCRs, transporters), this often means enrichment of preactivated or artificially stabilised states. Larger nanodiscs (≥18 nm) partially alleviate these effects by reducing lateral pressure, albeit at the cost of increased heterogeneity and potentially lower resolution. Comparative multi environment approaches can provide complementary structural snapshots that collectively represent a broader conformational landscape.

For functional and biophysical assays (e.g. SPR, MST, DSF, etc), the logic is inverted. Mimetics that pre-stabilise basal states or suppress conformational heterogeneity reduce the dynamic range of measurable response. Larger nanodiscs, liposomes, or fast exchange detergents (e.g. DM/DDM) perform better because they preserve the energetic balance required for ligands or substrates to redistribute populations across states. Crucially, assays that measure changes rather than absolute values are the most sensitive to distortions introduced by the choice of membrane mimetic.

Conclusion: the value of a balanced approach

In conclusion, mimetic choice is not simply about preserving compositional similarity to native membranes, or maximising stability of the sample. While the ideal system would preserve the native membrane environment for downstream analyses, it must also allow the protein to be extracted, stabilised, purified, and handled under experimental conditions. If the target protein proves tractable and the choice of membrane mimetics is opened to the scientist, careful consideration is necessary to avoid artificially biasing the protein towards a narrower conformational landscape. Moreover, due attention is to be given to the desired downstream applications for the sample and to selecting approaches that are most appropriate for addressing the research question of interest.

Bibliography

• Bower et al., 2026 – Protein Science
  Stabilization versus flexibility: Detergent‑dependent trade‑offs in neurotensin receptor 1 GPCR ensembles
  → Established the ΔG_conf vs ΔG_ligand partitioning and showed that high stability (LMNG) can suppress ligand‑induced stabilisation.
• Nouel Barreto et al., 2025 – FEBS Letters
  ABC transporter activity is affected by the size of lipid nanodiscs
  → Demonstrated that small MSP nanodiscs bias basal activity and dampen regulatory responses, while large nanodiscs approximate liposomes functionally.
• Arcario et al., 2025 – BBA Biomembranes
  Examining the thermotropic properties of large circularized nanodiscs
  → Quantified how increasing nanodisc diameter progressively recovers bulk bilayer properties, but never fully matches liposomes for complex lipid mixtures.
• Real Hernandez & Levental, 2023 – Biophysical Journal
  Lipid packing is disrupted in copolymeric nanodiscs compared with intact membranes
  → Showed that polymer nanodiscs alter lipid headgroup packing and lateralpressure, even when native lipids are retained.
• Johansen et al., 2023 – Biochimie
  Travel light: Essential packing for membrane proteins with an active lifestyle
  → Provided the thermodynamic and kinetic framework linking detergency, conformational bias, ligand retention, lipid exchange, and carrier choice.
• Hägg et al., 2026 – BBA – Biomembranes
  How receptor conformation depends on lipid nanodisc size: Adenosine A₂A receptor and implications for class A GPCR proteins
  → Demonstrated GPCR‑specific nanodisc‑size–dependent conformational bias
• Fernandes & Zoonens, 2026 – Current Opinion in Structural Biology
  Trends in the use of amphipathic environments and future perspectives for determining the structure of membrane proteins by cryo‑EM
  → Positioned detergents, nanodiscs, polymers, Salipro, and emerging systems within cryo‑EM practice, explicitly emphasising environment‑dependent structural bias and the need for cross‑validation.
• Harpole & Delemotte, 2018 – BBA – Biomembranes
  Conformational landscapes of membrane proteins delineated by enhanced sampling molecular dynamics simulations
• Hoffman et al., 2025 – Structure
  The ABC transporter MsbA in a dozen environments