Membrane proteins: integral versus peripheral, topology, and lipid-protein interactions

Intuition Beginner

Proteins embedded in or attached to the cell membrane carry out most of the membrane's specific functions. Without them, the lipid bilayer would be nothing more than a passive barrier — impermeable to ions, unresponsive to signals, and unable to transport nutrients. Membrane proteins give the bilayer its biological identity.

These proteins come in two broad types. Integral membrane proteins span the entire bilayer, with portions exposed on both sides of the membrane. They include channels that let ions pass through, transporters that move molecules across, and receptors that detect signals from outside the cell. Because they cross the hydrophobic core, extracting them requires detergents that replace the lipid environment.

Peripheral membrane proteins sit on one surface of the membrane, attached through electrostatic interactions, hydrogen bonds, or lipid anchors. They do not cross the bilayer. Many peripheral proteins serve as enzymes or scaffolding, organising other proteins into signalling complexes on the cytoplasmic face.

Some membrane proteins carry carbohydrate chains on their extracellular portions, forming glycoproteins. These sugar chains act as molecular identity tags, allowing cells to recognise each other and interact correctly. Blood type, for example, is determined by glycoprotein and glycolipid carbohydrate chains on the surface of red blood cells.

Visual Beginner

The diagram shows the two major classes of membrane protein in relation to the bilayer:

• The integral protein spans the full bilayer with alpha-helical transmembrane domains (shown as cylinders). Its extracellular domain carries carbohydrate chains forming the glycocalyx. Its cytoplasmic domain interacts with intracellular signalling molecules.
• The peripheral protein sits on the cytoplasmic face, attached through electrostatic interactions with lipid head groups and with the integral protein's cytoplasmic domain.
• A GPI-anchored protein is attached to the outer leaflet through a glycosylphosphatidylinositol lipid anchor, floating above the membrane surface without spanning it.

Worked example Beginner

Glycophorin A is a single-pass integral membrane protein found in red blood cells. Let us identify its structural features and explain how it sits in the membrane.

Step 1. Locate the transmembrane domain. Glycophorin A has a single stretch of approximately 20 hydrophobic amino acids (residues 25–45) that forms an alpha helix. This length matches the $\sim 30$ Å thickness of the hydrophobic core of the bilayer, since each residue adds $\sim 1.5$ Å along the helix axis.

Step 2. Identify the two flanking domains. The N-terminal portion (residues 1–24) is on the extracellular side and carries numerous O-linked carbohydrate chains — this is the glycosylated domain. The C-terminal portion (residues 46–131) is on the cytoplasmic side and interacts with the cytoskeleton beneath the membrane.

Step 3. Explain the orientation. The positively charged residues (arginine and lysine) cluster on the cytoplasmic side of the transmembrane helix, following the positive-inside rule: the side of the membrane with more positive charges stays in the cytoplasm. This rule determines the protein's orientation during insertion into the ER membrane.

Check your understanding Beginner

Formal definition Intermediate+

Membrane proteins are classified by their mode and depth of association with the lipid bilayer. The classification reflects both structural topology and the energetic cost of removing the protein from the membrane.

Integral (transmembrane) proteins have one or more segments that traverse the hydrophobic core of the bilayer. They are operationally defined as requiring detergent or organic solvent for extraction. Two structural motifs account for nearly all integral membrane proteins:

• Alpha-helical bundles. Each transmembrane segment is an alpha helix of approximately 20 hydrophobic residues, with a span of $\sim 30$ Å matching the hydrophobic core thickness. Single-pass proteins (type I: N-extracellular, C-cytoplasmic; type II: N-cytoplasmic, C-extracellular) contain one helix. Multi-pass proteins (type III and type IV) contain two or more helices arranged as a bundle. The alpha-helical motif dominates in bacterial inner membranes and in all eukaryotic membranes.
• Beta-barrels. Antiparallel beta-strands (typically 8–24) form a closed cylindrical sheet with hydrophobic residues facing the lipid and hydrophilic residues lining the central pore. Beta-barrels are characteristic of bacterial outer membranes, the mitochondrial outer membrane, and the chloroplast outer envelope. Each strand crosses the bilayer at a tilt of approximately $45$ degrees, with alternating hydrophobic and hydrophilic residues.

Peripheral proteins associate with one face of the membrane through noncovalent interactions or lipid anchors and can be removed by changes in pH, ionic strength, or chelating agents:

• Electrostatic association with lipid head groups (e.g., spectrin on the cytoplasmic face of red blood cells).
• $N$-myristoylation (covalent attachment of a 14-carbon myristate to an N-terminal glycine; irreversible; e.g., Src-family kinases).
• $S$-palmitoylation (covalent attachment of a 16-carbon palmitate to a cysteine thiol; reversible; e.g., H-Ras, G-alpha subunits).
• Prenylation (covalent attachment of a 15-carbon farnesyl or 20-carbon geranylgeranyl to a C-terminal CaaX motif; e.g., Ras, Rho).
• GPI anchors (glycosylphosphatidylinositol covalently attached to the C-terminus; extracellular face only; e.g., alkaline phosphatase, Thy-1/CD90).

Topology of an integral membrane protein refers to the number of transmembrane segments and the orientation of the N- and C-termini relative to the membrane. The positive-inside rule (von Heijne 1986) provides a statistical predictor: cytoplasmic loops contain more arginine and lysine residues than extracellular or lumenal loops, across both prokaryotic and eukaryotic membrane proteins. The rule predicts topology from sequence with approximately 85% accuracy.

Lipid-protein interactions occur at multiple levels. The annular lipid shell is the first layer of lipid molecules in direct contact with the transmembrane surface of an integral protein. These lipids have restricted mobility (residence times of $1$–$100$ ns on the protein surface, versus $\sim 1$ ns for bulk lipid-lipid exchange), altered acyl-chain order, and can mediate specific regulatory interactions (e.g., PIP${}_2$ binding to inward-rectifier potassium channels). The annular shell also buffers hydrophobic mismatch: when the hydrophobic thickness of the protein's transmembrane domain differs from the hydrophobic thickness of the surrounding bilayer, annular lipids adjust their acyl-chain order to reduce the energetic penalty.

Key mechanism Intermediate+

Mechanism: The biological hydrophobicity scale and transmembrane helix insertion.

Whether a polypeptide segment inserts into the bilayer during co-translational translocation through the Sec61 translocon is determined by the free energy of transfer from the aqueous translocon channel to the lipid environment. Hessa, von Heijne, and colleagues (2005 Nature 433, 377-381) measured this transfer energy for each amino acid using a model system in which engineered segments of defined composition were assayed for membrane integration.

The apparent free energy of insertion for a segment of $n$ residues is:

$$\Delta G_{\text{app}}=\sum _{i=1}^n\Delta G_i^{\text{aa}}+\Delta G_{\text{termini}}$$

where $\Delta G_i^{\text{aa}}$ is the per-residue contribution of amino acid $i$ from the biological hydrophobicity scale, and $\Delta G_{\text{termini}}$ accounts for the energetic cost of bringing the polypeptide backbone across the membrane-water interface. A segment inserts as a transmembrane helix when $\Delta G_{\text{app}}$ falls below a threshold of approximately $+0.5$ kcal/mol under the assay conditions; above this threshold the segment is translocated through the channel into the lumen instead.

The scale quantifies what was previously understood qualitatively: hydrophobic residues (Ile, Leu, Val, Phe, Trp) favour insertion ($\Delta G_i^{\text{aa}}<0$), charged residues (Arg, Lys, Asp, Glu) oppose insertion ($\Delta G_i^{\text{aa}}>0$), and the contribution of each residue depends on its position within the segment (residues near the helix ends experience the interfacial environment rather than the bilayer interior). The interfacial region has its own hydrophobicity scale, the Wimley-White interfacial hydrophobicity scale (Wimley and White 1996 Nat. Struct. Biol. 3, 842-848), which differs from the bilayer-interior scale because tryptophan and tyrosine residues are strongly favoured at the interface — their indole and phenol rings partition preferentially into the head-group region.

Positive-inside rule as a thermodynamic consequence. The clustering of arginine and lysine on cytoplasmic loops follows from the insertion energetics. Each positively charged residue contributes $+2$ to $+3$ kcal/mol to $\Delta G_{\text{app}}$ when placed in a transmembrane context, making it energetically costly to translocate them across the membrane. The translocon's lateral gate releases the transmembrane helix into the bilayer while retaining the positively charged flanking segment on the cytoplasmic side — the lowest-energy configuration. The rule is therefore not a separate biological constraint but a thermodynamic consequence of the insertion energy scale.

Hydrophobic mismatch. The hydrophobic-matching principle (Mouritsen and Bloom 1984 Biophys. J. 46, 141-153) states that the bilayer locally adjusts to match the hydrophobic thickness of a transmembrane protein, and vice versa. When a transmembrane helix of hydrophobic length $d_p$ is embedded in a bilayer of hydrophobic thickness $d_l$, the free-energy penalty per unit length of mismatch is approximately:

$$\Delta G_{\text{mismatch}}\approx K_{\text{m}}\cdot (d_p-d_l{)}^2\cdot \ell$$

where $K_{\text{m}}$ is an elastic modulus for hydrophobic mismatch (of order $10$–$30$ cal/(mol $\cdot$ Å${}^2$)) and $\ell$ is the perimeter of the transmembrane domain in contact with the bilayer. For a single alpha helix of circumference $\sim 12$ Å, a mismatch of $5$ Å produces a penalty of roughly $3$–$9$ kcal/mol — enough to affect protein folding stability, oligomerisation, and lateral diffusion. The cell exploits this mismatch to sort membrane proteins along the secretory pathway: the ER membrane is thinner ($\sim 27$ Å hydrophobic core) than the plasma membrane ($\sim 32$ Å), and proteins with short transmembrane segments are retained in the ER while those with longer segments traffic forward.

Exercises Intermediate+

The free energy landscape of membrane protein folding and stability Master

The stability of an integral membrane protein in the bilayer is governed by a free energy landscape qualitatively different from that of a soluble protein. Soluble proteins fold in a uniform aqueous medium where the hydrophobic effect drives core formation and hydrogen bonding stabilises secondary structure. Membrane proteins fold in an anisotropic environment: the bilayer presents a nonpolar core flanked by polar interfacial regions, and the free-energy contributions of hydrogen bonding, van der Waals packing, and electrostatic interactions change sign or magnitude depending on where in the bilayer the residue sits.

The two-stage model of membrane protein folding (Popot and Engelman 1990 Biochemistry 29, 4031-4037) separates the process into two thermodynamically distinct steps. In the first stage, each hydrophobic segment independently inserts into the bilayer as a transmembrane alpha helix. The driving force is the burial of hydrophobic side chains away from water — the same hydrophobic effect that drives soluble protein folding, but here the "core" is the lipid bilayer itself. The helix forms because the intramolecular hydrogen bonds of the alpha helix replace the hydrogen bonds that the peptide backbone would form with water in solution; inside the low-dielectric bilayer interior, there is no water to hydrogen-bond with, so backbone hydrogen bonding becomes a dominant stabilising interaction. In the second stage, the independently inserted helices associate laterally within the bilayer to form the native helical bundle. The driving forces for helix-helix association are van der Waals packing interactions between complementary surfaces and, in some cases, specific side-chain interactions such as the GxxxG motif first identified in glycophorin A dimerisation (Russ and Engelman 2000 J. Mol. Biol. 297, 375-382).

The two-stage model has substantial experimental support. Individual transmembrane helices can be synthesised chemically and inserted into lipid bilayers independently; when mixed, they assemble into native-like oligomers with correct stoichiometry. The glycophorin A transmembrane helix dimerises through its GxxxG motif in detergent micelles, in bilayers, and even in bacterial membranes, demonstrating that the helix-helix association energy is intrinsic to the sequence and does not require cellular chaperones.

The biological hydrophobicity scale measured by Hessa et al. (2005) and the related Wimley-White whole-residue hydrophobicity scales (Wimley, Creamer, and White 1996 Biochemistry 35, 5109-5124; Wimley and White 1996 Nat. Struct. Biol. 3, 842-848) provide the quantitative foundation for predicting transmembrane segments from sequence. The Wimley-White framework distinguishes two scales: the octanol scale, which measures the free energy of transferring a peptide from water to a membrane-mimetic n-octanol phase (approximating the bilayer interior), and the interfacial scale, which measures transfer from water to the membrane-water interface (approximating the head-group region). The difference between the two scales encodes the position dependence of each amino acid's preference: tryptophan and tyrosine are strongly favoured at the interface but only moderately favoured in the bilayer interior; arginine and lysine are strongly disfavoured in both environments but slightly less so at the interface.

White and Wimley (1999) ^{[White & Wimley 1999]} reviewed the experimental basis for these scales and identified a critical puzzle: the total hydrophobic effect stabilising a membrane protein in the bilayer is considerably smaller than the hydrophobic effect stabilising a soluble protein of comparable size in water. The per-residue transfer free energies from water to bilayer are typically $0.5$–$1.0$ kcal/mol per residue (favourable), compared to $1.5$–$2.5$ kcal/mol per residue for soluble protein core burial. This means that membrane proteins have marginal thermodynamic stability — typically $\Delta G_{\text{folding}}$ of only $-2$ to $-7$ kcal/mol for the overall fold, compared to $-5$ to $-15$ kcal/mol for soluble proteins of comparable size. The practical consequence is that membrane proteins are more sensitive to mutations that destabilise packing, more prone to misfolding, and more difficult to obtain in native conformation for structural studies.

Energetics of the peptide bond in the bilayer. A single unsatisfied hydrogen bond in the bilayer interior carries a penalty of approximately $+4$ to $+6$ kcal/mol — very costly in the low-dielectric environment where there is no competing water to form alternative hydrogen bonds. This is why transmembrane segments must be fully hydrogen-bonded (as alpha helices or beta barrels): an extended, non-hydrogen-bonded polypeptide backbone would be thermodynamically prohibitive in the bilayer interior. The requirement for complete backbone hydrogen bonding explains the near-universal adoption of alpha-helical or beta-barrel folds in transmembrane domains. A loop or turn that places even one peptide bond in the bilayer interior without hydrogen bonding creates an energetic penalty comparable to the entire stabilising hydrophobic effect of a transmembrane helix.

The lipid annular shell as a thermodynamic buffer. The lipid molecules in direct contact with the transmembrane surface of an integral protein form an annular shell whose properties differ from bulk bilayer lipid. Molecular dynamics simulations (reviewed in Lee 2003 Biochim. Biophys. Acta 1612, 1-18) and experimental measurements (detergent exchange rates, electron spin resonance order parameters) show that annular lipids have higher acyl-chain order parameters (more extended chains), reduced lateral diffusion (residence times $10$–$100$ times longer than bulk exchange), and composition that depends on the protein surface: positively charged residues at the interface prefer phosphatidylserine or phosphatidylglycerol; hydrophobic grooves on the protein surface can accommodate specific lipid head groups or acyl chains.

The annular shell serves three functions. First, it buffers hydrophobic mismatch by locally adjusting the effective bilayer thickness around the protein: if the transmembrane domain is shorter than the bilayer, annular lipids compress their chains (increasing order) to thin the local region; if the domain is longer, chains stretch. Second, it provides a solvation environment that lowers the activation barrier for conformational transitions: a channel that opens by tilting its helices (e.g., MscL) changes its transmembrane perimeter, and the annular lipids redistribute to accommodate the new surface without exposing hydrophobic protein surface to water. Third, specific lipid-protein interactions within the annular shell can regulate protein function directly. The inward-rectifier potassium channel Kir2.1 requires PIP${}_2$ binding within its annular shell for activity; phospholipase C-mediated depletion of PIP${}_2$ causes rapid channel closure, coupling a lipid-signalling event to ion-channel gating through a specific annular-lipid interaction.

Membrane protein misfolding and disease. The marginal stability of membrane proteins makes them susceptible to misfolding when mutations destabilise the native fold. Cystic fibrosis, caused by the $\Delta$F508 mutation in the CFTR chloride channel, is the best-characterised example: deletion of phenylalanine 508 from the first nucleotide-binding domain reduces the thermodynamic stability of the full-length protein by approximately $2$–$3$ kcal/mol, which is sufficient to cause ER-associated degradation instead of proper folding and trafficking to the plasma membrane. The misfolded protein is recognised by the ER quality-control machinery and targeted for proteasomal degradation. Other membrane protein misfolding diseases include retinitis pigmentosa (mutations in rhodopsin that cause misfolding and ER retention) and certain forms of Charcot-Marie-Tooth neuropathy (mutations in the PMP22 peripheral myelin protein that cause misfolding and aggregation).

Structural biology of membrane proteins: crystallography, cryo-EM, and the resolution revolution Master

The structural determination of integral membrane proteins has been one of the slowest and most technically demanding areas of structural biology, for reasons that follow directly from their physical properties. Membrane proteins are unstable outside the bilayer, aggregate in aqueous solution, require detergent solubilisation that introduces heterogeneity, and resist crystallisation because their transmembrane surfaces are featureless and hydrophobic. From the first atomic-resolution membrane protein structure (the photosynthetic reaction centre of Rhodopseudomonas viridis, Deisenhofer et al. 1985 Nature 318, 618-624 — Nobel Prize 1988) to the explosion of cryo-electron microscopy structures in the 2010s, the field has undergone three methodological revolutions.

X-ray crystallography of membrane proteins requires that the protein be solubilised in detergent micelles or lipidic cubic phases, then crystallised in two or three dimensions. The detergent or lipid surrounding the transmembrane domain makes the crystals weakly diffracting and difficult to grow. The lipidic cubic phase (LCP) method, developed by Landau and Rosenbusch (1996 Proc. Natl. Acad. Sci. 93, 14532-14535), embeds the protein in a bicontinuous lipid-water matrix that mimics the bilayer environment and provides a more native-like crystallisation scaffold. LCP crystallisation was instrumental in solving the structure of bovine rhodopsin (Palczewski et al. 2000 Science 289, 739-745) and the ${\beta }_2$-adrenergic receptor (Cherezov et al. 2007 Science 318, 1258-1265). The 2012 Nobel Prize in chemistry to Kobilka and Lefkowitz recognised the structural-biology programme that produced these GPCR structures.

The pace of membrane protein crystallography was limited by two bottlenecks. First, membrane proteins are difficult to express in the quantities needed for crystallisation (typically milligrams of pure protein). Eukaryotic membrane proteins often require eukaryotic expression systems (insect cells, mammalian cells) that produce low yields. Second, the conformational flexibility of many membrane proteins (GPCRs, ion channels, transporters) means that the protein exists as a mixture of conformations in solution, and crystallisation captures only one. Conformational heterogeneity is the enemy of crystallisation: it broadens diffraction spots and reduces resolution.

Cryo-electron microscopy (cryo-EM) has largely supplanted crystallography for membrane protein structure determination since the "resolution revolution" of 2013-2015 (Kühlbrandt 2014 Science 343, 1443-1444). In single-particle cryo-EM, the protein is frozen in vitreous ice and imaged directly; computational classification of particle images separates different conformational states without requiring a homogeneous sample. This eliminates both the crystallisation bottleneck and the conformational-homogeneity requirement. Direct electron detectors (the K2 Summit camera, introduced in 2013) and improved image-processing algorithms (Bayesian particle picking, maximum-likelihood classification, gold-standard FSC resolution estimation) pushed achievable resolution from $\sim 10$ Å to $\sim 3$ Å for favourable samples.

The impact on membrane protein structural biology has been transformative. Structures that resisted crystallography for decades — the TRPV1 ion channel (Liao et al. 2013 Nature 504, 107-112, the first high-resolution cryo-EM structure of a membrane protein), the PIEZO1 mechanosensitive channel (Saotome et al. 2018 Nature 554, 481-486), the eukaryotic voltage-gated sodium channel (Shen et al. 2017 Science 355, eaal4326), and the full mitochondrial respirasome supercomplex (Guo et al. 2017 Science 357, 836-840) — were solved by cryo-EM within a few years. The method is especially powerful for large membrane protein complexes ($>150$ kDa) where the high molecular weight provides sufficient contrast in the micrographs. For smaller membrane proteins ($<100$ kDa), crystallography and NMR in detergent micelles or bicelles remain competitive.

The lipid environment matters for structure and function. A persistent theme across both crystallographic and cryo-EM studies is that the detergent or lipid environment used for structural studies affects the observed conformation. Detergent micelles differ from bilayers in curvature, lateral pressure profile, and thickness; structures solved in detergent may capture non-native conformations. The use of lipid nanodiscs (discoidal bilayer patches stabilised by membrane scaffold proteins or synthetic polymers) and saposin-lipoprotein particles (Salipro) has enabled structural studies in a more bilayer-like environment. Structures of channels and transporters in nanodiscs versus detergent show differences in the conformation of transmembrane helices near the lipid-facing surface, confirming that the lipid environment modulates protein conformation — exactly as predicted by the hydrophobic-matching and annular-lipid frameworks.

Membrane protein partitioning, lipid rafts, and the role of lipid modification Master

The spatial distribution of membrane proteins across the bilayer surface is not random. Proteins partition between coexisting lipid phases (liquid-ordered versus liquid-disordered) according to the compatibility of their transmembrane domains and lipid modifications with the physical properties of each phase. This partitioning underlies the raft hypothesis and has functional consequences for signal transduction, membrane trafficking, and pathogen entry.

Transmembrane domain partitioning. A transmembrane helix enriched in saturated-like residues (Leu, Ile, Val, Phe) and depleted in bulky or polar residues (Trp, Tyr at the interface excepted) partitions preferentially into the liquid-ordered phase, where the tightly packed, ordered acyl chains provide favourable van der Waals contacts. A helix containing unsaturated or branched residues, or one with charged residues near the bilayer centre, partitions preferentially into the liquid-disordered phase. Single-pass transmembrane proteins with long, saturated helices are therefore raft-associated; those with shorter or more heterogeneous helices are raft-excluded. The palmitoylation state of the juxtamembrane cysteines can switch a protein between raft and non-raft compartments: palmitoylation adds a saturated 16-carbon chain that partitions into the liquid-ordered phase, while depalmitoylation releases the protein into the disordered phase.

GPI-anchored proteins and rafts. GPI anchors carry two saturated acyl chains (typically C16:0 and C18:0) attached to the glycan core, which partition strongly into the liquid-ordered phase. GPI-anchored proteins are therefore constitutively raft-associated at the nanoscale level. At the mesoscale, they cluster into larger assemblies upon cross-linking (antibody-induced, ligand-induced, or cholesterol-dependent), providing the experimental basis for the raft hypothesis. The clustering is mediated not by protein-protein interactions but by the lipid-driven phase coalescence of their GPI anchor acyl chains: removing cholesterol (with methyl-$\beta$-cyclodextrin) or replacing the saturated acyl chains with unsaturated ones abolishes clustering. The cluster state regulates signalling: many GPI-anchored immunoregulatory proteins (Thy-1/CD90, CD14, CD59) signal only when clustered in rafts that recruit Src-family kinases to the inner leaflet.

Lipid modifications and raft targeting. The four major lipid modifications have distinct raft-targeting properties:

• $S$-palmitoylation targets proteins to rafts. The saturated 16-carbon palmitate partitions into the liquid-ordered phase. Reversible palmitoylation provides a switch for raft entry and exit.
• $N$-myristoylation alone provides weak membrane association but does not target to rafts; the unsaturated 14-carbon myristate has moderate preference for the disordered phase. Dual myristoylation + palmitoylation (as in Src-family kinases) provides strong raft targeting.
• Prenylation (farnesyl, geranylgeranyl) targets proteins to the disordered phase. The unsaturated isoprenoid chains have poor packing compatibility with the ordered raft environment. H-Ras (farnesylated + palmitoylated) partitions between raft and non-raft compartments depending on its palmitoylation state; K-Ras (farnesylated only, with a polybasic membrane-targeting sequence) is constitutively non-raft.
• The polybasic sequence (a cluster of lysine residues adjacent to the prenylated C-terminus, as in K-Ras4B) provides electrostatic targeting to the negatively charged inner leaflet (enriched in phosphatidylserine) but does not target to rafts.

The interplay between lipid modification and raft partitioning has been quantified using fluorescence recovery after photobleaching (FRAP), fluorescence resonance energy transfer (FRET), and single-molecule tracking. FRAP recovery half-times for raft-localised proteins are longer than for non-raft proteins (reflecting restricted diffusion in the ordered phase), and cholesterol depletion accelerates recovery by dissolving the raft domains. FRET between raft-marker pairs decreases upon cholesterol depletion, providing a distance-sensitive readout of raft-mediated co-localisation.

Pathogen exploitation of raft partitioning. Several viruses and bacterial toxins exploit raft partitioning for cell entry. Influenza virus buds from the apical surface of polarised epithelial cells through raft domains; the viral envelope is enriched in cholesterol and sphingolipids, and the haemagglutinin and neuraminidase glycoproteins are raft-targeted by their transmembrane domains. Cholera toxin binds GM1 ganglioside (a raft lipid) on the cell surface, enters through clathrin-independent endocytosis of raft domains, and traffics retrogradely to the ER where the catalytic subunit is translocated to the cytoplasm. HIV incorporates raft lipids and raft-targeted host proteins (including GPI-anchored proteins) into its envelope during budding. The therapeutic implication is that disrupting raft integrity (cholesterol depletion, sphingolipid synthesis inhibition) can block pathogen entry, though the systemic toxicity of lipid-targeting drugs has limited clinical application.

Connections Master

• Cell membranes: structure 17.02.01. The bilayer structure and the fluid mosaic model established in the prerequisite unit provide the physical substrate in which membrane proteins operate. The lipid composition (phospholipid head groups, acyl chain saturation, cholesterol content) determines the hydrophobic thickness, fluidity, and phase behaviour that constrain membrane protein folding, stability, and partitioning.

• Membrane transport 17.02.02. The integral membrane protein classification developed here — single-pass, multi-pass alpha-helical, and beta-barrel — directly feeds into the transport chapter. Ion channels are multi-pass alpha-helical proteins; porins are beta-barrels; carrier proteins alternate between conformational states through transmembrane helix rearrangements. The positive-inside rule and topology prediction are essential for annotating transporter sequences from genomes.

• Vesicle trafficking 17.02.04 pending. The sorting signals, lipid modifications, and hydrophobic-mismatch-based organelle retention introduced here determine which membrane proteins are packaged into transport vesicles and where they are delivered. Palmitoylation state, transmembrane domain length, and raft partitioning all contribute to sorting decisions in the secretory pathway.

• Cell signalling 17.07.01. The GPCR seven-transmembrane topology, the GPI-anchored protein signalling mechanism, the raft partitioning of Src-family kinases, and the lipid-mediated regulation of ion channels all originate as structural features in this unit and become functional signalling mechanisms in the signalling chapter. The membrane protein typology is the structural vocabulary of signal transduction.

• Thermodynamics and free energy 14.06.01 (chemistry section). The biological hydrophobicity scale, the insertion free energy threshold, the hydrophobic mismatch penalty, and the marginal stability of membrane protein folds all draw on the free-energy framework of the chemical-thermodynamics chapter. The $\Delta G_{\text{app}}$ threshold for transmembrane insertion is a direct application of equilibrium thermodynamics to a molecular-scale biophysical process.

• Protein structure and folding 17.01.02 pending (proposed). The two-stage model of membrane protein folding parallels the hierarchical folding model for soluble proteins, with the added constraint of an anisotropic solvent environment. The marginal stability and misfolding sensitivity of membrane proteins connect to the protein-folding quality-control machinery (ER-associated degradation, chaperone-mediated folding) treated in the protein-folding unit.

Historical notes Master

The recognition that proteins are integral components of the cell membrane evolved alongside the membrane structure field. Danielli and Davson (1935) proposed protein layers coating the bilayer surfaces, but this "sandwich" model could not account for membrane transport or for the evidence of proteins spanning the bilayer. The Singer-Nicolson fluid mosaic model (1972) ^{[Singer & Nicolson 1972]} introduced the concept of proteins as dispersed elements embedded in the bilayer, with integral proteins spanning the hydrophobic core. The model resolved the thermodynamic inconsistency of the sandwich (hydrophilic protein surfaces binding to hydrophobic acyl chains) by recognising that transmembrane proteins are themselves amphipathic.

The biochemical classification of membrane proteins into integral and peripheral types emerged from detergent-extraction experiments in the 1960s and 1970s. Proteins that required detergent for solubilisation (integral) were distinguished from those released by salt washes or chelating agents (peripheral). The operational definition remains in use.

The signal hypothesis of Blobel and Dobberstein (1975 J. Cell Biol. 67, 835-851) established the mechanism by which integral membrane proteins are inserted co-translationally: a hydrophobic signal sequence directs the nascent polypeptide to the ER membrane, where the Sec61 translocon mediates either complete translocation (for soluble proteins) or lateral partitioning of transmembrane segments into the bilayer (for membrane proteins). Blobel received the 1999 Nobel Prize in physiology or medicine for this work.

The positive-inside rule was identified by von Heijne (1986 EMBO J. 5, 3021-3027) through statistical analysis of charged-residue distributions in bacterial inner membrane proteins with known topology. The rule was subsequently validated in eukaryotic membrane proteins and remains the most reliable sequence-based predictor of membrane protein topology, forming the basis of modern topology-prediction algorithms (TMHMM, MEMSAT, Phobius).

The Wimley-White hydrophobicity scales (1996) and the biological hydrophobicity scale of Hessa and von Heijne (2005) provided the first quantitative, experimentally measured free-energy scales for amino acid transfer between water, the membrane interface, and the bilayer interior. These scales replaced the earlier Kyte-Doolittle (1982 J. Mol. Biol. 157, 105-132) and Engelman-Steitz (1984) scales, which were derived from theoretical considerations and vapour-phase transfer data rather than from measurements in actual membrane-mimetic environments. The quantitative scales made transmembrane segment prediction a thermodynamic calculation rather than a heuristic exercise.

Structural biology of membrane proteins progressed from the first high-resolution structure (the photosynthetic reaction centre, Deisenhofer et al. 1985, Nobel Prize 1988) through the slow accumulation of crystallographic structures in the 1990s and 2000s, to the cryo-EM revolution of the 2010s that democratized membrane protein structure determination. The 2017 Nobel Prize in chemistry to Dubochet, Frank, and Henderson recognised cryo-EM as a general method for high-resolution structure determination, and its application to membrane proteins has been the most transformative consequence.

Bibliography Master

Primary literature.

Singer, S. J. & Nicolson, G. L., "The fluid mosaic model of the structure of cell membranes," Science 175 (1972), 720-731.

Blobel, G. & Dobberstein, B., "Transfer of proteins across membranes I. Presence of proteolytically processed and unprocessed nascent immunoglobulin light chains on membrane-bound ribosomes of murine myeloma," J. Cell Biol. 67 (1975), 835-851.

von Heijne, G., "The distribution of positively charged residues in bacterial inner membrane proteins correlates with the trans-membrane topology," EMBO J. 5 (1986), 3021-3027.

Deisenhofer, J., Epp, O., Miki, K., Huber, R. & Michel, H., "Structure of the protein subunits in the photosynthetic reaction centre of Rhodopseudomonas viridis at 3 Å resolution," Nature 318 (1985), 618-624.

Kyte, J. & Doolittle, R. F., "A simple method for displaying the hydropathic character of a protein," J. Mol. Biol. 157 (1982), 105-132.

Wimley, W. C. & White, S. H., "Experimentally determined hydrophobicity scale for proteins at membrane interfaces," Nat. Struct. Biol. 3 (1996), 842-848.

Wimley, W. C., Creamer, T. P. & White, S. H., "Experimental measurements of the hydrophobic effect in protein folding: the solvation of aromatic and aliphatic hydrocarbons," Biochemistry 35 (1996), 5109-5124.

White, S. H. & Wimley, W. C., "Membrane protein folding and stability: physical principles," Annu. Rev. Biophys. Biomol. Struct. 28 (1999), 319-365.

Popot, J.-L. & Engelman, D. M., "Membrane protein folding and oligomerization: the two-stage model," Biochemistry 29 (1990), 4031-4037.

Mouritsen, O. G. & Bloom, M., "Mattress model of lipid-protein interactions in membranes," Biophys. J. 46 (1984), 141-153.

Engelman, D. M. & Steitz, T. A., "The spontaneous insertion of proteins into and across membranes: the helical hairpin hypothesis," Cell 23 (1981), 411-422.

Hessa, T., Kim, H., Bihlmaier, K., Lundin, C., Boekel, J., Andersson, H., Nilsson, I., White, S. H. & von Heijne, G., "Recognition of transmembrane helices by the endoplasmic reticulum translocon," Nature 433 (2005), 377-381.

Russ, W. P. & Engelman, D. M., "The GxxxG motif: a framework for transmembrane helix-helix association," J. Mol. Biol. 296 (2000), 911-919.

Landau, E. M. & Rosenbusch, J. P., "Lipidic cubic phases: a novel concept for the crystallization of membrane proteins," Proc. Natl. Acad. Sci. USA 93 (1996), 14532-14535.

Palczewski, K. et al., "Crystal structure of rhodopsin: a G protein-coupled receptor," Science 289 (2000), 739-745.

Cherezov, V. et al., "High-resolution crystal structure of an engineered human ${\beta }_2$-adrenergic G protein-coupled receptor," Science 318 (2007), 1258-1265.

Liao, M., Cao, E., Julius, D. & Cheng, Y., "Structure of the TRPV1 ion channel determined by electron cryo-microscopy," Nature 504 (2013), 107-112.

Saotome, K., Murthy, S. E., Kefauver, J. M., Whitwam, T., Patapoutian, A. & Ward, A. B., "Structure of the mechanically activated ion channel Piezo1," Nature 554 (2018), 481-486.

Lee, A. G., "Lipid-protein interactions in biological membranes: a structural perspective," Biochim. Biophys. Acta 1612 (2003), 1-18.

Textbook and monograph.

Alberts, B. et al., Molecular Biology of the Cell, 7th ed. (Garland Science, 2022), Chs. 10-11 Membrane Structure; Membrane Transport.

Lodish, H. et al., Molecular Cell Biology, 9th ed. (W. H. Freeman, 2021), Ch. 10 Membrane Proteins.

White, S. H. & Wimley, W. C., "Membrane protein folding and stability: physical principles," Annu. Rev. Biophys. Biomol. Struct. 28 (1999), 319-365.

Gennis, R. B., Biomembranes: Molecular Structure and Function (Springer, 1989).

Hille, B., Ion Channels of Excitable Membranes, 3rd ed. (Sinauer, 2001).