Membrane Structure

Introduction

Every living cell is defined by its plasma membrane — a thin, flexible barrier only ~5 nm thick that separates the controlled chemical environment of the cytoplasm from the unpredictable exterior. But cell membranes are far more than passive wrappers. They are dynamic, asymmetric, two-dimensional fluids studded with proteins that carry out transport, signaling, adhesion, and enzymatic catalysis. Internal membranes further divide the eukaryotic cell into compartments — the endoplasmic reticulum, Golgi apparatus, mitochondria, lysosomes — each with a distinct lipid and protein composition tailored to its function.

This lesson examines membrane structure in depth: the chemistry of the lipid bilayer, the physical principles governing its fluidity and organization, the diverse ways in which proteins associate with membranes, and the computational tools used to analyze membrane composition and predict membrane protein topology. Understanding membrane architecture is essential for everything that follows — from membrane transport and signaling to vesicle trafficking and energy conversion.

10.1 — The Lipid Bilayer

Phosphoglycerides, Sphingolipids, and Sterols Are the Major Membrane Lipids

Cell membranes contain three major classes of lipid molecules, each contributing distinct physical and functional properties:

Phosphoglycerides (also called glycerophospholipids) account for more than half of the lipid molecules in most cell membranes. Each has a glycerol backbone with two fatty acid chains esterified at positions 1 and 2, and a phosphate-linked head group at position 3. The identity of the head group — choline, ethanolamine, serine, or inositol — determines the specific phospholipid and influences its interactions with proteins and other lipids. Phosphatidylinositol (PI) is particularly important in cell signaling: kinases phosphorylate its inositol ring to produce phosphoinositides (such as PIP₂ and PIP₃) that recruit signaling proteins to specific membrane surfaces.

Sphingolipids share the general amphipathic architecture but are built on a sphingosine backbone rather than glycerol. Sphingomyelin, the most common sphingolipid, has a phosphocholine head group and resembles phosphatidylcholine in shape. Glycosphingolipids carry one or more sugar residues and are found exclusively in the outer leaflet of the plasma membrane, where they contribute to the glycocalyx and participate in cell recognition. Gangliosides — glycosphingolipids with sialic acid residues — are particularly abundant in neuronal membranes.

Cholesterol constitutes up to 20–25% of the lipid molecules in animal cell plasma membranes. Its rigid steroid ring structure is oriented with the small hydroxyl group near the polar head groups of neighboring phospholipids, while its hydrocarbon tail extends into the hydrophobic core. This positioning allows cholesterol to modulate the physical properties of the bilayer in critical ways, as we shall see.

Phospholipids Spontaneously Form Bilayers

The defining property of membrane lipids is that they are amphipathic — each molecule has a hydrophilic (water-loving) head and hydrophobic (water-fearing) tails. When placed in water, phospholipids spontaneously self-assemble into structures that shield their hydrophobic tails from water. The thermodynamic driving force is the hydrophobic effect: burying the nonpolar tails increases the entropy of the surrounding water molecules, which would otherwise be forced into ordered cage-like arrangements around exposed hydrophobic surfaces.

The geometry of phospholipids — a relatively small head group and two bulky fatty acid tails — favors the formation of bilayers (flat sheets) rather than micelles (spheres formed by single-tailed lipids like detergents). In a bilayer, the hydrophobic tails face inward, forming an oily interior ~3 nm thick, while the hydrophilic heads face the aqueous environment on both sides. Bilayers spontaneously seal to form closed compartments, because a free edge would expose the hydrophobic interior to water — an energetically unfavorable state.

This self-sealing property means that membranes can be torn and will spontaneously reseal, and that vesicles can bud from one compartment and fuse with another without ever breaching the permeability barrier. It is a fundamental requirement for the compartmentalization of the cell.

The Lipid Bilayer Is a Two-Dimensional Fluid

The bilayer is not a rigid structure. Individual lipid molecules move constantly within the plane of the membrane by lateral diffusion — a single phospholipid can traverse the length of a bacterial cell (~1 μm) in about one second. Lipids also rotate rapidly around their long axis and flex their hydrocarbon tails.

However, the spontaneous movement of a lipid molecule from one leaflet to the other — called transverse diffusion or flip-flop — is extremely rare without enzymatic assistance. Flipping a polar head group through the hydrophobic interior of the bilayer is energetically costly. Specific enzymes catalyze this movement:

• Flippases — move specific phospholipids from the outer to the inner leaflet (ATP-dependent)
• Floppases — move phospholipids from the inner to the outer leaflet (ATP-dependent)
• Scramblases — move phospholipids in both directions, randomizing the distribution (Ca²⁺-activated, energy-independent)

This combination of rapid lateral diffusion and restricted transverse diffusion is what makes the bilayer a two-dimensional fluid — liquid within each leaflet, but with a stable distinction between the two leaflets.

Fluidity Depends on Composition

The fluidity of a membrane — how easily its lipid molecules move relative to one another — is determined by several factors:

Fatty acid chain length. Shorter chains interact less with neighboring lipids, increasing fluidity. Most membrane fatty acids are 16–18 carbons long.

Degree of unsaturation. Each cis double bond introduces a kink in the hydrocarbon chain, preventing tight packing and increasing fluidity. Polyunsaturated fatty acids (with multiple double bonds) have a particularly strong fluidizing effect. Bacterial membranes adjust their ratio of saturated to unsaturated fatty acids in response to temperature to maintain optimal fluidity — a process called homeoviscous adaptation.

Cholesterol. Cholesterol has a dual effect. At physiological temperatures (37°C), the rigid steroid ring restricts the movement of neighboring phospholipid tails, reducing fluidity and making the membrane less permeable to small water-soluble molecules. At low temperatures, cholesterol prevents solidification by disrupting the regular packing of fatty acid chains. In both cases, cholesterol acts as a fluidity buffer, broadening the temperature range over which the membrane remains in a functional liquid-crystalline state.

Lipid Bilayers Can Form Domains of Different Compositions

Cell membranes are not homogeneous mixtures. Different lipids are not uniformly distributed but can organize into microdomains with distinct compositions and physical properties. The most studied are lipid rafts — transient, nanoscale assemblies enriched in cholesterol and sphingolipids, which pack together more tightly than the surrounding glycerophospholipid-rich membrane.

Lipid rafts are more ordered (less fluid) than the bulk membrane and are sometimes described as existing in a liquid-ordered (Lo) phase, distinct from the liquid-disordered (Ld) phase of the surrounding membrane. They serve as platforms for organizing membrane proteins involved in signal transduction, membrane trafficking, and pathogen entry. GPI-anchored proteins and certain transmembrane proteins preferentially associate with rafts.

The existence, size, and lifetime of lipid rafts in living cells remain topics of active research. While artificial membranes clearly separate into Lo and Ld domains, the situation in living cells is more complex, with rafts likely being small (<200 nm), short-lived, and dynamically regulated by the cortical cytoskeleton.

Lipid Droplets Are Surrounded by a Lipid Monolayer

Not all lipid-bounded structures in the cell are enclosed by bilayers. Lipid droplets — the cell’s primary energy storage organelles — consist of a core of neutral lipids (triacylglycerols and cholesterol esters) surrounded by a phospholipid monolayer with the hydrophobic tails facing inward toward the neutral lipid core and the hydrophilic heads facing the cytoplasm. This monolayer is studded with specific proteins (the perilipins and lipid-metabolizing enzymes) that regulate lipid storage and mobilization.

Lipid droplets originate from the endoplasmic reticulum, where neutral lipids accumulate between the two leaflets of the ER membrane and bud off as monolayer-enclosed droplets. They are now recognized as dynamic organelles that interact with mitochondria, peroxisomes, and autophagosomes.

Asymmetry of the Lipid Bilayer Is Functionally Important

The two leaflets of the plasma membrane have markedly different lipid compositions, an asymmetry that is established during biosynthesis in the ER and actively maintained by flippases and floppases:

This asymmetry has profound functional consequences. Phosphatidylserine is normally confined to the inner leaflet by flippases. During apoptosis, scramblases are activated and flippases are inactivated, exposing PS on the outer surface. Phagocytes express receptors that recognize exposed PS as an “eat me” signal, triggering engulfment of the dying cell. PS exposure is also essential for blood coagulation, where it provides a negatively charged surface for the assembly of clotting factor complexes.

Phosphatidylinositol on the inner leaflet is the substrate for PI kinases that generate phosphoinositide signaling lipids. PI(4,5)P₂, for example, is cleaved by phospholipase C to produce the second messengers IP₃ and diacylglycerol (DAG), while PI(3,4,5)P₃ recruits signaling proteins containing PH domains to the membrane — a critical step in the PI3K/Akt signaling pathway.

Lipidomics and Membrane Modeling

Lipidomics Data Analysis

Lipidomics is the comprehensive analysis of all lipid species in a cell, tissue, or organism. Modern lipidomics relies on mass spectrometry (MS), typically coupled with liquid chromatography (LC-MS/MS), to identify and quantify hundreds to thousands of distinct lipid species in a single experiment.

A typical lipidomics workflow involves lipid extraction (using organic solvents like the Bligh-Dyer or Matyash methods), chromatographic separation, mass spectrometric detection (identifying lipids by their mass-to-charge ratio and fragmentation patterns), and computational analysis. The data reveal the lipid composition of specific membranes, changes in lipid profiles during disease or development, and the effects of drugs or dietary interventions on membrane composition.

Two major databases organize lipid structural and classification information:

Membrane Simulations

Computational modeling of membranes has become a powerful tool for understanding bilayer behavior at the molecular level. All-atom molecular dynamics (MD) simulations can model every atom in a lipid bilayer, capturing detailed interactions but limited to small systems and short timescales (microseconds).

Coarse-grained (CG) simulations, particularly those using the Martini force field, group several atoms into single interaction sites, reducing computational cost by orders of magnitude. CG simulations can model large membrane patches (thousands of lipids), membrane protein insertion, lipid raft formation, and vesicle budding on timescales of milliseconds. The CHARMM-GUI Membrane Builder is a widely used web tool for setting up membrane simulations with specific lipid compositions.

Membrane Protein Topology Prediction

A central challenge in membrane bioinformatics is predicting which segments of a protein span the membrane. Several tools address this:

• TMHMM — a hidden Markov model that identifies transmembrane α-helical segments from sequence; one of the most widely used topology predictors
• Phobius — combines transmembrane topology and signal peptide prediction in a single model, avoiding the common error of confusing signal peptides with transmembrane helices
• DeepTMHMM — a deep-learning successor to TMHMM with improved accuracy, distinguishing α-helical and β-barrel membrane proteins and handling signal peptides

These tools take a protein’s amino acid sequence as input and output a per-residue prediction of whether each position is inside the cell, in the membrane, or outside the cell. The predictions are based on the characteristic pattern of ~20 hydrophobic amino acids that form a membrane-spanning α helix.

Membrane protein sequences are enriched in hydrophobic amino acids (leucine, isoleucine, valine, phenylalanine, alanine) within their transmembrane segments. We can use sequence analysis to explore this composition:

let tm_helix = "LLILALLAVFAGAILIDPKRRIAQ"
let soluble = "EDKATLQNELKAQFGEYVNPEHWD"
let tm_props = Struct.protein_props(tm_helix)
let sol_props = Struct.protein_props(soluble)
print("Transmembrane helix properties:")
print(tm_props)
print("Soluble domain properties:")
print(sol_props)

The transmembrane segment encodes a stretch of predominantly hydrophobic amino acids (Leu, Ile, Val, Ala) — the hallmark pattern that topology prediction algorithms detect.

10.2 — Membrane Proteins

Membrane Proteins Can Be Associated with the Lipid Bilayer in Various Ways

Proteins interact with membranes through three fundamentally different mechanisms:

About 30% of all genes in a typical genome encode membrane proteins, underscoring their biological importance. Membrane proteins carry out nearly all of the dynamic functions of membranes: selective transport of solutes, reception and transduction of extracellular signals, cell-cell adhesion, enzymatic catalysis, and anchoring of the cytoskeleton.

Lipid Anchors Control Membrane Localization

Some proteins are attached to the membrane not by a transmembrane segment but by a covalently linked lipid anchor:

GPI-anchored proteins are found exclusively on the extracellular surface and can be released by the enzyme phospholipase C, which cleaves the GPI linkage. Many GPI-anchored proteins associate with lipid rafts. Palmitoylation is unique among lipid modifications in being reversible, allowing dynamic regulation of membrane association — a protein can be palmitoylated to recruit it to the membrane and depalmitoylated to release it.

Transmembrane Proteins Cross as α Helices or β Barrels

Transmembrane proteins span the bilayer using one of two structural motifs:

α-Helical transmembrane proteins are by far the most common. A membrane-spanning α helix requires approximately 20–25 hydrophobic amino acids to cross the ~3 nm hydrophobic core of the bilayer (3.6 residues per turn × 0.15 nm rise per residue × 20 residues ≈ 3 nm). The side chains of these residues face outward, interacting with the hydrocarbon tails of the lipid bilayer. Proteins may cross the membrane once (single-pass) or multiple times (multi-pass).

β-Barrel transmembrane proteins are found in the outer membranes of bacteria, mitochondria, and chloroplasts. They consist of antiparallel β strands arranged in a cylindrical barrel, with hydrophobic residues on the exterior (facing lipids) and hydrophilic residues lining the interior channel. β barrels always have an even number of strands (typically 8–22).

let phosphatidylcholine = "CCOP(=O)([O-])OCC(COC(=O)CCCCCCCCCCCCCCC)OC(=O)CCCCCCC/C=C\\CCCCCCCC"
let cholesterol = "CC(C)CCCC(C)C1CCC2C1(CCC3C2CC=C4C3(CCC(C4)O)C)C"
print("Phosphatidylcholine properties:")
print(Chem.properties(phosphatidylcholine))
print("Cholesterol properties:")
print(Chem.properties(cholesterol))

Phospholipids and cholesterol have dramatically different molecular properties — phospholipids are larger, more flexible amphipathic molecules, while cholesterol’s rigid steroid ring gives it distinctive physical characteristics that explain its role as a fluidity buffer.

Transmembrane α Helices Often Interact with One Another

In multi-pass transmembrane proteins, the α helices within the membrane are not isolated rods — they interact with each other through van der Waals contacts, hydrogen bonds between polar side chains buried in the bilayer, and specific packing motifs. The most common helix-helix interaction motif is the GxxxG motif (glycine-any-any-any-glycine), where the small glycine residues create a flat surface that allows two helices to pack closely together.

These helix-helix interactions determine the three-dimensional structure of multi-pass proteins within the membrane. They create binding sites for substrates and ligands, form transport channels, and enable the conformational changes that drive protein function.

Some β Barrels Form Large Channels

While α-helical transmembrane proteins typically form narrow, selective channels, β-barrel proteins can form much larger pores. Porins in the outer membranes of Gram-negative bacteria contain β barrels with 16–18 strands that create water-filled channels ~1 nm in diameter, large enough for small molecules (sugars, amino acids, ions) to diffuse through passively.

The largest β-barrel channels include TOM40 (the main protein import channel in the mitochondrial outer membrane, with 19 β strands) and bacterial autotransporter domains. The voltage-dependent anion channel (VDAC) in the mitochondrial outer membrane is also a β barrel and is the primary pathway for metabolites crossing the outer mitochondrial membrane.

Many Membrane Proteins Are Glycosylated

Most proteins exposed on the outer surface of the plasma membrane carry oligosaccharide chains attached to them — they are glycoproteins. Glycosylation occurs in the endoplasmic reticulum and Golgi apparatus and comes in two forms:

• N-linked glycosylation — oligosaccharides attached to the nitrogen of asparagine (Asn) side chains, at the consensus sequence Asn-X-Ser/Thr (where X is any amino acid except proline)
• O-linked glycosylation — oligosaccharides attached to the oxygen of serine (Ser) or threonine (Thr) side chains; no strict consensus sequence

Glycosylation occurs exclusively on the extracellular (or lumenal) side of the membrane, never on the cytoplasmic side. This reflects the topology of the biosynthetic pathway: sugars are added in the ER lumen and Golgi lumen, which are topologically equivalent to the cell exterior.

Membrane Proteins Can Be Solubilized in Detergents

Studying membrane proteins is technically challenging because they are insoluble in aqueous solution once removed from the lipid bilayer. Detergents — amphipathic molecules with a single hydrophobic tail and a hydrophilic head — solve this problem by replacing the lipid annulus around the protein with a ring of detergent molecules, forming a water-soluble protein-detergent micelle.

Commonly used detergents include Triton X-100 (nonionic, mild), SDS (ionic, denaturing), and DDM (n-dodecyl-β-D-maltoside) (nonionic, mild, widely used for structural studies). The choice of detergent is critical: too harsh a detergent (like SDS) denatures the protein, while too mild a detergent may fail to extract it from the membrane. Modern approaches using nanodiscs (small lipid bilayer patches enclosed by scaffold proteins) and styrene-maleic acid lipid particles (SMALPs) allow membrane proteins to be solubilized while remaining embedded in native-like lipid environments.

Bacteriorhodopsin: A Model Multi-Pass Transmembrane Protein

Bacteriorhodopsin from the archaeon Halobacterium salinarum was the first membrane protein whose three-dimensional structure was determined at high resolution. It traverses the membrane as seven α helices arranged in a bundle, with the light-absorbing chromophore retinal bound covalently in the center of the bundle.

Bacteriorhodopsin functions as a light-driven proton pump: when retinal absorbs a photon of green light, it undergoes a conformational change (photoisomerization from all-trans to 13-cis) that triggers a series of proton transfer steps through the protein, pumping one H⁺ from the cytoplasm to the extracellular space per photon absorbed. The resulting proton gradient drives ATP synthesis.

The seven-helix bundle architecture of bacteriorhodopsin is shared by the enormous family of G protein–coupled receptors (GPCRs), which include receptors for hormones, neurotransmitters, odors, and light (rhodopsin in vertebrate vision). GPCRs constitute the largest family of cell-surface receptors and the target of approximately 35% of approved drugs.

Membrane Proteins Often Function as Large Complexes

Many membrane proteins do not work alone but assemble into multi-subunit complexes. The respiratory chain complexes in the inner mitochondrial membrane (Complex I through Complex V) contain dozens of subunits each, both nuclear- and mitochondrially-encoded. The photosystem complexes in chloroplast thylakoid membranes are similarly elaborate.

Even simpler membrane proteins often form functional dimers or oligomers. Receptor tyrosine kinases dimerize upon ligand binding, bringing their intracellular kinase domains into proximity for cross-phosphorylation. Ion channels are typically tetramers (K⁺ channels) or pentamers (ligand-gated channels), with each subunit contributing to the central pore.

Many Membrane Proteins Diffuse in the Plane of the Membrane

Like lipids, many membrane proteins undergo lateral diffusion within the plane of the bilayer, although they move more slowly than lipids (roughly 10–100 times slower) because of their larger size. This mobility was first demonstrated by the Frye-Sendai experiment (1970), in which human and mouse cells were fused: within minutes, human and mouse membrane proteins intermixed across the surface of the hybrid cell.

However, not all membrane proteins are freely mobile. Their diffusion can be restricted by several mechanisms:

• Tethering to the cortical cytoskeleton (actin filaments beneath the membrane)
• Tethering to the extracellular matrix
• Tethering to proteins on adjacent cells (at cell junctions)
• Confinement by diffusion barriers (such as the tight junctions that separate apical from basolateral membrane domains in epithelial cells)

Cells Can Confine Proteins to Specific Membrane Domains

Epithelial cells provide the clearest example of membrane domain organization. These cells have an apical surface (facing the lumen of a duct or the exterior) and a basolateral surface (facing the underlying tissue and neighboring cells). The two domains have different protein and lipid compositions, maintained by tight junctions that form a seal between adjacent cells, preventing the lateral diffusion of membrane proteins and lipids between the apical and basolateral domains.

This domain organization is essential for epithelial function. The apical surface of an intestinal epithelial cell carries nutrient transporters that absorb sugars and amino acids from the lumen, while the basolateral surface carries different transporters that export these nutrients into the bloodstream. If the two sets of transporters could mix freely, directional transport would be impossible.

The Cortical Cytoskeleton Gives Mechanical Strength

The plasma membrane by itself is mechanically fragile — a pure lipid bilayer is easily deformed and ruptured. Cells reinforce their membranes with a cortical cytoskeleton — a meshwork of actin filaments and associated proteins just beneath the plasma membrane.

The best-studied example is the spectrin-based cytoskeleton of the red blood cell. A meshwork of spectrin tetramers is connected to the membrane through adaptor proteins (ankyrin and band 4.1) that link spectrin to integral membrane proteins (band 3 and glycophorin). This flexible scaffold gives the red blood cell its remarkable ability to deform as it squeezes through narrow capillaries while springing back to its biconcave disc shape.

Defects in spectrin, ankyrin, or band 4.1 cause hereditary spherocytosis and hereditary elliptocytosis — diseases in which red blood cells lose their normal shape and are prematurely destroyed in the spleen.

The Cell Surface Is Coated with Sugar Residues

The outer surface of virtually all eukaryotic cells is coated with a carbohydrate-rich layer called the glycocalyx (literally “sugar coat”). This layer is composed of the oligosaccharide side chains of glycoproteins and glycolipids, plus proteoglycans (proteins with long glycosaminoglycan chains). In some cells, the glycocalyx extends hundreds of nanometers from the membrane surface.

The glycocalyx serves multiple functions: it protects the cell surface from mechanical and chemical damage, it creates a hydration shell that lubricates cell surfaces (important for blood cells moving through vessels), and it mediates cell-cell recognition. The ABO blood group antigens are glycolipids and glycoproteins in the red blood cell glycocalyx, and differences in their sugar structures determine blood type compatibility.

Lectins — proteins that bind specific sugar structures — are key mediators of cell recognition. Selectins are lectins on endothelial cells that bind sugar residues on white blood cells, initiating the rolling and adhesion steps that allow leukocytes to exit the bloodstream at sites of inflammation.

Membrane Protein Bioinformatics

Transmembrane Helix and Signal Peptide Prediction

Predicting which segments of a protein cross the membrane is one of the most practically useful bioinformatics tasks. Since transmembrane α helices require ~20 consecutive hydrophobic residues, prediction algorithms search for such hydrophobic stretches in the amino acid sequence.

Hydropathy plots (Kyte-Doolittle) were the first approach: a sliding window averages the hydrophobicity of each amino acid, and peaks above a threshold suggest transmembrane segments. Modern tools use sophisticated statistical and machine-learning models:

A common pitfall is confusing signal peptides (short hydrophobic N-terminal sequences that target proteins to the ER for secretion or membrane insertion) with transmembrane helices. Signal peptides are cleaved off after targeting and do not remain in the mature protein. Phobius and DeepTMHMM specifically address this problem.

Membrane Protein Structure Databases

Experimentally determined membrane protein structures are cataloged in specialized databases:

• mpstruc — curated list of all unique membrane protein structures determined by X-ray crystallography and cryo-EM, maintained by Stephen White’s laboratory
• OPM (Orientations of Proteins in Membranes) — provides the calculated position and orientation of each membrane protein within the lipid bilayer, including the tilt angle of transmembrane helices and the boundaries of the hydrophobic core
• PDBTM — automatically identifies and classifies transmembrane proteins in the PDB

These databases are essential for understanding how membrane proteins sit within the bilayer and for validating topology predictions.

GPI-Anchor and Glycosylation Site Prediction

Computational tools predict post-translational modifications relevant to membrane association:

• PredGPI and big-PI predict GPI-anchor attachment sites (ω-sites) from the C-terminal sequence. The signal for GPI anchoring is a C-terminal hydrophobic stretch preceded by small amino acids (Gly, Ala, Ser, Asn, Asp, Cys) at the ω-site.
• NetNGlyc predicts N-linked glycosylation sites by evaluating the consensus sequence Asn-X-Ser/Thr in the context of the surrounding sequence.
• NetOGlyc predicts O-linked glycosylation sites (primarily on Ser and Thr), which lack a strict consensus and are more difficult to predict.

These predictions help researchers determine which parts of a protein are exposed on the cell surface and how they are modified.

Membrane Protein Homology Modeling Challenges

Predicting the three-dimensional structures of membrane proteins poses special challenges. Membrane proteins are underrepresented in the PDB because they are difficult to crystallize and require detergents for solubilization. While the PDB contains over 200,000 structures, only a few thousand are membrane proteins.

Key challenges include modeling the lipid-exposed surface (which has different amino acid preferences from the water-exposed surface), the tilt and positioning of transmembrane helices within the bilayer, and the oligomeric state (many membrane proteins function as dimers or higher-order complexes). Tools like SWISS-MODEL and AlphaFold have made significant progress, but membrane protein structure prediction remains more challenging than for soluble proteins, particularly for β-barrel proteins and large multi-pass complexes.

Let us examine how coding sequences for membrane-associated versus soluble proteins differ in their GC content, reflecting different amino acid compositions:

let lipid_data = '[{"label": "PC", "value": 40}, {"label": "PE", "value": 25}, {"label": "PS", "value": 10}, {"label": "SM", "value": 10}, {"label": "Cholesterol", "value": 15}]'
let chart = Viz.bar(lipid_data, '{"title": "Typical Plasma Membrane Lipid Composition (%)", "color": "#F59E0B"}')
print(chart)

The plasma membrane is a complex mixture of lipids, with phosphatidylcholine (PC) typically the most abundant species. The precise composition varies between cell types and between the inner and outer leaflets.

Exercise: Analyze Lipid Properties

Different membrane lipids have different molecular properties that determine their behavior in the bilayer. Compare phosphatidylcholine and sphingomyelin — how similar are they structurally?

let pc = "CCOP(=O)([O-])OCC(COC(=O)CCCCCCCCCCCCCCC)OC(=O)CCCCCCC/C=C\\CCCCCCCC"
let sm = "CCCCCCCCCCCCCCCCC(=O)NC(COP(=O)([O-])OCC[N+](C)(C)C)C(O)/C=C/CCCCCCCCCCCCC"
let similarity = Chem.tanimoto(pc, sm)
print("PC vs Sphingomyelin similarity: " + similarity)
// Are they structurally similar or dissimilar?
let answer = "similar"
print(answer)

Exercise: Transmembrane Helix Properties

A key prediction task in membrane bioinformatics is distinguishing transmembrane helices from soluble protein regions. Analyze the amino acid properties of three peptide segments and identify which is the transmembrane helix:

let seg_a = "EDKATLQNELKAQFGEYVNP"
let seg_b = "LAVIFAGAILLLLVPALLIV"
let seg_c = "RKKDEGQPRNITAVHSGLRS"
print("Segment A properties:")
print(Struct.protein_props(seg_a))
print("Segment B properties:")
print(Struct.protein_props(seg_b))
print("Segment C properties:")
print(Struct.protein_props(seg_c))
// Which segment is the transmembrane helix?
let answer = "segment_b"
print(answer)

Exercise: Membrane Composition Visualization

Build a visualization comparing the lipid compositions of the inner and outer leaflets of the plasma membrane. The asymmetry of PS (phosphatidylserine) distribution is a key signal in apoptosis:

let outer = '[{"label": "PC", "value": 45}, {"label": "SM", "value": 20}, {"label": "PE", "value": 10}, {"label": "PS", "value": 2}]'
let inner = '[{"label": "PC", "value": 15}, {"label": "SM", "value": 5}, {"label": "PE", "value": 35}, {"label": "PS", "value": 25}]'
print("Outer leaflet composition:")
let outer_chart = Viz.bar(outer, '{"title": "Outer Leaflet (%)", "color": "#3B82F6"}')
print(outer_chart)
print("Inner leaflet composition:")
let inner_chart = Viz.bar(inner, '{"title": "Inner Leaflet (%)", "color": "#EF4444"}')
print(inner_chart)
// In which leaflet is PS concentrated?
let answer = "inner"
print(answer)

Knowledge Check

Summary

In this lesson you covered membrane structure in depth, from lipid chemistry to membrane protein architecture and computational analysis:

• Three lipid classes build cell membranes: phosphoglycerides (most abundant, glycerol backbone), sphingolipids (sphingosine backbone, sugar-bearing glycosphingolipids), and sterols (cholesterol buffers fluidity)
• Phospholipids spontaneously form bilayers through the hydrophobic effect, creating self-sealing compartments without requiring energy input
• The lipid bilayer is a two-dimensional fluid — lipids diffuse laterally within each leaflet but rarely flip between leaflets without enzymatic help (flippases, floppases, scramblases)
• Membrane fluidity depends on fatty acid chain length, degree of unsaturation (cis double bonds introduce kinks), cholesterol content, and temperature
• Lipid rafts are cholesterol- and sphingolipid-enriched microdomains in the liquid-ordered phase that organize signaling and trafficking proteins
• Lipid droplets are bounded by a phospholipid monolayer, not a bilayer, and store neutral lipids
• Bilayer asymmetry is functionally important: PS on the inner leaflet is an apoptotic signal when exposed; PI on the inner leaflet generates signaling lipids (PIP₂, PIP₃)
• Membrane proteins associate via transmembrane segments (α helices or β barrels), lipid anchors (GPI, myristoyl, palmitoyl, prenyl), or peripheral interactions
• Transmembrane α helices require ~20 hydrophobic residues; they interact via GxxxG motifs and other packing interfaces
• β-Barrel proteins in outer membranes form large channels (porins, VDAC, TOM40)
• Glycosylation (N-linked at Asn-X-Ser/Thr, O-linked at Ser/Thr) occurs exclusively on the extracellular surface
• Detergents, nanodiscs, and SMALPs are used to solubilize membrane proteins for study
• Bacteriorhodopsin and GPCRs share the seven-helix bundle architecture; GPCRs are the target of ~35% of approved drugs
• Membrane protein mobility is regulated by the cortical cytoskeleton (spectrin mesh in red blood cells), cell junctions, and extracellular matrix attachments
• The glycocalyx coats cell surfaces with sugars that protect, lubricate, and mediate recognition (ABO blood groups, selectin-mediated leukocyte adhesion)
• Lipidomics uses LC-MS/MS to profile membrane lipid composition; databases include LIPID MAPS and SwissLipids
• Topology prediction tools (TMHMM, Phobius, DeepTMHMM) identify transmembrane segments from sequence; Phobius distinguishes signal peptides from TM helices
• Membrane protein structure databases (mpstruc, OPM, PDBTM) catalog experimentally determined structures and their orientations in the bilayer
• Post-translational modification predictors (PredGPI, NetNGlyc, NetOGlyc) identify GPI-anchor and glycosylation sites from sequence