The Fluid-Mosaic Model of Membranes

The plasma membrane and the internal membranes of organelles are built around a <a class="wiki-link" href="/wiki/bio-cell-phospholipid-bilayer" data-track="wiki_article.body.wiki_link">lipid bilayer</a>: two sheets of amphipathic phospholipid molecules arranged tail-to-tail, with hydrophilic phosphate head groups facing the aqueous cytosol and extracellular fluid, and hydrophobic fatty-acyl tails buried in the interior. This much was established by the 1920s–1960s through experiments like Gorter and Grendel's lipid-monolayer measurements. What the fluid-mosaic model added was a precise statement about the physical state of this bilayer and about where proteins sit relative to it. Two claims define the model. First, the lipid bilayer behaves as a two-dimensional fluid at physiological temperature: individual phospholipids rotate rapidly around their own axis, flex their tails, and diffuse laterally within their own leaflet at rates on the order of 1–2 µm/second, while flipping from one leaflet to the other (flip-flop) is very slow, often requiring minutes to hours without enzymatic help, because it means dragging a polar head group through the hydrophobic core. Second, membrane proteins are not spread as continuous surface sheets but exist as discrete particles — a mosaic — either embedded partway into the bilayer, spanning it completely (integral or transmembrane proteins), or attached loosely to one surface (peripheral proteins). Most of these protein particles are free to diffuse and rotate within the plane of the membrane, giving the whole structure the character of a viscous two-dimensional liquid with protein icebergs floating in a lipid sea. Crucially, "fluid" does not mean "unstructured" or "unconstrained." Real membranes are heterogeneous and asymmetric: the two leaflets have different lipid and protein compositions (membrane asymmetry), certain lipids cluster into more ordered microdomains often called lipid rafts, and the underlying cortical cytoskeleton creates fences that confine and slow the diffusion of many proteins into "corrals" tens to hundreds of nanometers wide. The modern view, sometimes called the fluid-mosaic model with lipid rafts and cytoskeletal corralling, keeps the essential Singer-Nicolson insight — proteins float and move — while adding decades of refinement about how that motion is organized and restricted rather than random.

Step 1 — the bilayer as solvent. Phospholipids self-assemble into a bilayer because their fatty-acyl tails are hydrophobic and their phosphate-containing heads are hydrophilic; in water, this amphipathic geometry spontaneously minimizes exposure of hydrocarbon tails, driven by the hydrophobic effect. The resulting sheet is only about 4–5 nanometers thick but can extend over tens of micrometers, behaving mechanically like a two-dimensional oil film sandwiched between two thin sheets of water-loving head groups.

Step 2 — fluidity is a phase, and it is tunable. At body temperature, the hydrocarbon tails are in a liquid-disordered state, jostling and bending much like the tails of molecules in liquid oil. Below a characteristic transition temperature (Tm), the bilayer can freeze into a more rigid gel phase in which tails pack tightly and ordered. Three factors set Tm and hence fluidity at a given temperature: 1. Tail length — longer saturated tails pack more tightly and raise Tm (less fluid at a given temperature). 2. Degree of unsaturation — cis double bonds introduce kinks that prevent tight packing, lowering Tm and increasing fluidity. 3. Cholesterol content — in animal cell membranes, cholesterol acts as a fluidity buffer: at high temperature it restrains excessive motion (making the membrane less fluid), while at low temperature it wedges between phospholipids and prevents them from packing into a rigid gel (keeping the membrane more fluid). This dual, temperature-dependent buffering action is why cholesterol-rich membranes show a broadened, less sharp phase transition rather than an abrupt gel-to-fluid switch.

Step 3 — proteins are inserted, not layered. Integral membrane proteins are synthesized and inserted co-translationally into the endoplasmic reticulum membrane via the translocon, folding so that hydrophobic alpha-helical or beta-barrel segments (often 20–25 nonpolar residues per transmembrane helix) reside within the hydrophobic core, while hydrophilic loops and termini protrude into the aqueous compartments on either side. This gives every integral protein a fixed, permanent orientation (topology) that is preserved as the protein is trafficked to the plasma membrane — a membrane protein is never later "flipped." Peripheral proteins, by contrast, associate with the membrane surface or with the head groups or with integral proteins through electrostatic and hydrogen-bonding interactions, and can be stripped off with mild treatments like high salt or altered pH, unlike integral proteins, which require detergents to dissolve the bilayer itself.

Step 4 — lateral diffusion. Once inserted, most transmembrane proteins are free to diffuse laterally within the plane of the membrane by random thermal (Brownian) motion, colliding with and being jostled by surrounding lipids, at diffusion coefficients roughly 1000-fold slower than the lipids themselves (protein diffusion coefficients are typically in the range of 10⁻¹⁰ to 10⁻¹² cm²/s, versus roughly 10⁻⁸ cm²/s for lipids), because proteins are far bulkier and often interact with cytoskeletal elements. Rotational diffusion also occurs, allowing many receptors to tumble around an axis perpendicular to the membrane plane.

Step 5 — constraints on fluidity. Several structures locally restrict this free diffusion, all discovered after the original 1972 model and now considered essential refinements rather than contradictions:

• Membrane-skeleton fences: the cortical actin cytoskeleton, attached to the cytoplasmic face via anchor proteins, partitions the membrane into corrals roughly 30–250 nm across; proteins hop between corrals less often than they diffuse within one, producing "hop diffusion" rather than pure free diffusion.
• Lipid rafts: cholesterol- and sphingolipid-rich microdomains that are more ordered ("liquid-ordered" phase) than the surrounding liquid-disordered bilayer, concentrating specific signaling proteins (e.g., GPI-anchored proteins) and slowing their diffusion.
• Protein-protein oligomerization and cytoskeletal tethering, which anchor some proteins (e.g., spectrin-linked band 3 in red blood cells) almost immobile relative to freely diffusing counterparts.

1. Freeze-fracture electron microscopy: proteins live inside the bilayer, not on its surface. Developed in the 1950s–60s and applied decisively to membranes in the late 1960s (notably by Daniel Branton and colleagues), this technique rapidly freezes a membrane, then fractures the frozen block with a knife. The fracture plane preferentially splits the bilayer down the middle, between the two leaflets, because that hydrophobic interface is the weakest plane in the structure. When the exposed fracture faces are shadowed with metal and viewed by electron microscopy, they reveal a cobblestone-like surface studded with intramembrane particles, 5–8.5 nm in diameter, embedded within the plane of the split bilayer itself. This was the single most important piece of physical evidence against the Davson-Danielli model, because that model predicted a smooth lipid surface with proteins only coating the outside — freeze-fracture instead showed particles literally lodged inside the membrane interior, exactly as the fluid-mosaic model required for integral, transmembrane proteins.

2. The Frye-Edidin heterokaryon (cell fusion) experiment (1970). Larry Frye and Michael Edidin fused mouse and human cells together using inactivated Sendai virus, which promotes cell-cell fusion without immediately killing the cells, producing a single heterokaryon cell with both a mouse-derived and a human-derived plasma membrane region. They labeled mouse-specific membrane proteins with one fluorescent antibody (green) and human-specific membrane proteins with a different fluorescent antibody (red), then examined many such fused cells at various times after fusion.

• At time zero, the fused cell showed a sharp boundary: green fluorescence confined to the mouse half, red confined to the human half.
• Within about 40 minutes at 37°C, the two fluorescent labels had become completely intermixed over the entire cell surface, producing a uniform yellow (green+red) glow across the whole membrane.
• When the same experiment was performed at low temperature (around 0-4°C) or with metabolic poisons that block ATP production, intermixing was dramatically slowed or prevented, showing that the mixing depended on the physical fluidity of the membrane (temperature-sensitive, consistent with a lipid phase transition) rather than requiring continuous new protein synthesis.

This experiment provided direct, visual, quantitative proof that membrane proteins are not permanently fixed in one location but diffuse freely across the whole membrane surface, exactly the "protein iceberg drifting in a lipid sea" behavior the fluid-mosaic model proposed. It remains one of the most frequently cited direct demonstrations of membrane protein mobility.

3. Fluorescence recovery after photobleaching (FRAP). Developed in the 1970s (Axelrod, Koppel, Schlessinger, Elson, Webb and colleagues), FRAP labels membrane components with a fluorescent tag, then uses an intense, narrowly focused laser pulse to irreversibly bleach the fluorophores in a small spot on the membrane. If the surrounding, unbleached fluorescent molecules can diffuse laterally, they will move into the bleached spot over time while bleached molecules diffuse away, and the fluorescence intensity in the spot will recover. The rate of recovery gives a diffusion coefficient, and the fraction of fluorescence that ultimately recovers (the mobile fraction) reveals what proportion of the labeled molecules are free to move versus permanently immobilized (e.g., tethered to the cytoskeleton or aggregated). FRAP experiments confirmed quantitatively that lipids diffuse fast (~10⁻⁸ cm²/s) and most membrane proteins diffuse 100- to 1000-fold slower, and — importantly — revealed that mobile fractions are often less than 100%, providing the first quantitative evidence for the cytoskeletal corralling and protein immobilization that refined the original, simpler fluid-mosaic picture.

4. Lipid monolayer and X-ray diffraction studies (earlier groundwork). Gorter and Grendel's 1925 experiment — extracting lipids from a known number of red blood cells, spreading them as a monolayer at an air-water interface, and finding the monolayer's area was almost exactly twice the calculated surface area of the cells — provided the founding evidence for a lipid bilayer in the first place, the substrate onto which the fluid-mosaic model of protein mobility was later built.

Example 1 — Interpreting a FRAP experiment quantitatively. A researcher photobleaches a 2-µm-diameter spot on a cell expressing a GFP-tagged receptor and follows fluorescence recovery.

• Prebleach intensity: 1000 arbitrary units (a.u.)
• Immediately after bleach: 50 a.u.
• Plateau after full recovery: 700 a.u.
• Half-time to recovery (t½): 8 seconds

Mobile fraction is calculated as: Mobile fraction = (F∞ − F₀) / (Fprebleach − F₀) = (700 − 50) / (1000 − 50) = 650 / 950 ≈ 0.68, or 68% This tells us 32% of the receptor population is immobile on the experimental timescale — likely tethered to the cytoskeleton, clustered in an oligomer, or trapped in a corral, not simply "stuck" by damage (a properly controlled experiment rules out photodamage by confirming recovery reaches a stable, reproducible plateau in repeat trials). Using the bleach spot radius (w = 1 µm) and t½, an approximate diffusion coefficient can be estimated as: D ≈ 0.224 × w² / t½ = 0.224 × (1 µm)² / 8 s ≈ 0.028 µm²/s Comparing this to a lipid diffusion coefficient of roughly 1 µm²/s (measured with a lipid dye under the same conditions) shows the receptor diffuses roughly 35-fold slower than bulk lipid, consistent with it being a bulky transmembrane protein experiencing cytoskeletal drag and possibly weak oligomerization — a quantitative signature of the "mosaic" (protein) moving through the "fluid" (lipid) at a very different rate than the solvent itself.

Example 2 — Predicting relative membrane fluidity from lipid composition. Two hypothetical organisms living at different environmental temperatures must maintain a fluid plasma membrane to survive (a process called homeoviscous adaptation, foreshadowing Membrane Fluidity and Its Regulation). Consider their membrane lipid compositions:

Reasoning through this: shorter chains pack less tightly (raising fluidity), and cis double bonds put kinks in the tails that prevent close packing (further raising fluidity) — both features would be expected in an organism that must keep its membrane from freezing solid at 4 °C. Organism B's longer, more saturated tails would pack tightly and risk gel-phase rigidity at low temperature, but at 37 °C this tighter packing is exactly what is needed to prevent the membrane from becoming dangerously "leaky" and over-fluid; the added cholesterol further buffers Organism B's membrane against becoming too fluid at the higher physiological temperature. If Organism A's membrane composition were transplanted unchanged into a 37 °C environment, the double bonds and short chains that kept it appropriately fluid at 4 °C would instead make it excessively fluid and mechanically unstable at 37 °C — illustrating that the "fluid" in fluid-mosaic is a tuned, adaptive property, not a fixed universal constant.

Seymour Jonathan Singer trained as a physical chemist before turning to immunology and membrane biology at the University of California, San Diego. Working alongside graduate student and postdoctoral collaborator Garth L. Nicolson, Singer synthesized an enormous body of then-recent physical, biochemical, and microscopic evidence — thermodynamic arguments about amphipathic protein folding, freeze-fracture images, and early diffusion measurements — into a single coherent, falsifiable model, published in Science in 1972 under the title "The Fluid Mosaic Model of the Structure of Cell Membranes." Singer's key theoretical contribution was insisting that membrane proteins, like water-soluble globular proteins, must fold to bury their hydrophobic residues away from water — in a membrane protein's case, this meant burying them within the lipid bilayer rather than exposing them at an aqueous surface, a thermodynamic argument that directly predicted proteins must be embedded in, not layered on, the bilayer.

The fluid-mosaic model is not a historical curiosity; it is the operating framework behind much of modern cell biology, pharmacology, and medicine. Drug design depends on knowing that receptors are mobile, independently folded proteins embedded in a fluid bilayer, rather than fixed points on a rigid surface — this underlies how researchers design small molecules to dock into G-protein-coupled receptors (GPCRs), which must diffuse, cluster, and change conformation within the membrane to transmit a signal. Vaccine and antibody engineering relies on the mobility and clustering behavior of membrane antigens; the way B-cell receptors cluster upon antigen binding, driving immune activation, is a direct consequence of lateral diffusion and corralling first characterized by Frye-Edidin-style experiments and FRAP. In membrane fluidity disorders, understanding cholesterol's dual buffering role explains conditions ranging from atherosclerosis (where cholesterol handling goes awry) to certain lipid storage diseases, and informs how statins and other cholesterol-modulating drugs might indirectly affect membrane-dependent processes. Cryobiology and food science use homeoviscous adaptation principles directly: cryoprotectants and controlled cooling rates for preserving cells, gametes, and organs for transplantation are designed around preventing lethal gel-phase transitions in cell membranes, the same phase behavior described by Singer and Nicolson's fluid framework. In biotechnology, liposome and lipid-nanoparticle drug delivery systems (including those used in mRNA vaccines) are engineered by manipulating exactly the variables the fluid-mosaic model identifies as governing bilayer behavior — chain saturation, cholesterol content, and headgroup chemistry — to control fluidity, stability, and fusion with target cell membranes. And in basic research, techniques descended directly from FRAP (including single-particle tracking and super-resolution microscopy) continue to refine the model, revealing ever-finer detail about Membrane Proteins organization, corralling, and the regulation described more fully under Membrane Fluidity and Its Regulation.

Ask the tutor: How do lipid rafts and cytoskeletal corralling change the "mosaic" part of the model without breaking the "fluid" part?