Membrane Organization and Topology

Digest away the outer leaflet’s lipids of a human red blood cell and analyse what is left, and the two halves of its membrane turn out to be different. The outer half’s phospholipids carry about 1.6 double bonds each, the inner half’s about 3.4.

That measurement, made in 2020 on intact human erythrocytes, is one recent addition to an old finding: a cell membrane is not the same on both sides. Its lipids differ between the halves, and so do the parts of every protein it carries. The cell spends energy to keep it that way, and it moves membrane from compartment to compartment without ever turning it inside out. This lesson asks three questions. Who puts lipids where? Which way does a protein face, and why does traffic never flip it? And how do you measure how fast molecules move within the plane of the membrane?

The idea: a membrane has two different faces

The bilayer is two sheets of lipid, each sheet called a leaflet, with tails meeting in an oily core. Its chemistry and physics are covered in lipids and membrane structure and function. Here you need one consequence of that core: a polar head group, or a polar stretch of protein, crosses it only rarely. Lipids and proteins can slide sideways within their own leaflet but cannot easily flip to the other one.

Think of a jacket with a lining. You can slide a badge anywhere on the outside, but it never ends up on the lining unless someone unpicks it and sews it back.

Every membrane in the cell therefore has two faces. One faces the cytosol, the fluid that fills the cell outside the organelles. The other faces either the outside of the cell or the lumen, the space enclosed by an organelle such as the ER or Golgi. That second face is called the exoplasmic or noncytosolic face, and it is the key to the second half of this lesson.

Asymmetry: who puts lipids where

In human erythrocytes, studies summarized in a recent review place phosphatidylserine (PS) and phosphatidylethanolamine (PE) mainly in the cytoplasmic leaflet, and phosphatidylcholine (PC) and sphingomyelin mainly in the outer leaflet. Lorent and colleagues added the unsaturation numbers above. These are erythrocyte numbers. Other cells and organelles have their own, less completely measured.

Left alone, a phospholipid moves between leaflets slowly, on a scale of hours or longer for most species. That slowness lets asymmetry last, but something has to create it. Two classes of protein act on it in opposite directions.

• Flippases use ATP to move chosen lipids, chiefly PS and PE, from the outer to the cytoplasmic leaflet, against their gradient. Most are P4-type ATPases. In a haploid genetic screen in human cells, Segawa and colleagues found ATP11C and its partner CDC50A required for this inward movement.
• Scramblases let lipids cross in both directions down their gradients and, unlike flippases, generally need no energy input. Switching one on collapses asymmetry far faster than spontaneous flip-flop would.

TMEM16F is the best-studied calcium-activated scramblase. Alvadia and colleagues purified the mouse protein and showed that, in artificial vesicles, it scrambled both labelled PS and labelled PE when calcium was added. Its activity can expose PS on the surface of activated platelets, where the negatively charged lipid supports blood clotting. A different route exposes PS on dying cells, where it serves as a signal for engulfment; that pathway belongs to death receptors, BCL-2 and caspases.

So asymmetry is a steady state: pumps push one way, slow leakage and regulated scramblases push the other. Cholesterol, which flips quickly, is a harder case, and how it divides between the leaflets of the plasma membrane is still argued.

Topology and why vesicles preserve it

A membrane protein’s topology is which of its parts face the cytosol and which face the other side. Four common arrangements:

• Single-pass type I: one membrane-spanning segment, N terminus in the lumen or outside, C terminus in the cytosol.
• Single-pass type II: one segment, N terminus in the cytosol, C terminus in the lumen or outside.
• Multipass: several segments, with loops alternating between the two faces.
• GPI-anchored: no spanning segment; the protein is attached by a glycolipid, glycosylphosphatidylinositol (GPI), to the lumenal leaflet. Proteins held by myristoyl or prenyl groups on the cytosolic face are a different class, covered in lipids as anchors and addresses.

The rule that makes these predictable: a lumen is topologically equivalent to the outside of the cell. A transport vesicle buds from one membrane and fuses with another, and its lumen is continuous first with the donor lumen and then with the target lumen or the outside. Nothing crosses a membrane during the trip, so the face that pointed at the cytosol keeps pointing at the cytosol.

Orientation is set once, when the protein enters the ER membrane through the translocon, the protein channel that threads new chains into or across the membrane. How the signals for that decision work is the job of protein sorting. The decision can be close. Positive charges are statistically enriched on the cytosolic side of membrane segments, a bias first described for bacterial inner-membrane proteins; it is a tendency, not a law. Goder, Bieri and Spiess showed in COS-1 cells that engineered chains with competing signals can reorient within the translocon, and that glycosylation can shift the outcome. Reorientation happens at insertion. A finished protein is not turned over by traffic.

Worked problem, part 1: trace a protein. A type I protein is made in the ER of a cultured mammalian cell. It has an N-linked glycan on its N-terminal domain, and it reaches the plasma membrane, then is taken back in by endocytosis.

1. In the ER: N terminus in the lumen, C terminus in the cytosol. The glycan was added in the lumen, because the oligosaccharyl transferase has its active site on the lumenal side.
2. Through the Golgi by vesicles: the lumenal face stays lumenal.
3. A secretory vesicle fuses with the plasma membrane. Its lumen opens to the outside. N terminus and glycan now face the extracellular fluid; C terminus stays in the cytosol.
4. Endocytosis pinches the plasma membrane inward. The outside face becomes the endosome’s lumen. N terminus and glycan face the endosomal lumen; C terminus still faces the cytosol.

At no step did the C terminus leave the cytosol. That is how a sugar made in the ER ends up on the cell surface.

Measuring motion: FRAP and domains

In fluorescence recovery after photobleaching (FRAP), a laser destroys the fluorescence of labelled molecules in a small spot. If they are mobile, unbleached ones diffuse in and the spot brightens. The bleached molecules are still there and still moving; they are simply dark. The imaging side of the method is covered in microscopy and quantitative images.

Worked problem, part 2: read a FRAP curve. In this made-up data set, a fluorescent protein in the plasma membrane of a cultured cell is bleached in a circular spot of radius w = 2.0 µm. Assume the bleach is instantaneous and uniform, the membrane is flat, the unbleached reservoir is large, and recovery is by two-dimensional diffusion alone. The values are already corrected for background and for bleaching during imaging.

• Before the bleach, Fi = 100. Immediately after, F0 = 20. At 4.0 s, 50. Final plateau, F∞ = 80.

Step 1, mobile fraction:

mobile fraction = (F∞ - F0) / (Fi - F0) = (80 - 20) / (100 - 20) = 60 / 80 = 0.75

The denominator is 80, not 100: only what was bleached can recover. The immobile fraction is 1 - 0.75 = 0.25.

Step 2, half-time. Half the recoverable increment is 60 / 2 = 30, so half recovery is reached at 20 + 30 = 50. The curve reads 50 at 4.0 s, so t½ = 4.0 s. It is not the time to reach half the prebleach value.

Step 3, diffusion coefficient, valid only for this model:

D ≈ 0.224 w² / t½ = 0.224 × (2.0 µm)² / 4.0 s = 0.224 × 4.0 / 4.0 = 0.22 µm²/s

That sits in the range tabulated for membrane-bound molecules; Day and colleagues list about 0.2 µm²/s for cholera toxin B bound to the plasma membrane of COS-7 and mouse embryonic fibroblast cells at 37 °C. EGFP free in the cytoplasm of the same cells moves more than a hundred times faster, at about 40 µm²/s.

Two cautions. The formula assumes a circular, instantaneous bleach; a scanning confocal bleach lets molecules move during the bleach and needs a different analysis. And a 25% immobile fraction does not say why. Anchoring to the cytoskeleton, bleaching a large share of the cell’s total pool, or bleaching during imaging can each lower the plateau.

Recovery is also not always free diffusion. The plasma membrane has domains, regions that differ in composition or mobility. Tight junctions separate apical from basolateral membrane in epithelia. In a FRAP study of cholera toxin B on cultured cells, Day and Kenworthy found its slow diffusion was impeded by actin- and ATP-dependent processes, as expected if an actin meshwork restrains it (PLoS ONE 2012).

Lipid rafts, proposed ordered domains rich in cholesterol and sphingolipid, are where care matters most. Membrane material that resists cold detergent was long called raft material, but a detergent-resistant fraction is not a native raft; detergent can create or rearrange domains. Reviews now describe the existence, size and composition of rafts in living cells as unresolved, not as one settled universal structure. Stone and colleagues, using super-resolution imaging in the CH27 mouse B-cell lymphoma line, found that a peptide probe preferring ordered membrane gathered near clustered B-cell receptors and one preferring disordered membrane was excluded, in live and fixed cells; fixed primary mouse B cells agreed. That supports phase-like sorting in that system. It does not give a raft’s composition, and it does not show that every cell has them.

Design problem

You have cloned T, a new protein with one predicted membrane-spanning segment. Determine which way it faces in the ER and plasma membrane of a cultured mammalian cell, and confirm that it is integrated in the membrane rather than stuck to it.