Endocytosis, Recycling and Degradation

Cells from one patient with familial hypercholesterolaemia carry LDL receptors that grip LDL normally, yet the LDL stays outside the cell.

LDL, low-density lipoprotein, is the particle that carries much of the cholesterol in human blood, and cells take it in through a receptor. In the 1970s Michael Brown and Joseph Goldstein grew skin fibroblasts from people with familial hypercholesterolaemia (FH), an inherited disease of very high blood LDL. The homozygous cultures they first tested lacked normal high-affinity LDL binding. Cells from a patient known as J.D. were different: they bound LDL but did not take it in. A receptor has to do more than bind. It must be gathered into the right patch of membrane, carried inside, let go of its cargo and come back.

This lesson follows that cycle: the doors into a cell, the building of a clathrin-coated vesicle, the sorting stations inside, and FH cell data that separate one step from another.

Ways into a cell: several doors, each with its own machinery

Endocytosis is the taking-in of outside material in membrane-bound carriers made from the plasma membrane. The cargo sits in the carrier’s lumen, the space enclosed by its membrane, which is topologically outside the cytosol, the fluid part of the cytoplasm. Nothing has yet crossed a membrane. (The membrane topology lesson sets out this logic.)

Cells have several doors.

• Phagocytosis takes large particles, such as bacteria or dead cells, into phagosomes. In animals it is mostly the work of specialized cells such as macrophages and neutrophils, and actin drives it.
• Macropinocytosis makes large vacuoles when sheet-like membrane projections fold back onto the cell surface and enclose a gulp of fluid. Growth factors stimulate it in many mammalian cells.
• Clathrin-mediated endocytosis concentrates chosen receptors into small coated pits. It is the best-understood route for receptors such as the LDL receptor and the transferrin receptor.
• Caveolae are small flask-shaped pits built around the protein caveolin. They are often drawn as a major entry route, but a 2020 review cautions that few if any cargoes are shown to rely on caveolae to enter.
• FEME (fast endophilin-mediated endocytosis) is triggered when ligands activate selected surface receptors, and it needs dynamin. The CLIC/GEEC route does not need dynamin.

One trap: clathrin-independent does not mean caveolar. FEME and CLIC/GEEC are separate routes with their own machinery.

Building and cutting a clathrin-coated vesicle

A coat is a protein lattice that assembles on one face of a membrane, bends it and selects cargo. Here the coat is clathrin. Its unit is the triskelion: three heavy chains and three light chains joined at a hub, with three legs that interlock with neighbours into a cage of hexagons and pentagons. Picture a geodesic dome built from three-legged pieces.

Clathrin binds neither membrane nor cargo. Adaptors do that. In mammalian cells the steps run:

1. Initiation. The adaptor complex AP2 and other clathrin-associated sorting proteins bind PI(4,5)P2, a phosphoinositide in the inner leaflet of the plasma membrane (see lipids).
2. Cargo capture. Adaptors read short sorting motifs in cargo tails. AP2 binds YXXΦ motifs (tyrosine, any two residues, then a bulky hydrophobic residue), such as the one in the transferrin receptor. The LDL receptor carries an NPXY motif, which the adaptors Dab2 and ARH can recognize.
3. Coat assembly. Membrane-bound adaptors recruit clathrin triskelia. The lattice grows and the membrane curves into a pit.
4. Scission. Dynamin, a GTPase, assembles as a collar around the narrow neck and, using GTP hydrolysis, drives the membrane to split.
5. Uncoating. The chaperone Hsc70, recruited by auxilin, pulls the lattice apart, so the triskelia can be reused and the vesicle can fuse with an endosome.

Not every clathrin cargo leans on AP2 the same way. In HeLaM cells depleted of AP2, a chimeric protein carrying the LDL receptor’s tail was still taken up, consistent with other adaptors reading its NPXY motif. That result is about one construct in one cell line; it does not make AP2 dispensable in general.

How the coat bends is still argued. Live imaging in human SK-MEL-2 cells found curvature appearing early in some selected events and only after much of the clathrin had arrived in others. A 2023 analysis of fixed coats in three cell lines inferred that coats, on average, gain about half their final area before bending, with a fitted preferred radius of about 72 nm in SK-MEL-2 cells. The vesicle traffic lesson treats coats in general.

Sorting stations: maturing endosomes, recycling and the lysosome

Uncoated vesicles fuse with early endosomes, whose lumen is mildly acid, roughly pH 6 (textbooks give about 6.0 to 6.2; proton pumps set it, as the organelles lesson explains). For LDL, acid is the sorting signal: at this pH LDL comes off its receptor. A crystal structure of the receptor’s outer part at acid pH suggests one way. The ligand-binding repeats fold back over a propeller-shaped domain, which binds two of them and so takes the place of LDL.

Receptor and cargo now part.

The receptor recycles. Empty receptors collect in narrow tubules that pinch off and return to the plasma membrane, directly or through a recycling endosome. These compartments are labelled by Rabs. A Rab is a small GTPase that, bound to GTP, sits on one membrane and marks its identity by recruiting the tethers and enzymes that belong there. In human A431 cells, Rab4-positive recycling tubules were seen to separate from Rab5-positive early endosomes by fission.

LDL goes to the lysosome. It stays in the lumen while the endosome matures. Lysosomal enzymes digest its protein and hydrolyse its cholesteryl esters, releasing cholesterol. That cholesterol in turn damps the cell’s own synthesis through HMG-CoA reductase, as the lipid synthesis lesson describes.

How does an early endosome become a late one? Either cargo moves between permanent compartments, or the compartment itself changes. In 2005 Marino Zerial’s laboratory imaged A431 cells expressing fluorescent Rab5 and Rab7 as they took up LDL. On single endosomes carrying LDL, Rab5 was replaced by Rab7, which the authors read as Rab conversion: the compartment matures. Recycling tubules still left along the way, so maturation and carriers coexist.

Membrane proteins meant for destruction take one more step. Late endosomes become multivesicular bodies, where patches of the limiting membrane bud inward into the lumen. ESCRT complexes sort proteins tagged with ubiquitin into those inner vesicles; in yeast, ESCRT-I recognizes the ubiquitinated cargo, and related complexes work in mammalian cells. When the multivesicular body fuses with a lysosome, the inner vesicles are digested. This is how many activated growth-factor receptors are switched off, covered in turning signals off, receptor tyrosine kinases and signalling endosomes.

A receptor can also be diverted. PCSK9, a secreted protein, binds the LDL receptor and in the cultured human cells tested sends it to the lysosome instead of back to the surface.

Familial hypercholesterolaemia: three defects, three experiments

FH is usually inherited as a dominant trait, and people with two defective copies are rarer and more severely affected (see pedigrees). Patients’ cells split the receptor cycle into steps, because different mutations break different steps.

Binding. In 1974 Brown and Goldstein reported that fibroblasts from five unrelated homozygous FH subjects lacked the high-affinity LDL binding of normal fibroblasts.

Entry. In 1977 Richard Anderson, with Brown and Goldstein, gave cultured fibroblasts ferritin-LDL at 4 °C and found 60 to 70 percent of the bound label over coated segments covering no more than 2 percent of the surface; on warming, it entered through coated vesicles. In J.D.’s cells the receptors failed to become incorporated into coated pits. In 1986 Davis and colleagues traced J.D.’s receptor to a tyrosine-to-cysteine substitution at residue 807 (in the numbering of that paper) in the tail. Rebuilt in receptor-deficient hamster cells, it bound LDL normally but entered slowly.

Return. A receptor engineered without its EGF-precursor-like domain could not release ligand at acid pH and recycled poorly. A patient receptor lacking exons 7 to 14 failed to recycle efficiently, but it also could not bind LDL; it was tested with beta-VLDL, which it still bound. It is a mixed defect, not a pure recycling one.

Worked problem: reading binding and entry

These numbers are illustrative, made up for practice. Cultured fibroblasts bind 125I-LDL at 4 °C, where nothing enters. They are washed and warmed to 37 °C for 10 minutes. An acid or heparin wash then strips LDL still on the surface; what remains counts as internal. Assume the strip removes all surface LDL and none of the internal LDL, and that nothing is degraded in 10 minutes.

Cells	Bound at 4 °C	Surface after warming	Internal after warming
Normal	40	10	30
Strain X	1.0
Strain Y	38	35	3

(ng LDL per mg cell protein)

binding relative to normal = bound (strain) ÷ bound (normal)

fraction internalized = internal ÷ bound at 4 °C

Step 1. Strain X binds 1.0 ÷ 40 = 2.5 percent of normal. This is a binding defect: the receptor is absent from the surface or cannot bind.

Step 2. Normal cells internalize 30 ÷ 40 = 75 percent in 10 minutes. Strain Y binds nearly normally (38) but internalizes 3 ÷ 38 = 7.9 percent. This is an entry defect, like J.D.’s receptor. Check: 35 + 3 = 38, so no LDL went missing.

Cells

Normal Strain X Strain Y

Step

Bound at 4 °C After 10 min at 37 °C, then strip

Binding relative to normal

Fraction internalized

Binding relative to normal = bound (strain) ÷ bound (normal); fraction internalized = internal ÷ bound at 4 °C. A low value of the first is a binding defect; normal binding with a low value of the second is an entry defect. The numbers are illustrative, made up for practice in the lesson.

Step 3. Electron microscopy of cold-bound ferritin-LDL: suppose 60 percent of the label lies in coated pits covering 2 percent of the surface. Set total label and area to 1.

density in pits ÷ whole-surface density = 0.60 ÷ 0.02 = 30

uncoated density = 0.40 ÷ 0.98 = 0.408

density in pits ÷ uncoated density = 30 ÷ 0.408 = 73.5

Check: 0.02 × 30 + 0.98 × 0.408 = 0.60 + 0.40 = 1. The two answers differ only in the reference. These are calculated ratios, not binding constants.

Design problem

A patient’s receptor appears to reach the cell surface, but the cells lose surface receptor after repeated rounds of uptake. Design experiments to show a recycling defect.

One way to do it

First choose a ligand the mutant still binds. If it cannot bind LDL, as with the receptor lacking exons 7 to 14, test beta-VLDL. Measure surface receptor with an antibody to its outer part, surface biotinylation or Pronase sensitivity, and total receptor by immunoblot.

Add cycloheximide to block protein synthesis, so new receptor cannot refill the surface. Then run rounds of ligand uptake and measure surface receptor after each. A recycling defect predicts normal entry with surface receptor falling round by round, while normal cells hold it steady. Measure release of bound ligand at acid pH as well, since failure to release can keep the receptor from returning.

Controls: normal cells, mutant and normal cells matched for starting surface receptor, and a binding-defective line, which should show no ligand-driven loss.