Vesicular Transport and Coat Proteins

A coat protein complex is a cytosolic protein assembly that is recruited transiently to a specific donor membrane, where it performs two coupled jobs: it mechanically deforms the flat bilayer into a curved bud, and it selects specific transmembrane and membrane-associated cargo proteins for inclusion in that bud by binding short sorting signals in their cytosolic tails. The three principal coats in the Secretory Pathway and endocytic system are COPII (coat protein complex II), COPI (coat protein complex I), and clathrin, each dedicated to a different transport step and each recruited by its own small GTPase switch. The unifying molecular logic is a GTPase cycle. Small GTPases of the Arf/Sar1 family exist in a GDP-bound, cytosolic, inactive form and a GTP-bound, membrane-anchored, active form. A guanine nucleotide exchange factor (GEF) on the donor membrane catalyzes GDP release and GTP loading, which exposes an amphipathic helix that inserts into the bilayer and anchors the GTPase there. The activated GTPase then recruits coat subunits from the cytosol. Coat polymerization proceeds until a bud has fully formed and pinched off as a vesicle; shortly afterward, GTP hydrolysis (stimulated by a GTPase-activating protein, GAP) releases the GTPase from the membrane, which destabilizes the coat lattice and allows it to be shed. This timed disassembly is essential because a coated vesicle cannot fuse with its target—the coat physically occludes the fusion machinery—so uncoating is a prerequisite for the next step, vesicle fusion. Each coat recognizes cargo through distinct short linear motifs, usually four to seven residues long, located in the cytosolic tails of transmembrane cargo proteins. COPII recognizes di-acidic (DXE) and other export motifs; COPI recognizes the C-terminal dilysine motif KKXX (or KXKX) on ER-resident membrane proteins, which specifies retrieval; clathrin, via adaptor proteins, recognizes tyrosine-based (YXXΦ) and dileucine ([DE]XXXL[LI]) motifs on receptors destined for endocytosis or lysosomal targeting. Soluble ER-resident proteins carry a distinct C-terminal signal, KDEL, recognized not by a coat directly but by the KDEL receptor, a transmembrane protein that itself becomes COPI cargo when it captures an escaped KDEL protein in the Golgi. This distinction—direct coat recognition of transmembrane cargo versus receptor-mediated capture of luminal cargo—is a recurring theme worth remembering when comparing coats.

Step-by-step: COPII budding at the ER

1. GTPase activation. Sar1-GDP, cytosolic and inactive, encounters Sec12, a GEF embedded in the ER membrane at specialized ribosome-free patches called ER exit sites (ERES). Sec12 catalyzes GDP-to-GTP exchange. 2. Membrane insertion. Sar1-GTP undergoes a conformational change that exposes an N-terminal amphipathic helix, which inserts into the cytosolic leaflet of the ER membrane, locally increasing curvature and anchoring Sar1 there. 3. Inner coat recruitment. Membrane-bound Sar1-GTP recruits the Sec23/Sec24 heterodimer. Sec23 is a GAP-binding scaffold; Sec24 contains multiple distinct binding sites that directly read cargo sorting signals (di-acidic DXE motifs, ER export signals on SNAREs, and others), concentrating specific transmembrane cargo into the nascent bud. 4. Outer coat polymerization. Sec13/Sec31 heterotetramers are recruited on top of the Sar1–Sec23–Sec24 "pre-budding complex." Sec13/Sec31 self-assembles into a cage-like lattice of cuboctahedral geometry, and this polymerization drives further membrane curvature, growing the bud. 5. Scission. The neck of the bud constricts; Sar1's amphipathic helix, concentrated at the highly curved neck, together with the mechanical strain of the polymerized coat, promotes membrane fission, releasing a ~60-80 nm COPII-coated vesicle. 6. Hydrolysis and uncoating. Sec23 acts as the GAP for Sar1, stimulating GTP hydrolysis. This is normally restrained until cargo is properly loaded (a proofreading step), after which Sar1 releases from the membrane, destabilizing the coat, which disassembles and returns its subunits to the cytosol for reuse.

Step-by-step: COPI budding at the Golgi

1. GTPase activation. Arf1-GDP is recruited to Golgi membranes where a GBF1-type GEF catalyzes GTP loading, exposing Arf1's N-terminal myristoylated amphipathic helix, which inserts into the membrane. 2. Coatomer recruitment. Arf1-GTP recruits coatomer, a stable heptameric complex (subunits α, β, β′, γ, δ, ε, ζ-COP) that is pre-assembled in the cytosol, unlike the two-step COPII coat. Coatomer binds Arf1-GTP and simultaneously engages cargo. 3. Cargo selection. The β′- and α-COP subunits contain WD40-domain pockets that directly bind the C-terminal KKXX/KXKX dilysine motif found on ER-resident membrane proteins (such as the KDEL receptor and ER-resident enzymes), driving their retrieval from the Golgi (and the ER-Golgi intermediate compartment, ERGIC) back to the ER. 4. Polymerization and budding. Coatomer polymerizes into a lattice analogous to clathrin's, curving the membrane into a bud that pinches off as a COPI-coated vesicle. 5. Hydrolysis and uncoating. ArfGAP1 (and related ArfGAPs) stimulate GTP hydrolysis on Arf1 once the vesicle has budded, triggering coat release. COPI operates bidirectionally in function even though it is mechanistically one coat: it mediates retrograde Golgi-to-ER transport (retrieving KKXX-tagged ER residents and recycling escaped SNAREs) and also intra-Golgi retrograde transport between cisternae, helping maintain the distinct enzyme composition of each Golgi cisterna as cargo matures through the stack—a theme picked up in Golgi Structure and Glycosylation.

Step-by-step: Clathrin budding at the plasma membrane and TGN

1. GTPase/adaptor recruitment. At the plasma membrane, the phosphoinositide PI(4,5)P₂ and activated Arf6 (or, at the TGN, Arf1) recruit heterotetrameric adaptor protein (AP) complexes—AP-2 at the plasma membrane, AP-1 at the trans-Golgi network/endosomes. 2. Cargo capture. AP complexes bind cytosolic tails of transmembrane receptors bearing tyrosine-based (YXXΦ, where Φ is a bulky hydrophobic residue) or dileucine-based sorting motifs. AP-2 recruitment also depends on direct PI(4,5)P₂ binding by its μ2 and α subunits. 3. Clathrin triskelion recruitment and lattice assembly. Clathrin is a three-legged triskelion, each leg built from one heavy chain (~190 kDa) and one light chain. Triskelions bind AP complexes and polymerize into a curved hexagonal/pentagonal basketwork lattice (like a geodesic dome), mechanically bending the membrane. 4. Accessory protein-assisted curvature. BAR-domain proteins and epsin insert amphipathic wedges or impose intrinsic curvature to help drive the deep invagination characteristic of clathrin-coated pits. 5. Scission by dynamin. The large GTPase dynamin polymerizes as a collar around the neck of the deeply invaginated pit; GTP hydrolysis by dynamin drives a conformational change that constricts and severs the neck, releasing a clathrin-coated vesicle (~100 nm). 6. Uncoating. The chaperone Hsc70, together with its cofactor auxilin, uses ATP hydrolysis to actively disassemble the clathrin lattice from the vesicle surface immediately after scission—mechanistically distinct from the GTP-hydrolysis-triggered disassembly of COPI/COPII, since clathrin itself has no GTPase.

Sorting-signal logic, compared

The coat-protein model was not deduced from structure alone; it was built from decades of biochemistry and genetics that converged on the same conclusion from independent directions.

Randy Schekman's yeast secretory mutants (late 1970s–1980s). Schekman screened temperature-sensitive yeast mutants for defects in protein secretion, identifying a set of sec genes. At the restrictive temperature, sec mutants accumulated distinct morphological blocks visible by electron microscopy: some mutants ballooned the ER with unexported cargo, others accumulated vesicles that failed to fuse with the Golgi, and others accumulated Golgi that could not export to the plasma membrane. This genetic dissection defined an ordered pathway of discrete, gene-controlled steps rather than a single continuous flow, and it identified SEC12, SEC23, SEC24, SEC13, and SEC31 as the genes later shown to encode the COPII machinery. Crucially, this showed that vesicle budding is a genetically separable, essential process—cells cannot secrete without these specific gene products.

James Rothman's cell-free reconstitution of Golgi transport (1980s). Rothman's lab, working initially in mammalian cells, developed an in vitro assay using Golgi membranes from two mutant cell lines: one lacking an enzyme needed to add N-acetylglucosamine to a viral glycoprotein (VSV G protein), the other lacking a different enzyme. When Golgi stacks from the two mutants were mixed with cytosol and ATP, transport vesicles budded from one Golgi population and fused with the other, transferring the glycoprotein and allowing both modifications to occur on the same molecule—detectable as a shift in size on a gel. This assay proved biochemically that vesicular transport between Golgi compartments could be reconstituted outside the cell, and it became the functional assay used to purify the coat proteins (initially called "coatomer") biochemically, subunit by subunit, based on which cytosolic fractions were required to support transport.

Balch, Orci, Schekman, and Rothman's convergence. William Balch and Lelio Orci contributed detailed electron microscopy showing coated vesicular intermediates budding from the Golgi in the reconstituted system, directly visualizing the coats predicted by the biochemistry. Schekman's genetic SEC gene products were then shown, protein by protein, to correspond to the biochemically purified coat subunits identified in Rothman's cell-free system—the same molecules identified by entirely independent methods (genetics in yeast, biochemistry in mammalian cell extracts) turned out to be the same proteins. This convergence of two independent lines of evidence onto one mechanism is exactly why Schekman and Rothman (together with Thomas Südhof, for the fusion machinery) shared the 2013 Nobel Prize in Physiology or Medicine.

Purified-protein reconstitution: sufficiency, not just necessity. The most rigorous demonstration came from experiments reconstituting COPII budding using only purified, recombinant components: synthetic liposomes of defined lipid composition, purified Sar1, purified Sec23/Sec24, purified Sec13/Sec31, and GTP (or the non-hydrolyzable analog GMP-PNP to trap intermediates). This minimal system was sufficient to generate uniformly sized ~60–80 nm vesicles budding from the liposomes, visualized directly by electron microscopy. Because no other cytosolic proteins were present, this proved that these five components (Sar1, Sec23, Sec24, Sec13, Sec31) are sufficient for membrane deformation and vesicle scission—not merely correlated with budding, but mechanistically causal for it. Parallel reconstitutions established the same sufficiency for COPI (Arf1 + coatomer + GTP + liposomes) and for clathrin/AP-2 coated pits (clathrin + AP-2 + PI(4,5)P₂-containing liposomes), generalizing the conclusion across all three coats.

Dynamin and endocytosis: the shibire mutant. A complementary line of evidence for clathrin-mediated endocytosis came from the Drosophila temperature-sensitive mutant shibire, which paralyzes flies at restrictive temperature because synaptic vesicle recycling fails. Electron microscopy of shibire neurons at the restrictive temperature revealed deeply invaginated clathrin-coated pits with long, uncut necks decorated by a collar of protein, physically stuck at the point of scission. The shibire gene was shown to encode dynamin, directly demonstrating that dynamin's GTPase activity is required for the final membrane-severing step of clathrin-coated vesicle formation—visualized as an accumulated intermediate exactly where the model predicted a block would appear.

Example 1 — Diagnosing a coat defect from a trafficking phenotype. A researcher generates a temperature-sensitive yeast mutant. At 37°C (restrictive), electron microscopy shows a massively distended ER studded with ribosomes, but no accumulation of 60–80 nm vesicles near ER exit sites; pulse-chase labeling shows that a secreted invertase reporter remains in an ER-glycosylated (core-glycosylated, not Golgi-modified) form and never reaches the cell surface. Reasoning: because vesicles never accumulate at all (rather than accumulating and failing to fuse), the block must be upstream of budding itself, not at a fusion step. The core-glycosylation pattern confirms cargo never left the ER (Golgi enzymes never modified it). This phenotype is diagnostic of a COPII coat-assembly defect—for example, a mutation in SEC23 that prevents Sec23/Sec24 recruitment to Sar1-GTP, so ER exit sites cannot even initiate budding. Contrast this with a mutant that does accumulate 50 nm vesicles in the cytoplasm but fails to deliver cargo to the Golgi: that phenotype would implicate a fusion-stage defect (e.g., a SNARE or tethering factor), not a coat defect, because budding itself succeeded.

Example 2 — Predicting a cargo's itinerary from its sorting motif. A membrane protein X has, on its short cytosolic tail, the sequence …-Lys-Lys-Thr-Asn-COOH (a C-terminal KKXX-like motif) and is known to reside in the ER lumen-facing membrane, functioning as a chaperone-interacting enzyme. A second membrane protein Y has a cytosolic tail sequence …-Tyr-Gln-Thr-Ile-… (a YXXΦ motif, where Φ = Ile, a bulky hydrophobic residue) and is known to be a nutrient receptor normally displayed at the plasma membrane. Predict each protein's expected trafficking behavior if its sorting motif is deleted by mutagenesis. Reasoning for protein X: the KKXX motif is the classic COPI-recognized ER-retrieval signal, read directly by the α/β′-COP subunits of coatomer. If protein X leaks out of the ER into the Golgi (a certain amount of leakage happens for any ER protein because bulk flow moves membrane forward by default), COPI recognizes its KKXX tail and packages it into retrograde vesicles back to the ER, maintaining ER residency at steady state. If the KKXX motif is deleted, protein X can no longer be retrieved once it escapes the ER; it will be carried forward by bulk anterograde flow and progressively mislocalize to the Golgi and then the plasma membrane, since nothing pulls it back. Reasoning for protein Y: the YXXΦ motif is read by the μ subunit of AP-2 (at the plasma membrane) and AP-1 (at the TGN), directing active recruitment into clathrin-coated pits for internalization (and, in the biosynthetic pathway, for TGN-to-endosome sorting). If this motif is deleted, protein Y can still reach the plasma membrane normally via default bulk secretory flow, but it will no longer be efficiently endocytosed by clathrin-coated pits. The experimentally observed consequence in real receptors bearing such mutations (e.g., mutant LDL receptors in familial hypercholesterolemia, class 2 internalization-defective alleles) is a receptor that sits at the cell surface, ligand-binding-competent, but fails to be internalized—raising surface half-life but crippling receptor-mediated uptake of its ligand.

Randy Schekman began, in the mid-1970s, with a deceptively simple genetic question: could yeast secretion be dissected the way Benzer and others had dissected bacteriophage assembly—by isolating conditional mutants that stall the pathway at defined points? Working at Berkeley, Schekman and his students screened for yeast that, at elevated temperature, could no longer secrete invertase and instead accumulated aberrant internal membranes. The electron micrographs from this screen—showing distinct morphological "traffic jams" at the ER, at 50 nm vesicles, and at the Golgi depending on which gene was mutated—became some of the most cited images in cell biology, because they showed, visually, that secretion is built from separable, genetically encoded steps. James Rothman took a complementary biochemical path. Rather than mutating genes, he asked whether the transport step between Golgi compartments could be broken and reconstituted entirely in a test tube, using cell-free extracts and an assay sensitive enough to detect a single sequential pair of enzymatic modifications on one glycoprotein. This assay let his lab purify, fraction by fraction, the cytosolic proteins required for vesicle budding and fusion—work that identified coatomer and, alongside parallel work on SNARE proteins, established the biochemical machinery underlying the genetic pathway Schekman had mapped. Their independent routes converged, gene for protein, into a single unified mechanism.

Coat protein biology is directly clinically relevant. Mutations in SAR1B cause chylomicron retention disease (Anderson disease), in which intestinal cells cannot package the largest dietary lipoprotein cargoes into COPII vesicles efficiently, causing fat malabsorption and failure to thrive in infants—a direct demonstration that COPII coat capacity can become cargo-limiting for unusually large cargo. Mutations in COPA (an α-COP subunit) cause a familial autoinflammatory lung and joint disease (COPA syndrome), because defective COPI retrieval of the STING protein back from the Golgi causes STING to accumulate and constitutively activate innate immune signaling. Defects in clathrin-adaptor recognition of the LDL receptor's internalization signal cause a subclass of familial hypercholesterolemia, where receptors reach the cell surface normally but cannot be endocytosed, so LDL clearance fails despite normal receptor expression and ligand binding. Coat proteins are also central experimental tools: fluorescently tagged Sec23, coatomer, or clathrin light chain are routinely used to live-image budding events in real time, and dominant-negative GTPase mutants (Sar1-H79G, a GTP-hydrolysis-deficient mutant that locks COPII coats onto membranes) are workhorse reagents for trapping and purifying budding intermediates. In biotechnology, understanding COPII cargo-capacity limits informs the engineering of mammalian cell lines for high-yield recombinant protein and antibody production, since ER export can become rate-limiting when a cell is engineered to overexpress a therapeutic protein. In virology, many enveloped and non-enveloped viruses hijack clathrin or COPI machinery for entry or egress, making coat-protein-interacting steps candidate antiviral targets. Finally, coat protein logic is the conceptual bridge to the next steps in the pathway: once a coated vesicle buds and uncoats, its fate is decided by Rab GTPases that mark it for a specific destination and by SNARE proteins that catalyze the final membrane fusion event—and the compositional changes that COPI-mediated retrograde traffic maintains across cisternae underlie the entire organization discussed in Golgi Structure and Glycosylation.

Ask the tutor: How does Sec24's ability to bind multiple different cargo signals simultaneously allow a single COPII vesicle to package many different kinds of cargo proteins at once?