# Vesicle Budding, Targeting and Fusion

> A vesicle delivers cargo in three handoffs that can each fail on its own: a coat buds it, a Rab and a tether catch it, and SNARE proteins, which NSF later pulls apart, fuse it. This lesson follows one vesicle through budding, capture and fusion. Afterwards you can read yeast sec-mutant and cell-free data to put the steps in order and to tell budding from targeting and tethering from fusion.

Course: Cell Biology (Traffic and movement). Level: core university (third year). About 24 minutes.
Lesson page (link this one when citing): https://stemhop.com/lessons/vesicle-traffic/

Builds on: Protein Sorting into Compartments · Membrane Organization and Topology · Experiments That Distinguish Mechanisms

## What you'll learn

- Explain how COPII, COPI and clathrin with AP2 each select cargo and bend membrane, and name the small GTPase or lipid that starts each coat.
- Describe how Rab GTPases and tethers capture a vesicle, and how Qa, Qb, Qc and R SNAREs zipper into a four-helix bundle that NSF and SNAP later take apart.
- Calculate a fraction secreted and a recovery fraction from invertase data, and order sec mutants from single and double-mutant phenotypes.
- Design a cell-free experiment that separates budding, docking and fusion and uses timed inhibitor addition to place a step.

**A yeast mutant called sec1 grows at 25 °C, but at 37 °C it keeps making secretory enzymes and cannot release them.** Novick and Schekman reported it in 1979. The first film above follows those enzymes along the route and through the vesicles that never fuse with the surface; the three jobs it names are **budding** (pick up cargo and pinch off), **targeting** (be caught by the right membrane) and **fusion** (merge with it).

**Predict.** In sec1 cells at 37 °C, which of the three jobs worked for the last vesicle on the route, and which could have failed?

> **Check your prediction.** Budding worked: the vesicles exist. Targeting or fusion failed, and a pile of vesicles alone cannot say which.

> **The one idea to carry through this lesson.** A vesicle’s trip is three handoffs, and each can fail on its own. A **coat** buds the vesicle and picks its cargo. A **Rab-GTP with a tether** catches it at the right membrane (targeting). A **SNARE bundle, guided by an SM protein**, fuses it. On a linear route, when one handoff fails, cargo piles up in the last place it reached, and a fast, specific inhibitor stops working once the state it targets is past.

*Figure (AI-generated illustration).* A vesicle's trip is three separate handoffs: a coat buds it and picks its cargo, a Rab-GTP and a tether catch it at the right membrane, and a SNARE bundle guided by an SM protein fuses it. Each can fail on its own, which is why mutants and inhibitors can tell them apart. Image: a three-panel strip. Left, a coated bud pinches off a donor membrane carrying cargo. Middle, an uncoated vesicle is held above a target membrane by a long tether rising from a Rab-GTP, its SNARE not yet touching the target's SNAREs. Right, the vesicle sits on the target membrane on a wound SNARE bundle, with an SM protein beside it.

This picture fits the coat-dependent routes in this lesson. Some endocytosis uses other budding machinery, and tubular carriers and lipid transfer without a vesicle fall outside it.

A few terms. The **cytosol** is the cell’s fluid outside organelles; the **lumen** is the inside of a membrane-bound compartment. A **vesicle** is a small membrane sphere, often 50 to 100 nm across, that carries lumenal cargo and membrane proteins from a donor compartment to a target. Its lumen never mixes with the cytosol (see [membrane topology](https://stemhop.com/lessons/membrane-organization-topology/)), so every protein that steers it works on its cytosolic face. A coat grabs a membrane protein’s cytosolic tail directly. A soluble cargo, such as the secretory enzyme invertase measured below, cannot touch the coat: it is caught by a membrane-spanning receptor that the coat grabs (Erv29 in yeast, ERGIC-53 in mammals), or rides along in the fluid a bud encloses.

## Step 1, budding: a coat picks the cargo and bends the membrane

A **coat** is a lattice of cytosolic proteins that assembles on a patch of membrane. The film’s cookie cutter with hands is a simplification: the cutter shapes the bud and the hands grab the cargo, but different parts of a coat often do each job.

Take **COPII**, the ER’s coat. It is started by a switch, **Sar1**, a small GTPase (see [GTPase switches](https://stemhop.com/lessons/gtpase-switches/)). Sar1-GTP grabs the ER membrane and pulls in an inner layer, **Sec23/24**, which holds cargo and in turn recruits an outer cage, **Sec13/31**. Cargo choice sits mainly in Sec24: yeast mutant studies support a model in which Sec24 has several independent cargo-binding sites. Sar1’s GTP is also a timer. When Sar1 hydrolyzes it, the coat begins to let go, and the vesicle can go on to be caught and fused.

| Coat | Started by | Where it works | Cargo signal it reads |
| --- | --- | --- | --- |
| COPII | Sar1-GTP | ER to Golgi | signals read by Sec24 |
| COPI | Arf1-GTP, recruiting coatomer (seven subunits) | within the Golgi, and Golgi back to ER | dilysine (KKXX) tails, read by α- and β′-COP |
| Clathrin with AP2 | the lipid PI(4,5)P2, which opens AP2 | plasma membrane | motifs read by AP2; clathrin binds no cargo |

A KKXX tail is a return label: COPI fetches escaped ER membrane proteins back from the Golgi. [Crystal structures](https://stemhop.com/methods/x-ray-crystallography/) of recombinant AP2 show it locked, part of its β2 subunit covering both cargo sites; it opens on membranes with PI(4,5)P2, a phosphoinositide concentrated in the plasma membrane (see [lipid anchors and addresses](https://stemhop.com/lessons/lipid-anchors-and-addresses/)), so the lipid and the cargo together tell it where it is (compare the locked and open structures on this page). **Dynamin**, a GTPase, pinches the finished clathrin bud off.

*Figure (AI-generated illustration).* Each coat is started by a switch or a lipid (Sar1-GTP, Arf1-GTP, PI(4,5)P2) and does two jobs, choosing cargo and bending the membrane. In COPII, Sec23/24 holds cargo under a Sec13/31 cage; coatomer grips KKXX return tails; AP2 grips cargo while clathrin only builds the cage, and dynamin pinches the bud off. Image: three membrane buds in schematic style. A large ER bud shows a row of Sar1-GTP on the membrane under an inner Sec23/24 layer that holds cargo tails and a cargo receptor carrying soluble cargo, all under an outer Sec13/31 cage (COPII). A smaller Golgi bud shows Arf1-GTP under a coatomer layer gripping a KKXX tail (COPI). A plasma-membrane bud shows AP2 on a PI(4,5)P2 patch gripping cargo, a clathrin cage outside it, and a dynamin ring at the neck.

**Predict.** In 1994 Barlowe and colleagues made ER vesicles in a test tube with GMP-PNP, a GTP analogue that cannot be hydrolyzed. Do the vesicles bud? Can they fuse with the Golgi?

> **Check your prediction.** They bud, but the timer never runs out. The vesicles kept their coat proteins and failed to target and fuse with the Golgi. Making a vesicle and delivering it are separate steps, and the coat must start to come off before the next handoff.

Real coats and ER exit sites are messier than this picture (Going deeper, “Coats that curve late, and tubes as well as spheres”).

## Step 2, targeting: a Rab and a tether catch the vesicle

A budded vesicle must find its target among many membranes. Part of the answer is **Rab** proteins, a large family of small GTPases. A Rab has a lipid tail (prenyl groups) that can anchor it in a membrane. When the Rab carries GDP, a protein called **GDI** (GDP dissociation inhibitor) can wrap that tail and carry the Rab in the cytosol. At the right membrane the Rab leaves GDI, its tail enters the bilayer, and a guanine-nucleotide exchange factor (GEF) (an exchange factor) there loads GTP, which holds it on. Rab-GTP recruits proteins called effectors, which generally prefer the active, GTP-bound form. The film’s Rab “anchored in a membrane” and “held in the cytosol by GDI” are these two states, simplified: a Rab-GDP can also sit briefly on a membrane.

Among the effectors are **tethers**: long coiled-coil proteins or large complexes that capture a vesicle before its SNAREs engage. In yeast, the tether Uso1p works with the Rab Ypt1p on ER-to-Golgi vesicles. On the yeast vacuole, a six-subunit tether called HOPS binds the Rab Ypt7.

*Figure (AI-generated illustration).* GDI carries Rab-GDP in the cytosol; at the right membrane a GEF loads GTP, and Rab-GTP recruits effectors such as a tether that catches the vesicle. No single molecule is the address: the Rab, the membrane's lipid and the SNAREs each narrow the choice. Image: on a target membrane, an active Rab-GTP anchored by its lipid tail recruits a long tether that holds an incoming vesicle, next to a phosphoinositide patch and target SNAREs. An inset shows a Rab-GDP with its lipid tail wrapped by GDI in the cytosol, and a dashed arrow through a GEF to the membrane-bound Rab-GTP.

The film’s lit sign over a loading bay is one of several: the tether, the membrane’s phosphoinositide and the SNAREs also narrow the choice, and the evidence fits a combination better than any one label.

## Step 3, fusion: SNAREs zip, an SM protein guides them, NSF resets them

A caught vesicle still has to merge, and bilayers resist merging: their water-coated surfaces repel at close range. **SNAREs** supply the pull. Most have a **SNARE motif** of about 60 to 70 amino acids next to a membrane anchor; **v-SNAREs** sit in the vesicle and **t-SNAREs** in the target. A lone motif is floppy. Four of the right kind wind into a parallel four-helix bundle. The strands are named by the amino acid each places at the bundle’s centre: three give a glutamine (Q) and one an arginine (R), so they are **Qa, Qb, Qc and R**. Often the vesicle gives R and the target gives the three Q strands, though not at every step. At a nerve terminal, syntaxin-1 gives Qa, SNAP-25 gives both Qb and Qc, and synaptobrevin on the vesicle gives R.

The film shows the zip, which runs from the N-terminal end toward the membrane anchors. Four strands are not four anchors: SNAP-25 gives two strands and attaches by lipid chains on the loop between them. While the strands come from two membranes, the bundle is a **trans-SNARE complex**. Afterwards the whole bundle sits in one membrane: a **cis-SNARE complex**.

**Predict.** Does the moment of fusion use ATP?

> **Check your prediction.** No. Four floppy strands folding into one tight rope releases free energy, and the zip turns it into a pull on the membranes. The zip itself uses no ATP; ATP pays to get SNAREs ready. A cis bundle is stable and useless, since both ends are already in one membrane. **NSF** (Sec18 in yeast), an ATPase, binds cis bundles through its adaptor **SNAP** (Sec17) and spends ATP to pull the strands apart, after one fusion and before the next. NSF cocks the spring that the next zip releases.

*Figure (AI-generated illustration).* SNARE strands from two membranes zip into one four-helix bundle, and the zip itself pulls the bilayers together without ATP. The spent cis bundle sits in one membrane until NSF, through SNAP, spends ATP to pull it apart for the next round. Image: a three-frame strip in which one vesicle SNARE and three target SNARE strands wind together between two membranes with an SM protein beside them, the membranes merge, and the finished bundle lies in one membrane. An inset shows NSF, bound through SNAP adaptors, pulling the cis bundle apart, with ATP used only at that step.

In a cell, SNAREs do not work alone. The main fusion steps that have been studied also need an **SM protein**, named for its first two members, Sec1 and Munc18, which binds SNAREs and helps the right four zip together. On the yeast vacuole the SM protein Vps33 is built into the tether HOPS, so one complex both catches and helps fuse.

That answers the opening puzzle. **Sec1 is the SM protein for the last step of secretion: fusion of secretory vesicles with the plasma membrane.** Upstream, invertase-laden vesicles still budded from the Golgi at 37 °C. They piled up, then fused after Sec1 worked again at 25 °C.

Sec1’s known job, not the pile, names fusion. The same pile of 80 to 100 nm vesicles appears when the secretory-vesicle Rab (Sec4), its tether (the exocyst) or a SNARE (Sec9, the yeast relative of SNAP-25) fails; a pile cannot tell a vesicle never caught from one caught but unable to fuse. Test tubes split them; imaging single vesicles in living cells can too. Before that, the steps had to be found and put in order.

The second film picks up here: how a pile of stuck cargo puts the handoffs in order, and how test tubes split catching from fusing.

**Film transcript: Reading Where Cargo Stops**

### Film transcript: Reading Where Cargo Stops

Chapter one followed a vesicle through three handoffs: bud, catch, fuse. But how did anyone put those handoffs in order? What can a pile of stuck cargo tell you? First, measure the block. Yeast secrete an enzyme called invertase. The fraction secreted is the enzyme outside the cell, divided by all of it. Is the trapped enzyme dead, or only waiting?

Return the cells to 25 °C, with new protein synthesis blocked. If the trapped enzyme moves outside, it was waiting to be delivered. At 37 °C, these secretion-blocked cells become dense. In 1980, Novick, Field and Schekman used that to enrich new mutants.

Electron microscopy then showed what each one piled up. Chapter one's rule applies here: on a linear route, cargo piles up in the last place it reached. In 1981, double mutants put the blocks in a broad order. Now a harder case. An enzyme called NSF resets SNAREs at many steps along the route. Where would a heat-sensitive NSF mutant pile up its cargo?

At its earliest blocked step. In classical sec18 experiments, ER and small ER-derived vesicles piled up, not late secretory vesicles. A pile shows where transport stops, not when or where else a protein acts. Test tubes add what a pile cannot. Cell-free assays mix donor and acceptor membranes with cytosol and ATP. Then each step can be timed, removed or rebuilt. Is catching enough to fuse? Yeast HOPS is a tether that binds the Rab Ypt7. Separate forms missing the beta propeller of Vps11 or Vps18 still tethered Ypt7-bearing lipid spheres. But they failed to promote fusion. So read stuck cargo in two ways. Where cargo stops gives the order; test tubes split catch from fuse.

From the film

- The fraction secreted is the enzyme outside the cell divided by all of it, so a block can be measured.
- If trapped enzyme leaves once the cells return to 25 °C with synthesis blocked, it was waiting to be delivered, not dead.
- Double mutants put the blocks in order, but a pile shows where transport stops, not when or where else a protein acts.
- Because NSF acts at several steps, a sec18 mutant reveals only its earliest blocked step.
- HOPS forms lacking the Vps11 or Vps18 beta propeller still caught Ypt7-bearing lipid spheres but failed to promote fusion.

## Putting the handoffs in order: mutants

In 1980 the density screen from the second film used a Ludox gradient, and Novick, Field and Schekman sorted 188 clones that trapped invertase at 37 °C, the restrictive temperature, into 23 complementation groups (see [complementation and pathways](https://stemhop.com/lessons/complementation-and-pathways/)): the screen’s yield, not the number of genes secretion needs. The SEC numbers carry no information about order; order came from [double mutants](https://stemhop.com/methods/epistasis-analysis/) in 1981.

Now the numbers, made up for practice. Invertase is measured in arbitrary units after 1 h at 37 °C.

> **fraction secreted = external / (external + internal)**

**Step 1.** Wild type: external 180, internal 20, total 200, so 180/200 = 90% secreted. Mutant secX: external 12, internal 188, total 200, so 12/200 = 6%. The block is real.

**Step 2.** The recovery test: shift secX back to 25 °C with cycloheximide to stop new synthesis. After 60 min external is 150 and internal 50, total still 200. The pool released is (188 − 50)/188 = 138/188 = 73.4%. External rose by the same 138 units, so the trapped material was deliverable.

**Step 3.** [Electron microscopy](https://stemhop.com/methods/electron-microscopy/) at 37 °C shows secX accumulates ER, secY accumulates Golgi-like stacks (Berkeley bodies), and secZ accumulates 80 to 100 nm vesicles. The secX secZ double mutant accumulates ER. The secY secZ double accumulates Golgi-like membranes.

**Predict.** Put X, Y and Z in order. Which double mutant would test the one link the data leave open?

> **Check your prediction.** Cargo piles up at the first block it meets, so a double mutant looks like the earlier single block: X acts before Z, and Y before Z. ER precedes Golgi on the secretory route, which places X before Y; a secX secY double accumulating ER would test that directly. Order: ER (X), Golgi (Y), vesicles (Z), surface.

The rule assumes a linear pathway with arrest states you can tell apart. It orders where transport is blocked, not when a protein acts: a gene product could still act early and be needed only later. A gene needed at several steps shows only its earliest blocked one. NSF resets SNAREs at many steps, and in classical sec18 experiments at 37 °C, ER and small ER-derived vesicles piled up, not late secretory vesicles.

*Interactive on the lesson page:* Yeast sec mutants: read the secretion data

## Splitting catch from fuse: test tubes

Mutants show where cargo stops. [Cell-free](https://stemhop.com/methods/in-vitro-reconstitution/) assays mix donor membranes, acceptor membranes, cytosol and ATP so that individual steps can be timed, removed or rebuilt. In the assays of the second film, docking means vesicles bound to the acceptor membrane, the tethered state.

**Predict.** In 1998 Cao, Ballew and Barlowe ran a yeast ER-to-Golgi assay and added GDI, which pulls Rabs off membranes, at different times. Added at the start, GDI blocked fusion. What happens to its effect if it is added after the vesicles have docked?

> **Check your prediction.** It fades. Once vesicles had docked, fusion was no longer sensitive to GDI, so the GDI-sensitive step is passed by docking. That does not show the Rab left or stopped mattering: GDI pulls off GDP-bound Rabs, and an active Rab-GTP resists it. In general, losing sensitivity to a fast, specific inhibitor shows that an inhibitor-sensitive state has been passed; it times the end of a whole step only if the inhibitor can still reach and act on its target later, which is why budding, docking and fusion are measured separately. The authors proposed that this first docking needs the tether Uso1p and the Rab Ypt1p but not SNARE proteins.

Purified yeast HOPS shows that catching is not fusing: separate forms lacking the beta-propeller domain of Vps11 or Vps18 still tethered Ypt7-bearing liposomes but failed to promote fusion.

### Design problem

You have this cell-free assay, with yeast ER as donor and Golgi as acceptor. You want to know whether GDI acts before or after vesicles dock, and whether a purified tether is enough for fusion. Design the experiment.

> **One way to do it.** Use three readouts. Budding: vesicles released from ER, collected after spinning out large membranes. Docking: vesicles that sediment with the acceptor Golgi. Fusion: cargo that gains an acceptor-only sugar, or lumenal content mixing.
>
> Start the reaction and add GDI at 0, 5, 10, 15 and 20 minutes. Plot fusion still blocked against time of addition, with docking measured on the same schedule. If GDI acts at docking, sensitivity is lost as docking completes, as Cao, Ballew and Barlowe found. If it acts at fusion itself, additions made after docking but before fusion still block. Measure when fusion itself happens on the same schedule, or hold vesicles docked but unfused before adding GDI, to tell the two apart (Going deeper, “Why the fusion time course matters”).
>
> For the tether, run docking and fusion separately with the purified tether, matched SNARE proteoliposomes and an active Rab, since tethering without fusion is possible. Controls: no ATP, no cytosol, no tether, and a leakage check so that content released by broken membranes is not read as fusion.

In one line: coat buds, Rab and tether catch, SNAREs with an SM protein fuse, NSF resets; where cargo stops and when an inhibitor stops mattering point to the failed handoff.

## Going deeper

### A coat from three parts

With purified yeast proteins, Matsuoka and colleagues showed in 1998 that Sar1, Sec23/24 and Sec13/31 alone bud coated vesicles from chemically defined liposomes that contain PI4P or PI(4,5)P2, so the coat can bend membrane without help from other ER proteins. After Sar1 hydrolyzes GTP, parts of the coat can stay on long enough to help the vesicle be caught at the Golgi.

### Coats that curve late, and tubes as well as spheres

In cultured BSC-1 cells, clathrin lattices began to curve only at around 70% of their final clathrin content, and raised membrane tension stalled the change, so a coat does not always curve from the start. And in mammalian cells, whole-cell imaging in 2021 found ER export through an interwoven tubular network, not only through separate small spheres.

### HOPS up close

The purified six-subunit HOPS complex is about 43 nm tall, and a structural working model places its Vps39 and Vps41 subunits at the Ypt7 contacts. Its SM protein, Vps33, is held in the complex by Vps16 and binds SNAREs.

### Paying for fusion

Assembly of one neuronal SNARE bundle releases about 65 kT, about 170 kJ/mol at body temperature (Gao and colleagues, 2012). In a single-molecule assay, NSF took one complex apart in one round of hydrolysis of the ATP already bound to it, at most six (Ryu and colleagues, 2015); other assays report more ATP per complex. As an illustration, six ATP at 50 to 57 kJ/mol each supply 300 to 342 kJ/mol, about 116 to 133 kT at 310 K, more than this construct’s assembly releases. Reversing the same assembly under the same conditions must cost at least what it released, but these separate experiments do not measure a matched reset cycle or a universal ATP budget.

### Can SNAREs fuse membranes alone?

In 1998 Weber and colleagues put recombinant v-SNAREs and t-SNAREs into separate liposomes, and the docked liposomes mixed their lipids at physiological temperature, which the authors read as fusion. That shows SNAREs can merge those bilayers on their own, though lipid mixing alone cannot tell full fusion from hemifusion. It does not show they supply specificity or speed in a cell.

### What NSF might do before fusion

NSF may do more than recycle SNAREs. In a 2024 yeast vacuole [reconstitution](https://stemhop.com/methods/in-vitro-reconstitution/), Sec18 still promoted fusion with Sec17 fixed to the membrane; whether this holds at other fusion steps is not yet known.

### Why the fusion time course matters

Take a simple model: a batch of vesicles, all undocked at time zero, docks with a first-order half-time of 10 min (75% docked by 20 min), each docked vesicle then fuses with its own first-order half-time, and each inhibitor instantly and completely blocks its step. If the fusion half-time is about 2 min, a docking inhibitor and a fusion inhibitor added at 20 min block about 25% and 31% of eventual fusion: nearly the same. With a 15 min fusion half-time they block about 25% and 69%, and the steps separate clearly.

---

From STEMhop (https://stemhop.com), free lessons by Christopher Deppmann. Lesson text © the authors. Practice questions: https://stemhop.com/lessons/vesicle-traffic/
