# Adaptors, Docking Proteins and Scaffolds

> Binding proteins route signals by controlling which partners meet, where they meet, and when they can act. Trace the Grb2 and Sos link from receptor to Ras, distinguish an adaptor from a docking protein and a scaffold, and explain how one kinase can serve several pathways without mixing their messages.

Course: Cell Signaling (The parts: switches and wiring). Level: advanced university and graduate. About 13 minutes.
Lesson page (link this one when citing): https://stemhop.com/lessons/adaptors-and-scaffolds/

Builds on: Phosphorylation: Kinases and Phosphatases · GTPase Switches: The Timer · Protein Modules: SH2, SH3, PTB, PDZ and PH

## What you'll learn

- Tell apart an adaptor, a docking protein and a scaffold by what each holds and whether it has enzyme activity.
- Explain how Grb2 and Sos connect an activated receptor to Ras, and how fly genetics and biochemistry found the same link.
- State what a scaffold can do for a cell (more productive encounters, the right output for a shared kinase, a place in the cell) and why more scaffold can mean less signal.
- Explain how one kinase can serve two pathways without mixing their messages.

**A growth-factor receptor can activate Ras without touching it. The bridge is assembly: an adaptor with no enzyme activity, Grb2, brings the exchange factor Sos to Ras.**

> **The one idea.** Binding proteins route signals by controlling which partners meet, where they meet, and when they can act.
>
> *Scope:* this rule describes the adaptor, docking-protein and scaffold examples here. Binding alone may not activate a pathway, not every scaffold lacks enzyme activity, and more scaffold does not always mean more signal.

## Three ways to organize a signal

A **phosphotyrosine** is a tyrosine carrying a phosphate. In a matching local sequence, it can become a docking site for an **SH2 domain**, a folded module that recognizes phosphotyrosine and nearby amino acids.

- An **adaptor** has no enzyme activity and supplies a link between partners. Grb2 is the model.
- A **docking protein** is itself phosphorylated at several sites and supplies a regulated platform for several partners. IRS-1 is the model.
- A **scaffold** organizes several successive members of one route. Ste5 in yeast is the model.

Partner count is only a shortcut: an adaptor typically bridges two partners, and a scaffold holds several, often three. The better test is what the protein supplies. KSR plays a scaffold-like role in mammals, but this pseudokinase changes its partners’ shapes and has some kinase activity of its own.

**Quick check.** A receptor phosphorylates many tyrosines on a non-enzyme protein, and different sites recruit different effectors. Which category fits best?

> **Check your prediction.** A docking protein. It supplies several regulated phosphosites itself; holding several partners would not identify that role.

*Figure (AI-generated illustration).* An adaptor supplies a link, a docking protein supplies regulated phosphosites, and a scaffold organizes successive members of one pathway. Image: three side-by-side schematics. An adaptor holds two partners, a receptor tail and Sos. A docking protein is a flat platform with four phosphorylated sites, each holding a different partner. A scaffold is a long bar holding three kinases.

## Grb2 routes a receptor signal to Ras

The first film follows the relay, **receptor phosphotyrosine → Grb2 → Sos → Ras-GTP**. Ras is a small switch protein anchored to the inner face of the plasma membrane, so bringing Sos to the membrane brings it within reach. Sos also has a second Ras-binding site, away from its exchange site; Ras-GTP bound there can speed Sos, so each Ras it activates helps it activate more.

Grb2’s two **SH3 domains** grip short proline-rich stretches, here the tail of Sos. Its SH2 domain reads more than the phosphate: nearby amino acids help decide which SH2 domain fits, so two phosphorylated tyrosines need not recruit the same SH2 protein.

**Predict.** Two receptor tails are phosphorylated equally, but only one has the local sequence recognized by Grb2’s SH2 domain. Which one recruits the Grb2–Sos complex?

> **Check your prediction.** Only the matching one should recruit it efficiently. The phosphate is necessary for SH2 recognition here; the surrounding residues help provide specificity.

### How experiments found the chain

A proposed order makes a genetic prediction: switching on a downstream part should bypass a missing upstream link.

**Predict.** Fly eye cells lacking a gene X never become the R7 photoreceptor. Suppose X acted downstream of Ras rather than upstream. What would always-active Ras do in those cells?

> **Check your prediction.** Nothing: R7 would still fail, because a part after Ras would still be missing.

In the fly eye, the receptor tyrosine kinase sevenless decides whether one cell in each unit becomes R7. In a sensitized screen, Michael Simon and Gerald Rubin weakened sevenless signaling, then found mutations in fly ras and in Sos, a relative of the yeast Ras activator CDC25. In nematodes, genetics placed the EGF-receptor-like let-23 before let-60 Ras, with sem-5 on the same pathway.

Biochemistry supplied the physical link. In 1992 Eve Lowenstein and colleagues in Joseph Schlessinger’s laboratory isolated Grb2, showed that its SH2 domain associates with activated epidermal growth factor (EGF) and platelet-derived growth factor (PDGF) receptors, and found that Grb2 plus Ras, but neither alone, drove resting rat fibroblasts to copy DNA. In 1993 several groups, including Tony Pawson’s and David Bowtell’s, showed that Grb2’s SH3 domains bind mammalian Sos and that growth factor recruits the pair to the receptor.

Grb2, fly Drk and worm Sem-5 are orthologs, as are Ras, fly Ras1 and worm let-60; the chain predates the split between those lineages. What follows Ras differs: three of six cells beside the worm gonad become vulval tissue, fly cells adopt R7 fate and fibroblasts copy DNA.

## IRS-1 turns one input into several branches

A docking protein expands the available surface. Insulin receptor substrate 1, **IRS-1**, reaches the membrane through a PH domain that binds membrane lipids and a PTB domain that recognizes a receptor phosphotyrosine motif. The insulin receptor then phosphorylates many tyrosines on IRS-1. The amino acids beside each phosphotyrosine decide which SH2 protein binds there, so different sites recruit PI 3-kinase, Grb2, the phosphatase Shp-2 and the adaptor Nck, sending the signal down several branches at once. Two of them:

> **Grb2 → Sos → Ras**
>
> **PI 3-kinase → PDK1 → AKT**

AKT helps regulate metabolism.

*Figure (AI-generated illustration).* Different phosphotyrosines on IRS-1 recruit different partners, allowing one insulin signal to enter several branches. The arrows summarize pathway order; they do not specify every intermediate or a direct phosphorylation reaction. Image: two schematic panels. A gold IRS-1 docking protein carries four illustrative phosphotyrosine sites bound by PI 3-kinase, Grb2, Shp-2 and Nck. A separate pathway overview shows Grb2 leading to Sos and Ras, and PI 3-kinase leading through an abbreviated pathway to PDK1 and AKT. Membrane attachment is not depicted.

Grb2 now sits in two routes that both lead to Ras. Harder is a kinase shared by routes that must reach different outputs, the problem the second film takes into a yeast scaffold.

**Film transcript: How a Scaffold Keeps Signals on Route**

### Film transcript: How a Scaffold Keeps Signals on Route

The first film used Grb2 to route a receptor signal toward Ras. But what if two pathways share the same kinase? Follow the same colors as one orange kinase faces two possible purple outputs. Yeast mating uses a kinase route ending in Fus3. Some upstream kinases also serve other responses. A pheromone input must reach the mating output without making the cell follow every route that shares a part. A shared kinase does not carry a label saying which response should follow. Its meaning depends on the partners around it. The cell therefore needs an organizer that assembles the mating route under the mating input. Ste5 is a green scaffold with no enzyme activity.

Different regions bind Ste11, Ste7 and Fus3, so one Ste5 can hold all three. This assembly can make productive encounters more likely. Within the assembled route, Ste11 activates Ste7, and Ste7 activates Fus3. Fus3 helps prepare the cell for mating. The scaffold does not perform those kinase reactions. It organizes the proteins that do. Proximity is only the first model. Ste5 responds to pheromone-linked recruitment at the membrane. Its shape opens there. A Ste5 domain also helps Ste7 activate Fus3, which is otherwise a poor substrate. The scaffold therefore helps select an output. It controls partners, place and activation, rather than locking every shared kinase away. The same shared parts can behave differently inside a different organized context.

The filament-growth route uses Ste11 and Ste7 but ends at Kss1. One salt-stress branch uses Ste11, then a different middle kinase, and ends at Hog1. Only the mating route has Ste5, so only the mating route switches on Fus3. Suppose the cell keeps adding empty Ste5 scaffolds. Must the number of complete kinase assemblies keep rising? Watch where each kinase might land, then make your prediction.

No. In the simplest model, all three kinases must share one scaffold to signal. With scaffold in excess, they spread across different scaffold molecules. Fewer scaffolds are complete, so predicted output falls. This model effect is called combinatorial inhibition. Experiments support both layers. Ste5 binds the mating kinases on different regions. Later work showed that a Ste5 domain makes Fus3 easier for Ste7 to activate. Assembly and selective activation work together. Return to the fork. A shared kinase does not choose an output alone. The organizer around it changes which partners meet, where the complex forms and whether the last step can proceed.

From the film

- Scaffolds route shared parts by controlling partners, place and activation.
- Ste5 organizes Ste11, Ste7 and Fus3 in the yeast mating pathway.
- Ste5 helps Ste7 activate Fus3; proximity alone does not explain specificity.
- In the simplest model, excess scaffold splits partners across incomplete complexes and lowers output.

## Ste5 routes shared kinases

The second film shows Ste5 controlling partners, place and activation on the mating route. The second film traced the shared parts to Kss1 and Hog1; of the two branches of the salt-stress cascade, Ste11 feeds only one. Fus3 helps halt the cell cycle for fusion, and without Ste5 a yeast cell cannot mate.

In 1994 Choi, Satterberg, Lyons and Elaine Elion showed that the three kinases bind different regions of the non-enzyme protein Ste5.

Scaffolds are often pictured as passive assembly platforms. A kinase and its target normally meet by random drift through a crowded cytoplasm, so holding both can increase productive encounters. That model is a useful start, but incomplete.

The first specificity idea was sequestration: a pathway-dedicated scaffold keeps kinase modules from talking to each other. Wendell Lim’s laboratory later found that Ste5 does not insulate by locking its kinases away; membrane gating and Fus3 unlocking do the work. Restraint also runs the other way: Hog1 helps keep high salt from activating the mating route.

Place has clearer cases elsewhere. InaD uses **PDZ domains**, modules that grip the last few amino acids of a partner, to gather light-response proteins in a fly photoreceptor. Paxillin gathers Ras-pathway proteins at focal adhesions, where a cell grips its surroundings.

### More scaffold can mean less signal

The second film’s excess-scaffold result comes from a model. Levchenko, Bruck and Sternberg modeled a generic scaffolded kinase cascade in 2000, found that output rose and then fell as scaffold increased, and called the effect **combinatorial inhibition**. The model assumes fixed kinase amounts and signaling only when all three share one scaffold; cooperative assembly or other regulation can change the result.

**Predict.** A Ste5 fragment promotes Fus3 autophosphorylation at one of its two activating sites. Must that raise the pathway’s final output?

> **Check your prediction.** No. The network effect must be measured. In the 2006 experiment from Wendell Lim’s laboratory, Fus3 made this way was only partly active and lowered pathway output, so Ste5 tuned the signal rather than simply accelerating it.

Ste5 has no human counterpart. KSR in the Ras–mitogen-activated protein kinase (MAPK) pathway and JIP-1 organize cascade kinases comparably but are not descended from Ste5; sequence similarity to Ste5 would not have found them.

## Local signals and rewiring

A kinase reads a messenger, and an enzyme destroys it. What holds the two close enough for a local pool of messenger to exist? An anchoring protein. In John Scott’s laboratory, Dodge and colleagues showed that muscle-selective A-kinase anchoring protein, **mAKAP**, holds protein kinase A and PDE4, a phosphodiesterase that destroys cyclic AMP. The kinase therefore responds especially to cyclic AMP made nearby, before it spreads away.

Separate binding surfaces also make scaffolds rewritable: swap one surface for another that recognizes a different partner, and an input can reach a new output. Wendell Lim’s laboratory has used that logic to build signaling proteins no organism has had.

### Design problem

Two pathways share kinase K. Input A must give only output A, and input B only output B. Design two ways to keep the messages apart.

> **One way to do it.** **Separate assemblies.** Scaffold A holds K with A’s activator and target; scaffold B holds K with B’s partners. This is the original sequestration proposal for Ste5; a stronger design also gates each scaffold by its input and helps activate the final kinase. Its weak point is dose: excess scaffold can split partners across incomplete complexes.
>
> **Separate places.** Anchor K at two locations. Input A makes the messenger at site A, where K sits beside target A; input B makes it at site B, beside target B. Each anchor also holds an enzyme that clears the messenger before it reaches the other site. This is the mAKAP logic. Its weak point is local clearance: remove the clearing enzyme and the signal can grow and spread.
>
> Other answers can work. A good design states what enforces specificity and what failure would break it.

## Going deeper

### Discovery order

The fly screen appeared in November 1991; Utpal Banerjee’s laboratory had isolated Sos earlier that year. The worm ordering came from Paul Sternberg’s group; Stuart Clark, Michael Stern and Robert Horvitz cloned sem-5 in March 1992, and Lowenstein’s Grb2 paper linked Grb2 to it in August. Across about nine months, a fly screen, worm clone and rat-cell experiment converged on one receptor-to-Ras chain.

The 1994 Ste5 result surprised because a protein essential for mating appeared to catalyze nothing. Elion’s group combined two-hybrid mapping with co-purification; Michael Wigler’s group (PNAS, 2 August) and George Sprague’s (Genetics, November) reached the same result with the then-new [yeast two-hybrid](https://stemhop.com/methods/immunoprecipitation-and-binding-assays/) method. Marcus and colleagues proposed that a dedicated scaffold might sequester a kinase module.

### A debated cardiac target

The mAKAP complex also holds a target. Marx and colleagues reported that protein kinase A phosphorylates the cardiac ryanodine receptor, the calcium-release channel for each heartbeat, causing FKBP12.6 to leave. They also reported defective regulation in failing hearts. Both findings remain debated, and some groups have not reproduced FKBP12.6 loss.

---

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