Protein-Protein Interactions

Experimental Techniques

These help discover protein-protein interactions.

Two-hybrid screening

Also known as yeast two-hybrid system (Y2H).

This tests for physical interaction or binding between two proteins.

In yeast there is a transcription factor GAL4, which allows growth when histidine is not present.

GAL4 is made up to two domains, a DNA-binding domain (DBD) and an activation domain (AD). These must be linked for the report gene to be transcribed.

Instead of linking DBD and AD together, protein A is fused to DBD, and protein B is fused to AD. If protein A binds to protein B then the link is established, gene will be transcribed and can be detected. Protein A is the bait, and protein B is the prey.

If binding happens then the link is DBD-BAIT-PREY-AD and transcription occurs.

If no binding happens then the link is DBD-BAIT and separately PREY-AD, and no transcription occurs.

ref: Protein-Protein Interactions and Networks Panchenko Przytycka (Springer 2008) pp5


Crystallography, NMR, mass spec




An interactome is a collection of all protein-protein interactions within an organism.

It can be represented as a network or graph of proteins (nodes) and interactions (edges).

Graphs and Networks

n is the total number of nodes

degree, for node k, is the number of edges or links it has to other nodes.

degree distribution p(k), is the probability of a node having k edges or links.

Random network

degree distribution is a poisson distribution.

Scale-free network

power law distribution. k^-γ

Most protein-protein interactions are like this.

Some nodes are busy, and are called hubs. Most only have a few edges.

Features include 'small worlds network', the number of edges in a path between nodes is small.

Robust, there are redundant edges or paths. Still functional when a node is removed or edge fails.

If multiple hub nodes are removed then failure will occur.


This is a form of quaternary structure. They are linked by non-covalent protein-protein interactions.

homodimer, heterodimer, hormone (ligand) receptor, DNA transcription factor, RNA binding protein


If a protein is unstable on its own, it is obligate. An example of hemoglobin, which requires 2αand 2β subunits to be stable and have cooperative effect when binding oxygen.

Stable vs Transcient

Stable: formed and remains formed

Transient: associates and disassociates

Binding Affinity

Dissociation constant can be calculated between the subunits.

$$ K_{d} = \frac{[{proteinA}][{proteinB}]}{[{proteinABcomplex}]} $$

where Kd is the dissociation constant.

The equilibrium constant K, is 1/Kd

As more complex is formed, for the dissociation constant Kd, the denominator is getting larger, and the numerator is getting smaller.

Values found in nature are:

Kd ≈ 10-6M, micro transient (e.g. signal transduction)
Kd ≈ 10-9M, nano stable (e.g. Ab-Ag (antibody antigen), or hemoglobin)
Kd ≈ 10-12M to 10-15M, pico and femto highly stable (e.g enzyme inhibitor)

Affinity is the binding strength or how tightly the ligand is bound to a protein.


The preference to interact with one partner over another.

e.g. fibroblast growth factor and receptors

Electrostatic complementarity contribute

Shape complementarity contribute

Binding types

Between folded domains. Kd ranges from 10-6 to 10-12M

Even after binding, the hydrophobic core is maintained. On the surface are polar residues.

Aromatic residues are typically found in the core.

Distinct patches of hydrophobic and polar residues are used for binding between proteins.

Some binding sites are dry, but other will use water for hydrogen bonds.

To increase affinity, a multiple small peptide segment recognition domains are used in nature.

Non-contiguous regions in the sequence may form the binding site.

Peptide recognition domain

A peptide recognition domain binds to a small sequence of peptides. e.g. SH2 and SH3

Multiple peptide recognition domains can increase affinity and specificity

Binding free energy

Most of the free energy is contributed by about 25% of the interface residues.

Half make no contribution to affinity

But residues with no affinity may be important for specificity.

Upon binding conformation change may occur, and there will be less total solvent accessible surface area (SASA).