Recent Changes - Search:








edit SideBar


Structure of Protein-Protein Interactions

Learn to use PyMol and examine some typical interactions of regulatory proteins.

The structures used in the exercises need to be downloaded from the Protein Data Bank (PDB) or Macromolecular Structure Database (MSD).

PyMOL should already be present on the teaching computers - try typing 'pymol' at the command-line. For later use, you can download it from here. There are some tutorials and documentation. Some pages that will be particularly useful cover selecting parts of structures (here and here), combining selections, and displaying selections, all on the PyMOL command-line. It's sometimes also useful to be able to label your selections on screen.

Here is one command line tip that anyone who spends time looking at structures needs to know. It is often useful to present the surface of one molecule with the interacting molecule shown in stick format. To draw the surface of just one molecule, it must be defined as a separate object:

  • create obj_a, chain A
  • show surface, obj_a

This will show the surface of chain A. Don't forget the comma in the command!


1) Recognition of a phosphopeptide

P27kip1 bound to Skp1-Skp2-Cks1, PDB 2AST

Look at this protein complex from the perspective of the binding of the P27kip1 phosphopeptide.

  • Display the various protein components to best view the interactions with the phosphopeptide.
  • Determine how many globular domains interact with the peptide. Identify these proteins.
  • Examine the interactions of the phosphorylated residue - compare the hydrophobic and polar interactions. Decide whether or not both are required.
  • Examine the interactions of the buried Glu residue. What is special about the binding cavity?
  • Is there a candidate KEN box motif in Skp1? KEN boxes do not seem to be found in globular domains. What about this one?
  • Think about why it should be necessary to use 3 protein domains to bind a ~10 AA peptide.

[Note: As yet there is no motif in ELM corresponding to the phosphopeptide of p21kip1]

2) Aurora B kinase - examining structural context of hits to linear motif regular expressions

PDBs 2BFY and 2BFX (neither are in complex with an ELM-interactor)

The protein contains hits to some APCC-binding Destruction motifs in ELM

One of these, toward the C-terminus, is described as being functional in a 2005 article while a second article also published in 2005 disagreed. Your challenge is to decide who is right? Run ELM with the sequence to get an overview of motifs and their structural context. (You can get the sequence from the PDB pages.)

Look at the structure in PyMol - and focus on any regions that contain a match to the LM.

Based on structural context, is there any location that is likely to be a functional instance of the motif?

Do you think the motif could be rearranged to become accessible without affecting the integrity of the kinase domain structure?

3) Cooperative phosphorylation in the human Antigen KI-67


This structure shows a triply-phosphorylated peptide interacting with an FHA domain.

Which phosphosites directly interact with the FHA domain? Which is the most important? (Check the reference below.) How do other phospho-sites contribute to affinity/stability?

Byeon IJ et al, Sequential phosphorylation and multisite interactions characterize specific target recognition by the fha domain of ki67, Nat Struct Mol Biol. 2005 Nov;12(11):987-93.

[Note - this type of interaction is hard to capture in ELM]

4) Beta-strand augmentation

In many cases, ELMs interact with globular domains by a process known as "beta-strand augmentation". This involves the linear motif adopting a structure that adds an additional beta-strand to a beta-sheet already existing in the protein. See the following citation for more details.

Remaut H. and Waksman G, Protein-protein interaction through beta-strand addition, Trends Biochem Sci. 2006 Aug;31(8):436-44.

Examine the structure 2G6Q. This shows the PHD finger of mouse Ing2 binding to a methylated histone H3 tail.

Find the backbone interaction between domain and peptide.

Prepare a surface view that reveals the methyl-lysine binding pocket.

Check individual motif residues for their specific interactions. Is this a highly sequence specific interaction?

Peņa PV et al, Molecular mechanism of histone H3K4me3 recognition by plant homeodomain of ING2, Nature. 2006 Jul 6;442(7098):100-3. Epub 2006 May 21.

If you have time, look at another example in PDB 1OBX: A LIG_PDZ_3 instance in IL5RA_HUMAN (chain B) that interacts with a PDZ domain in SDCB1_HUMAN (chain A).

(Beta-strand augmentation also occurs in PDB 2AFF from the previous exercise.)

5) A linear motif which binds as a helix


The Mdm2 protein targets P53 for destruction by binding to a helical motif in the N-terminus. Is the binding site in a groove? (Use a surface display to see this best.) Are the interacting residues hydrophobic or hydrophilic?

Kussie PH et al, Structure of the MDM2 oncoprotein bound to the p53 tumor suppressor transactivation domain, Science. 1996 Nov 8;274(5289):948-53.


In this structure the P53 peptide-binding cavity of Mdm2 is occupied by a small molecule inhibitor. What would be the effect of blocking this interaction? Is this a potential drug target?

Grasberger BL et al, Discovery and cocrystal structure of benzodiazepinedione hdm2 antagonists that activate p53 in cells, J.MED.CHEM. 48, 909.

6) A linear motif with moderate sequence conservation


HAUSP/USP7 binds short motif from P53. There are more backbone interactions than sidechain ones.

Which residue must be Ser? Could it be phosphorylated and still bind? Could Thr substitute? Which residue prefers to be Pro? Which residue(s) can never be Pro?

Sheng Y et al, Molecular recognition of p53 and MDM2 by USP7/HAUSP, Nat.Struct.Mol.Biol. v13 pp.285-291, 2006

7) Domain-domain interactions - the exosome

So far we have looked on the characteristics of interactions focused on a linear motif.

For contrast, look at the structure of the exosome (PDB 2ba0) - a large stable protein complex that is based on a large number of globular-domain globular-domain interactions. The exosome degrades various types of RNA.

Focus on examining individual domain-domain interactions. To do this, try displaying only a pair of adjacent chains (perhaps look at several of these). Roughly how many residues are involved in each interaction? How does this compare to the previous linear-motif/domain interactions? What does this imply about the relative stability of the two types of interactions?

Edit - History - Print - Recent Changes - Search
Page last modified on December 05, 2008, at 09:03 AM CET