SH3 Domain of Tyrosine Kinase

The proto-oncogene cCrk contains an SH3 (Src Homology 3) domain that binds tightly to proline-rich peptides. Like the SH2 domain, the SH3 domain acts as an adapter molecule by recruiting downstream proteins in cellular signaling pathways. SH3 domains are modules of ~60 residues that bind to proline-rich peptides in a sequence-specific manner. The structure shown here contains the short polyproline peptide (Pro-Pro-Pro-Ala-Leu-Pro-Pro-Lys-Lys-Arg) in complex with the SH3 domain from cCrk. The peptide forms a left-handed polyproline helix that binds in an extended conformation across the surface of the SH3 domain. Interestingly, its placement on the surface of the SH3 domain appears very superficial with the prolines interacting with the aromatic residues on the hydrophobic face of the SH3 domain. This superficial contact is in contrast to the deeper binding groove used in the binding of the phospho-tyrosyl peptide on the surface of the SH2 domain.

We constructed 3 protein surfaces. Here is a description of these surfaces:

The input file used is a pdb file (1ckaaH.pdb). For each surface, we remove water from the pdb file and set the solvent radius to zero (i.e. no solvent). The remaining operations differ for each surface.

1. 1ckaaH.nW: Use ProtOr (vanderwaals) radii for the atoms while computing the list of atoms.

2. 1ckaaH.nW.nS: Remove the peptide substrate from the pdb file and use ProtOr (vanderwaals) radii for the atoms.

3. 1ckaaH.nW.subs: Extract the substrate by computing the difference of the above two files.

NOTE: The ProtOr radii is described by Gerstein, Tsai and others. See
Gerstein, et al. 1995, JMB, 249:955-966 and Tsai et. al. 1999, JMB, 290:253:266.

Skin Surfaces


1ckaa with Substrate	1ckaa without Substrate	Substrate
Enlargement \| STL File	Enlagement \| STL File	Enlagement \| STL File

The above pictures show one view of the protein (with and without the substrate) and the substrate by itself. Look at the corresponding alpha complexes too. For reasons of visibility in these images, we scaled the substrate differently from the protein model.

Alpha Shapes


1ckaa with Substrate	1ckaa without Substrate	Substrate
Enlargement	Enlargement	Enlargement

Pockets

We computed the pockets in the protein using the Alpha Shapes again. Many of them were too small to be likely active sites. However, there were two pockets large enough to be good candidates. We show pictures of these pockets here (both alpha complex and skin surface).

Skin Surfaces:


Pocket 1	Pocket 2
Enlargement \| STL File	Enlargement \| STL File

Alpha Shapes:


Pocket 1	Pocket 2

Relationship between Skin Surfaces and Alpha Shapes

Think of the Alpha Shape as the skeleton and the Skin Surface as the outer appearance that surrounds the Alpha Shape. For a pocket, the relationship is the other way around because it is a local complement of both structures. In other words, the pocket in the Alpha Shapes model always contains the corresponding pocket in Skin Surface description.

By the way, the Skin Surface of a pocket is locally the same as the Skin Surface of the complementary protein. Only their respective volumes lie on opposite sides of that surface.