Protein domains can often be identified from their amino acid sequence, and their function deduced from similar, better characterized, proteins. Hence, when a new signalling molecule is identified, it is now often possible to predict, in general terms, from its sequence what it is likely to bind to, and what type of binding domains the signalling molecule contains. It is important to note that whereas the function of a binding domain may sometimes be predicted by the sequence (SH2 domains always bind phosphorylated tyrosines), protein–protein interactions are highly specific — that is, not all phosphorylated tyrosines are recognized by a particular SH2 domain.
The selectivity of recognition of a motif by a binding domain such as SH2 is conferred by the amino acid sequence adjacent to the phosphorylated residue. We shall illustrate this principle with the SH2 domain of, the tyrosine kinase Src, which has both SH2 and SH3 domains, and a kinase domain (Figure 14a).
The core structural elements of its SH2 domain comprise a central hydrophobic antiparallel β sheet, flanked by two short α helices (Figure 14b), which together form a compact flattened hemisphere with two surface pockets. The SH2 domain binds the phosphotyrosine-containing polypeptide substrate via these surface pockets (Figure 15). One pocket (phosphotyrosine pocket) represents the binding site for phosphotyrosine, whereas the specificity pocket allows interaction with residues that are distinct from the phosphotyrosine, in particular the third residue on the C-terminal side of the phosphotyrosine.
So, for example, the SH2 domain of Src recognizes the sequence pYXXI, where X is a hydrophilic amino acid, I is isoleucine and pY is phosphorylated tyrosine. Note that all proteins that contain this sequence of amino acids are putative binding partners for the SH2 domain of Src, including the C-terminal phosphotyrosine (pY 527) of Src itself.
The SH3 domain has a characteristic fold consisting of five β strands, arranged as two tightly packed antiparallel β sheets (Figure 14c). The surface of the SH3 domain bears a flat, hydrophobic ligand-binding pocket, which consists of three shallow grooves defined by aromatic amino acid residues, which determine specificity. In all cases, the region bound by the SH3 domain is proline-rich, and contains the sequence PXXP as a conserved binding motif (where X in this case is any amino acid).
There are various ways of assaying whether signalling proteins interact with each other through their binding domains such as co-immunoprecipitation, yeast two-hybrid screening, proteomics and FRET. Box 2 describes another technique used to analyse protein–protein interactions that you will use in Experimental investigation 3 at the end of this chapter.