Abstract
Author SummarySignificant diversity exists in protein structure and function, yet certain structural domains are used repeatedly across species to execute similar functions. The SH3 domain is one such common structural domain. It is found in signaling proteins and mediates protein–protein interactions by binding to short peptide sequences generally composed of proline. To investigate both the generality and selectivity of peptide binding by SH3 domains, we examined peptide specificity for almost all SH3 domains encoded within the proteome of the budding yeast, Saccharomyces cerevisiae, using a range of experimental methods. We found that although most of the intrinsic binding specificity for SH3 domains can be summarized by the two previously described canonical binding modes, each individual SH3 domain that we studied utilizes unique features of its cognate ligand to achieve binding selectivity. Moreover, some domains exhibit binding specificities that are distinct from the two canonical classes. We integrated peptide-SH3 domain binding data from three complementary screening techniques using a Bayesian statistical model to generate a protein–protein interaction network for the budding yeast SH3 domain family. This network was highly enriched in endocytosis proteins and their interactions. By examining these interactions in detail, we show that our SH3 domain network can be used to predict the temporal localization of several previously uncharacterized proteins to dynamic complexes that orchestrate the process of endocytosis.
Highlights
Families of peptide recognition modules (PRMs), such as PDZ (PSD-95/Discs-large/ZO-1), SH2 (Src homology 2), and SH3 (Src homology 3) domains bind peptide motifs within proteins to mediate protein–protein interactions required for the assembly of stable or transient biological complexes [1]
The SH3 domain is one such common structural domain. It is found in signaling proteins and mediates protein–protein interactions by binding to short peptide sequences generally composed of proline
To investigate both the generality and selectivity of peptide binding by SH3 domains, we examined peptide specificity for almost all SH3 domains encoded within the proteome of the budding yeast, Saccharomyces cerevisiae, using a range of experimental methods
Summary
Cloning and Protein Expression For cloning, the SH3 domain boundaries were defined as the union of the domain regions identified by BLAST [12], PFAM [13], and SMART [14], plus an additional ten amino acids (where applicable) on either side as described previously [3]. All domains were first screened using a random decapeptide library (X10, where X is any amino acid). Domains that failed to select peptides with the decapeptide library were subsequently screened using a biased peptide library (X6-PXXP-X6, where P is proline). Three SH3 domain proteins (Cyk, Lsb, and Sla1-1/2) were tested using a biased library containing a fixed positive charge (X7-R/K-X7, where R and K are arginine or lysine, respectively). Individual binding clones were tested for positive interactions with cognate yeast SH3 domains by phage ELISA as described [44]. The phage library used to select peptides for each domain is indicated in Tables S1 and S2. As some peptide files contain peptides selected from different libraries (Cyk3-class II, Lsb, and Sla1-1/2-class II), the library from which each peptide was isolated is indicated in each sequence file
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
