Final Report Summary - ORDER IN DISORDER (Order in Disorder: Context-dependent strategies for integrating peptide-mediated interactions)
Many protein interactions are mediated by short motifs (peptides). These interactions are usually transient and weak, and considerably modulated by their context. The aim of this project was to elucidate how individual, mostly weak, multiple interactions are integrated into a signal – what are the strategies used by nature for decision making processes such as proliferation vs growth arrest at the level of individual interactions?
To address this aim, we used diverse directions, in particular: (1) We developed novel approaches, and significantly extended and improved our existing ones for the structural modeling of peptide-protein interactions – to provide a structural basis for the study of these interactions, including modeling of binding and design of peptidic modulators, including their flanking regions; (2) We established a lab to provide a framework for the experimental characterization and manipulation of peptide-mediated interactions, as well as assessment of phenotypic effects of domain repeats; (3) We compiled an annotated and periodically updated database of peptide-mediated interactions, to provide a source of analysis of prevalent use of motifs and their binding domains, and used this database for global analyses, and to select specific systems.
Our main achievements are:
(1) A breakthrough in peptide-protein modeling: Analysis of fragments extracted from solved structures (based on sequence and predicted structure similarity) suggested that the free peptide samples bound-like conformations. This observation allowed for the decoupling of folding and binding, and permitted the establishment of the first efficient and accurate global blind peptide docking protocol (PiperFlexPepDock). Improvements and extensions of our Rosetta FlexPepDock suite for the accurate prediction of substrates for peptide binding and modifying proteins (FlexPepBind), and their experimental validation, revealed new network motifs regulated via post-translational modification (e.g. histone deacetylase modifies not only histones, but also upstream regulatory proteins (such as histone acetylase).
(2) We chose the WW domains and their proline rich motif containing partners as an example system for the in depth study of integration strategies of multiple peptide-mediated interactions among partners. Analysis of our annotated database of peptide binding domains, and corresponding peptide motifs in partners, revealed that many WW domains occur in tandem. And show different mutual effects on binding, including cooperativity in affinity, as well as change in specificity. We then focused on the tandem WW domains of the WWOX protein for in depth experimental and computational characterization. Experimental measures of protein stability of single and tandem domains, as well as binding affinity to peptide substrates, (including Urea denaturation, CD, NMR, ITC, and more) established a model that can explain the exceptional mode of cooperativity in this protein.
(3) Can phenotypic changes be achieved by domain repeat number variation? Certain proteins vary in the number of repeats of domain-coding sequences, which suggested to us that such variation in context could result in functional variation. We studied PIR2 proteins (which contain a variable number of repeats in different yeast strains), to detect phenotypic changes, e.g. growth rate, that would correlate with number of domain repeats. These experiments were not conclusive as to the phenotype change, but they revealed that the number of repeats as determined by us using PCR differed dramatically from the number of repeats in the reported genomes. This motivated us to devise a new approach for the more accurate definition of repeat numbers from NGS data.
To address this aim, we used diverse directions, in particular: (1) We developed novel approaches, and significantly extended and improved our existing ones for the structural modeling of peptide-protein interactions – to provide a structural basis for the study of these interactions, including modeling of binding and design of peptidic modulators, including their flanking regions; (2) We established a lab to provide a framework for the experimental characterization and manipulation of peptide-mediated interactions, as well as assessment of phenotypic effects of domain repeats; (3) We compiled an annotated and periodically updated database of peptide-mediated interactions, to provide a source of analysis of prevalent use of motifs and their binding domains, and used this database for global analyses, and to select specific systems.
Our main achievements are:
(1) A breakthrough in peptide-protein modeling: Analysis of fragments extracted from solved structures (based on sequence and predicted structure similarity) suggested that the free peptide samples bound-like conformations. This observation allowed for the decoupling of folding and binding, and permitted the establishment of the first efficient and accurate global blind peptide docking protocol (PiperFlexPepDock). Improvements and extensions of our Rosetta FlexPepDock suite for the accurate prediction of substrates for peptide binding and modifying proteins (FlexPepBind), and their experimental validation, revealed new network motifs regulated via post-translational modification (e.g. histone deacetylase modifies not only histones, but also upstream regulatory proteins (such as histone acetylase).
(2) We chose the WW domains and their proline rich motif containing partners as an example system for the in depth study of integration strategies of multiple peptide-mediated interactions among partners. Analysis of our annotated database of peptide binding domains, and corresponding peptide motifs in partners, revealed that many WW domains occur in tandem. And show different mutual effects on binding, including cooperativity in affinity, as well as change in specificity. We then focused on the tandem WW domains of the WWOX protein for in depth experimental and computational characterization. Experimental measures of protein stability of single and tandem domains, as well as binding affinity to peptide substrates, (including Urea denaturation, CD, NMR, ITC, and more) established a model that can explain the exceptional mode of cooperativity in this protein.
(3) Can phenotypic changes be achieved by domain repeat number variation? Certain proteins vary in the number of repeats of domain-coding sequences, which suggested to us that such variation in context could result in functional variation. We studied PIR2 proteins (which contain a variable number of repeats in different yeast strains), to detect phenotypic changes, e.g. growth rate, that would correlate with number of domain repeats. These experiments were not conclusive as to the phenotype change, but they revealed that the number of repeats as determined by us using PCR differed dramatically from the number of repeats in the reported genomes. This motivated us to devise a new approach for the more accurate definition of repeat numbers from NGS data.