The calcium-dependent interactions of p53 and other intrinsically disordered peptides with members of the S100 protein family.

Abstract

Previous work has demonstrated that peptides sufficiently mimic protein binding sites since they can recapitulate the affinities and kinetics of larger, intact proteins. These finding have dramatically simplified the study of these systems, enabling peptides that match the binding sequence of larger proteins to be used instead of the full-length protein. In addition, the chemical synthesis and modification of synthetic peptides allows one to more easily probe the effect of individual amino acid residues on binding and examine the role of various structural conformations on the binding affinity. However, in order to address the questions related to protein-protein interactions, an appropriate model system needs to be selected. Sufficient sequence divergence is necessary in both the proteins and peptides to adequately address the binding specificity. Similarly, adequate structural information is required to rationally decipher the binding mechanism. Lastly, peptide sequences known to interact with at least one of the proteins under study should be selected to examine the specificity of the peptide for its binding partners.

Despite the large amount of data available for the S100 proteins and the CapZ, p53, NDR, HDM2 and HDM4 peptides, there remains ambiguity regarding the structural, thermodynamic, and kinetic basis of the interactions between them. This work will address the complexity of protein-protein interactions by attempting to answer specific questions regarding this model system. Can we discriminate between the kinetic and thermodynamic modes of binding for different peptides to the S100 proteins? If so, what are the implications? Is the consensus binding sequence for an individual S100 protein applicable to the entire S100 protein family? If not, what is the basis for the binding specificity? Finally, are S100 proteins functionally distinct from closely related EF-hand proteins, such as calmodulin?

In addition to their biological relevance, the S100 proteins are capable of discriminating between similar peptide targets despite having up to 93% sequence identity. Furthermore, even when they interact with the same binding partner, they often exhibit opposing or orthogonal effects. Structures for many members of the human S100 protein family and various orthologs from related species (Bos Taurus, Rattus norveggicus, Oryctolagus cuniculus, Mus musculus, Sus scrofa, and Danio rerio) are available in the apo, holo, and target-bound forms. Additionally, evidence suggests that there may be functional redundancy within the family. Previous studies indicate that several peptides interact with multiple S100 proteins and studies on MRP-14 (S100A9) deficient or calgizzarin-like gene (S100A11) deficient mice showed no obvious phenotypic abnormalities. Lastly, the association or dissociation of a particular peptide to or from the S100 proteins can be manipulated by the addition or chelation of the calcium cofactor.

The peptide binding partners of the S100 proteins were selected based on similar criteria. First, all peptides are derived from proteins with established functions in vivo. For example, the p53 peptide used in these studies is derived from the negative regulatory domain (NRD) of the tumor-suppressor p53 protein. Second, the peptides correspond to a known binding region in the full-length protein (e.g.- residues 367-388 of the p53 protein have been shown to bind several proteins). In addition, these peptides were previously shown to interact with at least one S100 protein, making them ideal candidates to examine the binding specificity of a given sequence to the entire S100 family. Last, these peptides cover a large amount of sequence space and there is structural information available for these peptide sequences in their full-length protein counterparts.

In this regard, we have chosen a model system for this project consisting of sixteen representative members of the human S100 protein family, as well as calmodulin, and peptides derived from CapZ, p53, NRD, HDM2 and HDM4. The S100 protein family is the largest subset of EF-hand proteins and consists of approximately twenty-five dimeric, calcium-binding proteins that are unique to vertebrates. These proteins are able to bind two calcium ions per monomer, which induce conformational changes that are required for most biological interactions. Calmodulin is a ubiquitous, calcium-binding protein that was used as a control for the specificity of peptide targets for the S100 proteins. The S100 proteins play a role in a variety of intracellular and extracellular functions and have been implicated in many types of cancers and neurological disorders. Furthermore, the S100 proteins have proven to be increasingly reliable biological markers for the prognosis of a variety of diseases.

Understanding the molecular recognition that occurs during protein-protein interactions is one of the fundamental goals of biochemistry. These interactions are of particular interest because of their role in signal transduction and regulation. Recently, it has become evident that many of these interactions are the result of large, globular proteins recognizing a short amino acid sequence in, or a structural motif on, their partner proteins. In some cases, the binding epitope has been shown to be as small as a few residues, such as binding of the SH2 or SH3 domains. This discovery was somewhat surprising considering the large quantity of solvent-accessible surface area available for binding on most proteins and has raised a number of key questions. How do small sequences encode binding specificity? How are homologous proteins, with similar sequences and tertiary structures, able to differentiate between binding partners? How do a sequence and its structure influence the thermodynamic and kinetic modes of binding?