The calcium-binding, vertebrate-specific S100 protein family consists of 20 paralogs in humans (referred as the S100ome), with several clinically important members. To assess their interactome, high-throughput, systematic analysis is indispensable, which allows one to get not only qualitative but quantitative insight into their protein-protein interactions (PPIs). We have chosen an unbiased assay, fluorescence polarization (FP) that revealed a partial functional redundancy when the complete S100ome (n=20) was tested against numerous model partners (n=13). Based on their specificity, the S100ome can be grouped into two distinct classes: promiscuous and orphan. In the first group, members bound to several ligands (>4-5) with comparable high affinity, while in the second one, the paralogs bound only one partner weakly, or no ligand was identified (orphan). Our results demonstrate that in vitro FP assays are highly suitable for quantitative ligand binding studies of selected protein families. Moreover, we provide evidence that PPI-based phenotypic characterization can complement the information obtained from the sequence-based phylogenetic analysis of the S100ome, an evolutionary young protein family.
Author summaryFunctional similarity among a protein family can be essential in order to understand proteomic data, to find biomarkers, or in inhibitor design. Proteins with similar functions can compensate the loss-offunction of the others, their expression can co-vary under pathological conditions, and simultaneous targeting can lead to better results in the clinics. To investigate this property one can use sequencebased approaches. However, this path can be difficult. In the case of the vertebrate specific, evolutionary young, S100 family, phylogenetic approaches lead to ambiguous results. To overcome this problem, we applied a high-throughput biochemical approach to experimentally measure the binding affinities of a large number of S100 interactions. We performed unbiased fluorescence polarization assay, involving the complete human S100ome (20 paralogs) and 13 known interaction partners. We used this measured 20x13 (260) protein-protein interaction array to reveal the functional relationships within the family. Our work provide a general framework for studies focusing on phenotype-based domain classification.