Computational Analysis Of Calcium Binding Sites In Proteins

As a secondary messenger in cellular signal transduction of eukaryotic cells, calcium is one of the most important metals for the life, and implicated in a variety of biological processes, such as cell division, differentiation and apoptosis. Calcium carries out its functions by binding to specific Calcium receptors or Calcium-binding proteins (CaBPs). Based on the3D structure, they were classified into two main subfamilies:the EF-hand CaBPs and the non-EF-hand CaBPs. The typical example of EF-hand CaBPs is calmodulin, which has been characterized by the presence of structural motifs called ’EF-hands’. Non EF-hand CaBPs do not use this structural motif to bind calcium. Identifying calcium-binding sites in proteins is fundamental for understanding the molecular mechanisms and regulatory roles of calcium. Identification of calcium ion binding sites with conventionally experimental approaches is time-consuming, labor-intensive and expensive. Prediction of calcium binding sites in proteins with computational methods can efficiently narrow down the potential candidates, and provide useful information for further experimental studies. Due to the complexity and diversity of calcium-binding sites, a fast and accurate method for predicting and identifying calcium-binding protein is urgent needed. However, most of current computational analysis are3D structure-based and can not be generally used.Here we developed a novel method to predict calcium-binding sites in proteins. From the scientific literature, we collected369experimentally identified calcium-binding proteins with986sites. The PSSM (Position-specific scoring matrix) algorithm was adopted, while the leave-one-out and n-fold cross-validations were performed to achieve the performance of sensitivity69.67%and specificity90%. Also, the statistical analysis were performed based on the GO (gene ontology) information for known calcium-binding proteins. From the results, we observed that calcium binding proteins preferentially locate in the cytoplasm (GO:0005829), cell membrane (the GO:0005886), and cells outside the region (such as organizations liquid)(the GO:0005576). Moreover, calcium-binding protein are preferentially involved in the physiological process of muscle contraction (GO: 0006936), synaptic signaling (GO:0,007268), and signal transduction (GO:0007165). In addition, the major molecular functions of calcium-binding proteins are calcium ion binding (GO:0005509), calcium-dependent protein binding (GO:0048306), and protein binding (GO:0005515). We downloaded the protein sequences of human, rat, Drosophila, C. elegans, yeast and Arabidopsis from the Uniprot database, and used our method to predict potential calcium binding proteins. In our results, the percentage of calcium-binding proteins is human76.4%, mouse68.6%, fly47.8%, yeast23.2%, nematode14.7%, and Arabidopsis thaliana53.5%. In this regard, calcium binding is ubiquitous in eukaryotes, while the number of calcium-binding proteins increases during evolution.Taken together, we conducted a systematic analysis of the calcium-binding proteins and the binding sites, which can be useful for further experimental consideration. We believe the computational prediction together with experimental validation can propel the research of calcium binding proteins into a new phase.