Identification of metal ion binding sites based on amino acid sequences

Abstract

The identification of metal ion binding sites is important for protein function annotation and the design of new drug molecules. This study presents an effective method of analyzing and identifying the binding residues of metal ions based solely on sequence information. Ten metal ions were extracted from the BioLip database: Zn2+, Cu2+, Fe2+, Fe3+, Ca2+, Mg2+, Mn2+, Na+, K+ and Co2+. The analysis showed that Zn2+, Cu2+, Fe2+, Fe3+, and Co2+ were sensitive to the conservation of amino acids at binding sites, and promising results can be achieved using the Position Weight Scoring Matrix algorithm, with an accuracy of over 79.9% and a Matthews correlation coefficient of over 0.6. The binding sites of other metals can also be accurately identified using the Support Vector Machine algorithm with multifeature parameters as input. In addition, we found that Ca2+ was insensitive to hydrophobicity and hydrophilicity information and Mn2+ was insensitive to polarization charge information. An online server was constructed based on the framework of the proposed method and is freely available at http://60.31.198.140:8081/metal/HomePage/HomePage.html.

The Identification of Metal Ion Ligand-Binding Residues by Adding the Reclassified Relative Solvent Accessibility

“…We therefore selected the predicted secondary structure information and RSA information. The prediction was helpful when simplifying the problem from the 3D structure to sequence information (Chen and Zhou, 2005;Lin et al, 2005;Hu et al, 2016a,b;Cao et al, 2017;Wang et al, 2019). In this paper, they were predicted by using ANGLOR software (Wu and Zhang, 2008).…”

“…We selected non-redundant datasets of metal ion-binding proteins that were constructed in our group (Cao et al, 2017;Wang et al, 2019). The benchmark datasets were entirely from the BioLip database (Yang et al, 2013a).…”

“…Thus, depending on the interaction between the metal ion ligands and specific binding residues, many metal ion ligands can affect the special protein functions (Caspers et al, 1990;Supek et al, 1997;Selvarengan and Kolandaivel, 2005). For instance, Mn 2+ is used as catalyst in photosynthesis (Degtyarenko, 2000;Reed and Poyner, 2000), Ca 2+ can lead to anxiety and Alzheimer's disease (Jiang et al, 2015;Cao et al, 2017), and Cu 2+ can cause Coronary Heart Disease (Sodhi et al, 2004;Lin et al, 2005). The basic principle of molecular drug design is that the interaction between the receptor and ligand must conform to the "Lock and Key Model, " and the interaction between the protein and ion ligands we studied also conforms to the "Lock and Key Model."…”

“…First, the databases were generally acquired from Protein Data Bank (PDB) (Tainer et al, 1991;Bernstein et al, 1997;Sodhi et al, 2004;Lin et al, 2005;Bordner, 2008;Babor et al, 2010;Lu et al, 2012), Structural Classification of Protein (SCOP) (Hubbard et al, 1997;Sodhi et al, 2004;Chauhan et al, 2010;Sobolev and Edelman, 2013), Ligand Protein Contact (LPC) (Sobolev et al, 1999;Chauhan et al, 2010), and BioLip (Yang et al, 2013a,b;Hu et al, 2016a,b, Wang et al, 2019. Second, the feature parameters generally contained the composition information of the amino acid (Cao et al, 2017;Wang et al, 2019), hydrophilicity-hydrophobicity (Lin et al, 2005;Lin et al, 2006;Cao et al, 2017), charge (Lin et al, 2005;Cao et al, 2017;Wang et al, 2019), position specific score matrix (PSSM) (Hu et al, 2016a), relative solvent accessibility (RSA) (Lin et al, 2006;Hu et al, 2016a;Cao et al, 2017;Wang et al, 2019) and three-dimensional structure information (Babor et al, 2010;Roy et al, 2012;Yang et al, 2015;Hu et al, 2016a). Finally, the classification algorithms used were artificial neural network (ANN) (Lin et al, 2005), Support Vector Machine (SVM) (Lin et al, 2006;Jiang et al, 2015;Cao et al, 2017;…”

“…Second, the feature parameters generally contained the composition information of the amino acid (Cao et al, 2017;Wang et al, 2019), hydrophilicity-hydrophobicity (Lin et al, 2005;Lin et al, 2006;Cao et al, 2017), charge (Lin et al, 2005;Cao et al, 2017;Wang et al, 2019), position specific score matrix (PSSM) (Hu et al, 2016a), relative solvent accessibility (RSA) (Lin et al, 2006;Hu et al, 2016a;Cao et al, 2017;Wang et al, 2019) and three-dimensional structure information (Babor et al, 2010;Roy et al, 2012;Yang et al, 2015;Hu et al, 2016a). Finally, the classification algorithms used were artificial neural network (ANN) (Lin et al, 2005), Support Vector Machine (SVM) (Lin et al, 2006;Jiang et al, 2015;Cao et al, 2017;Hu et al, 2016a), Naïve Bayes (Ebert and Altman, 2010), COFACTOR (Lin et al, 2006;Yang et al, 2015), TargetSeq, TargetCom (Hu et al, 2016b), COACH (Yang et al, 2015), and SMO (Wang et al, 2019). Among the three aspects in the prediction mentioned above, the key step of feature extraction was generated by one of two ways: (1) the three-dimensional structure information or (2) primary sequence information of the protein.…”