Prediction of protein‐glucose binding sites using support vector machines

Abstract

Glucose is a simple sugar that plays an essential role in many basic metabolic and signaling pathways. Many proteins have binding sites that are highly specific to glucose. The exponential increase of genomic data has revealed the identity of many proteins that seem to be central to biological processes, but whose exact functions are unknown. Many of these proteins seem to be associated with disease processes. Being able to predict glucose-specific binding sites in these proteins will greatly enhance our ability to annotate protein function and may significantly contribute to drug design. We hereby present the first glucose-binding site classifier algorithm. We consider the sugar-binding pocket as a spherical spatio-chemical environment and represent it as a vector of geometric and chemical features. We then perform Random Forests feature selection to identify key features and analyze them using support vector machines classification. Our work shows that glucose binding sites can be modeled effectively using a limited number of basic chemical and residue features. Using a leave-one-out cross-validation method, our classifier achieves a 8.11% error, a 89.66% sensitivity and a 93.33% specificity over our dataset. From a biochemical perspective, our results support the relevance of ordered water molecules and ions in determining glucose specificity. They also reveal the importance of carboxylate residues in glucose binding and the high concentration of negatively charged atoms in direct contact with the bound glucose molecule.

Identification of a Receptor Subunit and Putative Ligand-Binding Residues Involved in the Bacillus megaterium QM B1551 Spore Germination Response to Glucose

“…The decreases in the apparent affinity of the receptor when Glu196, Tyr191, and Phe192 are individually replaced are of magnitudes similar to those reported for various prokaryotic amino acid/sugar transporters where ligand-binding residues are replaced (11,26). Similarly, the chemical nature of the candidate ligand-binding residues, i.e., charged and/or aromatic amino acids, is consistent with the nature of residues observed to be directly involved in binding to ligands in various single-component transporters (10,21,34), although we acknowledge that any comparisons between spore germinant-receptor proteins and bacterial transporters have to be made with caution. Finally, amino acids at the equivalent positions in other germinant-receptor proteins are highly conserved, including the B. subtilis GerKB protein, which appears to be cognate for glucose (2,15) (Fig.…”

“…Instead, a survey of the literature pertaining to essential ligand-binding residues in a variety of integral membrane proteins involved in the transport of sugars, amino acids, and ions was conducted and revealed the recurring importance of charged and aromatic residues located on ␣ helices that span the membrane (TM helices) (8,9,21,25,27,34,37). Transmembrane-located proline residues, which serve to introduce distortions and/or regions of local flexibility in the alpha helix, are also often of functional importance, particularly where conformational changes at proline-containing molecular hinges are employed to transduce extracellular signals across the membrane (7,30,31).…”

SVM Prediction of Ligand-binding Sites in Bacterial Lipoproteins Employing Shape and Physio-chemical Descriptors

“…SVM also performs very efficiently for protein structural class prediction [77][78][79] as well as for prediction of binding sites in proteins [56].…”

“…The charge per atom can be positive, neutral, or negative; atoms are either able to form hydrogen bonds (hydrogen bonding) or not (nonhydrogen bonding); hydrophobicity of an atom is measured by considering it as hydrophobic, hydroneutral, or hydrophilic. Based on these values, the capability of an atom or molecule to bind to a particular ligand can be determined successfully [56]. There are several atoms or molecules present in the protein cavity, such as water, sulfate, phosphate, calcium, magnesium, and zinc, which are not part of the protein structure, but they may have an effect on ligand specificity [57].…”

Structural motif screening reveals a novel, conserved carbohydrate-binding surface in the pathogenesis-related protein PR-5d

“…Carbohydrate ligands are diverse in size, geometry and other physicochemical characteristics [2], and this diversity is mirrored in the features of carbohydrate-binding sites in proteins. A few recent studies have developed more targeted approaches that apply structure-based methods to specific classes of CBPs [10,11]. At a cost of lower generality, approaches that focus on structural motifs of particular functional classes of CBPs may achieve predictions with better ligand specificities and greater overall accuracies.…”