The N-terminal starch binding domain of Rhizopus oryzae glucoamylase (RoSBD) has a high binding affinity for raw starch. RoSBD has two ligand-binding sites, each containing a ligand-binding clamp: a polyN clamp residing near binding site I is unique in that it is expressed in only three members of carbohydrate binding module family 21 (CBM21) members, and a Y32/F58 clamp located at binding site II is conserved in several CBMs. Here we characterized different roles of these sites in the binding of insoluble and soluble starches using an amylose-iodine complex assay, atomic force microscopy, isothermal titration calorimetry, site-directed mutagenesis, and structural bioinformatics. RoSBD induced the release of iodine from the amylose helical cavity and disrupted the helical structure of amylose type III, thereby significantly diminishing the thickness and length of the amylose type III fibrils. A point mutation in the critical ligand-binding residues of sites I and II, however, reduced both the binding affinity and amylose helix disruption. This is the first molecular model for structure disruption of the amylose helix by a non-hydrolytic CBM21 member. RoSBD apparently twists the helical amylose strands apart to expose more ligand surface for further SBD binding. Repeating the process triggers the relaxation and unwinding of amylose helices to generate thinner and shorter amylose fibrils, which are more susceptible to hydrolysis by glucoamylase. This model aids in understanding the natural roles of CBMs in protein-glycan interactions and contributes to potential molecular engineering of CBMs.

Arabidopsis thaliana CBM45 and CBM53 were used to validate the FIA-based prediction model. The predicted ligand-binding residues residing on the surface in the hypothetical structures were verified to be ligand-binding residues. In the absence of 3D structural information, FIA demonstrated significant improvement in the estimation of sequence similarity and identity for a total of 808 sequences from 11 different CBM families as compared with six leading tools by Friedman rank test.

Carbohydrate binding modules (CBMs) are found in polysaccharide-targeting enzymes and increase catalytic efficiency. Because only a relatively small number of CBM structures have been solved, computational modeling represents an alternative approach in conjunction with experimental assessment of CBM functionality and ligand-binding properties. An accurate target-template sequence alignment is the crucial step during homology modeling. However, low sequence identities between target/template sequences can be a major bottleneck. We therefore incorporated the predicted hydrophilic aromatic residues (HARs) and secondary structure elements into our feature-incorporated alignment (FIA) algorithm to increase CBM alignment accuracy. An alignment performance comparison for FIA and six others was made, and the greatest average sequence identities and similarities were achieved by FIA. In addition, structure models were built for 817 representative CBMs. Our models possessed the smallest average surface-potential z scores. Besides, a large true positive value for liagnd-binding aromatic residue prediction was obtained by HAR identification. Finally, the pre-simulated CBM structures have been deposited in the Database of Simulated CBM structures (DS-CBMs). The web service is publicly available at http://dscbm.life.nthu.edu.tw/ and http://dscbm.cs.ntou.edu.tw/.

No abstract