From Protein Structure to Function with Bioinformatics.pdf

3 Comparative Protein Structure Modelling 71ing regions in the templates. Thus, no structural information about these inserted segmentscan be extracted from the template structures. These regions frequentlycorrespond to surface loops. Loops often play an important role in defining the functionalspecificity of a given protein framework, forming the functional sites such asantibody complementary determining regions (Rudolph et al. 2006), ligand bindingsites (for ATP (Saraste et al. 1990), calcium (Grabarek 2006), and NAD(P) (Lesk1995), for example), DNA binding sites (Tainer et al. 1995) or enzyme active sites(e.g. Ser-Thr kinases (Johnson et al. 1998) or Asp proteases (Wlodawer et al. 1989) ).The accuracy of loop modelling is a major factor determining the usefulness of comparativemodels in applications such as ligand docking or functional annotation (Fig.3.2). Loop modelling can be seen as a mini protein folding problem because the correctconformation of a given segment of a polypeptide chain has to be calculatedmainly from the sequence of the segment itself. However, loops are generally tooshort to provide sufficient information about their local fold – unless a very substantialpart of the fragments match sequentially and a known conformation – and on theother hand, the environment of each loop is uniquely defined by the solvent and theprotein that cradles it. In a few rare cases it was shown that even identical decapeptidesin different proteins do not always have the same conformation (Fernandez-Fuentes and Fiser 2006; Mezei 1998).There are two main classes of loop modelling methods: (i) the database searchapproaches, where a segment that fits on the anchor core regions is found in a databaseof all known protein structures (Chothia and Lesk 1987; Jones and Thirup1986) and (ii) the conformational search approaches (Bruccoleri and Karplus 1987;Moult and James 1986; Shenkin et al. 1987). There are also methods that combinethese two approaches (de Bakker et al. 2003; Deane and Blundell 2001; vanVlijmen and Karplus 1997).Fragment Based Approach to Loop ModellingThe database or fragment search approach to loop modelling is accurate and efficientwhen a database of specific loops is created to address the modelling of thesame class of loops, such as β-hairpins (Sibanda et al. 1989), or loops on a specificFig. 3.2 Examples of loops (rendered in yellow) that are responsible for functional specificitywithin protein superfamilies. From left to right: Flavodoxin, Immunoglobulin, Neuraminidasefrom, respectively, the α + β barrel, Ig and antiparallel β-barrel protein fold families