From Protein Structure to Function with Bioinformatics.pdf

304 I.A. Cymerman et al.alignment revealed that only the “PD” half-motif is nearly perfectly conserved,while a critical Lys residue is missing from the second half-motif “(D/E)XK”.Specifically, instead of the Lys residue most members of COG4636 possessed ahydrophobic amino-acid, such as Leu or Val. One possibility was therefore thatthis family was not related at all to PD-(D/E)XK proteins. Another possibility wasthat they are related to these nucleases, but they lost the active site residue andbecame catalytically inactive. A third possibility was that the function of the“missing” Lys residue was be taken over by another residue, but based only on thesequence alignment it was not possible to identify which of the other residuescould fulfil this role. Could structure predictions enable the true function ofCOG4636 to be determined?First, a fold-recognition analysis of COG4636 sequence supported the predictionthat they are indeed related to PD-(D/E)XK enzymes. A comparative modelwas then built based on a structure of a bonafide PD-(D/E)XK nuclease and analyzedfor the presence of spatially adjacent conserved residues. Analysis of themodel revealed that in COG4636 the missing Lys residue had been replaced byanother Lys residue that has appeared in a distinct region in the sequence (Fig.12.1c). The replacement Lys could place its functional group in the same spatialposition as the catalytic Lys residue of the templates thereby allowing the completionof the PD-(D/E)XK motif in three dimensions, despite the lack ofsequence conservation. This allowed for a strong prediction, unavailable frompurely sequence analyses, that COG4636 indeed contained active nucleases.Later on the correctness of the prediction of unusual configuration of the activesite was confirmed by crystallographic analysis of another member of theCOG4636 family (Fig. 12.1d; PDB code 1wdj) as well as by identification ofother bona fide nucleases with the same spatial rearrangement of the active site(Tamulaitiene et al. 2006).12.4.2 Mutation MappingRare mutations in important proteins underlie many genetic diseases. Similarly,allelic variations in drug targets can lead to differential drug binding and henceto different drug responses by patients. Structural mapping of mutations, a keyuse of molecular models, is therefore useful for understanding molecular mechanismsof disease as well as predicting patient responses as a step towards personalisedmedicine.ATP-sensitive potassium (K ATP) channels play key roles in many tissues by linkingcell metabolism to electrical activity. K ATPchannels are octameric complexes oftwo different proteins Kir6.2 and SUR. Binding of ATP or ADP to a K ATPchannelcauses its inhibition. The identification of the number of mutations in Kir6.2 leadingto reduced ATP sensitivity of the channel has turned out to be the cause of permanentneonatal diabetes (Hattersley and Ashcroft 2005). In pancreatic ß-cells the inhibitedK ATPchannel causes membrane hyperpolarization which in turn leads to a reduction