Computer modelling approach to study the modes of binding of α- and β -L -arabinose to L -arabinose-binding protein

The modes of binding of &#945;- and &#946; -<SUP>L</SUP> -arabinose to - <SUP>L</SUP>-arabinose-binding protein (ABP) have been studied using a computer modelling method. This method initially fixes six degrees of freedom (three rotational and three translational) that define the best relative positions of the ligand and the receptor, the basis of the selection being minimum steric overlap and good hydrogen bonding scheme. These sterically allowed fits were used as the starting points for subsequent energy minimization. Using this method, it was found that, starting from the low resolution (2.4 &#197;) X-ray data, it was possible to fit both the &#945; - and &#946; -anomers of <SUP>L</SUP> -arabinose in the binding site of ABP. The hydrogen bonding scheme also agreed better with that predicted from high resolution (1.7 &#197;) X-ray data. Thus, this method provides a way of using the low resolution X-ray data not only to fix the ligand unequivocally in the binding site but also to predict the molecular details of protein-ligand interactions and to generate other possible complexes. These studies also suggest that the higher binding energy of the &#946; -anomer of <SUP>L</SUP>-arabinose to ABP is offset by the preponderence of &#945;-anomer in the &#945;&#x21CB;&#946; equilibrium in solution and thus provides a theoretical explanation for the occurrence of both the &#945; - and &#946; -anomers to about the same extent in the complex