Protein modelling using molecular integral equation theory: applications to chymosin—κ-casein complexes

We discuss methods for modelling biomolecular complexes based on the Integral Equation Theory (IET) of Molecular Liquids. We begin by outlining recent advances in IET that have made it possible to use the theory in calculating solvation free energies, predicting small molecule binding sites on biomacromolecules, and computing absolute and relative host-guest binding affinities [1,2]. We use these IET methods (combined with standard molecular simulation tools) to study two homologous mammalian aspartic proteases (calf and camel chymosin) complexed with their native peptide ligands (cow and camel k-casein) [3,4,5]. The complexes are of industrial interest because camel chymosin has recently been marketed as an alternative to 
bovine chymosin as an enzyme to
 clot milk in the food industry. 
The camel enzyme has been shown
 to have 70% higher clotting activity
 and only 20% of the unspecific protease activity for bovine k-casein as 
compared to the bovine enzyme. Interestingly, bovine chymosin has
 a very low proteolytic rate for camel
 k-casein. The models provide putative atomic coordinates for the
 complexes, for which there are no 
available crystallographic or NMR
 structures, and suggest new avenues for experimental work. The IET methods are easily implemented using existing computational software and are shown to provide a useful complement to the standard molecular simulation toolbox.