A mechanistic study of the poorly understood pathway by which the inhibitor acarbose is enzymatically rearranged by human pancreatic alpha-amylase has been conducted by structurally examining the binding modes of the related inhibitors isoacarbose and acarviosine-glucose, and by novel kinetic measurements of all three inhibitors under conditions that demonstrate this rearrangement process. Unlike acarbose, isoacarbose has a unique terminal alpha-(1-6) linkage to glucose and is found to be resistant to enzymatic rearrangement. This terminal glucose unit is found to bind in the +3 subsite and for the first time reveals the interactions that occur in this part of the active site cleft with certainty. These results also suggest that the +3 binding subsite may be sufficiently flexible to bind the alpha-(1-6) branch points in polysaccharide substrates, and therefore may play a role in allowing efficient cleavage in the direct vicinity of such junctures. Also found to be resistant to enzymatic rearrangement was acarviosine-glucose, which has one fewer glucose unit than acarbose. Collectively, structural studies of all three inhibitors and the specific cleavage pattern of HPA make it possible to outline the simplest sequence of enzymatic reactions likely involved upon acarbose binding. Prominent features incorporated into the starting structure of acarbose to facilitate the synthesis of the final tightly bound pseudo-pentasaccharide product are the restricted availability of hydrolyzable bonds and the placement of the transition state-like acarviosine group. Additional "in situ" experiments designed to elongate and thereby optimize isoacarbose and acarviosine-glucose inhibition using the activated substrate alphaG3F demonstrate the feasibility of this approach and that the principles outlined for acarbose rearrangement can be used to predict the final products that were obtained.

A mechanistic study of the essential allosteric activation of human pancreatic alpha-amylase by chloride ion has been conducted by exploring a wide range of anion substitutions through kinetic and structural experiments. Surprisingly, kinetic studies indicate that the majority of these alternative anions can induce some level of enzymatic activity despite very different atomic geometries, sizes, and polyatomic natures. These data and subsequent structural studies attest to the remarkable plasticity of the chloride binding site, even though earlier structural studies of wild-type human pancreatic alpha-amylase suggested this site would likely be restricted to chloride binding. Notably, no apparent relationship is observed between anion binding affinity and relative activity, emphasizing the complexity of the relationship between chloride binding parameters and the activation mechanism that facilitates catalysis. Of the anions studied, particularly intriguing in terms of observed trends in substrate kinetics and their novel atomic compositions were the nitrite, nitrate, and azide anions, the latter of which was found to enhance the relative activity of human pancreatic alpha-amylase by nearly 5-fold. Structural studies have provided considerable insight into the nature of the interactions formed in the chloride binding site by the nitrite and nitrate anions. To probe the role such interactions play in allosteric activation, further structural analyses were conducted in the presence of acarbose, which served as a sensitive reporter molecule of the catalytic ability of these modified enzymes to carry out its expected rearrangement by human pancreatic alpha-amylase. These studies show that the largest anion of this group, nitrate, can comfortably fit in the chloride binding pocket, making all the necessary hydrogen bonds. Further, this anion has nearly the same ability to activate human pancreatic alpha-amylase and leads to the production of the same acarbose product. In contrast, while nitrite considerably boosts the relative activity of human pancreatic alpha-amylase, its presence leads to changes in the electrostatic environment and active site conformations that substantially modify catalytic parameters and produce a novel acarbose rearrangement product. In particular, nitrite-substituted human pancreatic alpha-amylase demonstrates the unique ability to cleave acarbose into its acarviosine and maltose parts and carry out a previously unseen product elongation. In a completely unexpected turn of events, structural studies show that in azide-bound human pancreatic alpha-amylase, the normally resident chloride ion is retained in its binding site and an azide anion is found bound in an embedded side pocket in the substrate binding cleft. These results clearly indicate that azide enzymatic activation occurs via a mechanism distinct from that of the nitrite and nitrate anions.

A new approach for the discovery and subsequent structural elucidation of oligosaccharide-based inhibitors of ␣-amylases based upon autoglucosylation of known ␣-glucosidase inhibitors is presented. This concept, highly analogous to what is hypothesized to occur with acarbose, is demonstrated with the known ␣-glucosidase inhibitor, D-gluconohydroximino-1,5-lactam. This was transformed from an inhibitor of human pancreatic ␣-amylase with a K i value of 18 mM to a trisaccharide analogue with a K i value of 25 M. The three-dimensional structure of this complex was determined by x-ray crystallography and represents the first such structure determined with this class of inhibitors in any ␣-glycosidase. This approach to the discovery and structural analysis of amylase inhibitors should be generally applicable to other endoglucosidases and readily adaptable to a high throughput format.

No abstract