Biochemical and biophysical investigations of the interaction between human glucokinase and pro-apoptotic BAD

The glycolytic enzyme glucokinase (GCK) and the pro-apoptotic protein BAD reportedly reside within a five-membered complex that localizes to the mitochondria of mammalian hepatocytes and pancreatic β-cells. Photochemical crosslinking studies using a synthetic analog of BAD's BH3 domain and in vitro transcription/translation experiments support a direct interaction between BAD and GCK. To investigate the biochemical and biophysical consequences of the BAD:GCK interaction, we developed a method for the production of recombinant human BAD. Consistent with published reports, recombinant BAD displays high affinity for Bcl-xL (KD = 7 nM), and phosphorylation of BAD at S118, within the BH3 domain, abolishes this interaction. Unexpectedly, we do not detect association of recombinant, full-length BAD with recombinant human pancreatic GCK over a range of protein concentrations using various biochemical methods including size-exclusion chromatography, chemical cross-linking, analytical ultracentrifugation, and isothermal titration calorimetry. Furthermore, fluorescence polarization assays and isothermal titration calorimetry detect no direct interaction between GCK and BAD BH3 peptides. Kinetic characterization of GCK in the presence of high concentrations of recombinant BAD show modest (<15%) increases in GCK activity, observable only at glucose concentrations well below the K0.5 value. GCK activity is unaffected by BAD BH3 peptides. These results raise questions as to the mechanism of action of stapled peptide analogs modeled after the BAD BH3 domain, which reportedly enhance the Vmax value of GCK and stimulate insulin release in BAD-deficient islets. Based on our results, we postulate that the BAD:GCK interaction, and any resultant regulatory effect(s) upon GCK activity, requires the participation of additional members of the mitochondrial complex.