Inhibition mode of a bisubstrate inhibitor of KDO8P synthase: a frequency-selective REDOR solid-state and solution...

In this report the mode of inhibition of mechanism-based inhibitor (2, K(i) = 0.4 microM) of 3-deoxy-d-manno-2-octulosonate-8-phosphate synthase (KDO8PS), which was designed to mimic the combined key features of its natural substrates arabinose-5-phosphate (A5P) and phoshoenolpyruvate (PEP) into a single molecule, was investigated. Our earlier solid-state NMR observations identified the inhibitor to bind in a way that partly mimics A5P, while the phosphonate moiety of its PEP-mimicking part exhibits no interactions with enzyme residues. This result was apparently in disagreement with the competitive inhibition of 2 against PEP and with the later solved crystal structure of KDO8PS-2 binary complex identifying the interactions of its PEP-mimicking part with the enzyme residues that were not detected by solid-state NMR. To solve this discrepancy, further solid-state REDOR NMR and (31)P solution NMR experiments were applied to a variety of enzyme complexes with the substrates and inhibitor. In particular, a novel frequency-selective REDOR experiment was developed and applied. Integration of the solution and solid-state NMR data clearly demonstrates that under conditions of stoichiometric enzyme-ligand ratio at thermodynamic equilibrium (a) PEP binding is unperturbed by the presence of 2 and (b) both PEP and 2 can bind simultaneously to the synthase, i.e., form a ternary complex with PEP occupying its own subsite and 2 occupying A5P's subsite. The latter observation suggests that under the conditions used in our NMR measurements, the inhibition pattern of 2 against PEP should have a mixed type character. Furthermore, the NMR data directly demonstrate the distinction between the relative binding strength of the two moieties of 2: enzyme interactions with PEP-mimicking moiety are much weaker than those with the A5P moiety. This observation is in agreement with KDO8PS-2 crystal structure showing only remote contacts of the phosphonate due to large structural changes of binding site residues. It is concluded that these phosphonate-enzyme interactions evidenced by both (31)P solution NMR and X-ray are too weak to be preserved under the lyophilization of KDO8PS-2 binary complex and therefore are not evidenced by the solid-state REDOR spectra.