The importance of domain-domain interactions in the regulation and activity of neuronal nitric oxide synthase

Abstract

Nitric oxide synthases (NOSs) catalyse the production of the physiological messenger molecule NO from L-arginine in a unique two step oxygenation reaction. Constitutive forms of NOS are activated by the binding of calmodulin (CaM) to a 20- amino-acid inter-domain region in the presence of calcium ions, causing inter-domain electron transfer to occur from FAD to heme via FMN. This electron transfer step involves a large scale movement of the FMN-binding domain and is influenced by a number of structural features unique to NOS which include an autoinhibitory loop, a C-terminal extension, a number of phosphorylation sites, and the hinge region that connects the FAD- and FMN- binding domains. X-ray crystallographic data indicate that all of these regulatory elements lie at the interface between the two domains, restricting their motion relative to each other. The importance of this interface region in the CaM-dependent activation mechanism of neuronal NOS (nNOS) was investigated by site-directed mutagenesis of interface residues in the isolated reductase domain (nNOSrd). A range of kinetic and thermodynamic analytical techniques were employed in order to determine which catalytic steps are affected by changes in domain mobility. The rate-limiting step in the turnover of nNOSrd has been speculated to be one of three catalytic events; the hydride transfer from NADPH to FAD, the electron transfer between the bound flavin cofactors, or the electron transfer between FMN and external electron acceptors such as cytochrome c. In each case, the binding of CaM enhances the rate of reduction. In wild-type nNOSrd, NADPH reduced the FAD by hydride transfer in what should have been a simple 1-step reaction but was in fact a biphasic process. Isotope effects and the use of differing ionic strength indicated that different conformations of enzyme have different rates of reaction with NADPH. The redox state of the FMN cofactor also influenced the rate of reduction of FAD, through the interaction with the peptide backbone in the interface region. A putative proton transfer pathway exists between the bound FAD cofactor and solvent, involving Ser1176, Asp1393, His1032 and Arg1229. Mutation of the former three residues diminished the catalytic activity, specifically focused on the rate of hydride transfer, while the R1229E mutation had a much more dramatic effect. Arg1229 forms one of only two electrostatic contacts between the FAD and FMNbinding domains in the interface region and the charge reversal substitution introduced a likely inter-domain repulsion. This was expected to cause the two domains to separate, favouring a hinged-open conformation. The hydride transfer step from NADPH to FAD was activated in the CaM-free enzyme, while FAD to FMN electron transfer was inhibited. Electron transfer from reduced FMN to the artificial electron acceptor horse-heart cytochrome c was also activated in the CaMfree state. The effect on the three catalytic events meant that during steady-state turnover with cytochrome c, CaM deactivated the enzyme and caused cytochrome cdependent inhibition. Evidently, domain-domain separation was large enough in the mutant to accommodate cytochrome c, a 12 kDa protein, in the space between the cofactors at the interface. The effects of this single charge-reversal on the three distinct catalytic events illustrated how each is differently dependent on the enzyme conformation. FAD to FMN electron transfer was shown to occur exclusively in the hinged-closed form, consistent with the crystal structure of nNOSrd. The remaining two events, hydride transfer from NADPH to FAD, and electron transfer from FMN to cytochrome c, occur in the hinged-open state. In the wild-type enzyme, the hinged-open and hinged-closed states are tightly regulated by a conformational equilibrium which is affected by CaM binding. It appears that CaM activates the enzyme by shifting this equilibrium to an open form.