Mammalian respiratory complex I (NADH:ubiquinone oxidoreductase) is an intricate multi-subunit, energy-transducing membrane protein that is essential for aerobic energy metabolism and NADH/NAD⁺ homeostasis. It couples the energy released from NADH oxidation and ubiquinone (Q) reduction to pump four protons across the inner mitochondrial membrane, contributing to the proton motive force used to synthesise ATP. Despite recent advances in structural knowledge and decades of biochemical investigations, the mechanism of redox-coupled proton translocation by complex I is still unknown. In the work presented in this thesis, electron cryomicroscopy (cryo-EM) was used as a primary tool to provide fundamental insight on this mechanism by generating three-dimensional reconstructions of complex I bound to substrates, ligands or inhibitors. Analyses of the structural data and further interrogation by complementary biochemical, biophysical, and computational approaches were used to develop an integrated understanding of substrate and inhibitor binding to complex I.

First, to investigate the mechanism of action of a drug in phase 1 clinical trials against cancers reliant on oxidative phosphorylation (IACS-010759), the structure of mouse complex I inhibited by IACS-2858 – a tighter binding derivative – was resolved to a global resolution of 3.0 Å. The inhibitor, which bears little resemblance to ubiquinone-10 (Q₁₀), occupies the entrance to the Q-binding channel in a ‘cork-in-bottle’ binding mode not previously observed for complex I. Key inhibitor-enzyme interactions were identified, providing a molecular basis for understanding cross-species differences in binding affinities. Modelling of kinetic data showed that IACS-2858 is a simple one-site competitive inhibitor, and the structural motif of a ‘chain’ of aromatic rings was proposed as a characteristic that promotes complex I inhibition.

Next, a strategy for reconstituting bovine complex I into lipid nanodiscs supplemented with exogenous Q₁₀ was devised to probe how the native substrate Q₁₀ binds to the ‘reactive site’ of the Q-binding channel. Five structurally and biochemically distinct conformational classes were identified at global resolutions up to 2.3 Å. These structures fall into three major states: an ‘active’ ready-to-catalyse state, a ‘deactive’ pronounced resting state, and a ‘slack’ state that appears partially disrupted and is of uncertain physiological and biochemical relevance. Comparisons of the deactive structures suggested how substrate/ligand binding restructures the Q-binding site and why both Q and NADH are required for reactivation. Importantly, a Q₁₀ molecule spanning the entirety of the Q-binding site was observed with the Q-headgroup close to its proposed ligating partners NDUFS2-His59 and NDUFS2-Tyr108. Combined with results from molecular dynamics simulations, these structures reveal how the charge states of key active-site residues influence the Q₁₀ binding pose. The bound Q₁₀ species is attributed to a quinone paused in a ‘pre-reactive’ conformation.