The structure of the entire respiratory complex I is now known at reasonably high resolution for many species – bacteria, yeast, and several mammals, including human. The structure reveals an almost 30 angstrom tunnel-like chamber for ubiquinone binding in the core part of the enzyme, at the joint between the membrane and hydrophilic arms of the enzyme. Here we characterize the geometric bottleneck forming the entrance of the quinone reaction chamber. Computer simulations of quinone/quinol passage through the bottleneck suggest that in all structures available, from bacterial to human, this bottleneck is too narrow for the quinone or quinol to pass and that a conformational change is required to open the channel. Moreover, the bottleneck is too narrow even for isoprenoid tail free passage. The closed structure can be an artifact of the crystallization packing forces, low temperature, or other unnatural conditions occurring in the structural data acquisition procedure that affect this flexible part of the enzyme. Two of the helices forming the bottleneck are in direct contact with the subunit (ND3) that was recently demonstrated to be involved in conformational changes during the redox proton pumping cycle, which indicates flexibility of that part of the enzyme. We conclude that the published structures are all locked in the unfunctional states and do not represent correctly the functional enzyme; we discuss possible ways to open the structure in the context of possible mechanisms of the enzyme.