(a) The table presents the cartoons of the closed angle22, Loose and Tight conformations23, showing the localization of the engineered bridges that either crosslink the upper D1 or the lower D2 lobe of the individual subunits (A in green, B in red, C in blue and D in yellow). The entries in the table indicate observed crosslinks for the different trapping mutations that were tested in the presence of KA, FW and glutamate. Full circles indicate trapping at saturating concentrations, whereas, half-filled circles indicate trapping at intermediate concentrations of the respective agonist. (b) Trapping profiles for all the mutants tested (different colours) in different partial agonists. Each plot compares the active fraction versus the time constant of recovery from trapping (τslow). All error bars are s.e.m. (n=5). (c) Each bridge has a different propensity to trap a range of distinct conformations (left panel). The crosslinking promiscuity (right panel), assessed over all agonist concentrations (filled symbols) or only for fully-bound LBD tetramers (open symbols) is plotted in relation to efficacy. A crosslink in saturating ligand was scored 1 unit, and in partially occupied LBD layers, a crosslink was scored as 0.25 units, because of the greater conformational variation in the latter case. (d) Cartoon illustrating the proposed mechanistic relation between conformational ensembles, efficacy and the number of different states trapped for the three classes of agonist.