Expression, purification, and reconstitution of Fluc channels were as described in detail27 (link),28 (link). In the final purification step, Fluc protein was collected from a S200 size-exclusion column equilibrated in 100 mM NaF (or NaCl for zero-F- preps), 10 mM Hepes pH 7.0, 5 mM n-decylmaltoside (DM). Bpe constructs carried two functionally neutral mutations to enhance expression, R29K/E94S or, for Hg labeling, R29K/E94C. Hg labeling was achieved by incubating Bpe with a 3-fold molar excess of Hg(II) acetate for 30 minutes between the Co-affinity and size exclusion columns. Ec2 constructs bore a single functionally neutral, expression-enhancing mutation, R25K, and, for selenomethionine incorporation, an additional methionine was introduced (A51M) to enhance phasing power. The C-terminal His6 tag was removed from Bpe by treatment with lysine endoproteinase C (Roche)27 (link), but was left on Ec2. Fluc protein was typically reconstituted into liposomes at low density (0.1-0.2 µg protein/mg lipid). For single-channel recording, liposomes were fused into planar lipid bilayers in symmetrical solutions of 300 mM NaF, 15 mM MOPS, pH 7.0, and channels were recorded at 200 mV holding voltage27 (link),29 . Monobodies were expressed in E. coli and purified as described28 (link). N-terminal His6 tags were removed while bound to Talon beads by 16-hr treatment with TEV protease also carrying a His tag; monobodies with cleaved His tags were eluted from the affinity column with 150 mM NaCl, 40 mM Tris-HCl pH 7.5. For the final purification step, the preparation was passed over a S75 size exclusion column in 100 mM NaF (or NaCl), 10 mM HEPES pH 7. Monobodies were used immediately for crystallization or stored in frozen aliquots for channel-blocking experiments. For crystallization from detergent micelles, Fluc protein in solution containing 5 mM DM was concentrated to 10 mg/mL, a maneuver that concentrates the detergent 5-10-fold. Monobody solution (10 mg/mL) was supplemented with 4 mM DM immediately before mixing with channel in a 1.2:1 molar ratio. This protein solution was then mixed with an equal volume of crystallization solutions (0.5 µL for sitting drops in 96-well plates or 1 µL for hanging drops in 24-well plates). Bpe-S7 crystals grew in 3-5 days in crystallization solutions of 36-41% (w/v) polyethylene glycol 550 MME, 0.2 M MgCl2 or CaCl2, 0.1 M Tris, pH 8.5-8.9. Ec2-S9 crystals grew in 10-14 days in crystallization solutions containing 28-32% (w/v) polyethylene glycol 550 MME, 0.05 M LiNO3, 0.1 M N-(2-Acetamido)iminodiacetic acid, pH 6.0-6.7. Crystals were frozen in liquid nitrogen for data collection. For lipidic cubic phase crystallization, Fluc protein concentrated to 10 mg/mL as above was dialyzed overnight to reduce the DM concentration to 10 mM. This was then mixed with monobody solution (10 mg/mL, with 4 mM DM) in a 1:1.2 molar ratio. The protein-laden mesophase was prepared by homogenizing 9.9 monoacylglycerol (monoolein) lipid with protein solution (10 mg/ml) at a weight ratio of 1:1.5 (protein:lipid) using a coupled syringe mixing device at 20°C30 (link). Crystallization trials were carried out at 19°C in 96-well glass sandwich plates with 50 nl mesophase and 0.8 µl precipitant solution using an in meso robot. Crystallization solutions consisted of 22–26% (v/v) polyethylene glycol 500 DME, 0.1 M Na-citrate pH 5.5 +/- 10 mM NaF. Surfboard shaped crystals grew to a maximum size of 100 × 50 × 5 µm in 5-10 days. Wells were opened using a tungsten–carbide glasscutter, and the crystals were harvested using 50-100 µm micromounts (MiTeGen). Crystals were snap-cooled directly in liquid nitrogen prior to data collection on the Diamond Light Source beamlines I24 or I04.