Phosphoinositide binding

Apparent K_d interaction data of phosphoinositides was determined on human apo-SF-1 LBD by a previously described native gel shift assay (21) to monitor equilibrium binding to each phosphoinositide. Briefly, apo-SF-1 ligand binding domain was generated using an on-column washout procedure to remove bound phospholipid, wherein pure SF-1 bound to dipalmitoyl PIP3 was loaded onto a Mono Q column and washed with 1.0 L of 20 mM Hepes (8.0), 1 mM EDTA with 2 mM CHAPS at 2 ml/min to remove all bound phospholipid from SF-1 by on-column dilution. The buffer was then changed into 20 mM Hepes (8.0), 1 mM EDTA to remove detergents, and apo-SF-1 eluted using a ammonium acetate gradient, pooled, protein concentration quantitated, and then used in the phosphoinositide binding assay. For the phosphoinositide binding assay, each phosphoinositide was resuspended in water, sonicated, diluted, and stored under nitrogen at 4°C. Phosphoinositides were always re-sonicated immediately before use in binding assays. Two microliters of serially diluted phosphoinositide were added to binding reactions of 25 μl total volume, containing 1 μM final concentration of apo-SF-1 LBD in 20 mM Hepes (8.0), 1 mM EDTA, and 10 mM ammonium acetate in 200 μl polypropylene PCR strip tubes. Binding reactions were incubated at 37°C for 1 h in a PCR thermocycler with a heated reaction cover, the reaction mixed with 3 μl native loading buffer (40% glycerol, 0.005% Ponceau S) and run on a 4% to 16% polyacrylamide Bis-Tris NativePage™ gel (Invitrogen). After fixation and silver staining (BioRad), gels were scanned, the phosphoinositide-shifted band quantitated using NIH image, and apparent K_d determined by nonlinear curve fit to a single-site binding model in GraphPad Prism. Data represent at least three independent replicates.