Testing of the portal hypothesis: analysis of a V32G, F57G, K58G mutant of the fatty acid binding protein of the murine adipocyte

Abstract

The portal region of fatty acid binding proteins is hypothesized to function as a dynamic aperture, controlling accessibility of external ligands to the internal fatty acid binding cavity. To test this hypothesis, a triple mutant of the murine FABP4 has been developed (V32G, F57G, K58G, referred to as the portal mutant) that is predicted to constitutively enlarge the opening due to a reduction in the molecular dimensions of the side chains of key portal amino acids. The portal mutant was purified from expressing Escherichia coli, its stability was evaluated, and the thermodynamics and kinetics of ligand binding were compared to that of wild-type protein. Introduction of the three amino acid substitutions caused no significant change in the stability of the protein with a free energy of unfolding of 13.7 kJ/mol as compared to 14.0 kJ/mol for the wild-type protein. The portal mutant exhibited a modest decrease (4-fold) in ligand binding affinity using the fluorescent probe 1-anilinonaphthalene-8-sulfonic acid (1,8-ANS) as a surrogate ligand. 1,8-ANS displacement assays revealed that the binding affinity for oleate increased from a K0.5 of 196 +/- 15 nM for the wild-type protein to 165 +/- 8 for the portal mutant, while that for arachidonate decreased from the wild type of 186 +/- 11 nM to 418 +/- 26 nM for the portal mutant. To evaluate cavity accessibility, rate of 1,8-ANS binding was assessed between the portal and wild-type protein. Using equimolar amounts of ligand and protein at 4 degrees, 1,8-ANS bound within the cavity to 95% saturation (t0.95) in 750 ms, while the mutant protein was fully modified in less than 1.4 ms. To independently evaluate cavity accessibility, modification of the sole protein cysteine residue, C117 residing within the cavity near C2-C4 of the bound ligand, was monitored using 5,5'-dithio-bis(2-nitrobenzoic acid) modification. The half time for modification (t0.5) for the wild-type protein was approximately 20 s, while that for V32G F57G K58G occurred in less than a second. As such, enlargement of the portal region of FABP4 markedly increased the accessibility of ligands to the cavity while having only modest effects on ligand affinity. Taken together, these data provide support for the portal region hypothesis and suggest dynamic fluctuations in this region regulate cavity access, but not ligand affinity or selectivity.