The role of intracellular binding proteins in drug distribution and drug targeting to the nucleus

It has traditionally been believed that the function of intracellular lipid binding proteins (iLBPs) is to solubilize and chaperone endogenous ligands in aqueous spaces and to facilitate their transport across the cytosol. In recent years, however, it has become increasingly clear that iLBPs are also able to bind to a diverse range of poorly water soluble compounds and are therefore likely to be involved in a range of additional functions. Notably, iLBPs have been shown to play specific roles in regulating the biological activities and metabolic properties of their ligands. For example, a number of proteins of the fatty acid binding protein (FABP) class of iLBPs have been shown to promote the transport of their ligands to the nucleus and to facilitate the activation of specific nuclear hormone receptors (NHRs). There are nine different subtypes of human FABPs. The current studies have focussed on liver FABP (hLFABP) and intestine FABP (hIFABP), since these iLBPs are highly expressed in organs that play significant roles in lipid processing and are important target organs in metabolic disease. The ability of hLFABP and hIFABP to bind a broad range of compounds including long chain fatty acids (LCFAs), peroxisome proliferator-activated receptors alpha (PPARα) agonists and other drug molecules has been studied using fluorescence spectroscopy. The ligand binding abilities of these FABPs has been further characterized by isothermal titration calorimetry (ITC) and a structural insight into the binding relationships has been probed by resolving the structure of several different protein-ligand complexes. hIFABP ligand binding studies were undertaken using a fluorescence displacement assay and showed that ketorolac is able to displace the fluorescent probe 11-dansylamino undecanoic acid (DAUDA), but not an alternate fluorescent probe, 1-anilinonapthalene 8- sulfonic acid (ANS). The thermodynamics of binding was investigated using ITC, which revealed differences in the mode of binding of these ligands. Thus, ketorolac displayed a unique thermodynamic profile that involved an entropy driven, endothermic binding interaction with hIFABP, in contrast to the more commonly observed exothermic and enthalpy driven interaction for most other ligands (including LCFA, most drugs and the model fluorophores) with hIFABP. The X-ray crystal structure of the DAUDA-hIFABP complex was subsequently determined and revealed a FA-like binding mode where the carboxylate of DAUDA formed a network of hydrogen bonds with residues at the bottom of the binding cavity and the dansyl group interacted with residues in the portal region. For ANS, NMR chemical shift perturbation (CSP) data also indicated binding deep inside the β-barrel, whereas ketorolac appeared to occupy the portal region of hIFABP. The CSP data further suggested that both ANS and ketorolac were able to bind simultaneously to hIFABP, consistent with the lack of displacement observed by fluorescence. The NMR solution structure of the ketorolachIFABP complex was subsequently determined and confirmed a newly characterized, hydrophobic ligand binding site in the portal region of hIFABP. The binding of endogenous ligands to hIFABP was further characterized via a detailed evaluation of oleate binding using ITC and NMR spectrometry. The ITC data showed that oleate binding occurred with low micromolar affinity (KD: 0.8 μM) and via an enthalpy driven process. NMR titration studies subsequently revealed that the complex was in slow exchange on the NMR time scale. The overall binding pose of oleate was consistent with previous data describing the rat IFABP (rIFABP) oleate complex. A similar ion pair interaction between the carboxylate of oleate with Arg106 and Trp82 of hIFABP was consistent with the enthalpy driven binding. Two populations of oleate methyl protons were apparent in the bound state compared to a single population in free oleate. The presence of major and minor populations suggests that there are two distinct binding modes – although the stoichiometry of the interaction was found to be 1:1. This is most consistent with a model in which oleate is bound at a single site in hIFABP but can adopt two different conformations at that site. The fact that distinct resonances were only observed for the methyl protons of oleate and there were no clearly observed peaks reflecting a minor conformation either for other oleate resonances, or for protein resonances in hIFABP, suggests that the two bound conformations are highly similar. Cell based assays were subsequently employed to probe the impact of ligand binding on FABP cellular distribution and in particular the ability of hIFABP and hLFABP to promote nuclear localization and ligand activation of PPARα. Interestingly, while all the PPARα agonists examined bound to hI and hLFABP, the ability of these drugs to stimulate nuclear redistribution and PPAR activation was ligand and FABP specific. Thus, in the presence of GW7647, hLFABP, but not hIFABP, preferentially redistributed to the nucleus, and enhanced ligand activation of PPARα (when compared to the absence of hLFABP). In contrast, nuclear localization of hIFABP (but not hLFABP) and PPAR activation was enhanced in the presence of fenofibrate. Interestingly, nuclear localization of hIFABP and hLFABP did not appear to be mediated via binding to importins, in contrast to previous data obtained with adipocyte FABP (AFABP). The binding of GW7647 to hLFABP was subsequently examined and revealed a nanomolar binding affinity (KD: 115 nM), favorable enthalpy and entropy of binding and a binding stoichiometry of 1:1, instead of the widely reported 1:2 binding stoichiometry for hLFABP binding to LCFA. We hypothesize that ligand binding to FABPs results in a conformational change that stabilizes a direct protein-protein interaction between PPAR and hI and hLFABP and that this is the mechanisms of nuclear localization and PPAR activation. To support this suggestion, the structure of GW7647 bound to hLFABP was solved in an attempt to provide evidence of conformational change on binding. The data suggest that ligand binding does indeed result in conformational change and redistribution of charge on the protein surface. Specifically, GW7647 binding led to opening of the portal cavity due to movement of key internal side chain residues (I52, M74 and R122), a shift in the position of β strands βD and βF and movement in the helix-turn-helix lid of holo hLFABP. Significant conformational change on one side of holo hLFABP was also observed including protrusion of K57, E77, F95 and K96, making them more accessible to surface interaction. This raises the possibility that this is a site of protein-protein interaction with PPARα and will be the subject of future studies. Together these studies advance knowledge of the structure and mechanisms of ligand binding to FABP. They also provide an insight into a potential role for FABP in regulating the intracellular activity of its ligands.