Native protein mass spectrometry of nuclear receptor-ligand and enzyme-substrate complexes

Abstract

Nuclear receptors (NRs) are transcription factors that are activated by the binding of low molecular weight ligands. The NR superfamily includes receptors for steroid hormones, anti-inflammatory and lipid-lowering drugs as well as endogenous compounds such as bile acids and fatty acids.

Several receptors in this family lack identified ligands and are thus termed orphan receptors. Identification of orphan receptor ligands is an area of intense research in the pharmacological industry. However, many methods for ligand identification are dependent on competition with known ligands or the recruitment of co-activator proteins. Mass spectrometry has developed in the last decade to allow the analysis of non-covalent complexes. The work in this thesis focuses on the use of mass spectrometry as a method to detect binding of low molecular weight compounds to proteins such as nuclear receptors and enzymes.

The retinoid X receptor alpha (RXRalpha, NR2B1) is a receptor for 9-cis-retinoic acid and has important functions during embryonic development. It was recently shown that a polyunsaturated fatty acid, cis-4, 7, 10, 13, 16, 19docosahexaenoic acid (DHA) is a low-efficacy agonist for RXR. Here it is shown that several other C18-C22 unsaturated fatty acids including arachidonic acid are also RXR agonists. Further, in vitro binding between high and low affinity agonists and the RXRalpha ligand binding domain (LBD) can be observed using mass spectrometry. The binding observed by electrospray mass spectrometry showed a good qualitative correlation between receptor activation in transfected cells and gas-phase binding, but quantitatively the responses did not correlate. Bile acid compounds showed a strong, non-specific binding to the RXRalpha LBD in the gas-phase. The stoichiometry of ligand binding was drastically different for bile acids and true RXRalpha agonist ligands. Binding stoichiometry is proposed as one way to identify non-specific gas-phase binding.

Further, agonist-induced changes in oligomerisation of the RXRalpha LBD protein, i.e. an increase in dimer content, could be monitored by mass spectrometry. Such changes in quaternary structure can be analysed by mass spectrometry and serve as another way to differentiate between specific and non-specific ligand binding.

Using the RXRalpha LBD and fatty agonists as a test system, the accuracy in direct measurements of the mass of ligands in complex with receptor proteins was investigated. Although the study was performed on an instrument capable of only medium resolution, amass error of better than 0.2 Da was obtained. This is sufficient to identify the degree of unsaturation or cyclisation if the compound class, for instance fatty acid, is known.

The methodology developed for investigation of protein-ligand interactions in the work on RXRalpha was extended to analysis of integral membrane proteins. The microsomal glutathione transferase 1 (MGST1) is a detoxification enzyme that is highly abundant in the liver, present at -3% of total microsomal protein. The enzyme catalyses the conjugation of reduced glutathione (GSH) to hydrophobic electrophilic substrates. Besides detoxification, a role has been proposed for MGST1 in protection against oxidative stress. MGST1 (molecular mass 17.3 kDa) has four transmembrane alpha-helices, forming a left-handed helical bundle.

Electrospray mass spectrometry was performed on the protein solubilised in a minimal amount of detergent (Triton X-100) compatible with enzyme activity. Using this approach, the oligomeric organisation of the enzyme was confirmed, showing that the functional unit of the enzyme is a trimer of -52 kDa. The mass spectrum showed monomer and trimer ion populations, that could be identified by mass labelling/activation of the enzyme using N-ethylmaleimide. Trimer ions could be dissociated within the mass spectrometer to form dimer.

Signal-to-noise ratios for the trimer substantially improved when nanoelectrospray was used instead of pneumatically assisted electrospray. This enabled the determination of substrate (GSH) binding stoichiometry. It was shown that the trimer binds three molecules of GSH. This finding was in apparent contradiction to published reports in which stopped flow-kinetic and equilibrium dialysis experiments clearly show utilisation/binding of a single GSH molecule. However, the results presented here and previous data are in support of a model for substrate binding consisting of three separate active sites displaying third of the sites reactivity.

In conclusion, the methods presented enhance the possibilities to investigate both membrane and soluble proteins and their interactions with substrates and ligands.