Differentiating specific and non-specific protein-metabolite interactions using gradient open port probe electrospray ionization mass spectrometry

ABSTRACTNative electrospray ionization mass spectrometry (ESI-MS) using nano ESI or desorption electrospray ionization (DESI) has been widely used to study interactions between macromolecules and ligands, usually protein-metabolite interactions (PMIs). In MS spectra the charge state distributions (CSD) of proteins differ between native and non-native conditions and based on this, we report a method that can differentiate specific protein-metabolite interactions from non-specific binding. Our approach is based on a 3D-printed open port probe electrospray system (gOPP-ESI) using mobile phase gradients (aqueous to methanol) for the ionization of protein/protein-metabolite complex. Notably, we found that a true protein-metabolite complex is more resistant to the denaturing effect of methanol compared to the free protein. This is corroborated by the observation that forming high charge states of protein-metabolite complexes requires higher proportions of methanol than free protein while, for non-specific complexes, there is no obvious difference in the CSD. Therefore, by comparing the changes in the CSD of free protein and protein-metabolite complex versus the increase of methanol, we can distinguish metabolites that specifically interact with the target protein. The approach is evaluated with well-characterized protein-ligand pairs, and we confirmed that cytidine phosphates,N, N′, N″-triacetylchitotriose, and fluvastatin are specific ligands for ribonuclease A, lysozyme, and beta-lactoglobulin respectively. However, cytidine-5’-triphosphate (CTP) interacts non-specifically with lysozyme and beta-lactoglobulin. We believe that after first-round native-MS assays to identify which metabolites cause mass shifts to the free protein, the gOPP-ESI-MS could be used as a quick second-round check to exclude non-specific binding and discover metabolites truly interacting with the protein of interest, reducing the number of candidates for subsequent validation experiments.Abstract Figure