Collisional cooling enhances the ability to observe non-covalent interactions within the inducible nitric oxide synthase oxygenase domain: Dimerization, complexation, and dissociation

Jeffrey C. Smith 0 1 2 3 K.W. Michael Siu 0 1 2 3 Steven P. Rafferty 0 1 2 3 0 Published online February 27, 2004 Address reprint requests to Dr. K. W. M. Siu, Department of Chemistry, York University , 4700 Keele Street, Toronto , ON M3J 1P3, Canada 1 Chemistry Department, Trent University , Peterborough, Ontario, Canada 2 Department of Chemistry and Centre for Research in Mass Spectrometry, York University , Toronto , Ontario, Canada 3 Received October 29, 2003 Revised December 19, 2003 Accepted January 5, 2004 The investigation of protein quaternary structure, protein-cofactor, and protein-ligand interactions by mass spectrometry is often limited by the fragility of such interactions under experimental conditions. To develop more gentle conditions of perhaps general use, we used as a model for study the oxygenase domain of murine inducible nitric oxide synthase (iNOS), which is homodimeric, binds heme and tetrahydrobiopterin H4B cofactors, and the substrate L-arginine. The energetics of the collisions in q2 and in the lens region of the mass spectrometer were manipulated for varying the degree of solvation around the non-covalently bound ions. Furthermore, the number of low-energy collisions in the collision cell of the instrument was varied, focusing and dampening the ion beam. Under gentle source collision conditions, and using multiple low-energy collisions in the collision cell of the mass spectrometer, dimers of the iNOS oxygenase domain containing heme, H4B, and arginine were observed intact after electrospraying at pH values near neutrality; a mutant of this protein (Trp188 3 Phe) was monomeric and did not bind cofactors. The pH dependence of the iNOS oxygenase domain under acidic conditions was also studied; while heme remained bound to the protein between pH 2.5 and 4.0, the dimeric structure was disrupted. Our findings confirm that non-covalently bound macromolecular complexes are retained and observable using electrospray mass spectrometry under the appropriate experimental conditions. (J Am Soc Mass Spectrom 2004, 15, 629 - 638) 2004 American Society for Mass Spectrometry - Tappropriate focus for studying non-covalent inhe enzyme nitric oxide synthase (NOS) is an teractions as few proteins rival it for the richness in variety in which it partakes. These include the binding of four cofactors (two flavins, a heme, and a pterin), three substrates (L-arginine, NADPH, and oxygen), and at least two proteinprotein interactions (between subunits of this homodimeric enzyme and interactions with the small calcium-sensing protein calmodulin, which regulates enzyme activity). In spite of the many functional differences, the NOS isotypes produce NO by the same mechanism: 3 NADP 3 NADPH 3 2 citrulline This similarity in mechanism is a consequence of structural homology within the NOS isotypes. All isotypes have a common domain architecture consisting of an oxygenase domain that binds heme and tetrahydrobiopterin (H4B) linked by a short calmodulin binding site to a carboxy-terminal reductase domain which, in turn, binds flavin cofactors [1]. Electrons are transferred from NADPH via the reductase domain to the oxygenase domain, which is the site of oxygen binding and NO production. Furthermore, NOS is active only as a dimer; dimerization requires the presence of the heme cofactor and is stabilized further by binding of H4B and arginine [2]. NOS dimerization occurs primarily through interactions between the oxygenase domains on the two subunits [3]. The organization of NOS into distinct structural domains lends itself to their study as independent species, obtained either by limited proteolysis of the complete protein or by recombinant expression. For example, the independently expressed oxygenase domain of NOS retains the capacity to bind cofactors, substrate, and to dimerize, and thus reflects the structural and functional properties of this domain in the intact protein [4]. Establishing a gas-phase representation of the equilibrium between oxygenase monomers and dimers will allow a standard to be set to which future modifications of the protein (i.e., point mutations) may be compared. As dimerization directly correlates to the proteins ability to be active, studying factors that affect this equilibrium may allow the elucidation of key stabilizing residues or inhibitory substances. Mutational studies commonly allow insight into the intra-physiological processes and key residues within a protein. Recently, it has been reported that a mutant form of the inducible NOS (iNOS) oxygenase domain has been created, and that this mutant lacks the ability to bind the cofactor heme [5]. In this mutant, the 188th amino acid residue, tryptophan, is replaced by phenylalanine. It is thought that this tryptophan residue in the wild type plays a major role in stabilization of the heme in the heme-binding pocket; it has been shown that, through a mutation to a phenylalanine residue, the heme binding becomes severely destabilized despite the protein remaining folded [5]. The present study also examines this mutant (known as the W188F) in order to observe its gas-phase behavior and any non-covalent interactions that may be present. With the advent of soft ionization techniques, such as electrospray ionization (ESI) and matrix-assisted laser desorption ionization (MALDI), large biomolecules are now accessible to investigation using mass spectrometry (MS). It has been stated that nanospray, a lowflow subclass of ESI, is even softer [6, 7] and has a greater tolerance for non-volatile salts [8] than conventional electrospray, thereby making it even more suitable for the study of non-covalent interactions. There are now a number of reports that detail the observation of non-covalent interactions using mass spectrometry, and this topic has been recently reviewed by several groups [9 13]. Most of these studies dealt with rather strong, electrostatic interactions that are arguably enhanced in the absence of solvent, [14] e.g., protein/ DNA complexes [15], DNA/ligand complexes [16], heme/protein and other interactions [1719]. Extensive time and effort have also gone into studying protein/ protein interactions through mass spectrometry [14, 20 26]. A few groups have reported the ability to observe hydrophobic non-covalent interactions by mass spectrometry including leucine zipper peptides [27] as well as acyl CoA derivatives to acyl CoA binding protein [28]. Both studies reported a very low relative abundance of non-covalently bound species ( 10 20% or less), which was somewhat expected as hydrophobic interaction is likely to diminish in the absence of solvent. The degree of hydrophobicity of a particular non-covalent interaction was rarely discussed, perhaps because it is not fully known, or due to the fact that, upon transfer to the gas phase, hydrophobic stabilization effects diminish, while electrostatic effects are generally enhanced. It has been reported that, in the endothelial NOS (...truncated)