Bicarbonate buffers can promote crosslinking and alternative gas-phase dissociation pathways for multiprotein complexes

Highlights

• Adding ammonium bicarbonate to samples prior to native MS promotes the formation of large sub-complexes during CID.

• Bicarbonate buffers can catalyze the formation of disuflide bonds and other chemical cross-links within protein complexes.

• Bicarbonate buffer additives could enable future protein-protein contact mapping efforts.

Abstract

Previously, we have reported the stabilization effect of Hofmeister salts for multiprotein complexes (MPC) in the absence of bulk solvent (J. Am. Chem. Soc. 2011, 133 (29), 11,358–11367; Angew. Chem. Int. Ed. 2012, 51 (23), 5692–5695; Angew. Chem. Int. Ed. 2013, 52 (32), 8329–8332.). Our efforts sought to bridge the gap between gas-phase protein structures and those found in solution. To reveal more detailed MPC topology information, native ion mobility-mass spectrometry (IM-MS) measurements are often combined with gas-phase activation methods. Conventional activation methods, including collision induced dissociation/unfolding (CID/CIU), however, primarily report information focused on monomeric subunits within the MPC, limiting the topological information obtained. Herein, we describe a simple buffer-doping method that promotes an alternative MPC CID pathway which readily produces product ions that correspond to larger sub-complexes from within some parent assemblies. Interestingly, tetramers exhibiting a dimer of dimers quaternary structure (e.g. hemoglobin and concanavalin A) produce dimeric product ions upon collisional activation following ionization from bicarbonate buffer, in contrast to the commonly observed monomer-ejection CID pathway. In order to both further investigate and validate our native IM-MS, we performed bottom-up proteomics experiments on MPCs housed in bicarbonate buffer. Our efforts revealed evidence of bicarbonate-mediated disulfide bond formation in proximal Cystine residues. We close by discussing the applications for these observations in the context of MPC structure determination by native IM-MS.

Introduction

Higher order structures of multiprotein complexes (MPCs) are a key determinative factor for their functions in vivo. For example, hemoglobin (Hb), a centrally-important heme-containing MPC involved in oxygen transport and renal dysfunction following rhabdomyolysis, is highly conserved [1,2]. As is typical for MPCs, the dynamic assembly and disassembly of Hb in solution involves several comprehensive and reversible steps. It is argued that hetero-tetrameric Hb dissociates into hetero-dimeric globin (αβ) in acetate buffer at a pH above 6.5, which is followed by a further dissociation of dimeric globin into single chain (α- and β-globin) concomitant with heme group detachment at a pH below 5 [3]. Konermann et al. observed a similarly dimeric globin-driven dissociation pathway in solution from freshly obtained Hb which is free of oxidative damage [4]. These solution architectures, combined with mass spectrometry (MS)-derived residue-level noncovalent contact information and many other structural biology tools including X-ray crystallography and cryoEM, provides a structural basis for Hb biological function of oxygen transport. Despite these and other efforts aimed at elucidating Hb dynamics [[5], [6], [7]], there remain many unanswered questions surrounding the interplay between Hb structure, function and subunit exchange.

Native ion mobility-mass spectrometry (IM-MS) is rapidly becoming a versatile tool for accessing protein assembly processes and topological information [[8], [9], [10], [11]]. To explore the structure of proteins and MPCs, desolvation is a required step. In general, this step is accomplished using nano-electrospray ionization (nESI), and is believed to generate MPC ions that resemble their native states in the solution phase. In addition to direct measurements of MPC mass (accomplished using MS) and collision cross-section (CCS) (accomplished using IM) of MPCs, additional organizational and sequence-level information can be obtained using gas-phase protein activation. For example, collision induced dissociation (CID) and collision induced unfolding (CIU) have emerged as widely used methods for detailed MPC structure interrogation [[12], [13], [14]]. The resultant MPC dissociation and unfolding patterns can often be correlated to native protein structures and properties [13,15,16].

Although multiple strategies have emerged for utilizing gas-phase information in the construction of native-state models of MPCs [6,[17], [18], [19], [20], [21], [22]], the dissociation pathways accessed by MPC ions upon CID contrast sharply with the known solution structures of such assemblies. For example, CID of Hb, a 64 kDa hetero-tetramer comprised of homo-dimers, exhibits evidence of classical asymmetric ejection of unfolded monomeric units, in contrast to expectations derived from the native quaternary structure of the MPC, which is organized as a dimer-of-dimers [3]. Notably, MPC charge state has been demonstrated experimentally and theoretically to have a profound effect on gas-phase dissociation and unfolding [15,23]. To pave the way for the development of native IM-MS for the characterization of MPCs of increasing size and complexity, activation techniques must be improved in order to provide needed details surrounding MPC structure and dynamics that remain refractory to other structural probes [24].

Chemical crosslinking, in combination with MS (especially in the context of bottom-up proteomics), has emerged as a useful strategy to capture transient protein-protein interactions and to map protein-protein interaction interfaces at an unprecedented scale [25,26]. Previous efforts have been devoted to developing multi-functional MS-cleavable crosslinkers with different spacer arms ranging from zero to more than 24 Å in length [[27], [28], [29]]. For example, bismaleimide sulfoxide reagents with spacer arms of 24.2 Å target cystine residues and permit the assessment of protein-protein interactions with distance constraints as long as ∼45 Å to be captured as a result of the protein backbone flexibility [30]. Our previous work has revealed the stabilization effects of MS-cleavable cross-linkers [31], tags [32,33], anions [19], and cations [18] on MPCs, as well as their influence on MPC CID [17]. Here, we describe a simple buffer-doping approach to promote alternative CID pathways for MPCs. In our experiments, ammonium bicarbonate (NH₄HCO₃) is added to native protein solutions prior to nESI. We explore the effect of this additive both in solution and in the gas phase using top-down/tandem MS, bottom-up LC-MS/MS, CIU fingerprints, pH measurements and circular dichroism (CD) spectroscopy. Based on this collection of evidence, we propose a tentative mechanism that relies upon the promotion of disulfide bonds in proximal Cysteine residues within bicarbonate buffers. These additional non-specific yet structurally-informative disulfide bonds promote both more compact MPC structures in the gas-phase and alternative CID pathways. For example, disulfide-linked dimerization between α- and β-chain monomers within Hb was observed experimentally using LC-MS/MS. Such precursors go on to produce dimeric product ions upon CID. We conclude by exploring the implications of our observations for native IM-MS in the context of MPC structure discovery.