Mass Spectrometry
Online ISSN : 2186-5116
Print ISSN : 2187-137X
ISSN-L : 2186-5116
Special Issue: Proceedings of 19th International Mass Spectrometry Conference
Integrating Native Mass Spectrometry and Top-Down MS for Defining Protein Interactions Important in Biology and Medicine
Joseph A. Loo Sabrina A. BenchaarJiang Zhang
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML

2013 Volume 2 Issue Special_Issue Pages S0013

Details
Abstract

Native protein mass spectrometry (MS), the measurement of proteins and protein complexes from non-denaturing solutions, with electrospray ionization (ESI) has utility in the biological sciences. Protein complexes exceeding 1 MDa have been measured by MS and ion mobility spectrometry (IMS), and the data yields information not only regarding size, but structural details can be revealed also. ESI-IMS allows the relative stability of protein–ligand binding to be measured. Top-down MS, the direct dissociation of the intact gas phase biomolecule, can generate sequence and identity information for monomeric (denatured) proteins, and topology information for noncovalent protein complexes. For protein complexes with small molecule ligands, i.e., drugs, cofactors, metals, etc., top-down MS with electron capture dissociation can be used to elucidate the site(s) of ligand binding. Increasing protein ESI charging, e.g., supercharging, enhances the efficiency for dissociation of protein complexes.

INTRODUCTION

The application of mass spectrometry (MS) and ion mobility spectrometry (IMS) for studying proteins and protein complexes has utility in structural biology and biomedical research. Proteins and other biomolecules such as small molecule ligands and proteins interact to form functional molecular partners. The assessment of protein interactions can address the functional role of proteins and protein complexes. The pharmaceutical industry depends on tools for screening millions of compounds to identify potent molecules that target specific proteins. Potentially native protein MS, i.e., the MS measurement of proteins and protein complexes from non-denaturing analyte solutions, with electrospray ionization (ESI) can be used as a tool for screening small molecule compound mixtures to identify low-to-medium affinity ligands.1,2) Ligand binding stoichiometry can be measured well by ESI-MS, but stability and affinity are relatively difficult issues to address.

Measuring the molecular masses of protein complexes reveals information on stoichiometry of binding and identification of interacting partners.1,3) MS and ion mobility measurements can supply further information on the structural aspects of gas-phase/solution-phase protein conformations and supramolecular assemblies in excess of 1 MDa.4,5) However, MS can advance potentially an additional layer of information that is critical to structural biology: location (or topology) and dynamics. Recent developments in protein MS and tandem MS (MS/MS, or top-down MS) to define the structures of native protein complexes, including the sites of ligand binding, have been made.6) Collisionally activated dissociation (CAD) of gas-phase stable noncovalent complexes resulting from electrostatic interactions can be used to probe ligand-binding sites (e.g., nucleotides, metals).7) Electron transfer dissociation (ETD) and electron capture dissociation (ECD) can probe the binding sites of weakly bound ligands and the topology of protein–protein complexes.8) For example, we are using ECD-Fourier transform ion cyclotron resonance (FT-ICR) MS to investigate the molecular action of compounds that prevent amyloid fibril formation in neurodegenerative diseases such as Alzheimer’s and Parkinson’s disease.9)

The role of ESI-MS and ESI-IMS-MS for elucidating the relative stability imparted to a protein upon ligand binding and the potential role of IMS-MS for ligand screening will be described. Native proteins and complexes are not necessarily static entities, as they are in constant motion and perhaps changing their partners during the “dance” that defines their roles in biological processes. We are using native MS and top-down MS to decipher the dynamics of protein interactions.

EXPERIMENTAL

Proteins were analyzed in native solution conditions. Proteins were desalted and recovered in 20 mM ammonium acetate buffer (pH 6.8) using centrifugal filter devices (Microcon, Millipore, Billerica, MA, USA). The final protein concentration used in the experiments was between 2.5–10 μM. Mass spectra were acquired in positive ion mode with a nano-ESI interface using Au/Pd coated borosilicate glass capillaries (Proxeon Biosystems, Odense, Denmark) and with a low flow rate (30–50 nL/min). Experiments were run on a hybrid IMS-quadrupole time-of-flight (Q-TOF) mass spectrometer (Waters Synapt HDMS G1, Manchester, UK). Top-down MS and MS/MS of protein samples were performed on a 15-Tesla Bruker SolariX hybrid Qq-FTICR mass spectrometer (Bruker Daltonics, Billerica, MA, USA). ECD was employed using the Bruker FT-ICR. FT-ICR data were processed with DataAnalysis and Biotools software (Bruker) and annotated manually for accurate ion assignment.

RESULTS AND DISCUSSION

The history of applying ESI-MS for noncovalent protein complexes has been well documented.1,3) Certainly, the advancement of nanoelectrospray sources to generate smaller charged analyte droplets from small analyte volumes and more sensitive mass spectrometers with extended mass-to-charge (m/z) ranges, such as time-of-flight (TOF) analyzers, has allowed the measurement of native protein complexes to be nearly standard in practice for many laboratories. Figure 1 shows ESI mass spectra for two relatively small protein complexes, 174 kDa homodimeric heat shock protein 90 (Hsp90) and 232 kDa homotetrameric pyruvate kinase.10) Hsp90 is a heat shock protein and a molecular chaperone that stabilizes the final stages of folding of a wide range of proteins in cells under stress. The protein is important for the survival and growth of cancer cells. The mass spectra shows the typical pattern from multiple charging of a noncovalent complex. Compared to ESI mass spectra of denatured proteins (e.g., solutions containing acetonitrile and organic acid at low solution pH), relatively few peaks are observed in the native protein mass spectra. Multiple charging is sufficient to measure a spectrum, but the availability of an analyzer such as a TOF instrument is useful because of its higher efficiency for detection of ions above m/z 6000.

Fig. 1. NanoESI-MS of (A) 232 kDa rabbit pyruvate kinase and (B) 174 kDa human Hsp90 (20 mM ammonium acetate, pH 6.6) measured with a Q-TOF mass spectrometer.

MS and IMS of protein–ligand complexes

Binding of a ligand to the folded state of a protein can result in thermodynamic stabilization of the complex. In solution, increasing melting temperature is observed for protein–ligand binding, and this can be monitored by techniques such as isothermal calorimetry. Ligand binding can also protect a protein from proteolysis by stabilizing the folded state.

For example, a recently developed method for small-molecule target identification termed “DARTS” (Drug Affinity Responsive Target Stability) takes advantage of changes in the protease sensitivity of the target protein upon binding by a small-molecule ligand.11) Given that a protein might become less susceptible to proteolysis when it is drug-bound compared to its drug-free state,12) this phenomenon could be exploited for target identification by using gel electrophoresis as the probe to screen for proteins that become resistant to proteolysis in the presence of the drug, an effect which can be demonstrated both with pure proteins and in complex mixtures such as whole cell lysates. Protection from proteolysis is specific to the target protein(s) while proteolysis of non-target proteins is unchanged. A key advantage of DARTS is that the compound does not need to be labeled, tagged, or immobilized. We and others have found that ligand binding stabilizes the gas phase protein complex as well, and this can be probed by IMS-MS.13,14)

Enolase is a well-studied metalloenzyme that catalyzes the dehydration of 2-phospho-d-glycerate in the glycolytic pathway. Yeast enolase has a molecular mass of 93 kDa for the homodimeric complex. It requires divalent metal ions and first row divalent transition metals (e.g., Mg2+, Mn2+) for enzymatic activity. Two metal ions are bound to the protein dimer with a solution equilibrium dissociation constant (KD) in the nanomolar range.

Our previous work demonstrated that metal-binding stabilizes the gas phase enolase dimer complex relative to the apo-form.15) CAD-MS experiments by increasing the energy in the region of the ESI atmospheric pressure/vacuum interface (i.e., in-source CAD) showed that Mg2+/Mn2+-binding stabilized the enolase dimer from dissociating to the monomer state. This enhanced stability of the gas phase metal-bound complexes is consistent with the solution phase structure of enolase, as binding Mg2+/Mn2+ strengthens subunit interactions. There are differences in conformation between Mg2+/Mn2+-enolase and apo-enolase that affect regions of the protein involved in subunit interactions.

The relative stability of the folded conformation of the enolase dimer complex (Fig. 2) can be monitored as it travels through the traveling-wave (T-wave) ion guide of the IMS cell found in the Synapt G1 instrument. Prior to entering the T-wave region, ions are stored in an ion trapping region (Trap) where they can undergo ion activation, i.e., heating.16) With a Vtrap potential of +32 V, the enolase dimer complex is intact and the 20+-charged dimer molecule shows a single conformer at 44 ms mobility drift time. Raising Vtrap to +70 V results in an additional conformer peak in the ion mobility profile at a longer drift time (60 ms). However, metal binding stabilizes the enolase dimer from unfolding upon activation in the Trap, as the relative proportion of the longer drift time unfolded conformer is reduced for the holo-protein relative to the apo-enolase (Fig. 2).

Fig. 2. IMS of yeast enolase dimer. The top panel depicts the drift-scope representation of the mobility profile for Mn2+-bound enolase. The ESI mass spectrum to the right of the drift-scope output is aligned with the mobility profile. The bottom panel plots the drift times of the 20+-charged dimer at two different Vtrap potentials and shows the enhanced stabilization of enolase with metal binding.

The stability of gas phase complexes depends highly on the nature of the noncovalent forces. We and others have demonstrated previously that electrostatic interactions are significantly enhanced in the absence of solvent.7,1719) With covalent-like strength, gas phase electrostatic interactions are often retained even in the “harsh” environment of CAD-MS/MS experiments. For example, top-down mass spectrometry of adenosine 5-triphosphate (ATP) bound to a protein kinase, adenylate kinase (AK), showed that the intrinsic stability of the electrostatic interactions in the gas phase allows the diphosphate group of ATP to remain noncovalently bound to the protein during CAD. This feature was exploited to yield positional information on the site of ATP-binding on adenylate kinase.7)

Moreover, increasing the charge carried by analytes (i.e., “supercharging”) is desirable because higher charged molecules are more effectively dissociated by tandem MS and because they reduce analyzer m/z requirements.2022) This feature is especially beneficial for the analysis of large, noncovalent complexes. Previously, we demonstrated that multiple charging of native proteins and noncovalent protein complexes could be increased in ESI-MS when spectra are obtained from non-denaturing protein solutions (e.g., 10–50 mM ammonium acetate, pH 6.8) containing up to 0.5% (v/v) m-nitrobenzyl alcohol (m-NBA)22); charge increased by +8% to +48% for the native proteins measured. More recently, we presented other compounds as potent as m-NBA (e.g., sulfolane and 3-(trifluoromethyl)-benzyl alcohol) for supercharging under both native and denaturing conditions.21)

Supercharging noncovalent proteins increases the efficiency of top-down MS.23) For 29 kDa zinc-bound carbonic anhydrase, the addition of m-nitrobenzyl alcohol increases charging by nearly 50%. CAD of the 11+ Zn–protein complex yields product ions that do not cover the zinc-binding domain (His-94, His-96, His-119). However, CAD of the supercharged 15+ Zn–protein complexes generates the y675+/(b192+Zn)9+ complementary product ion pair, with the b192 fragment retaining zinc and is consistent with the expected binding sites. Similar enhanced charging and dissociation of native protein complexes has been observed for a variety of other supercharging reagents, including sulfolane (Fig. 3).

Fig. 3. (A) ESI mass spectra of adenylate kinase in the presence of ATP (10 mM ammonium acetate, pH 6.6) (top) without and (bottom) with 200 mM sulfolane. The peaks marked with open circles represent multiply charged molecules of the apo-protein, and those marked with filled circles represent the holo-protein. (The corresponding mass spectra converted to the mass domain are shown in the insets.) The number of product ions generated by CAD and ECD of the protein–ATP complex is plotted in (B). MS/MS of the higher charged precursor yield a high proportion of holo-products (percentage listed above the bars). (Figure adapted from ref. 23 with permission from the publisher.)

Comparison of the CAD and ECD data generated for the 10+ (without sulfolane) and the 15+ (with sulfolane) AK-ATP complex showed a much higher proportion of holo-product formation (Fig. 3B). The proportion of CAD/ECD product ions that retain either the diphosphate group or the intact ATP ligand was significantly enhanced with the higher charged 15+ complex (generated by supercharging). For example, 48% of the 80 total product ions from ECD of 15+ AK-ATP retained either diphosphate or ATP, compared to only 11% for the 10+ complex. Moreover, most of the ECD-formed holo-products for the 15+ complex retained the intact ATP, whereas no intact ATP was observed for ECD of the 10+ complex.

Kinase enzymes transfer a phosphoryl group from ATP to the hydroxyl groups of proteins or mononucleotides. Protein kinases represent the largest mammalian enzyme family, with more than 500 members in the human proteome; the ATP-binding sites of protein kinases are highly conserved. Our CAD and ECD data for adenylate kinase-ATP suggested that the ATP binding site of the gas phase protein is consistent with the solution phase structure.23)

Likewise, the stability of the solution phase complex and gas phase complex can be probed by a proteolysis protection assay and by IMS, respectively. AK with and without ATP were digested with trypsin and monitored over time. By SDS-PAGE, the 22 kDa band for intact AK with ATP added can be observed to disappear over a 30 min digestion time, but remains over the same time period when bound to ATP (Fig. 4). ATP does not inhibit trypsin activity, and thus the data are consistent for the stabilization of the native conformation of AK in solution with ATP bound. Similarly, ion mobility spectrometry can monitor the enhanced stability in the presence of ligand binding. ESI-MS of a near neutral pH solution of AK with equimolar ATP added shows ions for both the apo- (21.6 kDa) and holo-forms (21.9 kDa), with a stoichiometry of 1 : 1 (protein : ATP) (Fig. 3A). Isolation of the 10+-charge molecules for the apo- and holo-forms in the T-wave ion mobility cell of the Synapt G1 instrument with different Vtrap potentials shows differential behavior for the two protein forms. With a Vtrap of +20 V, both apo- and holo-AK have a single relatively broad peak at 5.3 ms drift time. At Vtrap of +80 V, however, the broad peak is now split into two peaks that include a peak at longer drift time (6.8 ms). This suggests that a mixture of conformers is present originally at lower Vtrap potential. Heating the ions in the Trap region of the instrument destabilizes them. The apo-AK sample has more than 50% of its original population shifted to a more “unfolded” form upon Vtrap-induced heating, whereas the more holo-AK sample has retained more of the “folded” form after heating because of stabilization of the protein upon ligand binding (Fig. 5).

Fig. 4. Limited trypsin digestion of adenylate kinase. The protein was digested with trypsin in the absence (left 4 lanes) and presence (right 4 lanes) of ATP at room temperature. Digestion progress was monitored every 10 min, and an aliquot of the solution (quenched with PMSF) was loaded into a separate lane of the SDS-PAGE gel. After 30 min of trypsin digestion, most of the original protein was digested into products (as suggested by the disappearance of the 22 kDa band) for the apo-form, whereas ATP binding slows the rate of trypsin digestion.
Fig. 5. IMS of adenylate kinase (AK). Ion mobility profiles were acquired for the 10+-charged AK (red trace) and AK–ATP complex (green trace). The drift times were similar for both forms at Vtrap=+20 V (top). A second peak representing a more unfolded conformer appears at longer drift times with Vtrap=+80 V (bottom), but AK binding ATP reduces the amount of unfolding upon heating in the Trap cell.

Stabilization of proteins by ligand binding has been observed for several other systems, including Hsp90 protein dimer binding to ATP (data not shown). Monitoring the stabilization of proteins upon ligand binding by profiling activation-IMS fingerprints is analogous to thermal shift assays developed for solution phase melting temperature screens. This application of IMS-MS has the potential for screening compound mixtures for protein–ligand binding, even for hydrophobic protein–ligand interactions.

Tandem mass spectrometry of larger noncovalent protein complexes

Mass spectrometry studies have suggested that elements of the solution phase complex are preserved in the dehydrated complex.24) In general, a high correlation between the ESI-MS data and expectations from the solution state world has been found.25) The precise three-dimensional structure of the gas phase molecule or complex may or may not be the same as the solvated species; however, there may be some structural elements that are preserved upon lifting a biomolecule into the gaseous state. In some examples reported, other physical characteristics of the gas phase complex, such as their dissociation behavior and chemistries designed to probe topographical features, are consistent with the complex formed in solution. Collision cross-section measurements by ESI-ion mobility and ion scattering measurements have suggested that different conformers can be measured for gas-phase proteins.2628) The conserved bioactivity of a virus sample collected post-ESI mass spectrometry demonstrated that the ESI and desolvation processes are not destructive, but do not indicate directly the structure of the gas phase (or solution phase) complex.29) Recent work by Wysocki’s group applying surface induced dissociation (SID) of large protein complexes suggest that probing the gas phase structure of biomolecules can be used to elucidate their solution phase geometries.30) However, questions remain regarding the general applicability of the method.

The role of protein assemblies in normal cellular processes and diseases warrants a practical method for the study of large complexes. Our laboratory has been studying whether the dissociation of specific protein subunits from larger assemblies in the gas phase can be related to the overall topology of the complex and its solution phase structure. The proteasome is the intracellular protease complex that is responsible for degradation of most damaged or misfolded proteins in the cytosol and in the nucleus and is of critical importance to cell cycle and cell survival.31) The 20S proteasome is composed of 4 heptameric rings (α7β7β7α7) stacked in a barrel fashion, forming an internal cavity. Previous measurements by our lab measured the full 20S proteasome from archaeon M. thermophile to yield a MW of 693.5 kDa.32) Dissociation of the gas phase complex generates fragments from the loss of only the α-subunits, consistent with the known architecture and topology of the complex, as the α-subunits are located on the outer rings.

Unlike the archaea complexes, for eukaryotic 20S proteasomes, such as from humans, each of the α-subunits and each of the β-subunits are unique, i.e., there are 28 different subunits (and each may carry post-translational modifications). ESI-MS of the human 20S proteasome shows broad peaks yielding an approximate molecular mass of 735 kDa (Fig. 6). CAD of the 62+ complex results in the loss of one 27 kDa α-subunit (to generate ions for the remaining 27-mer complex). From the measured molecular mass (27,312 Da) of the α-subunit that dissociated from the intact complex, the subunit was identified as the α6-subunit (27,310 Da from the sequence minus the initiator Met1 and acetylation of Ser2; manuscript in preparation). The loss of an outer ring α-subunit is consistent with the architecture and previous data for other 20S proteasomes. However, it is not clear why only the α6-subunit specifically is lost. Examination of the high-resolution structure of the 20S proteasome (from bovine, Protein Data Bank entry 1IRU) does not reveal any obvious reasons why the α6-subunit is preferentially lost upon CAD. Additional experiments and analyses are required to address this question.

Fig. 6. NanoESI-MS of the 28-mer human 20S proteasome complex (top). Q-TOF MS/MS of the 62+ complex generates products for the 27-mer and the “departing” α6-subunit (bottom).

Bacteria need iron from the host to establish infection. The blood infection from Staphylococcus aureus affects a large fraction of the human population and its pathological mechanism has not been fully elucidated.33) Infection from S. aureus causes life-threatening conditions, including abscesses, pneumonia, and toxic shock syndrome. Methicillin-resistant S. aureus (MRSA) is a highly-resistant strain of S. aureus; more people die in the US from the MRSA bacterium than from HIV. It has been postulated that a receptor on the surface of S. aureus, IsdH (iron-regulated surface determinant protein H), is involved in the capture of heme from hemoglobin (Hb) and this is the source of iron required for infection.34) However, the molecular mechanism of heme capture and iron release for pathogen survival remains to be determined.

We have been applying native ESI-MS and MS/MS to study the dynamics of noncovalent interactions between Hb and the IsdH protein receptor.35) The complex processes of site-specific protein recognition and heme transfer, including the kinetics of multiple binding equilibria, can be monitored by ESI-MS to provide important clues to the mechanism of S. aureus infection. High resolution FT-ICR MS and MS/MS of Hb and its complex with IsdH will be of utility to develop a molecular picture of how heme is captured and utilized by the bacteria. Preliminary high-resolution mass spectra of Hb have demonstrated the capabilities of high-field FT-ICR MS for accurately measuring molecular masses of large noncovalent protein complexes (Fig. 7). Future work will probe directly the structure of the IsdH-Hb complex by CAD and ECD. This information could be useful for designing new compounds and strategies for fighting and preventing potentially lethal S. aureus infections.

Fig. 7. High-resolution nanoESI-MS (15 Tesla FT-ICR MS) of human hemoglobin (20 mM ammonium acetate, pH 6.6). Ions for the intact α2β2 tetramer (peaks marked in red) and αβ dimer (blue labels) (and each subunit bound to one molecule of heme) are measured.

CONCLUSION

Native protein mass spectrometry can yield important information on the structural details of large, noncovalent protein complexes and assemblies. It should be considered as useful to the biological scientist as other more commonly employed biophysical techniques, but there are several factors that preclude its more widespread use. These include the relatively high cost of high performance mass spectrometers (with extended m/z range) and the belief of many that experts highly trained in the field of protein mass spectrometry are required. However, the cost and the laboratory footprint of mass spectrometers that are capable of native mass spectrometry experiments are dropping. Moreover, instruments are easier to operate today and procedures for native MS are well-established in the literature and are found in many labs worldwide. The ability of native MS and MS/MS for providing a view of the dynamics of protein–ligand assembly–disassembly is unparalleled and this picture would be difficult to obtain by other biophysical methods.

Whether the composition of protein complexes can be and should be addressed by mass spectrometry depends highly on the answer to the question, “How closely linked are the structures of the gas phase protein to their solution structures?” There is mounting evidence from the literature that supports the application of MS for studying the solution structure of proteins, but it is less clear how much structural detail can be gained by measuring the gas phase molecule. Only future research by the entire mass spectrometry community will be able to address this issue fully.

Acknowledgment

Support from the US National Institutes of Health (R01 GM103479, S10 RR023045, S10 RR028893) is acknowledged. The S. aureus/hemoglobin project is lead by Reza Malmirchegini and Dr. Robert Clubb (UCLA). Helpful advice and suggestions from Dr. Rachel Loo and other current and past members of the group at UCLA are acknowledged.

REFERENCES
 
© 2013 The Mass Spectrometry Society of Japan
feedback
Top