Elucidating the Site of Protein-ATP Binding by Top-Down Mass Spectrometry
Total Page:16
File Type:pdf, Size:1020Kb
FOCUS:TOP-DOWN MASS SPECTROMETRY Elucidating the Site of Protein-ATP Binding by Top-Down Mass Spectrometry Sheng Yina and Joseph A. Looa,b a Department of Chemistry and Biochemistry, University of California-Los Angeles, Los Angeles, California, USA b Department of Biological Chemistry, David Geffen School of Medicine, University of California-Los Angeles, Los Angeles, California, USA A Fourier-transform ion cyclotron resonance (FT-ICR) top-down mass spectrometry strategy for determining the adenosine triphosphate (ATP)-binding site on chicken adenylate kinase is described. Noncovalent protein–ligand complexes are readily detected by electrospray ioniza- tion mass spectrometry (ESI-MS), but the ability to detect protein–ligand complexes depends on their stability in the gas phase. Previously, we showed that collisionally activated dissociation (CAD) of protein–nucleotide triphosphate complexes yield products from the dissociation of a covalent phosphate bond of the nucleotide with subsequent release of the nucleotide monophosphate (Yin, S. et al., J. Am. Soc. Mass Spectrom. 2008, 19, 1199–1208). The intrinsic stability of electrostatic interactions in the gas phase allows the diphosphate group to remain noncovalently bound to the protein. This feature is exploited to yield positional information on the site of ATP-binding on adenylate kinase. CAD and electron capture dissociation (ECD) of the adenylate kinase–ATP complex generate product ions bearing mono- and diphosphate groups from regions previously suggested as the ATP-binding pocket by NMR and crystallographic techniques. Top-down MS may be a viable tool to determine the ATP-binding sites on protein kinases and identify previously unknown protein kinases in a functional proteomics study. (J Am Soc Mass Spectrom 2010, 21, 899–907) © 2010 American Society for Mass Spectrometry roteins function in biology through direct inter- the transition from solution to the gas phase that action with other molecules. A ligand can be a the complex size and binding stoichiometry can be Pmolecule, an atom, or an ion that binds to a measured [1–3]. specific site on the protein by a variety of means, Although binding stoichiometry and, in many exam- including hydrogen bonding, Van der Waals interac- ples, the relative and absolute solution binding affinities tions, and ionic forces. Determining the presence of can be measured by ESI-MS methodologies, determin- protein–ligand interactions and the structures of protein– ing the precise ligand binding site on a target protein is ligand complexes are of critical importance in biol- not easily tractable using MS directly. Tandem mass ogy, and this knowledge is often essential for efforts spectrometry (MS/MS) is widely applied to derive the to establish new molecular drug targets and to develop amino acid sequence of small polypeptides (for exam- more potent medicines. ple, in the so-called bottom-up proteomics approach), A number of biophysical tools are available to char- and MS/MS of intact proteins (i.e., top-down MS [4]) acterize the interaction between a protein and its li- show potential to be a tool for protein identification gand(s). However, mass spectrometry (MS) offers po- and elucidation of post-translational modifications [5]. tential advantages in sensitivity, specificity, and speed, However, the application of top-down MS to access a especially compared with high-resolution nuclear mag- protein’s primary structure while preserving noncova- netic resonance (NMR) spectroscopy and X-ray crystal- lent ligand binding would seem to be an insurmount- lography. With electrospray ionization (ESI) to ionize able task because of the lability of the protein–ligand macromolecules without disrupting covalent bonds interaction, both in solution and in the gas phase. while maintaining weak noncovalent interactions, the However, recent reports suggest that an MS ap- ESI-MS molecular mass measurement provides direct proach may be useful for pinpointing protein–ligand evidence for protein–ligand associations. The protein– binding. Hydrogen-deuterium exchange (HDX) exper- ligand interactions are often sufficiently retained upon iments can yield structural information on the regions of a protein involved in protein–protein and protein– ligand contacts [6]. Limited proteolysis of protein com- Address reprint requests to Dr. J. A. Loo, University of California-Los Angeles, Molecular Biology Institute, 402 Paul D. Boyer Hall, 405 Hilgard plexes can sometimes deliver clues on the sites of ligand Avenue, Los Angeles, CA 90095, USA. E-mail: [email protected] binding. For example, we used limited trypsinolysis Published online January 18, 2010 © 2010 American Society for Mass Spectrometry. Published by Elsevier Inc. Received November 16, 2009 1044-0305/10/$32.00 Revised January 8, 2010 doi:10.1016/j.jasms.2010.01.002 Accepted January 8, 2010 900 YIN AND LOO J Am Soc Mass Spectrom 2010, 21, 899–907 with ESI-MS to determine the sites of zinc- and oligo- actions with basic amino acids found in a number of nucleotide-binding to a nucleocapsid zinc finger protein peptide systems [16, 17, 19]. Our group recently showed [7]. Our laboratory first demonstrated the application of that CAD-MS/MS of the gas-phase RNase A-CTP com- top-down mass spectrometry, coupled with electron plex results in fragmentation of the CTP ligand and capture dissociation (ECD), for determining the binding formation of free CMP and RNase A firmly retaining a site of a small molecule ligand to a protein [8]. Earlier diphosphate group as the primary products [10]. The work by Zubarev and coworkers had suggested that electrostatic interactions between the diphosphate weak, noncovalent intermolecular bonds could be pre- group and RNase A are preserved. Such covalent-like served upon ECD [9]. We applied their initial findings strength of gas-phase electrostatic interactions provides to localize the binding site of a polyamine compound, us with potential opportunities to identify and charac- spermine, to 13 kDa ␣-synuclein [8]. Spermine had been terize binding sites of protein–ligand complexes that demonstrated previously to enhance the propensity of involve similarly strong electrostatic interactions. Here, ␣-synuclein aggregation in Parkinson’s disease. More we show how the stability of electrostatic interactions commonly employed collisionally activated dissocia- can be exploited by top-down mass spectrometry to tion (CAD) for top-down MS readily dissociated the determine the adenosine 5=-triphosphate (ATP) binding spermine ligand from the protein. However, using site of a protein kinase, adenylate kinase. ECD, the spermine ligand remains bound to specific c-/z·-product ions to localize spermine binding to the acidic C-terminal region of the protein. Thus, although Experimental ␣ the solution binding association for the -synuclein/ Materials ϳ Ϫ3 spermine complex is relatively weak (kd 10 M), ligand binding is retained in the gas phase and even Adenosine 5=-monophosphate (AMP), ATP, and ade- upon ECD [8]. Moreover, these data suggest that many nylate kinase (AK; myokinase, from chicken muscle, targeted aspects of the intermolecular associations product number M5520) were purchased from Sigma- formed initially in solution are preserved upon transi- Aldrich (St. Louis, MO, USA). All protein samples were tion to the gas phase; the spermine binding sites deter- desalted and concentrated with 10 mM ammonium mined by ESI-ECD-MS overlapped significantly with acetate buffer (pH 6.6), using centrifugal filter devices those measured by solution NMR. (10 kDa MWCO, Amicon Ultra; Millipore Corp., Bil- In this report, we extend our initial studies by lerica, MA, USA). After desalting, ATP was added to a demonstrating that top-down MS with both ECD and 5 M AK solution in ammonium acetate. The solution CAD can be used to determine the ligand binding sites protein and ligand ratio was kept at 1:1. for specific protein–ligand complexes that are particu- larly stable in the gas-phase, such as the interaction Mass Spectrometry Experiments between proteins and nucleotides. The stability of gas- phase noncovalent complexes depends highly on the A nanoESI source and Au/Pd coated borosilicate glass nature of the binding interactions [8, 10]. For example, capillaries (Proxeon Biosystems, Odense, Denmark), ribonuclease S (RNase S) is a protein complex com- with flow rate around 50 nL/min, were coupled to a posed of the 11.5 kDa S-protein and the 2.2 kDa 7-Tesla LTQ-FT Ultra mass spectrometry (Thermo S-peptide that are noncovalently bound through hydro- Fisher Scientific, San Jose, CA, USA) to acquire positive Ϫ9 phobic interactions in solution with a kd of ca. 10 M. ionization mode ESI-MS spectra. Maintaining such interactions in the gas phase is diffi- In the source region of the LTQ-FT Ultra, the capil- cult because of the extreme instability of hydrophobic lary temperature was set to 210 °C, the capillary voltage interactions in a waterless environment [11, 12]. How- was ϩ45 V, and the tube lens was set to ϩ225 V. The ever, through optimized instrumental parameters to resolution of the Fourier-transform ion cyclotron reso- minimize ion activation in the atmosphere/vacuum nance (FT-ICR) measurements was established to be interface of the ESI source, the S-protein/S-peptide 200,000 at 400 mass-to-charge ratio (m/z). The linear ion complex can be measured intact by ESI-MS [10, 13, 14]. trap (LTQ) tuning was optimized to maximize the On the other end of the stability scale, the strength of signal intensity of the ion at m/z 2160 for the 10ϩ- electrostatic interactions is significantly enhanced in the charged molecule of native AK. The target ion number absence of solvent [2, 10, 15–19]. The influence of for the LTQ and FT-ICR full MS scan was 3 ϫ 104 and solvent on electrostatic interactions and their role in 1 ϫ 105, respectively. For MS/MS experiments (e.g., noncovalent complexes has been described [11, 20, 21]. precursor ion isolation), the target ion number for the With covalent-like strength, gas-phase electrostatic in- LTQ and FT-ICR was 1 ϫ 105.