Critical evaluation of mass spectrometric measurement of dissociation constants: accuracy and cross-validation against surface plasmon resonance and circular dichroism for the calmodulin-melittin system

We present a comprehensive study for determining the binding affinity of a protein-ligand complex, using mass spectrometric methods. Mass spectrometry has been used to study noncovalent interactions for a number of years. However, the use of soft ionization mass spectrometry for quantitative analysis of noncovalently bound complexes is not widely accepted. This paper reports a comparison of MS methods against established methods such as surface plasmon resonance (SPR) and circular dichroism (CD) whose suitability for the quantitative assessment of noncovalent interactions is well known. ESI titration and MALDI-SUPREX were used as representative mass spectrometric methods for this work. We chose to study the calmodulin-melittin complex that presents three challenges: (i) it exhibits a high affinity (low nanomolar KD); (ii) complexes are formed only in the presence of a coactivator, calcium ions in this case; and (iii) the protein and the complex show a different ionization efficiency. Dissociation constants were obtained from each method for the selected system and compared thoroughly to elucidate pros and cons of the selected methodologies in terms of their ability for the determination of binding constants of protein-ligand complexes. ESI titration, SPR, CD and MALDI-SUPREX yielded KD values in the low nanomolar range that are in general agreement with an older value reported in the literature. We also critically evaluated the limitations in particular of the MS methods and the associated data evaluation procedures. We present an improved evaluation of SUPREX data, as well as a detailed error analysis for all methods used.     

Evaluation of automated nano-electrospray mass spectrometry in the determination of non-covalent protein-ligand complexes

The use of electrospray ionization mass spectrometry (ESI-MS) for studying non-covalent interactions between macromolecules and ligands is well established. ESI-MS can be a useful tool for the determination of dissociation constants between molecules in the gas phase. We validate this method by studying the binding of the catalytic domain of cellobiohydrolase I (CBH I) from Trichoderma reesei to the disaccharide inhibitor cellobiose. The method was further applied to study two newly synthesized cellobiose derivatives (m-iodobenzyl 2-deoxy-2-azido-beta-cellobioside and p-benzyloxybenzyl beta-cellobioside). In a titration experiment, peak areas of different charge states of the free enzyme and the complex were summed in order to determine the dissociation constant. For cellobiose and m-iodobenzyl 2-deoxy-2-azido-beta-cellobioside, the calculated values are in good agreement with those reported from either displacement titration or equilibrium binding experiments in solution. Due to non-specific binding, the dissociation constant of p-benzyloxybenzyl beta-cellobioside does not correspond with the solution-based value. Our results indicate the need for careful interpretation of data sets when using nanoESI to study non-covalent interactions.     

Application of RP-TLC technique for the determination of dissociation constants of 1-substituted pyrrolidin-2-one derivatives

The procedures for measuring dissociation constants (pK(a)) of twenty 1-substituted pyrrolidin-2-one derivatives are described. The dissociation constants of the compounds tested were determined using potentiometric titration, reversed-phase thin-layer chromatography (RP-TLC) and calculated using Pallas and Marvin programs. It was found that the RP-TLC method of determination of pK(a) could be considered as a feasible alternative to potentiometric titration. The Marvin program is a better tool for preliminary estimation of dissociation constant than the Pallas one.     

A simultaneous determination method for 5‐fluorouracil and its metabolites in human plasma with linear range adjusted by in‐source collision‐induced dissociation using hydrophilic interaction liquid chromatography–electrospray ionization–tandem mass spectrometry

We applied a new technique for quantitative linear range shift using in‐source collision‐induced dissociation (CID) to complex biological fluids to demonstrate its utility. The technique was used in a simultaneous quantitative determination method of 5‐fluorouracil (5‐FU), an anticancer drug for various solid tumors, and its metabolites in human plasma by liquid chromatography–electrospray ionization–tandem mass spectrometry (LC/ESI‐MS/MS). To control adverse effects after administration of 5‐FU, it is important to monitor the plasma concentration of 5‐FU and its metabolites; however, no simultaneous determination method has yet been reported because of vastly different physical and chemical properties of compounds. We developed a new analytical method for simultaneously determining 5‐FU and its metabolites in human plasma by LC/ESI‐MS/MS coupled with the technique for quantitative linear range shift using in‐source CID. Hydrophilic interaction liquid chromatography using a stationary phase with zwitterionic functional groups, phosphorylcholine, was suitable for separation of 5‐FU from its nucleoside and interfering endogenous materials. The addition of glycerin into acetonitrile‐rich eluent after LC separation improved the ESI‐MS response of high polar analytes. Based on the validation results, linear range shifts by in‐source CID is the reliable technique even with complex biological samples such as plasma. Copyright © 2016 John Wiley & Sons Ltd.     

Mass spectrometric studies of alkali metal ion binding on thrombin-binding aptamer DNA

The binding sites and consecutive binding constants of alkali metal ions, (M⁺ = Na⁺, K⁺, Rb⁺, and Cs⁺), to thrombin-binding aptamer (TBA) DNA were studied by Fourier-transform ion cyclotron resonance spectrometry. TBA-metal complexes were produced by electrospray ionization (ESI) and the ions of interest were mass-selected for further characterization. The structural motif of TBA in an ESI solution was checked by circular dichroism. The metal-binding constants and sites were determined by the titration method and infrared multiphoton dissociation (IRMPD), respectively. The binding constant of potassium is 5–8 times greater than those of other alkali metal ions, and the potassium binding site is different from other metal binding sites. In the 1:1 TBA-metal complex, potassium is coordinated between the bottom G-quartet and two adjacent TT loops of TBA. In the 1:2 TBA—metal complex, the second potassium ion binds at the TGT loop of TBA, which is in line with the antiparallel G-quadruplex structure of TBA. On the other hand, other alkali metal ions bind at the lateral TGT loop in both 1:1 and 1:2 complexes, presumably due to the formation of ion-pair adducts. IRMPD studies of the binding sites in combination with measurements of the consecutive binding constants help elucidate the binding modes of alkali metal ions on DNA aptamer at the molecular level.     

Quantitative evaluation of noncovalent chorismate mutase-inhibitor binding by ESI-MS

Electrospray time-of-flight mass spectrometry was used to quantitatively determine the dissociation constant of chorismate mutase and a transition state analogue inhibitor. This system presents a fairly complex stoichiometry because the native protein is a homotrimer with three equal and independent substrate binding sites. We can detect the chorismate mutase trimer as well as chorismate mutase-inhibitor complexes by choosing appropriate conditions in the ESI source. To verify that the protein-inhibitor complexes are specific, titration experiments with different enzyme variants and different inhibitors were performed. A plot of the number of bound inhibitors versus added inhibitor concentration revealed saturation behavior with 3:1 (inhibitor:functional trimer) stoichiometry for the TSA. The soft ESI conditions, the relatively high protein mass of 43.5 kDa, and the low charge state (high m/z) result in broad peaks, a typical problem in analyzing noncovalent protein complexes. Due to the low molecular weight of the TSA (226 Da) the peaks of the free protein and the protein with one, two or three inhibitors bound cannot be clearly resolved. For data analysis, relative peak areas of the deconvoluted spectra of chorismate mutase-inhibitor complexes were obtained by fitting appropriate peak shapes to the signals corresponding to the free enzyme and its complexes with one, two, or three inhibitor molecules. From the relative peak areas we were able to calculate a dissociation constant that agreed well with known solution-phase data. This method may be generally useful for interpreting mass spectra of noncovalent complexes that exhibit broad peaks in the high m/z range.     

Complexation of Hg2+ with α‐Lipoic and Dihydrolipoic Acids: Study by Differential Pulse Voltammetry on Rotating Au‐Disk Electrode and ESI‐MS

The complexation of the natural antioxidants α‐lipoic acid (ALA) and its reduced form dihydrolipoic acid (DHLA) with Hg²⁺ was investigated by a recently proposed differential pulse voltammetric (DPV) method using the rotating Au‐disk electrode. Complexation processes are proposed from the multivariate curve resolution by alternating least squares (MCR‐ALS) analysis of DPV titration data. Main complexes were both 1 : 1 Hg : ALA and Hg : DHLA, although the formation of 1 : 2 complexes can be also deduced. ALA and DHLA show different Hg²⁺‐binding patterns at different pH. Voltammetric findings are completed with the data obtained by electrospray ionization mass‐spectrometry (ESI‐MS), especially in negative mode.     

Optimization of Electrospray Ionization by Statistical Design of Experiments and Response Surface Methodology: Protein–Ligand Equilibrium Dissociation Constant Determinations

Electrospray ionization mass spectrometry (ESI-MS) binding studies between proteins and ligands under native conditions require that instrumental ESI source conditions are optimized if relative solution-phase equilibrium concentrations between the protein–ligand complex and free protein are to be retained. Instrumental ESI source conditions that simultaneously maximize the relative ionization efficiency of the protein–ligand complex over free protein and minimize the protein–ligand complex dissociation during the ESI process and the transfer from atmospheric pressure to vacuum are generally specific for each protein–ligand system and should be established when an accurate equilibrium dissociation constant (K_D) is to be determined via titration. In this paper, a straightforward and systematic approach for ESI source optimization is presented. The method uses statistical design of experiments (DOE) in conjunction with response surface methodology (RSM) and is demonstrated for the complexes between Plasmodium vivax guanylate kinase (PvGK) and two ligands: 5′-guanosine monophosphate (GMP) and 5′-guanosine diphosphate (GDP). It was verified that even though the ligands are structurally similar, the most appropriate ESI conditions for K_D determination by titration are different for each.

Open image in new window

Graphical Abstract ?

     

Potentiometric titrations using dual microband electrodes.

We present a novel method of analysis using potentiometric end-point detection and dual microband electrodes in generator-collector mode. The titrant is electrogenerated using either a constant current or a current that increases linearly with time, and the fluxes of reactive material rather than molar amounts are balanced. The advantage of the ramp current system over the constant current method is that all of the information needed for a full titration curve and a proper analysis can be obtained during a single scan. The method was applied to the determination of vitamin C with ferricyanide and to the determination of thiosulfate and sulfite with iodine using gold microband electrodes. As a new type of analysis, a potentiometric titration that uses dissolving silver microband electrodes in order to generate the titrant is demonstrated. The system was applied to the detection of chloride, iodide, thiosulfate and cyanide. The accuracy of the analysis is +/-10%, limited mainly by the present screen-printing process, and the method is well adapted for measurements on the millimolar scale.     

Significance of data treatment and experimental setup on the determination of copper complexing parameters by anodic stripping voltammetry

Different procedures of voltammetric peak intensities determination, as well as various experimental setups were systematically tested on simulated and real experimental data in order to identify critical points in the determination of copper complexation parameters (ligand concentration and conditional stability constant) by anodic stripping voltammetry (ASV). Varieties of titration data sets (Cu_measuredvs. Cu_total) were fitted by models encompassing discrete sites distribution of one-class and two-class of binding ligands (by PROSECE software). Examination of different procedures for peak intensities determination applied on voltammograms with known preset values revealed that tangent fit (TF) routine should be avoided, as for both simulated and experimental titration data it produced an additional class of strong ligand (actually not present). Peak intensities determination by fitting of the whole voltammogram was found to be the most appropriate, as it provided most reliable complexation parameters.Tests performed on real seawater samples under different experimental conditions revealed that in addition to importance of proper peak intensities determination, an accumulation time (control of the sensitivity) and an equilibration time needed for complete complexation of added copper during titration (control of complexation kinetics) are the keypoints to obtain reliable results free of artefacts.The consequence of overestimation and underestimation of complexing parameters is supported and illustrated by the example of free copper concentrations (the most bioavailable/toxic specie) calculated for all studied cases. Errors up to 80% of underestimation of free copper concentration and almost two orders of magnitude overestimation of conditional stability constant were registered for the simulated case with two ligands.     

New methods for data analysis of isothermal titration calorimetry

Heat divided by ligand concentration vs. heat, similar to the Scatchard plot, was introduced to obtain the equilibrium constant (K) and the enthalpy of binding (DH) using isothermal titration calorimetry data. Values of K and DH obtained by this linear pseudo-Scatchard plot for a system with a set of independent binding sites (such as binding fluoride ions on urease and monosaccharide methyl a-D-mannopyranoside on concavalin A) were remarkably like that obtained from a normal fitting Wiseman method and other our technical methods. On applying this graphical method to study the binding of copper ion on myelin basic protein (MBP), a concave downward curve obtained was consistent with the positive cooperativity in the binding. A graphical fitting by simple method for determination of thermodynamic parameters was also introduced. This method is general, without any assumption and restriction made in previous method. This general method was applied to the product inhibition study of adenosine deaminase. This revised version was published online in July 2006 with corrections to the Cover Date.     

Identification of the oxidation products of cysteamine and cystamine by CE-MS interfaced by a noncommercial electrospray ionization source

Sodium cysteamine phosphate is a prodrug derivative of cysteamine that can be used in cystinosis treatment. Although titrimetric assays are very well established and precise, iodimetric determination of sodium cysteamine phosphate requires considerably more carefulness and time from the analyst than usual. The possibility to assess sodium cysteamine phosphate by CE was evaluated by means of the quantification of its oxidation product, cystamine, which is a more suitable substance to be used as primary standard than sodium cysteamine phosphate. Apparently, this approach should be straightforward, but systematic differences between the results obtained with CE and titrimetric assays were noticed. MS and CE-MS were employed to aid in the investigation of the possible causes of imprecision of the sodium cysteamine phosphate titration and CE determination. For this purpose, a simple and inexpensive ESI source was constructed. It was observed that cystamine is not the final product of the cysteamine and/or sodium cysteamine phosphate iodine-oxidation and other species besides cystamine may be formed depending on the reaction conditions, which explains the difficulties observed in the sodium cysteamine phosphate quantification.     

Spectroscopic analysis of the interaction between bilirubin and bovine serum albumin

The interactions between bilirubin (BR) and bovine serum albumin (BSA) have been studied by fluorescence spectroscopy. The association constant between BR and BSA was obtained by fluorescence enhancement titration. Furthermore, fluorescence quenching was studied at different temperatures, and the binding constant was also determined by the method of fluorescence quenching. The two methods yielded similar results. It indicated that the former method could be successfully applied to the determination of BR. The results showed that the binding of BR to BSA induced conformational changes in BSA. Based on the theory of F?rster energy transfer, the distance between BR and protein were calculated. According to the thermodynamic parameters, the main binding force could be judged. The experimental results revealed that BSA and BR had strong interactions. The mechanism of quenching belonged to static quenching and the main sort of binding force was van der Waals interactions and hydrogen bonds.     

Application of a simple calorimetric data analysis on the binding study of calcium ions by human growth hormone

A simple graphical linear method was introduced for isothermal titration calorimetric data analysis in the protein-ligand interaction. The number of binding sites, the dissociation binding constant and the molar enthalpy of binding site can be obtained by using this new isothermal titration calorimetric data analysis method. The method was applied to the study of the interaction of human growth hormone (hGH) with divalent calcium ion at 27°C in NaCl solution, 50 mM. hGH has a set of three identical and independent binding sites for Ca ²⁺ . The intrinsic dissociation equilibrium constant and the molar enthalpy of binding are 52 μMand -17.4, respectively. Results obtained by this new calorimetric data analysis are in good agreement with results obtained using our previous method.     

Applications of electrospray ionization mass spectrometry to neutral organic molecules including fullerenes

The use of electrospray ionization mass spectrometry (ESI/MS) for the detection of neutral organic molecules becomes possible by their derivation with specific ESI/MS tagging reagents that have either proton or metal ion binding sites. We used the neutral crown ether group in several reagents to attach a metal binding site to substrate molecules. Application of this method to steroids, amino acids, vitamin D, fatty acids, and fullerenes is described. Besides characterization, tagged molecules can be used for studying organic reactions by ESI/MS. This work demonstrates that ESI/MS provides a unique window on fullerene solution chemistry. ESI/MS is not only an excellent tool for the analysis of biopolymers but is also useful for studying the organic chemistry of small neutral molecules.     

Protein-peptide affinity determination using an H/D exchange dilution strategy: Application to antigen-antibody interactions

A new methodology using hydrogen/deuterium amide exchange (HDX) to determine the binding affinity of protein-peptide interactions is reported. The method, based on our previously established approach, protein ligand interaction by mass spectrometry, titration, and H/D exchange (PLIMSTEX) [J. Am. Chem. Soc. 2003, 125, 5252–5253], makes use of a dilution strategy (dPLIMSTEX) for HDX, using the mass of the peptide ligand as readout. We employed dPLIMSTEX to study the interaction of calcium-saturated calmodulin with the opioid peptide β-endorphin as a model system; the affinity results are in good agreement with those from traditional PLIMSTEX and with literature values obtained by using other methods. We show that the dPLIMSTEX method is feasible to quantify an antigen-antibody interaction involving a 3-nitrotyrosine modified peptide in complex with a monoclonal anti-nitrotyrosine antibody. A dissociation constant in the low nanomolar range was determined, and a binding stoichiometry of antibody/peptide of 1:2 was confirmed. In addition, we determined that the epitope in the binding interface contains a minimum of five amino acids. The dPLIMSTEX approach is a sensitive and powerful tool for the quantitative determination of peptide affinities with antibodies, complementary to conventional immuno-analytical techniques.     

Quantitative analysis of carbohydrate-protein interactions using glycan microarrays: determination of surface and solution dissociation constants

Carbohydrate-protein interactions on surface and in solution were quantitatively measured by a glycan microarray. Assessing carbohydrate affinities is typically difficult due to weak affinities and limited sources of structurally complex glycans. We described here a sensitive, high-throughput, and convenient glycan microarray technology for the simultaneous determination of a wide variety of parameters in a single experiment using small amounts of materials. Assay systems based on this technology were developed to analyze multivalent interactions and determine the surface dissociation constant (KD,surf) for surface-coated mannose derivatives with mannose binding lectins and antibodies. Competition experiments that employed monovalent ligands in solution yielded KD and Ki values in solution similar to equilibrium binding constants obtained in titration microcalorimetry and surface plasmon resonance experiments.     

Self-assembled metallocycles with two interactive binding domains

Five metallocycles 1 a-e have been self-assembled from S-shaped bispyridyl ligands 2 a-e and a palladium complex, [Pd(dppp)(OTf)(2)] (dppp=1,3-bis(diphenylphosphanyl)propane), and have been characterized by elemental analysis and various spectroscopic methods including (1)H NMR spectroscopy and electrospray ionization (ESI) mass spectrometry. These metallocycles all are monocyclic compounds, but can fold to generate two binding domains bearing hydrogen-bonding sites based on pyridine-2,6-dicarboxamide units. The binding properties of the metallocycles with N,N,N',N'-tetramethylterephthalamide (G) have been probed by means of ESI mass spectrometry and (1)H NMR spectroscopy. The results both in the gas phase and in solution are consistent with the fact that the metallocycles accommodate two molecules of the guest G. Thus, the ESI mass spectra clearly show fragments corresponding to the 1:2 complexes in all cases. (1)H NMR studies on 1 a and G support the formation of a 1:2 complex in solution; the titration curves are nicely fitted to a 1:2 binding isotherm, but not to a 1:1 binding isotherm. In addition, a Job plot also suggests a 1:2 binding mode between 1 a and G, showing maximum complexation at approximately 0.33 mol fraction of the metallocycle 1 a in CDCl(3). The binding constants K(1) and K(2) are calculated to be 1600 and 1400 M(-1) (+/-10 %), respectively, at 25 degrees C in CDCl(3), indicative of positively cooperative binding. This positive cooperativity was confirmed by the Hill equation, affording a Hill coefficient of n = 1.6. Owing to insufficient solubility in CDCl(3), for comparison purposes the binding properties of the metallocycles 1 b-e were investigated in a more polar medium, 3 % CD(3)CN/CDCl(3). (1)H NMR titrations revealed that the metallocycles all bind two molecules of the guest G with Hill coefficients ranging from 1.4 to 1.8. This positive cooperativity may be attributed to a structural reorganization of the second binding cavity when the first guest binds to either one of the subcavities present in the metallocycles.     

Calorimetric determination of surfactant/polyelectrolyte binding isotherms

Mixing of oppositely charged surfactants and polyelectrolytes in aqueous solutions leads to cooperative surfactant adsorption onto the polyelectrolyte chains. Experimental determination of surfactant/polyelectrolyte binding isotherms is usually done using custom-built surfactant-ion-specific electrodes. As an alternative, we present an indirect isotherm approximation method that uses conventional isothermal titration calorimetry (ITC). The calorimetric data is fitted to the two-binding-state Satake-Yang adsorption model, which quantifies the extent of binding in terms of the binding constant (Ku) and the cooperativity parameter (u). This approach is investigated using two surfactant/polyelectrolyte mixtures: sodium perfluorooctanoate (FC7) and N,N,N-trimethylammonium derivatized hydroxyethyl cellulose (UCARE Polymer JR-400), whose binding behavior follows the Satake-Yang model, and dodecyltrimethylammonium bromide (DTAB) and poly(styrenesulfonate) (NaPSS), whose behavior deviates dramatically from the Satake-Yang model. These studies demonstrate that, in order to apply the indirect ITC method of binding isotherm determination, the surfactant/polyelectrolyte adsorption process must have no more than two dominant binding states. Thus, the technique works well for the FC7/JR-400 mixture. It fails in the case of the DTAB/NaPSS adsorption, but its mode of failure offers insight into the multiple-binding-state adsorption mechanism.     

The acid dissociation constants of 2,3-quinoxalinedithiol

Sodium salts of 2,3-quinoxalinedithiol have been prepared and used for determination of the acid ionization constants by potentiometric titration. The fully protonated form of the reagent was used in an alternative determination of the first ionization constant by a solubility method. The constants are pK(2), == 6.84 +/- 0.04 and pK(2) = 9.95 +/- 0.03.