Dissociation Kinetics of the Streptavidin–Biotin Interaction Measured Using Direct Electrospray Ionization Mass Spectrometry Analysis

Research Article


Dissociation rate constants (koff) for the model high affinity interaction between biotin (B) and the homotetramer of natural core streptavidin (S4) were measured at pH 7 and temperatures ranging from 15 to 45 °C using electrospray ionization mass spectrometry (ESI-MS). Two different approaches to data analysis were employed, one based on the initial rate of dissociation of the (S4  +  4B) complex, the other involving nonlinear fitting of the time-dependent relative abundances of the (S4  +  iB) species. The two methods were found to yield koff values that are in good agreement, within a factor of two. The Arrhenius parameters for the dissociation of the biotin–streptavidin interaction in solution were established from the koff values determined by ESI-MS and compared with values measured using a radiolabeled biotin assay. Importantly, the dissociation activation energies determined by ESI-MS agree, within 1 kcal mol–1, with the reported value. In addition to providing a quantitative measure of koff, the results of the ESI-MS measurements revealed that the apparent cooperative distribution of (S4  +  iB) species observed at short reaction times is of kinetic origin and that sequential binding of B to S4 occurs in a noncooperative fashion with the four ligand binding sites being kinetically and thermodynamically equivalent and independent.

Key words

Electrospray ionization mass spectrometry Rate constants Protein–ligand complexes 

1 Introduction

Noncovalent interactions between proteins and between proteins with other biopolymers, small molecules, or metal ions are critical to most cellular processes. The abundances of protein complexes and their lifetimes reflect the rates of the corresponding association and dissociation reactions. Quantification of the kinetic parameters—the association and dissociation rate constants (kon and koff, respectively)—under specific solution conditions (e.g., pH, temperature, ionic strength) is important in understanding the structure and function of protein complexes and is relevant to drug design [1, 2, 3]. There exist a number of established experimental techniques for measuring the rates of biochemical reactions, including association and dissociation reactions. These include surface plasmon resonance [4, 5], spectroscopic methods (e.g., atomic force spectroscopy, circular dichroism, or fluorescence-based approaches) [6, 7, 8, 9, 10], kinetic capillary electrophoresis [11, 12], radiolabeling combined with filtration/dialysis [13, 14] and NMR [15]. Many of these techniques require the labeling of one of the binding partners, their attachment to a surface or some other manipulation of the system, which can complicate the interpretation of the kinetic data and, in some instances, influence the rates of the reactions being investigated [16].

Electrospray ionization mass spectrometry (ESI-MS) has emerged as an important addition to the arsenal of techniques available for measuring the kinetics of chemical and biochemical reactions [17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28]. The ESI-MS approach is attractive as there is no requirement for labeling or immobilization since the identity of reactants and products and, possibly intermediates, can usually be established directly from the measured mass-to-charge ratios (m/z) [18, 29]. Moreover, ESI-MS analysis allows for multiple reactions to be monitored simultaneously, a feature not associated with most kinetic assays. The determination of reaction rates by ESI-MS analysis normally follows one of two general strategies: on-line (real-time) monitoring of the reaction mixture, and off-line analysis, usually following a quench step that stops the reaction. The advantage of the on-line approach is that it allows, in principle, for direct analysis of the time-dependent distribution of reactants, intermediates, and products. The minimum acquisition time for an ESI mass spectrum, which is typically in the s to min range (although it varies between instruments and the nature of the sample being analyzed), places restrictions on the speed of reactions that can be reliably analyzed using the on-line approach. For this reason, real-time ESI-MS kinetic measurements are most commonly applied to relatively slow reactions, with timescales > min. However, there are examples where ESI-MS has been successfully applied to relatively fast reactions, in the ms to s range [17, 18, 20, 22, 24, 25]. The measurement of fast kinetics typically requires the use of rapid mixing systems, such as a continuous-flow [17, 18, 20, 22], rapid quenched-flow apparatus [24], or stopped-flow [25, 26]. The off-line approach is generally easier to implement and affords greater flexibility in terms of the experimental conditions under which the reactions are carried out. For example, this approach is suitable for the analysis of enzyme kinetics under solution conditions that are not amenable to direct ESI-MS analysis, such as high concentrations of salts or non-volatile buffers (e.g., PBS, citrate, HEPES, or TRIS) that are commonly used to stabilize proteins and ensure relevance to physiological conditions. Following the quench step, the solvent composition can be altered in order to facilitate detection of reactants or products by ESI-MS. However, a limitation of the off-line approach is that information on the distribution of species present under the reaction conditions of interest may be lost.

ESI-MS has been used to study the reaction rates for a variety of non-covalent protein interactions, including protein-protein [19, 20, 30], protein–small molecule [31] and protein–metal ion complexes [32], as well as for other biological complexes, such as DNA duplexes [33, 34]. However, to the best of our knowledge, absolute values of kon and koff for protein–ligand interactions measured using this approach have not been previously reported. Here, we describe the application ESI-MS for quantifying koff for the high affinity interaction between biotin (B) and a truncated form (containing residues 13–139) of wild-type (WT) streptavidin. Streptavidin is a homotetrameric protein complex (S4) that is isolated from Streptomyces avidinii [35]. Each streptavidin subunit is organized into an eight-stranded β-barrel, with a binding site for B at one end [36]. The streptavidin–biotin interaction is one of the most stable in nature and the exceptionally high affinity (Ka of ~2.5  ×  1013 M–1 at pH 7.4 and 25 °C) arises from an unusually small dissociation rate constant (5.4  ×  10–6 s–1 at pH 7.4 and 25 °C) [13]. The origin of the slow dissociation kinetics has been the focus of many experimental and theoretical studies [13, 14, 37, 38, 39, 40]. For example, the temperature dependence of koff for the interaction between B and WT streptavidin has been compared with values measured for a variety of single site mutants in an effort to elucidate the influence of the specific amino acid side chains on the kinetic barrier to dissociation [13, 14, 41]. These data were measured using a radiolabeled B assay, whereby the release of bound B from (S4  +  4B) complex was monitored in the presence of an excess of unlabeled B. In the present study, values of koff for the sequential loss of B from the (S4  +  4B) complex, at pH 7 and temperatures ranging from 15 to 45 °C, were measured using ESI-MS. Two different approaches were used to analyze the ESI-MS data, one based on the initial rate of change in the relative abundance of the (S4  +  4B) species, the other based on nonlinear fitting of the time-dependent relative abundances of all free and B-bound S4 species. The Arrhenius parameters determined from the koff values measured by ESI-MS were compared with the values obtained using the radiolabeled B assay [14]. In addition to providing a quantitative measure of the dissociation rate constants, the direct ESI-MS measurements provide a definitive answer to the question of whether the sequential binding of B to S4 occurs in a cooperative or noncooperative fashion.

2 Experimental

2.1 Streptavidin and Biotin

The plasmid for natural core streptavidin (containing residues 13–139 of WT streptavidin, MW 13 271 Da) was a gift from Professor P. Stayton (University of Washington). Streptavidin was expressed in E. coli and purified using procedures described elsewhere [42]. Solutions of purified S4 were exchanged directly into 100 mM aqueous ammonium acetate buffer using an Amicon microconcentrator with a MW cut-off of 10 kDa and lyophilized. Stock solutions of S4 (100 μM) were prepared by dissolving a known amount of lyophilized streptavidin into 100 mM ammonium acetate and stored at –20 °C until needed. Biotin (B, MW 244.3 Da) was purchased from Sigma-Aldrich Canada (Oakville, Canada). The stock solution of B (800 μM) was prepared by dissolving B in Milli-Q water. All stock solutions were stored at –20 °C until needed.

2.2 Mass Spectrometry

All measurements were performed using an Apex Qe 9.4 T Fourier-transform ion cyclotron resonance (FTICR) mass spectrometer (Bruker, Billerica, MA, USA). Nanoflow ESI (nanoESI) was performed using borosilicate tubes (1.0 mm o.d., 0.68 mm i.d.), pulled to ~5 μm o.d. at one end using a P-97 micropipette puller (Sutter Instruments, Novato, CA, USA). The electric field required to spray the solution in positive ion mode was established by applying a voltage of 1.0–1.3 kV to a platinum wire inserted inside the glass tip. The solution flow rate was typically ~20 nL min–1. Details of the instrumental and experimental conditions typically used for quantifying protein–ligand interactions can be found elsewhere [43].

2.3 Kinetic Measurements

For the kinetic measurements, the reaction mixtures were prepared by mixing aliquots of the stock solutions to achieve the desired concentrations of S4 (10 μM), B (10–26 μM) and ammonium acetate (5 mM). The reaction mixtures were kept at constant temperature (15–45 °C) using a water bath (Colora, Germany). Aliquots of the reaction mixtures were removed at specific reaction times (t) and analyzed by ESI-MS. Three ESI mass spectra were measured for each aliquot. An acquisition time of approximately 1 min was used for each mass spectrum.

2.4 Data Analysis

Assuming that the four ligand binding sites of S4 are kinetically equivalent and independent, vide infra, the apparent rate constants for ligand association and dissociation reactions (kon,i and koff,(i + 1), respectively) for each (S4  +  iB) species Equation (1) are related to the intrinsic (microscopic) rate constants (kon and koff) through statistical factors, which reflect the number of free and occupied binding sites, Equations (2a) and (2b):
$$ \begin{array}{*{20}c} {{k_{on,i }}} \\ {\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)+\mathrm{B}\rightleftarrows \left( {{{\mathrm{S}}_4}+\left( {i+1} \right)\mathrm{B}} \right)} \\ {{k_{{off,\left( {i+1} \right)}}}} \\ \end{array} $$
$$ {k_{on,i }}=\left( {4-i} \right){k_{on }} $$
$$ {k_{{off,\left( {i+1} \right)}}}=\left( {i+1} \right){k_{off }} $$
It follows that the rate of change of the concentration of each (S4  +  iB) species can be described by Equations (3a)–(3e):
$$ \frac{{d\left( {\left[ {{{\mathrm{S}}_4}} \right]} \right)}}{dt }=-4{k_{on }}\left[ \mathrm{B} \right]\left[ {{{\mathrm{S}}_4}} \right]+{k_{off }}\left[ {{{\mathrm{S}}_4}+\mathrm{B}} \right] $$
$$ \frac{{d\left( {\left[ {{{\mathrm{S}}_4}+\mathrm{B}} \right]} \right)}}{dt }=-\left( {3{k_{on }}\left[ \mathrm{B} \right]+{k_{off }}} \right)\left[ {{{\mathrm{S}}_4}+\mathrm{B}} \right]+4{k_{on }}\left[ \mathrm{B} \right]\left[ {{{\mathrm{S}}_4}} \right]+2{k_{off }}\left[ {{{\mathrm{S}}_4}+2\mathrm{B}} \right] $$
$$ \frac{{d\left( {\left[ {{{\mathrm{S}}_4}+2\mathrm{B}} \right]} \right)}}{dt }=-\left( {2{k_{on }}\left[ \mathrm{B} \right]+2{k_{off }}} \right)\left[ {{{\mathrm{S}}_4}+2\mathrm{B}} \right]+3{k_{on }}\left[ \mathrm{B} \right]\left[ {{{\mathrm{S}}_4}+\mathrm{B}} \right]+3{k_{off }}\left[ {{{\mathrm{S}}_4}+3\mathrm{B}} \right] $$
$$ \frac{{d\left( {\left[ {{{\mathrm{S}}_4}+3\mathrm{B}} \right]} \right)}}{dt }=-\left( {{k_{on }}\left[ \mathrm{B} \right]+3{k_{off }}} \right)\left[ {{{\mathrm{S}}_4}+3\mathrm{B}} \right]+2{k_{on }}\left[ \mathrm{B} \right]\left[ {{{\mathrm{S}}_4}+2\mathrm{B}} \right]+4{k_{off }}\left[ {{{\mathrm{S}}_4}+4\mathrm{B}} \right] $$
$$ \frac{{d\left( {\left[ {{{\mathrm{S}}_4}+4\mathrm{B}} \right]} \right)}}{dt }=-4{k_{off }}\left[ {{{\mathrm{S}}_4}+4\mathrm{B}} \right]+{k_{on }}\left[ \mathrm{B} \right]\left[ {{{\mathrm{S}}_4}+3\mathrm{B}} \right] $$
where \( \left[ {{{\mathrm{S}}_4}+i\mathrm{B}} \right] \) is the concentration of the (S4  +  iB) species and [B] is the concentration of free B.
Because the rate of association of B to S4 is very high, [B] will be extremely small when substoichiometric amounts of ligand are used (which is the case in the present study). Consequently, it is reasonable to assume that [B] exists at a steady-state over the course of the entire reaction. Therefore, the ikon[B] terms can be approximated as ik1, where k1 is a pseudo first-order rate constant. Equations (3a)–(3e) can then be rewritten in terms of the normalized abundances \( \left( {{A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}}} \right) \) of the (S4  +  iB) species, Equations (4a)–(4e):
$$ \frac{{d{A_{{R\,{{\mathrm{S}}_4}}}}}}{dt }=-4{k_1}{A_{{R\,{{\mathrm{S}}_4}}}}+{k_{off }}{A_{{R({{\mathrm{S}}_4}+\mathrm{B})}}} $$
$$ \frac{{d{A_{{R\left( {{{\mathrm{S}}_4}+\mathrm{B}} \right)}}}}}{dt }=-\left( {3{k_1}+{k_{off }}} \right){A_{{R\left( {{{\mathrm{S}}_4}+\mathrm{B}} \right)}}}+4{k_1}{A_{{R{{\mathrm{S}}_4}}}}+2{k_{off }}{A_{{R\left( {{{\mathrm{S}}_4}+2\mathrm{B}} \right)}}} $$
$$ \frac{{d{A_{{R\left( {{{\mathrm{S}}_4}+2\mathrm{B}} \right)}}}}}{dt }=-\left( {2{k_1}+2{k_{off }}} \right){A_{{R\left( {{{\mathrm{S}}_4}+2\mathrm{B}} \right)}}}+3{k_1}{A_{{R\left( {{{\mathrm{S}}_4}+\mathrm{B}} \right)}}}+3{k_{off }}{A_{{R\left( {{{\mathrm{S}}_4}+3\mathrm{B}} \right)}}} $$
$$ \frac{{d{A_{{R\left( {{{\mathrm{S}}_4}+3\mathrm{B}} \right)}}}}}{dt }=-\left( {{k_1}+3{k_{off }}} \right){A_{{R\left( {{{\mathrm{S}}_4}+3\mathrm{B}} \right)}}}+2{k_1}{A_{{R\left( {{{\mathrm{S}}_4}+2\mathrm{B}} \right)}}}+4{k_{off }}{A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}} $$
$$ \frac{{d{A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}}}}{dt }=-4{k_{off }}{A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}}+{k_1}{A_{{R\left( {{{\mathrm{S}}_4}+3\mathrm{B}} \right)}}} $$
As described below, under the solution conditions used in the present study, there is a significant difference in the ligand association and dissociation rates immediately upon mixing S4 and B [9, 13, 14]. Because of this and the fact that the initial concentration of B is less than the total concentration of binding sites, mixing S4 with B initially produces a non-equilibrium distribution of (S4  +  iB), one that favors the S4 and (S4  +  4B) species. As the reaction proceeds, the relative abundance of (S4  +  4B) will decrease and system will eventually achieve an equilibrium distribution of (S4  +  iB) species. Because \( {A_{{R\left( {{{\mathrm{S}}_4}+3\mathrm{B}} \right)}}} \) is initially relatively small, the rate of change of \( {A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}} \), described by Equation (4e), can be approximated by Equation (5):
$$ \frac{{d{A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}}}}{dt}\approx -4{k_{off }}{A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}} $$
and koff can be evaluated from a linear least squares fit of the plot of the natural logarithm of \( {A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}} \) versus t, Equation (6):
$$ \ln {A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}}=-4{k_{off }}\times t+b $$
where b is a constant that is equal to \( \ln {A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}} \) at t  =  0. In this case, t  =  0 corresponds to the earliest reaction time for which ESI mass spectra were acquired. It should be noted that the magnitude of b depends on the initial concentrations of S4 and B.
An alternative approach to determining koff involves applying nonlinear regression analysis to the time-dependence of \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) for each (S4  +  iB) species. This approach is more general than the initial rate method described above since there are no simplifying assumptions needed and k1, in addition to koff, can be determined. Moreover, this approach is not limited to data measured early in the reaction. In fact, inclusion of data measured at longer times, where the system is approaching equilibrium, enhances the reliability of the fitting procedure. Expressions for the time-dependent\( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) for each (S4  +  iB) species are obtained by solving Equations (4a)–(4e) (as a system) using Maple 14 (Maplesoft, Waterloo, Canada). The experimental \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) values at t  =  0 (the earliest time point measured) served as a boundary conditions. Shown in Supplementary Data is a set of solutions (functions) corresponding to experimental data acquired at 44.8 °C. Origin (OriginLab, Northampton, MA, USA) was used to fit the functions (with koff and k1 as adjustable parameters) to the experimental breakdown curves (\( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) values plotted versus t). At each temperature, koff and k1 values were calculated for each (S4  +  iB) species; the reported koff and k1 values in Table 1 correspond to the average of these values.
Table 1

Microscopic Rate Constants (koff) for the Dissociation of the Streptavidin-Biotin Interaction at pH 7 and Temperatures Ranging from 15 to 45 °C Measured Using Direct ESI-MS Analysisa,b

[S4]o (μM)

[B]o (μM)

T (°C)

koff (s–1)a Linear fitting

koff (s–1) b Nonlinear fitting

k1 (s–1)b Nonlinear fitting




(1.1 ± 0.1) × 10–6

(1.3 ± 0.7) × 10–6

(2.3 ± 0.8) × 10–6




(4.1 ± 0.1) × 10–6

(5.5 ± 0.3) × 10–6

(2.8 ± 0.9) × 10–6




(1.8 ± 0.1) × 10–5

(2.1 ± 0.1) × 10–5

(8.7 ± 0.3) × 10–6




(5.0 ± 0.2) × 10–5

(5.1 ± 0.1) × 10–5

(1.8 ± 0.3) × 10–5




(1.6 ± 0.1) × 10–4

(2.5 ± 0.2) × 10–4

(8.4 ± 0.7) × 10–5

akoff was obtained from the slope (koff = 1/4 slope) of a linear least squares fit of the plot of the natural logarithm of \( {A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}} \) versus t. Reported errors correspond to one standard deviation

bAt each temperature, koff and k1 were obtained from nonlinear regression analysis of the time-dependent relative abundance of each of the (S4 + iB) species. The reported koff and k1 values correspond to the average of these values and errors are one standard deviation

The temperature dependence of the measured koff values was analyzed according to the Arrhenius equation, Equation (7):
$$ \ln \left( {{k_{off }}} \right)=-\frac{{{{\mathrm{E}}_{\mathrm{a}}}}}{\mathrm{RT}}+\ln \left( \mathrm{A} \right) $$

The activation energy (Ea) and pre-exponential factor (A) were calculated from the slope and intercept, respectively, of a linear least-squares fit of the plot of ln(koff) versus 1/T.

3 Results and Discussion

Shown in Figure 1a–c are representative ESI mass spectra acquired in positive ion mode for a solution (pH 7 and 22.1 °C) of S4 (10 μM) and B (14 μM) and ammonium acetate (5 mM) immediately after mixing (i.e., reaction time ~0 min), and after 112 min and 1602 min (1.1 d). In each mass spectrum, signals corresponding to the protonated (S4  +  iB)n+ ions with 0  ≤  i  ≤  4, at n  =  12 – 16, are evident. Initially, the S4n+ and (S4  +  4B)n+ ions represent the dominant species present (Figure 1a). The observation of predominantly free and fully ligand-bound protein shortly after mixing B with S4 is consistent with results of Sano and Cantor [44]. These authors used gel electrophoresis to analyze solutions of B and S4, immediately after mixing, and observed only two major bands, which corresponded to S4 and (S4  +  4B) [44]. The authors interpreted these results as evidence of cooperative ligand binding [44]. However, the observation of abundant S4 and (S4  +  4B) in the ESI mass spectra does not necessarily imply cooperative binding of B to S4. Instead, mixing of S4 and B could lead initially to a non-equilibrium distribution of streptavidin-biotin species due to the fast association kinetics and extremely slow dissociation kinetics. The latter explanation finds support in the observation that the distribution of (S4  +  iB) species changes at longer reaction times. For example, after 112 min, abundant signal is observed for all five of the (S4  +  iB) species (i.e., with i from 0 to 4) although S4 remains the dominant species detected (Figure 1b); after 1602 min, the (S4  +  B) species dominates (Figure 1c). At much longer times, a constant distribution of (S4  +  iB) species is observed, indicating that an equilibrium distribution of (S4  +  iB) species was reached. As an example, shown in Figure 1d is a representative ESI mass spectrum acquired after 13080 min (9 d). Notably, the measured distribution of (S4  +  iB) species agrees with the distribution expected in the case of four identical and independent ligand binding sites, each with a microscopic Ka of 2.5  ×  1013 M–1 (at pH 7.4 and 25 °C) (Figure 1e) [13]. These results establish, unambiguously, that B binding to S4 is not a cooperative process, in agreement with the findings of Jones and Kurzban [45] and Fidelio and co-workers [46], and that the binding sites are thermodynamically equivalent and independent. It is also concluded that the apparent cooperative distributions of the (S4 + iB) species observed at short reaction times are, in fact, of kinetic and not thermodynamic origin.
Figure 1

ESI mass spectra acquired for an aqueous ammonium acetate (5 mM) solution of S4 (10 μM) and B (14 μM) at 22.1 °C and pH 7 and different reaction times (a) 0 s, (b) 112 min, (c) 1602 min (1.1 d), and (d) 13080 min (9 d). (e) Normalized distribution of (S4 + i B) species, where i = 0 – 4, determined from the ESI mass spectrum shown in (d). The reported errors correspond to one standard deviation and were determined from three replicate measurements. Also shown is the calculated distribution for four equivalent ligand binding sites, each with a Ka of 2.5 × 1013 M–1

As noted in the Data Analysis section, two different approaches were used to quantify koff. One approach is based on the initial rate of change of \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \). Shown in Figure 2a are plots of the natural logarithm of \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) versus t, measured at the reaction temperatures indicated. Importantly, the plots exhibit excellent linearity. This result indicates that neglect of the ligand association reaction involving (S4  +  3B) to \( {A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}} \) in Equation (4e) is a reasonable assumption. The koff values calculated at each reaction temperature are listed in Table 1. Notably, the value of 5.0  ×  10–5 s–1 determined at 36.2 °C agrees very well with the reported value of 4.1  ×  10–5 s–1, which was measured at 37 °C [14]. An alternative approach used to determine koff involves nonlinear fitting of the solutions of Equations (4a)–(4e) to the time-dependent \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) values. Plotted in Figure 2b are the \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) values measured at 44.8 °C and the curves obtained from the nonlinear fitting procedure. It can be seen that the calculated curves describe the experimental data very well. The average kinetic parameters, determined at each temperature investigated, are summarized in Table 1. Notably, the koff values determined by the two different methods agree very well, within a factor of 2, at all of the temperatures investigated. For example, a koff value of 5.0  ×  10–5 s–1 was determined at 36.2 °C with the nonlinear fitting method, which is indistinguishable from the value of 5.1  ×  10–5 s–1 determined from the initial rates approach. With the nonlinear fitting method, the k1 (=kon[B]) terms were also established at each temperature investigated (Table 1). Although [B] is not accurately known (and, in fact, varies over the course of the reaction), it is nevertheless possible to estimate kon by assuming that [B] (at all reaction times) is similar in magnitude to the [B]eq, the concentration of free B at equilibrium. Following this approach, kon was estimated to be 1.3  ×  108 M–1 s–1 at 22.1 °C using a [B]eq of 2.2  ×  10–14 M, which was calculated for a solution of S4 (10 μM) and B (14 μM) and a microscopic Ka of 2.5  ×  1013 M–1 [13]. This value of kon agrees reasonably well with a value of 4.5  ×  107 M–1 s–1, which was determined from measurements carried out at ambient temperature (not specified) using droplet microfluidics integrated with a confocal fluorescence detection system [9].
Figure 2

(a) Plots of the natural logarithm of \( {A_{{R\left( {{{\mathrm{S}}_4}+4\mathrm{B}} \right)}}} \) versus reaction time measured by ESI-MS for neutral aqueous ammonium acetate (5 mM) solutions of S4 (10 μM) and B (10–20 μM) at 15.3 °C, 22.1 °C, 30.5 °C, 36.2 °C, and 44.8 °C. The solid curves represent linear least squares fits of the experimental data. (b) Plots of \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) versus reaction time measured by ESI-MS neutral aqueous ammonium acetate (5 mM) solution of S4 (10 μM), B (10 μM) at 44.8 °C. The solid curves were determined from nonlinear regression analysis to the time-dependent \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) values

Shown in Figure 3 are the Arrhenius plots constructed from the koff values determined from the ESI-MS data using the two different data analysis approaches. The corresponding Arrhenius parameters (Ea, A) are listed in Table 2. Also shown in Figure 3 is the calculated curve based on the reported activation enthalpy and entropy, for the loss of B from the (S4  +  4B) complex in aqueous solution at pH 7.4 [14]. Inspection of Figure 3 (and Table 2) reveals that the Arrhenius plots for the loss of B from the (S4  +  4B) complex from linear and nonlinear fitting are similar, with Ea values of 30.4  ±  0.7 and 31.7  ±  0.8 kcal mol–1, respectively. Moreover, the Ea values agree with the reported value of 31.0  ±  0.2 kcal mol–1 [14]. These findings indicate that both approaches to the analysis of the time-resolved ESI-MS data can provide a reliable determination of the temperature-dependence of koff. However, the nonlinear fitting approach is more general and, in principle, can be applied in cases where sequential ligand binding exhibits cooperativity or where multiple, nonequivalent binding sites are present.
Figure 3

Arrhenius plots the loss of B from the (S4 + 4B) complex at pH 7 constructed from koff values measured by ESI-MS [linear fitting (Open image in new window), nonlinear fitting (Open image in new window)] and corresponding plot (Open image in new window) calculated from the activation enthalpy and entropy reported in [14]

Table 2

Arrhenius Activation Parameters (Ea, A) for the Loss of B from the (S4 + 4B) Complex Determined from koff Values Measured by ESI-MS (Using Linear and Nonlinear Data Analysis Methods) at pH 7, and Arrhenius Parameters Measured at pH 7.4 Using a Radiolabeled Biotin Assay


Ea (kcal mol–1)

A (s–1)

ESI-MS (linear fitting)

30.4 ± 0.7a

1017.1 ± 0.5 a

ESI-MS (nonlinear fitting)

31.7 ± 0.8a

1018.2 ± 0.6 a

Radiolabeled biotin assay

31.0 ± 0.2b

1017.3 ± 0.1 b

aReported errors correspond to one standard deviation.

bValues calculated from the activation enthalpy and entropy reported in [14].

The streptavidin–biotin interaction is unusually kinetically stable and it takes several days to achieve an equilibrium distribution of (S4  +  iB) species at the temperatures investigated. However, it is important to note that this same experimental approach could be applied, in an on-line fashion, to determine koff and, in principle, kon for protein–ligand interactions that require less time to reach an equilibrium distribution. Given that ~1 min is typically required to acquire an ESI mass spectrum (with a high signal-to-noise ratio) for solutions of protein–ligand complexes [47], it should be possible to apply this approach to complexes that take >10 min (under the desired solution conditions) to reach equilibrium. Although this approach will not be suitable for all protein–ligand complexes (those that exhibit both fast association and dissociation kinetics), the kinetic parameters for many protein–ligand interactions are expected to be accessible with this technique [48, 49].

4 Conclusion

In summary, ESI-MS measurements have been used to quantify koff for the sequential loss of B from the (S4  +  4B) complex at pH 7 and temperatures ranging from 15 °C to 45 °C. Two different general strategies for data analysis were considered, one based on the initial rate of dissociation of the (S4  +  4B) complex, and the other employing nonlinear fitting of the time-dependent \( {A_{{R\left( {{{\mathrm{S}}_4}+i\mathrm{B}} \right)}}} \) values of the (S4  +  iB) species. The two methods were found to yield koff values that agree within a factor of two. Importantly, the dissociation Ea values measured by ESI-MS agree within 1 kcal mol-1 with the reported value, which was measured using a radiolabeled B assay. In addition to providing a quantitative measure of koff at the temperatures investigated, the ESI-MS measurements also revealed, unambiguously, that sequential B binding to S4 occurs in a non-cooperative fashion and that the four ligand binding sites are kinetically and thermodynamically equivalent and independent.



The authors are grateful for financial support provided by the Natural Sciences and Engineering Research Council of Canada and the Alberta Glycomics Centre. The authors acknowledge Professor P. Stayton (University of Washington) for generously providing the streptavidin plasmid.

Supplementary material

13361_2012_533_MOESM1_ESM.doc (71 kb)
ESM 1(DOC 71 kb)


  1. 1.
    Copeland, R.A., Pompliano, D.L., Meek, T.D.: Drug-target residence time and its implications for lead optimization. Nat. Rev. Drug Discov. 5, 730–739 (2006)CrossRefGoogle Scholar
  2. 2.
    Swinney, D.C.: The role of binding kinetics in therapeutically useful drug action. Curr. Opin. Drug Discovery Dev. 12, 31–39 (2009)Google Scholar
  3. 3.
    Callender, R., Dyer, R.B.: Advances in time-resolved approaches to characterize the dynamical nature of enzymatic catalysis. Chem. Rev. 106, 3031–3042 (2006)CrossRefGoogle Scholar
  4. 4.
    Daghestani, H.N., Day, B.W.: Theory and applications of surface plasmon resonance, resonant mirror, resonant waveguide grating, and dual polarization interferometry biosensors. Sensors 10, 9630–9646 (2010)CrossRefGoogle Scholar
  5. 5.
    De Crescenzo, G., Boucher, C., Durocher, Y., Jolicoeur, M.: Kinetic characterization by surface plasmon resonance-based biosensors: principle and emerging trends. Cell. Mol. Bioeng. 1, 204–215 (2008)CrossRefGoogle Scholar
  6. 6.
    Bizzarri, A.R., Cannistraro, S.: Atomic force spectroscopy in biological complex formation: strategies and perspectives. J. Phy. Chem. B 113, 16449–16464 (2009)CrossRefGoogle Scholar
  7. 7.
    Hill, J.J., Royer, C.A.: Fluorescence approaches to study of protein-nucleic acid complexation. Methods Enzymol. 278, 390–416 (1997)CrossRefGoogle Scholar
  8. 8.
    Málnási-Csizmadia, A., Pearson, D.S., Kovács, M., Woolley, R.J., Geeves, M.A., Bagshaw, C.R.: Kinetic resolution of a conformational transition and the atp hydrolysis step using relaxation methods with a dictyostelium myosin ii mutant containing a single tryptophan residue. Biochemistry 40, 12727–12737 (2001)CrossRefGoogle Scholar
  9. 9.
    Srisa-Art, M., Dyson, E.C., deMello, A.J., Edel, J.B.: Monitoring of real-time streptavidin − biotin binding kinetics using droplet microfluidics. Anal. Chem. 80, 7063–7067 (2008)Google Scholar
  10. 10.
    Sugawara, T., Kuwajima, K., Sugai, S.: Folding of staphylococcal nuclease A studied by equilibrium and kinetic circular dichroism spectra. Biochemistry 30, 2698–2706 (1991)CrossRefGoogle Scholar
  11. 11.
    Chu, Y.H., Cheng, C.C.: Affinity capillary electrophoresis in biomolecular recognition. Cell. Mol. Life Sci. 54, 663–683 (1998)CrossRefGoogle Scholar
  12. 12.
    Hall, D.R., Winzor, D.J.: Potential of biosensor technology for the characterization of interactions by quantitative affinity chromatography. J. Chromatogr. B 715, 163–181 (1998)CrossRefGoogle Scholar
  13. 13.
    Chilkoti, A., Stayton, P.S.: Molecular origins of the slow streptavidin-biotin dissociation kinetics. J. Am. Chem. Soc. 117, 10622–10628 (1995)CrossRefGoogle Scholar
  14. 14.
    Klumb, L.A., Chu, V., Stayton, P.S.: Energetic roles of hydrogen bonds at the ureido oxygen binding pocket in the streptavidin − biotin complex. Biochemistry 37, 7657–7663 (1998)CrossRefGoogle Scholar
  15. 15.
    Gizachew, D., Dratz, E.: Transferred NOESY NMR studies of biotin mimetic peptide (fshpqnt) bound to streptavidin: a structural model for studies of peptide–protein interactions. Chem. Biol. Drug Des. 78, 14–24 (2011)CrossRefGoogle Scholar
  16. 16.
    Bao, J., Krylova, S.M., Wilson, D.J., Reinstein, O., Johnson, P.E., Krylov, S.N.: Kinetic capillary electrophoresis with mass-spectrometry detection (KCE-MS) facilitates label-free solution-based kinetic analysis of protein–small molecule binding. Chem, BioChem 12, 2551–2554 (2011)Google Scholar
  17. 17.
    Konermann, L., Collings, B.A., Douglas, D.J.: Cytochrome c folding kinetics studied by time-resolved electrospray ionization mass spectrometry. Biochemistry 36, 5554–5559 (1997)CrossRefGoogle Scholar
  18. 18.
    Lee, V.W.S., Chen, Y.-L., Konermann, L.: Reconstitution of acid-denatured holomyoglobin studied by time-resolved electrospray ionization mass spectrometry. Anal. Chem. 71, 4154–4159 (1999)CrossRefGoogle Scholar
  19. 19.
    Sobott, F., Benesch, J.L.P., Vierling, E., Robinson, C.V.: Subunit exchange of multimeric protein complexes. Real-time monitoring of subunit exchange between small heat shock proteins by using electrospray mass spectrometry. J. Biol. Chem. 277, 38921–38929 (2002)CrossRefGoogle Scholar
  20. 20.
    Simmons, D.A., Wilson, D.J., Lajoie, G.A., Doherty-Kirby, A., Konermann, L.: Subunit disassembly and unfolding kinetics of hemoglobin studied by time-resolved electrospray mass spectrometry. Biochemistry 43, 14792–14801 (2004)CrossRefGoogle Scholar
  21. 21.
    Deng, G., Sanyal, G.: Applications of mass spectrometry in early stages of target based drug discovery. J. Pharm. Biomed. Anal. 40, 528–538 (2006)CrossRefGoogle Scholar
  22. 22.
    Pan, J., Rintala-Dempsey, A.C., Li, Y., Shaw, G.S., Konermann, L.: Folding kinetics of the S100A11 protein dimer studied by time-resolved electrospray mass spectrometry and pulsed hydrogen − deuterium exchange. Biochemistry 45, 3005–3013 (2006)CrossRefGoogle Scholar
  23. 23.
    Sharon, M., Robinson, C.V.: The role of mass spectrometry in structure elucidation of dynamic protein complexes. Annu. Rev. Biochem. 76, 167–193 (2007)CrossRefGoogle Scholar
  24. 24.
    Clarke, D.J., Stokes, A.A., Langridge-Smith, P., Mackay, C.L.: Online quench-flow electrospray ionization fourier transform ion cyclotron resonance mass spectrometry for elucidating kinetic and chemical enzymatic reaction mechanisms. Anal. Chem. 82, 1897–1904 (2010)CrossRefGoogle Scholar
  25. 25.
    Robbins, M.D., Yoon, O.K., Barbula, G.K., Zare, R.N.: Stopped-flow kinetic analysis using hadamard transform time-of-flight mass spectrometry. Anal. Chem. 82, 8650–8657 (2010)CrossRefGoogle Scholar
  26. 26.
    Miao, Z., Chen, H., Liu, P., Liu, Y.: Development of submillisecond time-resolved mass spectrometry using desorption electrospray ionization. Anal. Chem. 83, 3994–3997 (2011)CrossRefGoogle Scholar
  27. 27.
    Pacholarz, K.J., Garlish, R.A., Taylor, R.J., Barran, P.E.: Mass spectrometry based tools to investigate protein–ligand interactions for drug discovery. Chem. Soc. Rev. 41, 4335–4355 (2012)CrossRefGoogle Scholar
  28. 28.
    Goodlett, D.R., Ogorzalek Loo, R.R., Loo, J.A., Wahl, J.H., Udseth, H.R., Smith, R.D.: A study of the thermal denaturation of ribonuclease S by electrospray ionization mass spectrometry. J. Am. Soc. Mass Spectrom. 5, 614–622 (1994)CrossRefGoogle Scholar
  29. 29.
    Soya, N., Fang, Y., Palcic, M.M., Klassen, J.S.: Trapping and characterization of covalent intermediates of mutant retaining glycosyltransferases. Glycobiology 21, 547–552 (2011)CrossRefGoogle Scholar
  30. 30.
    Deroo, S., Hyung, S.-J., Marcoux, J., Gordiyenko, Y., Koripella, R.K., Sanyal, S., Robinson, C.V.: Mechanism and Rates of exchange of L7/L12 between ribosomes and the effects of binding EF-G. ACS Chem. Biol. 7, 1120–1127 (2012)CrossRefGoogle Scholar
  31. 31.
    Peleg-Shulman, T., Najajreh, Y., Gibson, D.: Interactions of cisplatin and transplatin with proteins: Comparison of binding kinetics, binding sites and reactivity of the Pt-protein adducts of cisplatin and transplatin towards biological nucleophiles. J. Inorg. Biochem. 91, 306–311 (2002)CrossRefGoogle Scholar
  32. 32.
    Pan, J., Konermann, L.: Calcium-induced structural transitions of the calmodulin − melittin system studied by electrospray mass spectrometry: conformational subpopulations and metal-unsaturated intermediates. Biochemistry 49, 3477–3486 (2010)CrossRefGoogle Scholar
  33. 33.
    Rosu, F., Gabelica, V., Shin-ya, K., De Pauw, E.: Telomestatin-induced stabilization of the human telomeric DNA quadruplex monitored by electrospray mass spectrometry. Chem. Commun. 2702–2703 (2003)Google Scholar
  34. 34.
    Gabelica, V., Rosu, F., De Pauw, E.: A simple method to determine electrospray response factors of noncovalent complexes. Anal. Chem. 81, 6708–6715 (2009)Google Scholar
  35. 35.
    Chaiet, L., Wolf, F.J.: The properties of streptavidin, a biotin-binding protein produced by Streptomycetes. Arch. Biochem. Biophys. 106, 1–5 (1964)CrossRefGoogle Scholar
  36. 36.
    Hendrickson, W.A., Pähler, A., Smith, J.L., Satow, Y., Merritt, E.A., Phizackerley, R.P.: Crystal structure of core streptavidin determined from multiwavelength anomalous diffraction of synchrotron radiation. Proc. Natl. Acad. Sci. U. S. A. 86, 2190–2194 (1989)CrossRefGoogle Scholar
  37. 37.
    Chu, V., Stayton, P.S., Freitag, S., Le Trong, I., Stenkamp, R.E.: Thermodynamic and structural consequences of flexible loop deletion by circular permutation in the streptavidin-biotin system. Protein Sci. 7, 848–859 (1998)CrossRefGoogle Scholar
  38. 38.
    Stayton, P.S., Freitag, S., Klumb, L.A., Chilkoti, A., Chu, V., Penzotti, J.E., To, R., Hyre, D., Le Trong, I., Lybrand, T.P., Stenkamp, R.E.: Streptavidin–biotin binding energetics. Biomol. Eng. 16, 39–44 (1999)CrossRefGoogle Scholar
  39. 39.
    Li, Q., Gusarov, S., Evoy, S., Kovalenko, A.: Electronic structure, binding energy, and solvation structure of the streptavidin − biotin supramolecular complex: ONIOM and 3D-RISM study. J. Phy. Chem. B 113, 9958–9967 (2009)CrossRefGoogle Scholar
  40. 40.
    Deng, L., Broom, A., Kitova, E.N., Richards, M.R., Zheng, R.B., Shoemaker, G.K., Meiering, E.M., Klassen, J.S.: Kinetic stability of the streptavidin–biotin interaction enhanced in the gas phase. J. Am. Chem. Soc. 134, 16586–16596 (2012)CrossRefGoogle Scholar
  41. 41.
    Piran, U., Riordan, W.J.: Dissociation rate constant of the biotin–streptavidin complex. J. Immunol. Methods 133, 141–143 (1990)CrossRefGoogle Scholar
  42. 42.
    Chilkoti, A., Tan, P.H., Stayton, P.S.: Site-directed mutagenesis studies of the high-affinity streptavidin-biotin complex: contributions of tryptophan residues 79, 108, and 120. Proc. Natl. Acad. Sci. 92, 1754–1758 (1995)CrossRefGoogle Scholar
  43. 43.
    El-Hawiet, A., Kitova, E.N., Kitov, P.I., Eugenio, L., Ng, K.K., Mulvey, G.L., Dingle, T.C., Szpacenko, A., Armstrong, G.D., Klassen, J.S.: Binding of clostridium difficile toxins to human milk oligosaccharides. Glycobiology 21, 1217–1227 (2011)CrossRefGoogle Scholar
  44. 44.
    Sano, T., Cantor, C.R.: Cooperative biotin binding by streptavidin. Electrophoretic behavior and subunit association of streptavidin in the presence of 6 M urea. J. Biol. Chem. 265, 3369–3373 (1990)Google Scholar
  45. 45.
    Jones, M.L., Kurzban, G.P.: Noncooperativity of biotin binding to tetrameric streptavidin. Biochemistry 34, 11750–11756 (1995)CrossRefGoogle Scholar
  46. 46.
    Gonzalez, M., Bagatolli, L., Echabe, I., Arrondo, J., Argarana, C., Cantor, C., Fidelio, G.: Interaction of biotin with streptavidin—thermostability and conformational changes upon binding. J. Biol. Chem. 272, 11288–11294 (1997)CrossRefGoogle Scholar
  47. 47.
    Kitova, E., El-Hawiet, A., Schnier, P., Klassen, J.: Reliable determinations of protein–ligand interactions by direct ESI-MS measurements. Are we there yet? J. Am. Soc. Mass Spectrom 23, 431–441 (2012)CrossRefGoogle Scholar
  48. 48.
    Nunez, S., Venhorst, J., Kruse, C.G.: Target-drug interactions: first principles and their application to drug discovery. Drug Discov. Today 17, 10–22 (2012)CrossRefGoogle Scholar
  49. 49.
    Tummino, P.J., Copeland, R.A.: Residence time of receptor-ligand complexes and its effect on biological function. Biochemistry 47, 5481–5492 (2008)CrossRefGoogle Scholar

Copyright information

© American Society for Mass Spectrometry 2012

Authors and Affiliations

  1. 1.Department of Chemistry and Alberta Glycomics CentreUniversity of AlbertaEdmontonCanada

Personalised recommendations