The ABRF-MIRG'02 study: assembly state, thermodynamic, and kinetic analysis of an enzyme/inhibitor interaction

Abstract

Fully characterizing the interactions involving biomolecules requires information on the assembly state, affinity, kinetics, and thermodynamics associated with complex formation. The analytical technologies often used to measure biomolecular interactions include analytical ultracentrifugation (AUC), isothermal titration calorimetry (ITC), and surface plasmon resonance (SPR). In order to evaluate the capabilities of core facilities to implement these technologies, the Association of Biomolecular Resource Facilities (ABRF) Molecular Interactions Research Group (MIRG) developed a standardized model system and distributed it to a panel of AUC, ITC, and SPR operators. The model system was composed of a well-characterized enzyme-inhibitor pair, namely bovine carbonic anhydrase II (CA II) and 4-carboxybenzenesulfonamide (CBS). Study participants were asked to measure one or more of the following: (1) the molecular mass, homogeneity, and assembly state of CA II by AUC; (2) the affinity and thermodynamics for complex formation by ITC; and (3) the affinity and kinetics of complex formation by SPR. The results from this study provide a benchmark for comparing the capabilities of individual laboratories and for defining the utility of the different instrumentation.

FIGURE 1

Processing SPR biosensor data. The example shows triplicate CBS binding responses collected across immobilized CA II. A: Raw data. B: Y-averaged to zero (red arrow). C: Cropped (inset: enlarged view of red box). D: X-aligned to zero (inset: enlarged view of green box). E: Referenced (inset: responses generated across an unmodified reference surface). F: Double referenced (inset: buffer responses).

FIGURE 2

Representative data sets demonstrating the use of complementary equilibrium and velocity sedimentation approaches in the molecular mass determination of CA II. Panels A and B depict sedimentation equilibrium data, C and D depict velocity whole boundary analysis, and E, F, and G represent velocity time derivative analysis. For velocity measurements, the sedimentation coefficient (s) was computed using the Svedberg equation, where 1 Svedberg unit (S) = 0.1 picoseconds (or 1 × 10⁻¹³ s). Data shown in C were analyzed using approximate solutions of the Lamm equation, whereas data shown in D were analyzed by fitting c(M) while floating f/f₀.

FIGURE 3

Molecular masses determined by sedimentation equilibrium and velocity approaches. A: Examples of equilibrium-based mass determinations by participants 1 (black) and 5 (red) for CA II samples analyzed at three concentrations (0.1, 0.3, and 0.9 mg mL⁻¹) and at three rotor speeds: 24,000 (triangles), 28,000 (squares), and 34,000 rpm (circles). B: Mean equilibrium-based masses obtained from five participants. Reported values for participants 1 and 5 each represent the mean of nine measurements (consistent with panel A), whereas a mean was derived for participant 2 from analyses performed at three different rotor speeds. Global best-fit values for data collected at various CA II concentrations and rotor speeds are shown for participants 3 and 4 (no standard errors were provided). The dotted line and gray band indicate the mean value obtained from four of the five participants (29 ± 1 kg mol⁻¹), excluding the outlying value provided by participant 2. C: Velocity-derived sedimentation coefficients measured at various CA II loading concentrations by three participants (1, 3, and 5) using several data analysis methods. Participant 1 used the LAMM method with absorbance data and considered both the whole boundary (solid squares) and the meniscus condition (open squares). Participant 3 used Svedberg with absorbance data (solid circles) and DCDT+ with interference data (open circles). Participant 5 used Sedfit (solid triangles), and reportedly deconvoluted the data into three molecular weight species, of which only the main species is plotted here. Results submitted from participant 2 were excluded because they appeared to be outliers (mean s_20,W = 3.68). Participant 4 did not report velocity measurements.

FIGURE 4

Calorimetric titration data for complex formation between CBS and CA II. For each analysis, top panels show the differential power signals recorded for CBS titrations into a cell containing CA II. Bottom panels show the same data integrated with respect to time. ΔH^ITC, K_A, and N were calculated from nonlinear least-squares analysis using a single-site binding isotherm. Participant numbering correlates with Table 2.

FIGURE 5

ITC characterization of the CBS/CA II interaction (based on Fig. 4 and Table 3). A: Stoichiometry (0.94 ± 0.15). B: Enthalpy upon binding (−10.4 ± 2.5 kcal mol⁻¹). C: Affinity [(1.00 ± 0.22) × 10⁶ L mol⁻¹]. D: molar extinction coefficient for CBS at 272 nm (1307 ± 126 L mol⁻¹ cm⁻¹). Mean values and the standard deviation for 14 determinations are represented by the horizontal dotted lines and gray bands. Error bars denote the standard deviation of nonlinear least squares analysis, except for participants 10 and 14 who reported standard deviations for replicate analyses. No errors were reported for the extinction coefficient determinations.

FIGURE 6

Correlation between the CBS/CA II binding enthalpy change measured by ITC and the calculated molar extinction coefficient for CBS at λ = 272 nm. The outlying data point at an extinction coefficient value of 1500 L mol⁻¹ cm⁻¹ was excluded from the drawing of the red trend line.

FIGURE 7

Correlation between the error in K_A and the C-value for the titration. C-values less than 20 appeared to result in greater uncertainty. The outlying value at a C-value of 70 was excluded from the drawing of the red trend line.

FIGURE 8

Twenty-four independent analyses of replicate buffer injections across unmodified sensor chips. Data were collected using BIACORE 2000 (three upper panels) and 3000 (three lower panels) platforms. Each plot shows data from three flow cells; the fourth served as a reference surface.

FIGURE 9

Examples of five independent CA II immobilization cycles using EDC/NHS-mediated amine coupling (see Methods).

FIGURE 10

Global analysis of a typical data set collected on a BIACORE 2000 instrument for CBS binding to immobilized CA II. A: Red lines represent data simulated using a simple bimolecular reaction mechanism that were superimposed onto measured responses (black lines). B: Residual plot showing the difference between measured and calculated responses.

FIGURE 11

Global analysis (red lines) of CBS/CA II binding responses (black lines). Triplicate injections of 0, 0.082, 0.25, 0.74, 2.22, 6.67, and 20.0 μmol L⁻¹ CBS were analyzed across differing capacity CA II surfaces. Data were collected on a range of BIACORE systems [S51 (1–3), 3000 (4–30), 2000 (31–55), 1000 (56–58), X (59)] and an IASYS biosensor (60).

FIGURE 12

Kinetic rate and equilibrium dissociation constants for CBS binding to immobilized CA II (based on Figure 11). Horizontal dotted lines and gray bands represent mean values ± standard deviations for 59 determinations. Global analysis yielded an affinity of 0.90 ± 0.22 μmol L⁻¹ from the kinetic rate constants, k_a = (4.0 ± 0.7)× 10⁴ L mol⁻¹ s⁻¹ and k_d = 0.036 ± 0.007 s⁻¹. IASYS data (white bars) were excluded due to an outlying value for k_d (0.07 ± 0.01 s⁻¹), which led to a large error in determining affinity (K_D = 2.3 ± 0.6 μmol L⁻¹; truncations are marked by asterisks). BIACORE instruments are color-coded: S51 (green), 3000 (black), 2000 (blue), 1000 (red), and X (yellow).

FIGURE 13

Correlation between the level of immobilized CA II and its binding capacity for CBS, as given by the R_max values obtained from data analysis (Fig. 11). The red trend line falls below the dotted line, which represents the theoretical correlation.

FIGURE 14

van’t Hoff analysis of SPR-derived affinities for the CBS/CA II interaction. A: A typical data set showing overlays of globally simulated data (red lines) and measured data (black lines) at various analysis temperatures. B: van’t Hoff analysis of six affinity determinations at each temperature yielded ΔH = −10.6 ± 1.4 kcal mol⁻¹ from the slope.

FIGURE 15

Comparative ITC- (red) and SPR- (green) derived affinities for the CBS/CA II binding interaction. Horizontal dotted lines and bands indicate affinities of 1.00 ± 0.22 μmol L⁻¹ (ITC, n = 14) and 0.90 ± 0.22 μmol L⁻¹ (SPR, n = 59), where standard deviations represent n replicate determinations.