Surface Plasmon Resonance and Isothermal Titration Calorimetry

Surface plasmon resonance (SPR) and isothermal titration calorimetry (ITC) are the two gold-standard orthogonal, label-free biophysical methodologies employed throughout drug discovery—from high-throughput screening hit confirmation and fragment-based drug discovery to hit-to-lead and lead optimization—to directly measure the physical interaction between a therapeutic ligand and its macromolecular target without fluorescent, radiometric, or enzymatic reporters. Whereas functional biochemical assays report a composite readout (IC₅₀) susceptible to optical artifacts and colloidal aggregation, SPR and ITC resolve the underlying physical mechanisms of molecular recognition.

Together, the two platforms provide a complete kinetic and thermodynamic profile of target engagement: SPR (alongside optical waveguide Grating-Coupled Interferometry [GCI] and Bio-Layer Interferometry [BLI]) measures real-time association and dissociation rate constants (k_on, k_off), drug–target residence time (τ), and equilibrium dissociation constants (K_d) using microgram quantities of immobilized protein, while ITC operates entirely in homogeneous solution to directly measure binding enthalpy (Δ*H*°), binding entropy (−*T*Δ*S*°), Gibbs free energy (Δ*G*°), affinity (K_d), and stoichiometry (n) in a single titration.

Kretschmann Evanescent-Field Optics and Resonance Units (RU)

In a Surface Plasmon Resonance (SPR) biosensor (pioneered commercially by Biacore in 1990), the macromolecular receptor ("ligand" in SPR terminology) is immobilized onto a 50 nm vapor-deposited gold film coated with a 100 nm carboxymethylated dextran hydrogel or streptavidin monolayer (capturing AviTag-biotinylated protein). Using the Kretschmann total internal reflection geometry, wedge-focused p-polarized monochromatic light strikes the glass–gold interface from the prism side above the critical angle. At a specific resonance angle (θ_SPR), incident photons couple into coherent collective oscillations of conduction-band electrons in the gold film—surface plasmon polaritons—producing a sharp dip in reflected light intensity.

The surface plasmon generates a non-propagating electromagnetic evanescent wave that decays with exponential distance into the aqueous microfluidic flow cell over a penetration depth of ~150 to 300 nm. When dissolved small-molecule or biologic "analyte" flows across the chip and binds to the immobilized protein, the local mass concentration within the evanescent field increases, raising the local refractive index (n_local) and shifting θ_SPR:

1 Resonance Unit (RU) = 0.0001° shift in θ_SPR ≈ 1 pg/mm² of bound surface mass

Because the theoretical maximum binding capacity (R_max) scales with the molecular weight ratio of analyte (MW_A) to immobilized protein (MW_P) and the immobilized protein density (R_L) according to R_max = (MW_A / MW_P) · R_L · v (where v is valency), immobilizing 5,000 RU of a 50 kDa protein yields a theoretical R_max of 30 RU for a 300 Da small molecule—well above modern optical noise floors (<0.03 RU in SPR and <0.001 pg/mm² in waveguide GCI).

Real-Time Binding Kinetics (k_on, k_off) and Copeland's Residence Time (τ)

Monitoring RU versus time (t) produces a real-time sensorgram comprising three phases:

1. Association Phase (analyte at concentration [L] injected at constant flow rate): Binding proceeds toward steady-state equilibrium according to the pseudo-first-order rate law:

d*R* / d*t* = k_on · [L] · (R_max − R) − k_off · R Integrating yields R(t) = R_eq · [1 − exp(−*k*_obs · t)], where the observed rate constant is k_obs = k_on · [L] + k_off.

1. Steady-State Plateau: At equilibrium (d*R*/d*t* = 0), fitting R_eq = (R_max · [L]) / (K_d + [L]) yields the equilibrium affinity for fast-exchanging fragments.
2. Dissociation Phase (switching flow back to pure running buffer, [L] = 0): Bound complex decays according to a single-exponential first-order decay:

R(t) = R₀ · exp(−*k*_off · t)

Simultaneous global nonlinear fitting of multi-concentration sensorgrams (or a single-cycle kinetic injection series) yields the association rate constant (k_on, in M⁻¹·s⁻¹)—bounded upperly by the diffusional encounter limit (~10⁸ to 10⁹ M⁻¹·s⁻¹) and slowed by conformational gating—and the dissociation rate constant (k_off, in s⁻¹), from which the equilibrium dissociation constant is:

K_d = k_off / k_on (with Complex Half-Life t_1/2 = ln 2 / k_off ≈ 0.693 / k_off)

As formulated by Robert A. Copeland, David Pompliano, and Thomas Meek (2006) in the Drug–Target Residence Time Model, the lifetime of the binary drug–target complex is quantified by the residence time (τ):

τ = 1 / k_off

In open in vivo physiological systems where systemic drug concentration in plasma fluctuates and is cleared by hepatic/renal elimination (pharmacokinetics), a drug with a slow off-rate (k_off ≤ 10⁻⁴ s⁻¹, corresponding to τ ≥ 2.8 hours)—often achieved via an induced-fit conformational lock—maintains prolonged target occupancy and pharmacodynamic efficacy long after free drug has vanished from the bloodstream (exemplified by the M₃ muscarinic antagonist tiotropium, τ > 24 h; the HER2/EGFR kinase inhibitor lapatinib, τ ≈ 300 min; and candesartan at AT₁ receptors).

Microcalorimetric Power Compensation and the Wiseman c-Value

Isothermal Titration Calorimetry (ITC) directly measures the heat (q_i) released (exothermic, Δ*H*° < 0) or absorbed (endothermic, Δ*H*° > 0) when a ligand solution in a computer-controlled rotating syringe (typically 100–300 μM) is injected in 15 to 25 small aliquots (1.5–2.0 μL each) into an adiabatic sample cell (200–1,400 μL) containing the target macromolecule (10–30 μM) held at constant temperature relative to a buffer-filled reference cell. A feedback power-compensation heater maintains zero temperature difference (Δ*T* ≈ 0) between the sample and reference cells; an exothermic binding event causes the feedback system to reduce electrical power applied to the sample cell, producing downward peaks of thermal power (μcal/s) versus time.

Integrating the area under each injection peak and normalizing by the moles of injected titrant yields a sigmoidal binding isotherm (kcal/mol of injectant versus molar ratio [L]_tot / [M]_tot). As shown by Thomas Wiseman and colleagues (1989), the shape of the ITC sigmoid is governed by the dimensionless Wiseman c-value:

c = n · [M]_tot / K_d = n · K_a · [M]_tot

where [M]_tot is the total macromolecule concentration in the cell and n is the binding stoichiometry:

• Optimal Window (10 ≤ c ≤ 500): Produces a well-defined sigmoidal curve where the initial plateau height equals Δ*H*°, the inflection midpoint on the molar-ratio axis gives the stoichiometry (n) (confirming 1:1 binding and active protein fraction), and the slope at the inflection point yields K_a = 1 / K_d.
• Very Tight Binding (c > 1,000, e.g., sub-nanomolar inhibitors): The curve becomes a vertical step function; Δ*H*° and n remain exact, while K_d is resolved by competitive displacement ITC—titrating the tight inhibitor into protein pre-saturated with a weak micromolar competitive ligand (K_d,weak), which shifts the apparent affinity into the measurable c window.

From the single titration, the complete thermodynamic signature is obtained via:

Δ*G*° = R · T · ln K_d = Δ*H*° − T · Δ*S*°

Freire's Enthalpic versus Entropic Optimization in Drug Design

By profiling historical series of FDA-approved drugs (such as HIV-1 protease inhibitors from saquinavir to darunavir, and statins from lovastatin to rosuvastatin) using ITC, biophysicist Ernesto Freire demonstrated a fundamental thermodynamic pattern in hit-to-lead and lead optimization:

• First-in-Class Drugs (Entropically Driven, −*T*Δ*S*° < 0): Early hits and first-generation clinical candidates typically achieve affinity via the hydrophobic effect—burying non-polar hydrocarbon surfaces to release ordered water molecules (favorable −*T*Δ*S*°), while exhibiting weak or even unfavorable binding enthalpy (Δ*H*° ≈ 0 or > 0). However, purely entropic hydrophobic optimization inflates lipophilicity (Log*P*), degrading aqueous solubility and target selectivity.
• Best-in-Class Drugs (Enthalpically Driven, Δ*H*° ≪ 0): Next-generation best-in-class medicines (such as darunavir, Δ*H*° = −12.7 kcal/mol, or rosuvastatin) achieve picomolar potency and high selectivity through favorable binding enthalpy driven by geometrically optimized hydrogen bonds, salt bridges, and van der Waals dispersion contacts, guided by crystallographic structure-based drug design and ITC thermodynamic profiling.

What is the "mass transport limitation" artifact in SPR, and how is it prevented?

When an immobilized protein possesses a very fast association rate constant (k_on > 10⁶ M⁻¹·s⁻¹) and is immobilized at excessively high surface density (R_L) on the sensor chip, analyte molecules near the gold surface bind faster than fresh analyte can diffuse across the microfluidic laminar boundary layer from bulk solution. During dissociation, freshly dissociated analyte molecules rebind to neighboring vacant receptors before escaping into the flow stream, artificially slowing both the apparent k_on and apparent k_off. Preventing mass transport limitation requires immobilizing the lowest possible protein density (R_max ≈ 20–50 RU) and operating at a high microfluidic flow rate (30–100 μL/min).

Why can the measured enthalpy in ITC (Δ*H*_obs) change when the assay buffer is switched from Tris to HEPES or phosphate?

If a ligand–protein binding event is coupled to the uptake or release of Δ*n*_H+ protons (for example, protonating a catalytic aspartate or ligand amine upon binding), those protons must be extracted from or donated to the conjugate acid/base of the buffer. Consequently, the observed calorimetric enthalpy includes the buffer's intrinsic enthalpy of ionization (Δ*H*_ion): Δ*H*_obs = Δ*H*_intrinsic + Δ*n*_H+ · Δ*H*_ion Because Tris has a large ionization enthalpy (Δ*H*_ion = +11.3 kcal/mol) whereas phosphate (+0.9 kcal/mol) and HEPES (+4.9 kcal/mol) have much smaller values, repeating ITC in buffers with different Δ*H*_ion values reveals the exact number of protons exchanged (Δ*n*_H+, the slope) and the true buffer-independent binding enthalpy (Δ*H*_intrinsic, the y-intercept).

How does SPR immediately distinguish a genuine stoichiometric small-molecule binder from a promiscuous colloidal aggregator?

In SPR, the theoretical maximum signal for a 1:1 interaction (R_max) is strictly bounded by the mass of immobilized protein and the molecular weight ratio of the small molecule to the protein. A genuine 1:1 ligand exhibits saturable Langmuir binding that plateaus at or below theoretical R_max with fast, clean dissociation. Conversely, a promiscuous colloidal aggregator or non-specific protein denaturant deposits multi-layered 100 nm particles onto the chip surface, generating super-stoichiometric responses (3- to 20-fold above theoretical R_max), non-saturating linear slopes, and irreversible baseline drift.

1. Copeland, R. A., Pompliano, D. L., & Meek, T. D. (2006). "Drug–Target Residence Time and Its Implications for Lead Optimization." Nature Reviews Drug Discovery, 5(9), 730–739. https://doi.org/10.1038/nrd2082
2. Wiseman, T., Williston, S., Brandts, J. F., & Lin, L.-N. (1989). "Rapid Measurement of Binding Constants and Heats of Binding Using a New Titration Calorimeter." Analytical Biochemistry, 179(1), 131–137. https://doi.org/10.1016/0003-2697(89)90213-390213-3)
3. Freire, E. (2008). "Do Enthalpy and Entropy Distinguish First in Class from Best in Class?" Drug Discovery Today, 13(19–20), 869–874. https://doi.org/10.1016/j.drudis.2008.07.005
4. Renaud, J.-P., Chung, C.-W., Danielson, U. H., et al. (2016). "Biophysics in Drug Discovery: Impact, Challenges and Opportunities." Nature Reviews Drug Discovery, 15(10), 679–698. https://doi.org/10.1038/nrd.2016.123