SPR kinetics
Surface plasmon resonance measures binding at a sensor surface in real time. One partner is attached to the chip, the other flows over it, and the instrument reports mass accumulating and leaving. Fitting the resulting curves gives an association rate, a dissociation rate, and their ratio, the equilibrium dissociation constant.
The appeal is that it separates the two halves of affinity. Two antibodies with the same KD can behave very differently if one binds fast and leaves fast while the other binds slowly and stays. For most therapeutic questions, the off-rate is the number that matters.
Orientation decides what you measure
The single most consequential choice is which partner goes on the surface.
Antigen on the surface, antibody flowing. A bivalent IgG flowing over a dense antigen surface can bind with both arms, which produces avidity and an off-rate that looks far slower than the intrinsic one. The fit will report a number, and the number will not be the affinity of one site.
Antibody captured, antigen flowing. Capturing the antibody, usually through an anti-Fc or a tag, and flowing monomeric antigen gives a genuinely monovalent interaction. For ranking a panel this is generally the cleaner configuration.
Whichever you choose, hold it constant across the panel. Kinetics collected in different orientations are not comparable, no matter how carefully each one was fit.
Getting a usable surface
- Keep density low. High surface density causes mass transport limitation, where the measured on-rate reflects diffusion to the surface rather than chemistry. It also promotes rebinding, which flattens the off-rate.
- Use a reference surface and subtract it. Bulk shifts and nonspecific binding are otherwise inside your data.
- Include buffer blanks in the cycle series and subtract them too.
- Check regeneration does not damage the captured protein. A capture level that drops across the run is a red flag for the whole dataset.
Fitting, and what the fit is telling you
A simple one-to-one model fits a monovalent interaction with a clean surface. When the data need a heterogeneous ligand or a bivalent analyte model to fit, that is information: it usually means the surface is heterogeneous or avidity is present, not that the molecule has two affinities.
Report the residuals, not just the constants. A fit that describes the data will have small, unstructured residuals. Systematic curvature in the residuals means the model is wrong, and the constants it produced do not mean what they are labeled.
Where the off-rate is too fast to measure, a steady-state analysis of the equilibrium response against concentration is the correct approach, and the result is a KD with no kinetics attached.
Biolayer interferometry
BLI measures a related signal on a dipped fiber optic tip rather than in a flow cell. It is faster to set up, more tolerant of crude samples, and generally less precise for fast kinetics. For ranking panels it is often the practical choice; for a final characterization, SPR remains the reference.