Binding Assay Formats: What Each One Measures

Three assays all described as binding assays measure three different things. One measures how much compound is bound at equilibrium, one measures how fast it associates and dissociates, and one measures whether it displaces something else. A single affinity figure quoted without its format has discarded the information needed to interpret it.

Saturation and competition

In a saturation experiment, increasing concentrations of labelled ligand are added to a fixed quantity of receptor until binding plateaus. The curve yields a dissociation constant and the total number of binding sites. It requires a labelled version of the compound itself.

In a competition experiment, a fixed concentration of a labelled reference ligand is displaced by increasing concentrations of the unlabelled test compound. The midpoint of the displacement curve yields an IC50, which is converted to an inhibition constant using the reference ligand’s own affinity and concentration.

Competition is what is usually run, because it requires no labelled version of the test compound. Its output therefore depends on a conversion that assumes simple competition at a single site — an assumption that fails where binding is allosteric or where more than one site exists, and the distinction between the resulting constants is the subject of Ki, IC50 and EC50.

Non-specific binding, which defines the answer

Labelled ligand sticks to the receptor, to everything else in the preparation, and to the tube. Specific binding is defined as total binding minus what remains in the presence of a large excess of unlabelled ligand.

That definition makes non-specific binding an operational quantity rather than a physical one, and it can be manipulated: the choice of blocking agent, of filter, and of wash conditions all change it. Where non-specific binding is a large proportion of the total, the specific signal is a small difference between two large numbers, and its uncertainty is correspondingly large.

Peptides are unusually prone to this because they adsorb readily, for the reasons in adsorptive loss to surfaces. A reported binding curve should be accompanied by what proportion of total binding was specific; below about half, the numbers deserve little weight.

Radioligand, fluorescence and surface-based formats

• Radioligand binding remains the reference method. The label is small and does not alter binding behaviour, which is its main advantage. It requires radioactive material handling and a separation step to remove unbound ligand.
• Fluorescence polarisation reads the change in tumbling rate when a small labelled ligand binds a large receptor. It needs no separation step, and it works best when the size difference is large — which makes it poorly suited to a small peptide binding a small target. The fluorophore is also large relative to a short peptide and can alter its binding.
• Surface plasmon resonance immobilises one partner and measures mass accumulating at the surface in real time, giving association and dissociation rates rather than only an equilibrium constant. The immobilisation is its limitation: attaching the receptor or the peptide at a particular point may block the surface that does the binding.

Why kinetics carry information an affinity figure does not

An equilibrium constant is a ratio of two rates. Two compounds with identical affinities can differ greatly in how long they remain bound, and a method that reports only the ratio cannot distinguish them.

This is worth knowing when comparing analogs, because a modification intended to improve binding may have changed the off-rate, the on-rate, or both, in ways an affinity figure conceals. It is also the reason a single constant from a competition assay is a thinner description of a compound than it appears.

What the preparation contributes

Binding is measured in membranes, in whole cells, or with purified receptor, and the answers differ. Membrane preparations lose the cellular context that influences receptor conformation. Whole cells retain it and add uptake and degradation. Purified receptor removes both and may not fold as it does in a membrane.

Buffer composition matters as much: ionic strength, pH, and the presence or absence of divalent cations and guanine nucleotides all shift apparent affinity at a G-protein-coupled receptor. Two laboratories using different buffers will obtain different constants from the same compound and the same receptor — a specific instance of the broader problem in why two laboratories get different results.

Equilibrium, which the experiment assumes

Every equilibrium constant assumes the reaction reached equilibrium before it was measured. That takes time, and the time required is set by the off-rate — a slowly dissociating compound may need hours.

An experiment stopped too early returns an apparent affinity that is weaker than the true one, and the error is larger for exactly the compounds that bind most tightly. The check is to confirm that binding no longer changes with further incubation, and a methods section that states an incubation time without showing it was sufficient has left the assumption untested.

What binding does not tell you

That a compound binds says nothing about what happens next. Agonists, antagonists and inverse agonists can all bind with similar affinity, and only a functional readout distinguishes them — which is why binding and function are complementary rather than alternative measurements, and why a compound characterised only by binding has been half-characterised.

Selectivity claims have the same limitation. A compound described as selective for one receptor subtype has usually been tested against a subset of the family, and the honest form of the claim names which subtypes were tested and which were not, as discussed for the melanocortin family in MC1R and MC4R selectivity.