Reading Peptide Receptor Assays: Kd, Ki, IC50 and EC50

Two kinds of experiment, four numbers

Studies of receptor-active peptides usually report two kinds of measurement. Binding experiments ask how tightly a peptide occupies a receptor and give a dissociation constant (Kd), or an inhibition constant (Ki) for an unlabelled competitor. Functional experiments ask how much of a cellular response a peptide produces at each concentration and give an EC50 (potency) and a maximal effect, Emax. The IC50 sits between them: it is the concentration that reduces a signal by half, and it can come from either kind of assay.

Pharmacology has standard definitions and symbols for these quantities, set out by the international committee on receptor nomenclature.^[1]

Saturation binding and Kd

At equilibrium, simple binding follows the law of mass action. The fraction of receptors occupied is [L] / ([L] + Kd), where [L] is the free ligand concentration. Kd is therefore the concentration at which half the receptors are occupied, and lower Kd means tighter binding.

A worked example: for a peptide with a Kd of 2 nM, 0.2 nM occupies about 9% of receptors, 2 nM occupies 50% and 18 nM occupies 90%. Going from 10% to 90% occupancy needs an 81-fold range, which is why curves use a log axis and a serial dilution across several orders of magnitude.

Two design errors distort Kd. If the receptor binds a large share of the added ligand, the free concentration is lower than the nominal one (ligand depletion); and if incubation is too short, the measurement is taken before equilibrium. Reviews of equilibrium binding methods describe how to detect and correct both, and note that buffer composition and temperature also change measured affinities.^[2] A practical essay on binding assays makes the same points: vary the concentration of one partner across a range around the Kd, and show by a time course that binding has reached equilibrium.^[3]

Competition binding: from IC50 to Ki

Many peptides have no labelled version of their own, so a fixed concentration of a labelled tracer is mixed with increasing concentrations of the unlabelled peptide. The IC50 is the concentration that displaces half the specific tracer binding. Because more tracer needs more competitor to displace it, the IC50 depends on the tracer concentration and is not a constant of the peptide.

For competitive binding, the Cheng-Prusoff relationship converts IC50 into Ki, the equilibrium constant of the competitor: Ki = IC50 / (1 + [L]/Kd), where [L] is the tracer concentration and Kd its dissociation constant.^[4]

A worked example: tracer at 0.5 nM with a Kd of 0.25 nM, and a measured IC50 of 30 nM, gives Ki = 30 / (1 + 2) = 10 nM. Run the same peptide with tracer at 0.025 nM and the IC50 would be about 11 nM; at 1 nM it would be about 50 nM. Ki, not IC50, is the number to compare. The correction assumes equilibrium, no ligand depletion and a competitive mechanism; the review cited above describes an extension of the assay that can show whether an interaction is competitive or allosteric.^[2]

Functional assays: EC50, Emax and the curve fit

In a functional assay, cells expressing the receptor are exposed to a concentration series and a response such as cAMP, calcium or beta-arrestin recruitment is measured. The data are usually fitted with a four-parameter logistic equation: a bottom plateau, a top plateau (Emax), a midpoint and a slope (Hill coefficient).

Guidelines for EC50 and IC50 estimation distinguish the relative value, the midpoint between the fitted plateaus, from the absolute value, the concentration that gives 50% of the control response. A relative value is reportable only when at least two tested concentrations lie beyond each bend of the curve, and an absolute value only when at least two predicted responses fall on each side of 50%.^[5] A curve that never levels off at the top gives an unreliable midpoint.

Why EC50 is not the same as Kd

A functional response passes through a chain of amplifying steps. The operational model of agonism describes this with an affinity constant (KA), the receptor concentration and a transducer ratio, tau, that captures how efficiently occupied receptors produce a response.^[6] For the simplest form of the model, EC50 = KA / (1 + tau). With tau = 9, the EC50 is one tenth of the KA: half the maximal response is reached when only about 10% of receptors are occupied.

The consequence is that EC50 depends on the cell system. The same peptide can look more potent in a cell line with high receptor expression or with a strongly amplified readout such as cAMP, and less potent in a weakly amplified readout. Potency comparisons are most meaningful side by side in the same assay.

Log units and averaging

Potency spans orders of magnitude, so values are often given as pEC50 or pKi, the negative base-10 logarithm of the molar value.^[1] 1 nM is a pEC50 of 9.0, 3 nM is 8.52 and 30 nM is 7.52; one log unit is a ten-fold difference.

Replicate experiments are averaged on the log scale. EC50 values of 1 nM and 10 nM have an arithmetic mean of 5.5 nM, but a mean pEC50 of 8.5, which corresponds to 3.2 nM (the geometric mean).

Peptide-specific pitfalls

Most errors in peptide potency data come from the actual concentration differing from the nominal one:

• Net peptide content. A stock made from weighed powder includes water and counter-ions. If the powder is 75% peptide by mass, every concentration is 25% lower than calculated and the reported EC50 is about 1.33-fold too high (0.12 log units). See purity versus net peptide content.
• Adsorption at low concentration. At picomolar to low nanomolar levels a large share of a peptide can be lost to plastic and glass, distorting the low end of the curve. See the labware adsorption guide; a carrier protein in the assay buffer is a common countermeasure.
• Buffer and albumin. Buffer composition and temperature change measured affinities.^[2] Peptides with a fatty-acid side chain, a design that promotes albumin binding, can appear less potent as the albumin content of the buffer rises, so the albumin or serum concentration belongs in the method.
• Stability during incubation. Proteases from cells or serum can degrade a peptide during long incubations. See stability in serum and plasma.

One peptide, several receptors and pathways

Multi-receptor peptides add another layer. A pharmacology study of Tirzepatide measured binding and signalling at the GIP and GLP-1 receptors and reported that it behaved like native GIP at the GIP receptor, while at the GLP-1 receptor it favoured cAMP generation over beta-arrestin recruitment and drove less receptor internalization than GLP-1.^[7] A single EC50 per receptor would miss this pathway difference, described as biased agonism. The same questions apply to Retatrutide, studied at three receptors, and to single-receptor peptides such as Kisspeptin-10.

A record checklist

For a binding or potency value to be interpretable, record:

• Material. Sequence, modifications, how the stock concentration was determined and the labware used.
• Assay. Cell line or membrane source, readout, tracer and its Kd and concentration, incubation time and temperature, buffer, carrier protein and albumin content.
• Analysis. Equation fitted, number of concentrations and replicates, any IC50-to-Ki conversion, and log-scale averages.

This guide describes general laboratory methods and published literature only. It is not a statement about the activity or behaviour of any particular material. Terms used here are defined in the glossary, and questions can be sent to support.