Dissociation Constants and Affinity Scales

Why Affinity Constants Need Careful Interpretation

A binding constant is a compact summary of how a ligand and target are distributed between unbound and bound forms at equilibrium. It is useful for comparing ligands, interpreting equilibrium occupancy, and communicating a concentration scale. But the number is meaningful only when its definition, units, and experimental context are clear. In particular, the concentration used in the equilibrium relation is the free ligand concentration, not automatically the amount of ligand added to an assay.

This tutorial considers reversible binding of one ligand to one target site. It focuses on the equilibrium dissociation constant \(K_d\), the association constant \(K_a\), and the limits of interpreting them as measures of affinity. These equilibrium constants do not by themselves describe how quickly binding occurs or how long a ligand remains bound.

Definitions and the Reciprocal Relationship

Let \(R\) denote unbound target, \(L\) unbound ligand, and \(RL\) the target–ligand complex. For a reversible 1:1 interaction at equilibrium, the concentration-based dissociation constant is defined by the equilibrium concentrations of these three species.

Here, brackets denote concentration, and \([L]_{\mathrm{free}}\) is the concentration of unbound ligand. With all concentrations expressed in molar units, \(K_d\) has units of concentration, such as M, μM, or nM. A smaller \(K_d\) means that a lower free ligand concentration is associated with a given equilibrium ratio of bound to unbound target; in the usual sense, it indicates higher affinity.

The association constant is the reciprocal concentration-based constant:

Under this convention, \(K_a\) has units of inverse concentration, such as M\(^{-1}\). A larger \(K_a\) indicates higher affinity. The units matter: a value reported as \(K_d=10\) nM is not numerically interchangeable with \(K_a=10\) M\(^{-1}\). Convert to a common concentration unit before taking a reciprocal or comparing values.

These definitions assume the system has reached equilibrium; the binding is reversible; the interaction is adequately represented by one ligand binding one target site; and concentrations represent the relevant binding species. The interpretation also assumes that experimental conditions such as temperature, pH, buffer, and target state are specified and sufficiently comparable. In practice, concentration-based constants treat concentrations as proxies for chemical activities. The approximation and the assay method can matter when comparing measurements across substantially different conditions.

For this simple binding model, fractional occupancy \(\theta\), the fraction of target sites occupied, is related to free ligand concentration by:

This expression makes the scale interpretation explicit: when free ligand concentration equals \(K_d\), occupancy is one-half. Thus \(K_d\) is the free ligand concentration associated with 50% occupancy in this equilibrium model. This does not mean every assay reporting a 50% effect measures \(K_d\); the measured quantity and model must actually correspond to equilibrium binding occupancy.

Converting Constants and Comparing Affinity

To convert \(K_d\) to \(K_a\), first express \(K_d\) in the concentration unit whose reciprocal is desired. For example, \(1\) nM is \(10^{-9}\) M, so a \(K_d\) written in nM must be converted to M before reporting \(K_a\) in M\(^{-1}\). Alternatively, retain reciprocal units explicitly: \(K_d=20\) nM corresponds to \(K_a=0.05\) nM\(^{-1}\), which is \(5.00\times10^7\) M\(^{-1}\).

Comparisons are easiest when all \(K_d\) values use the same unit. For example, \(0.2\) μM is \(200\) nM; it is not smaller than \(50\) nM simply because the number “0.2” is smaller than “50.” On the \(K_d\) scale, smaller values mean stronger affinity. On the \(K_a\) scale, the order reverses: larger values mean stronger affinity.

A logarithmic scale is also common. Define \(pK_d=-\log_{10}(K_d/\mathrm{M})\), where the ratio inside the logarithm is dimensionless. Because of the negative sign, larger \(pK_d\) means smaller \(K_d\), and therefore higher affinity. A one-unit increase in \(pK_d\) corresponds to a tenfold decrease in \(K_d\). The logarithmic scale can make comparisons across orders of magnitude easier, but it does not change the underlying definition.

Free Concentration, Total Concentration, and Occupancy

The binding equation uses \([L]_{\mathrm{free}}\), because only unbound ligand is available to bind at that instant. Total ligand concentration includes ligand in all relevant forms, including free and bound ligand. Consequently, the total concentration added to an assay need not equal the free concentration during the measurement. They may be close when binding removes only a negligible fraction of ligand from solution, but that cannot be assumed from the reported dose alone.

When free concentration is measured or otherwise appropriately estimated, it can be used in the occupancy equation. Substituting total concentration in its place without justification can produce a misleading occupancy estimate. A complete treatment of how free and total concentrations are linked requires a binding mass balance; the key point here is that they are distinct quantities.

Affinity Is Not Assay Potency

An assay potency value describes the concentration associated with a specified assay response, such as an \(IC_{50}\) or \(EC_{50}\). A binding \(K_d\), by contrast, describes an equilibrium relationship between free ligand, unbound target, and complex under a binding model. Potency and affinity may be related in a particular experimental system, but they are not interchangeable definitions.

A functional response can depend on more than target occupancy, including the relationship between occupancy and downstream response, assay conditions, and the response endpoint. A displacement assay’s \(IC_{50}\) is also an assay-dependent quantity; it should not be identified with \(K_d\) without the assumptions and conversion appropriate to that assay. Therefore, a potency ranking can differ from a binding-affinity ranking without either measurement being inherently wrong.

Diagnostics, Limitations, and a Common Pitfall

Before comparing affinity estimates, check that the values refer to the same target form and a comparable binding model, and that units, temperature, and relevant assay conditions are reported. Inspect whether the method estimates an equilibrium \(K_d\) or reports another endpoint. If the analysis relies on nominal or total ligand concentration, ask whether it is a reasonable proxy for free concentration; if not, the resulting estimate may not represent the defined equilibrium constant. A fitted value should also be interpreted cautiously if the data do not meaningfully constrain the concentration range around the binding transition.

The simple \(K_d\) relationship has limitations. It summarizes equilibrium binding under a specified model; it does not reveal association or dissociation rates, prove that a one-site model is adequate, or determine a downstream effect. Comparisons across different targets or assay environments may reflect differences in conditions or model assumptions as well as differences in ligand binding. For those reasons, report the measured quantity, units, experimental context, and model rather than presenting a bare number as a universal property.

Common mistake: “Compound A has a \(K_d\) of 5 nM and compound B has a \(K_d\) of 50 nM, so B has tenfold higher affinity because its number is larger.” This reverses the \(K_d\) scale. The correction is that smaller \(K_d\) means higher affinity: A’s \(K_d\) is tenfold lower. The opposite numerical direction applies to \(K_a\), where larger values indicate higher affinity.

A second frequent error is to calculate occupancy from the amount of ligand added, without establishing that this is the free concentration. Use \([L]_{\mathrm{free}}\) in the occupancy expression, and distinguish a binding constant from a functional \(EC_{50}\) or assay \(IC_{50}\). Keeping the definition attached to each number prevents most interpretation errors.