Ionization Services (pKa Analysis)

Ionization refers to the tendance of a compound to change its charge and this can arise through the reversible loss or gain of a proton (i.e., a hydrogen ion, H⁺) within a predominantly aqueous solvent system. Ionization occurs because the test compound contains at least one fragment that can chemically, and reversibly, release or gain a proton. Fragments that release a proton are categorised as acidic (e.g. carboxylic acid) and those that gain a proton are basic (e.g., primary amine). These fragments can have different tendencies (or strengths) to release or gain a proton and are quantified by measuring the associated acid dissociation equilibrium constant (K_a) which is typically reported as pK_a (i.e., -log₁₀K_a). A pK_a reflects the pH at which a particular ionizable fragment is 50% ionized. For example, benzoic acid has a carboxylic acid fragment whose pK_a is 4.2, so at pH 4.2, 50% of the benzoic acid will exist as its neutral form and 50% as the ionized form; at different pHs, these percentages vary reflecting the perturbation of the reversible chemical equilibrium due to changes in pH of the aqueous solvent system. Importantly, the reversible equilibrium reactant and product are chemically distinct identities (i.e., they differ by a single hydrogen proton) and will behave differently within a biological system, for example, a singly ionized fragment may form a stronger ligand-receptor interactions through a charged-reinforced bond compared to the neutral fragment form. Compounds that have one ionizable fragment are categorised as monoacidic or monobasic, but compounds can have multiple ionizable fragments of the same nature e.g., diacidic, triacidic, dibasic, tribasic, etc, or multiple ionizable fragments of differing nature e.g., zwitterionic or complex amphoteric.

From a drug discovery perspective, understanding a compound’s ionization profile is critical to understanding its affinity to its primary target, its affinity to off-targets and its pharmacokinetics. Typically, ionizable fragments within the pH range of approximately 2–12 are of interest — this may include acidic pK_as <12 and basic pK_as >2, although the concept of appreciably ionized is usually evoked and the cut-off of 7.4 in both cases is used to reflect whether the pK_a is significant. (Note, at physiological pH, acidic and basic pK_as <=7.4 and >=7.4, respectively, would be >=50% ionized.)

Cyprotex offers services to determine pK_as using potentiometric and spectrophotometric methods, or a combination to align to differing client needs and compound requirements.

The number of options reflect the difficulties in measuring pK_as for some compounds. The potentiometric method is applicable to a wider set of chemistries than spectrophotometric methods. However, it is an insensitive approach and requires larger samples (e.g., multiple milligrams) of the test compound to enable a measurement. Such quantity amounts can be prohibitive to a measurement in early drug discovery stages. In certain cases, spectrophotometric methods are more sensitive and rely on the molar extinction coefficient of a UV/visible chromophore structurally close to the ionizable fragment (e.g., within several bond widths). For ionizable fragments near conjugated double bonds, the release or gain of a proton will lead to a change in the said UV/visible chromophore permitting a pK_a to be determine via mathematical deconvolution of the spectrophotometric data. In situations where the ionizable fragment isn’t close to a UV/visible chromophore then a spectrophotometric method can’t be used. Where the use of a spectrophotometric method is viable, we offer two variants: a rapid and standard method. A critical difference is the number of pH data points collected during a measurement – the rapid approach collects less and the precision of a pK_a determination is lower than the standard approach.

At Cyprotex, we determine pK_as in water using a (swamping) constant ionic strength background (i.e., 0.15 M KCl) which is a pragmatic frame of reference that is generally accepted as standard within the pharmaceutical industry. (Note, this is distinct from the classical thermodynamic, zero ionic strength, frame of reference – the difference between the two frames of reference can be predicted from the Debye-Hückel equation.) A test compound needs to be soluble, in water containing 0.15 M KCl, at pHs either side of the pK_a being measured. However, the solubility of ionizable compounds is a function of pH and compounds can precipitate out of solution as the pH changes. Pragmatically, pK_a measurements need to balance the required test compound concentration to detect changes in potentiometric or spectrophotometric methods and the test compound’s intrinsic solubility (i.e., solubility of the compound's neutral form). Where aqueous solubility may be an issue, it is possible to use a co-solvent to raise the solubility of the compound to try and keep the compound in solution whilst a pK_a is being determined. This approach relies on the assumption of a linear relationship between the measured pK_a and the dielectric constant (i.e., relative permittivity) of the aqueous solvent system containing different percentages of co-solvent, enabling an extrapolation of a pK_a back to 0% co-solvent. This method is referred to as the Yasuda-Shedlovsky extrapolation approach and can be used with various organic co-solvents (note, the default used is methanol).

There is no generic approach for measuring pK_as of compounds, for certain classes of compounds spectrophotometric may work fine others may need a potentiometric approach, others may need a combination of both. It is usually possible to assess a compound’s molecular structure for ionizable fragments and the presence of associated UV/visible chromophores. However, the need to use a co-solvent is less easy to assess, which is why Cyprotex offers a pre-assessment step as part of a pK_a measurements (except for the Rapid UV method). Pre-assessments steps are intended to provide insight into the use of a co-solvent (note, the default used is methanol). Cyprotex also offers an advanced service whereby we select the appropriate method(s) to try and fully characterise a compound’s pK_a profile.

In trying to extract pK_a measurements from potentiometric and spectrophotometric data, there is a need to seed deconvolution algorithms with the correct number of pK_as (including the insignificant pK_as) along with charge type (i.e., acidic or basic). A chemical structure is not always needed but it becomes increasingly a necessity in cases where the test compound has 3, or more, ionizable centres regardless of their strength. Although a chemical structure is ideal, partial structures or a pK_a prediction may suffice albeit they may lead to data processing ambiguities that require more consideration. Where full chemical structures are provided, Cyprotex will attempt to assign measured pK_as to the fragments therein. In the scenario where Yasuda-Shedlovsky extrapolations are employed, Cyprotex will share the slope of the linear curve associated to each pK_a observed and in certain cases this slope may be indicative of a particular charge type (i.e., acid or base).