| MoKa manual | ||
|---|---|---|
| <<< Previous | Next >>> | |
Chapter 3. Background
The acid dissociation constant (pKa) measures the strength of an acid and is of general interest both in chemistry and biology. It has been suggested that 95% of all drugs are ionizable [Wells, J. I. Pharmaceutical Preformulation; Ellis Horwood Ltd.; London, 1998; p 25].
The degree of ionization controls lipophilicity and solubility, two properties widely used in pharmaceutical research to predict absorption and distribution of a compound. Therefore, the pKa is one of the most important physicochemical properties of a molecule. The ionization state is a key parameter in ADME profiling, but the effect of a charge on the biological behavior of a molecule can elicit also other types of effect. Ionizable groups affect the ability of a molecule to interact with a target. When the latter is a metabolic enzyme, pKa can be important in determining the rate and the site of metabolism. In drug formulation the ionization constant is important in choosing the correct excipient and counter-ion. In addition, pKa is often a relevant descriptor in QSAR models. All these factors explain why there is a growing interest in the development of better pKa prediction methods.
The pKa can be expressed as follows:
pKa = -log(Ka) (1) Ka = [H+][A-]/[AH] for an acid AH (2) Ka = [H+][B]/[BH+] for a base B (3) |
pH = - log10a[H+] (4) |
pH - pKa = log([A-] / [AH] ) = log( [B] / [HB+]) (5) |
- for acids, the deprotonated form (A-) will be the predominant one if pKa < pH
- for bases, the protonated form (BH+) will be the predominant one if pKa > pH
3.1. MoKa pKa prediction method
The MoKa pKa prediction method is based on a set of statistical models, and each model is specific to a different ionizable site. More than one hundred different ionizable species can be recognized and attributed to their corresponding model. The pKa values are predicted using proprietary descriptors.
The innovative procedure of MoKa, combined with a database of more than 25000 unique pKa values, is able to predict pKa with an accuracy of 0.4 pKa units for most compounds. Since MoKa was parameterized using pKa measured in the aqueous medium, the predicted pKas are water pKas.
3.2. Multiple ionization sites
MoKa handles multiprotic compounds and does not have any limits on the maximum number of ionizable sites in a molecule. Most drug-like compounds have more than one ionizable site between pH 2 and 12. MoKa is able to handle compounds with any number of ionizable sites.
When one group ionizes, the properties of other ionizable sites are affected. Thus, multiprotic compounds undergo a two steps procedure.
In the first step every pKa is computed by considering the whole molecule neutral. MoKa uses the results from this first prediction step to rank the ionizable groups of the molecule according to their increasing acidity/basicity.
In the second step, the correct ionization pattern of the molecule at different pH is known from the pKa values obtained from the first step. Thus, it is possible to include the presence of charges in different sites of a molecule when computing each pKa.
This procedure was tested both for multiprotic systems whose pKa values differ for several units and for system whose pKa values are very close (within 0.5 pKa units).
3.3. Tautomeric forms
MoKa warns of the presence of known unstable forms. Some organic compounds exhibit tautomerism, which involves a shift in the position of a hydrogen atom and of a double bond. Sometimes one form can be largely predominant (pyridones over hydroxypyridines), sometimes both forms exist in aqueous solution. Annular tautomerism is a form of tautomerism that involves the shift of a hydrogen atom bound to a ring. For example, imidazole derivatives are subjected to annular tautomerism, and the position of NH depends on the substituents in the ring.
MoKa handles tautomerism by correcting tautomeric forms that are unstable in the aqueous medium.
3.4. Covalent hydration
Covalent addition is a phenomenon whereby one molecule of water adds reversibly across C=C, C=N, or C=O bonds of a dissolved molecule. This phenomenon often involves C=N bonds of p-deficient heterocycles, such as quinoxalines and pteridines. When an ionizable compound is covalently hydrated there are important complications in the experimental determination of its pKa. MoKa warns against the possibility of covalent hydration by using a database of classes of compounds which are known to be covalently hydrated (Possible covalent hydration)
| <<< Previous | Home | Next >>> |
| Installation | MoKa graphical user interface (GUI) |