Tautomers need some love

Now here's a paper about something that every college student knows about and which is yet not considered by people who do drug design as often as it should- tautomers. Yvonne Martin (previously at Abbott) has a nice article about why tautomers are important in drug design and what are the continuing challenges in predicting and understanding them. This should be a good reminder for both experimentalists and theoreticians to consider tautomerism in their projects.

So why are tautomers important? For one thing, a particular tautomer of a drug molecule might be the one that binds to its protein target. More importantly, this tautomer might be the minor tautomer in solution, so knowing the major tautomer in solution may not always help determine the form bound to a protein. This bears analogy with conformational equilibria in which the conformer binding to a protein more often than not is a minor conformer. Martin illustrates some remarkable cases in which both tautomers of a particular kinase inhibitor were observed in the same crystal structure. In many cases, quantum chemical calculations indicate a considerable energy different between the minor protein-bound tautomer and its major counterpart. A further fundamental complication arises from the fact that solvent changes hugely impact tautomer equilibria, and not enough data is always available on tautomers in aqueous solution because of problems like solubility.

Thus, predicting tautomers is crucial if you want to deal with ligands bound to proteins. It is also important for predicting parameters like logP and blood brain barrier penetration which in turn depend on accurate estimations of hydrophobicity. Different tautomers have different hydrophobicities, and Martin indicates that different methods and programs can sometimes calculate different values of hydrophobicity for a given tautomer, which will directly impact important calculations of logP and blood-brain penetration. It will also affect computational calculations like docking and QSAR where tautomer state will be crucial.

Sadly, there is not enough experimental data on tautomer equilibria. Such data is also admittedly hard to obtain; the net pKa of a compound is a result of all tautomers contributing to its equilibrium, and the number of tautomers can sometimes be tremendous; for instance 8-oxoguanine which is a well known DNA lesion caused by radiation can exist in 100 or so ionic and neutral tautomers. Now let's say you want to dock this compound to a protein to predict a ligand orientation. Which tautomer on earth do you possibly choose?

Clearly calculating tautomers can be very important for drug design. As Martin mentions, more experimental as well cas theoretical data on tautomers is necessary; however such research, similar to solvation measurements discussed in a past post, usually falls under the title of "too basic" and therefore may not be funded by the NIH. But whether funded or not, successful ligand design cannot prevail without consideration of tautomers. What was that thing about basic research yielding its worth many times over in applications?

Martin, Y. (2009). Let’s not forget tautomers Journal of Computer-Aided Molecular Design DOI: 10.1007/s10822-009-9303-2