Every “history of pharmaceuticals” article ever written probably mentions quinine, and well they should. (I certainly reserved an entry for it while writing my own chemical history book). It’s a classic example of a natural product drug, one that was not known to the classical Mediterranean world but was only appreciated by Europeans after they came into contact with the New World civilizations. Malaria was most certainly known to them – it was a disease thought to be cause by harmful vapors (mal aria) particularly in warm, humid areas, but there was no effective treatment in the Old World pharmacopeia. The bark of the cinchona tree, though, was something new. It had been used by natives of the Inca Empire to treat shivering, and since that was a notable feature of malaria cases, the Spanish Jesuits who brought the bark to Europe tried it for that with notable success.
And as the idea of active natural products began to take hold, quinine itself was isolated as a pure substance in 1820. Later in the 19th century, with the Plasmodium parasites were identified as the cause of the disease, it was clear that quinine was affecting the organisms directly. But how? It may come as a surprise, but despite a great deal of work that’s still been unclear – if you’re looking for one, it’s a pretty good example of how tricky drug research can be. The mode of action of any of the antimalarial drugs (of any class) has been hard to pin down, for that matter, and it’s not for lack of trying. Chloroquine is perhaps the best worked-out of the bunch, but quinine doesn’t quite seem to share its proposed mode of action.
Here, though, is a new paper with what might be an answer. A Singapore-Karolinska collaboration tackled the problem with one of the best target ID technologies available, MS-CETSA. The Nordlund lab, which spans both locations, was the source of the first CETSA paper, and they’ve continued to develop the technique. The idea is that when a ligand binds to a protein target, it stabilizes it – the complex has to be lower-energy overall, or it wouldn’t form in the first place. And if you heat up a protein with and without such a ligand, you can see the effect on its melting and denaturation (that’s the DSF assay, differential scanning fluorimetry).
Wildly, this even works in living cells, which is the CETSA assay. Heat is a real stress situation on them, since proteins get misfolded/denatured under those conditions and tend to turn insoluble. If you lyse the cells, with and without compound treatment, and look for how much soluble protein target is left, it’s a measure of how much a compound stabilized it under the heat. Combining this with modern mass spec proteomics, you can do this trick and then monitor the whole proteome (OK, a lot of it) to see which proteins were stabilized, and thus identify targets from scratch: MS-CETSA. No labels needed, real cells, whatever dosages and time points you want to try.
Doing that with Plasmodium falciparum (and with infected red blood cells) after treatment with both quinine and mefloquine identified the organism’s purine nucleoside phosphorylase enzyme (PfPNP) as a target of both drugs. These assays weren’t trivial to set up, but after a good deal of optimization they’ve provided a new look at the proteomes of both the malaria parasites and of erythrocytes. That will be an excellent baseline for further CETSA work in either system. The troubleshooting and validation took place with pyrimethamine (famous these days as Daraprim), which is largely useless in the field these days against malaria, but whose target (dihydrofolate reductase) is well characterized.
Interestingly, resistance to mefloquine isn’t mediated by mutations in that enzyme; that’s instead via a changed transporter protein that becomes available to pump them out of the organism’s cytosol. And it turns out that mefloquine is a much weaker binder than quinine itself (40 micromolar versus 30 nanomolar, in an SPR assay), so the PNP enzyme is probably mainly relevant just for quinine’s effects. A DSF screen showed that quinidine (a quinine stereoisomer) is an even weaker binder, while chloroquine, primaquine and another antimalarial (lumefantrine) don’t even seem to bind to PfPNP at all. No wonder it’s been taking so long to unravel all this.
The team was able to obtain a very nice X-ray structure of quinine bound to PfPNP itself, and that picture alone (quinine bound to its target!) is something that many people in the field wondered when (and if) they’d ever see. As it turns out, though, it isn’t a new target itself. PNP inhibitors have already been pursued as antimalarials from first principles, with the idea of targeting the parasite’s purine salvage pathway. But no one realized that the first antimalarial of them all was working through the same mechanism! Quinine itself may still have some other modes of action (all of these compounds are a mess, and if you think this is bad you should see the MOA work on artemisinin). But this certainly appears to be the main one, and I think that case can finally be closed, after hundreds of years.