The field of late-stage modification of complex structures has seen several advances in the last few years, and this is just the sort of thing that medicinal chemists tend to be interested in. We like one-step transformations (which organic chemist doesn’t!) and we tend to have large collections of compounds that have already been made. The idea of running a bunch of these past some new reagent and producing a whole new pile of chemical matter is quite appealing.
For that to work out, several conditions have to be met, though. The reaction conditions have to be compatible with a huge variety of functional groups, and ideally they have to be able to transform a wide variety of molecules as well. And finally, these new reactions have to make something interesting. To pick a couple of examples, going through the collection with a big bottle of acetic anhydride is probably not the answer. You’ll acetylate OH and NH groups, which are pretty common, but the products you get are either not likely to be good drug leads (O-acetyls get cleaved readily) or are already represented in the collection (N-acetyls probably have amides of the same sort already made). On the other hand, if you have a reagent that selectively turns tertiary CH groups into CF, that’s worth a look. There are probably quite a few of those compounds in your collection, and the odds of their fluorinated analogs being in there are very small. Once fluorinated, the compound properties are likely to change quite a bit as well, so I would definitely be up for a mild, selective, high-yield fluorination pass.
This new paper (a free PDF may be available here) is a pretty wild addition to the toolbox. Thomas Hoye of Minnesota (and co-author Sean Ross) are applying the group’s favorite hexadehydro Diels-Alder benzyne reaction to natural product structures. Benzynes are pretty hot intermediates, and HDDA is an unusually mild route into them. The classic benzyne reactions are low-temperature affairs with honking strong bases, which is not what you’d want to expose a complex natural product to, but they can handle the thermal HDDA conditions. And being reactive creatures, the benzyne intermediates can form products with phenols, alcohols, amines, furans, carboxylic acids, and more, and they can also do interesting thing like unzip the DABCO bicyclic amine to a piperazine.
A couple of these products are shown from the reaction with estradiol and its acetate, with the HDDA additions in red. That piperazine one is actually produced in 80% yield, which is remarkable. Similar thinks happen with Vitamin E and other phenols. The real fireworks start when more complex natural products are exposed to the HDDA conditions, because all sorts of unusual ring openings and rearrangements start up. Reserpine, for example, is already a fairly large and complex alkaloid, but I can guarantee that you haven’t seen anything like the ten-membered ring that comes out of its reaction under these conditions. (“Nuc” is one of several traps, for example, a benzotriazole, and that product forms in about 35% yield). The paper shows unusual products from quinine, scopolamine, brucine and others, but I’d be all morning trying to draw them.
I think it’s safe to say that this paper only begins to show the variety of things that can be produced under such conditions – they would yield a compound library like no other, even if you don’t go up to the brucine-functionalization level. Just HDDA-ing a range of simpler substrates would give you a lot of odd variety, it seems safe to assume. Not all of them would be suitable additions to a serious drug screening library, but a surprising number of them would be worth a look, and you’d most certainly be exploring new chemical space. This work is another reminder that such space is large indeed.