The mention of automated organic synthesis in last month’s column has no doubt prompted some people to wonder what in the world I’m talking about. When I bring up that topic, some people always turn out not to have heard of anything of the kind, while others believe that it’s been going on for years, and has pretty well reached its maturity. Neither of those, I think, are quite right.
What got me thinking about this most recently was a paper from the Burke group at Illinois. It reports a very interesting way to do palladium-catalyzed boronate couplings in an automated flow-chemistry fashion. It relies on the serendipitous discovery of a solvent-switch catch-and-release purification method on silica gel cartridges, which was the main chemical innovation in the paper—the couplings and the flow setup were well precedented. That, I think, is what makes some people wonder what, exactly, the big deal is. We certainly do enough palladium-catalyzed couplings in medicinal chemistry already. Too many, probably—a colleague of mine has proposed, only half-jokingly, that we institute a cap-and-trade system for these reactions. Every chemist in the department would be issued with credits for X number of metal-catalyzed couplings per year, and if you want to exceed your quota, you’d have to make a deal with someone who has some left.

So no, the chemical world will not be revolutionized by a new way to do Suzuki-style reactions. But I saw this paper as a harbinger, because automating these couplings allows you to assemble a large library of building blocks and just assemble them iteratively. It’s like peptide synthesis, in other words. And the big thing about peptide synthesis is that no one makes a big deal out of it— the important thing is the product. You want the synthesis, the actual process of making the protein, to be as dull, uneventful, and fast as possible, so you can get on to what really matters. That’s where large areas of organic chemistry could be headed: dull, uneventful stuff that gets done as fast as possible.

Arguably, a fair amount of medicinal chemistry already fits this description. That’s something that every beginning researcher joining the drug industry has to come to terms with. Even if you come out of a high-powered synthetic organic group, doing all sorts of exotic transformations on your last 0.5mg of material, you will find yourself doing reactions that come right out of a sophomore organic chemistry course. Sometimes they’ll come right out of the first semester of that course, because those reactions tend to work—that’s how they became classics. You do those because the point of the work isn’t the tricky, intricate chemistry; the point is to make the compounds so you can see if they do what you want them to. That’s what really matters. I’ve often said, in a thought-experiment sort of way, that if drug companies could figure out a way to make organic compounds without hiring organic chemists to make them, they’d jump at the chance.

They may get their chance to jump. Burke’s synthesis machine by itself is not going to do the trick, but if the continuing efforts to, for example, realize metal-catalyzed couplings between saturated carbons bear fruit, then stand back. At that point, a very large part of routine organic chemistry is going to get a lot more routine, and coming up with a good old-fashioned retrosynthesis is going to become a lot less of an intellectual exercise. It’s not like there aren’t going to be a lot of tough molecules to synthesize, but the world may start dividing into those tough molecules in one pile and everything else in a very much larger pile. Sorting through that second pile, which is what medicinal chemistry spends most of its time doing, will be done largely by machine. Even now, a lot of routine med-chem gets done by someone standing in front of a fume hood, but that could be the first sort of thing to disappear. That chemist may be speeding things up with solid-phase workup cartridges or the like, but nothing repetitive that you do with your hands can stand up to the ability of a good machine to do the same thing.

This is not a vision of the future that warms everyone’s heart. The academic “art of organic synthesis” folks tend to yell the loudest, but honestly, they have the least to fear. There are still plenty of tough, important molecules that no one’s going to crank out on a synthesis machine in the foreseeable future. Want to make variations on taxotere or vancomycin? You’re going to have to take those on yourself, and it’ll need all the organic synthesis skill that you can muster. No, the high end of the synthesis world will still be around. It’s the rest of it that’s going to get dull, if your focus is just on the process of making the molecules.

That focus is what’s going to have to change. The rest of organic chemistry will have to start making the transition that medicinal chemists have already made: towards seeing the chemistry as a means to an end, not an end in itself. The field, in other words, is going to turn into even more of an applied science than it is already. There are still many reactions that we don’t have that need to be discovered, of course. But when they’re discovered, they’ll most likely be bolted on to the side of the existing largely automated techniques, which will then be able to crank out still more new kinds of molecules.

This process is not new. It hit analytical chemistry a long time ago. None of us worked in the era when analysis was a handcrafted art—that’s the subject of old quantitative and qualitative analysis texts, full of mortars and pestles, burets and Buchner funnels. Modern instrumental methods wiped that stuff from the face of the earth a long time ago, and by this point, no one misses it. It’s happened at the intersection of analytical and organic chemistry, too, with the advent of infrared spectroscopy and then especially NMR. The old art, and it was an art, of structure determination by degradation is gone forever. What used to be an entire well-earned PhD could be done in a few days or less, and there were people at the time who just could not stand to make the transition.
Similarly, in molecular biology, cloning, sequencing, and purification used to be enough work by themselves to keep a team busy for a long time, but things have moved on. Now when these become a rate-limiting step, it’s a higher-level problem—the routine stuff has become so routine that no one even notices it any more. Synthesizing molecules, at least a lot of molecules, will slide into the same category.

So that’s what I’m looking out for, the transition of organic chemistry from an art to a tool. For drug discovery, as I’ve mentioned, this transition has been underway for some time, so those of us in the field should have an easier time adjusting. We’re going to have to focus on the hard problems that are still out there, the ones about which molecules to make and why, rather than how to make them. That last part has been going away, and it’s leaving us even faster than ever. The rest of it, though, is just as important as it ever was, and we’re going to have to be ready for it to become even more so.