Skip to Content

Life in the Drug Labs

The Good New Days

I’ve been working on a longer post on antisense drugs (which, in case you’re wondering, are different – most of the time – from nonsense drugs), but it’s not quite ready yet. Home life and the Wonder Drug Factory are keeping me hopping these days.
The other day I was talking about old-fashioned reactions that we still use all the time, and I can testify to that from recent experience. I’ve been messing around with Grignard reagents all week, for one thing. But one of the comments to that post mentioned that I shouldn’t give people the impression that those reactions are all that we use, and that’s a good point.
For example, I’ve said before that if I had to pick one type of reaction that’s run every day now that wasn’t well-known when I was in graduate school, it would be a palladium-catalyzed coupling. The most commonly run is the Suzuki reaction, and we’ve been doing those all week, too. I would absolutely hate to do without this family of carbon-carbon bond forming methods – in fact, it’s hard to imagine how we ever did.
But even though it’s been around since 1979, not many people ran these reactions in the mid-1980s when I was in grad school. Palladium chemistry was seen as this exotic stuff that did weird things, and did them mostly in the hoods of organometallic chemists. As that decade wore on, though, the Suzuki and other such couplings became better known, and they just completely conquered the world in the 1990s. Now there are whole sections of the chemical catalogs devoted to them. You probably could buy about a half-dozen arylboronic acids back in 1985, but an entire industry has sprung up around such things now. And since new improvements and extensions keep coming all the time, the catalogs will surely look even stranger in another fifteen years.
My only real complaint about the Suzuki reaction is that there are so many ways to run it – solvents, catalysts, additives, temperature. You can generally get it to give you some product, no matter what conditions you choose. But in many cases, optimizing it to a reproducible high yield is like black magic. My pet theory is that any given palladium coupling reaction can be made to run in over 90% yield, if you’re just willing to devote enough of your life to finding out how. Most of the time, I take what they give me and move on.

11 comments on “The Good New Days”

  1. Grubbs the cat says:

    even worse is the Heck reaction – reactants and conditions are all over the place. I remember trying a few ‘orthogonal’ conditions for one transformation in my PhD thesis. None of them gave me anything so I decided the reaction would probably not stand a great chance (although I only tried maybe 0.1% of conditions out there).

  2. tom bartlett says:

    Kudos to Buchwald and Hartwig aminations– relatively fewer widely used variants and MUCH greater substrate independance/reliability than other Pd couplings, IMHO. A workhorse.

  3. Don B. says:

    How about asking your readers for a reference/citation to their most successful set of conditions?
    Don B.

  4. Chemgrad says:

    As an organometallic chemist, U have to get the conditions dead on to get a high yield in the coupling rxns.It does not matter if it is Heck, Suzuki, Buchwald-hartwig, Negishi, Sonogashira and so on. A lot of the time, when one tries to scale up, the yield of the reaction is greatly affected.Also, with so many variations of each coupling available, one has to scrutinize every new variation which is time consuming to say the least.My best bet is scifinder and look for the substrate (that u need) and search if it has been reported and if is not, then try the one that closely resembles it. Hey, if a coupling has been reported with a methyl substrate, then there is no reason why the HECK, an ethyl, propyl or an iso-propyl shud not work.

  5. Alex says:

    The Heck reaction was a workhorse of the synthesis that was my thesis project. I was (un)fortunate enough to have a substrate that, if modified slightly, required a completely different set of reaction conditions for optimisation, which were different enough from most literature conditions to require a lot of time and work. Rationalizing the result was even more frustrating. Even more fun is when people seem mystified (including my advisor and prospective employers) that I spent so much time working on optimizing a “textbook” reaction. The Heck is truly one of those reactions which doesn’t seem to be well understood.

  6. The Novice Chemist says:

    Any comments about Stille reactions? Or solvent quality? (DMF is both popular and problematic…)

  7. Derek Lowe says:

    We try not to run Stille reactions in the drug industry, because the tin is so hard to get rid of (and so undesirable as a residue in a drug). I don’t think I’ve ever run one, come to think of it, although I’ve done all the other common couplings.
    As for solvent quality, that’s another variable to chase down. We try to run things in fairly innocuous stuff like toluene/dioxane or toluene/ethanol – there’s not as much quality variation in those. But you’re right about DMF; it’s a fine solvent that varies all over the place in its water (and dimethylamine) content. Some people make it worse while attempting to make it better through too-vigorous distillation, too.

  8. Milo says:

    You are right about tin impurities being a pain in the rear to remove. I think I lost a good 2-3 months in grad school just trying to remove the tin by-products.

  9. Process Chemist says:

    “My pet theory is that any given palladium coupling reaction can be made to run in over 90% yield”
    I have yet to see an example where that theory fails. We just finished developing a palladium coupling that took two months to go from 30% yield at 110C to 90% yield at room temperature, just by tweaking ligands, metals and additives.
    My pet theory is that Med Chemists fall in love with the most elaborate set of conditions for any given reaction and keep using them all the time… until some even more elaborate conditions involving a more expensive reagent come up in JACS.
    Around here, Med Chemists seem to be in love with EDC couplings. Then it is our job to remind them that, most of the time, a methyl ester can be made by cooking the carboxylic acid in methanol.

  10. Derek Lowe says:

    Guilty as charged, PC!

Comments are closed.