Guest Post: The Medicinal Chemistry Reaction Cookbook- Packards and DeLoreans
We'd like to welcome the talented SeeArrOh to The Haystack as a guest blogger. A Ph.D. chemist working in industry, you might recognize SeeArrOh from the comment threads of your favorite pharma blogs, from Twitter, or from a recent Chemjobber guest post. SAO also enjoyed the paper on what medicinal chemists actually make that caught Derek Lowe's eye yesterday. You'll notice a few similar ideas to Derek's in SAO's commentary, but a few different insights as well.
Every so often, it comes up: Is there anything really new in med-chem? Is everything just a re-hash of time-honored reactions, set up with closed eyes to produce yellow oils and white solids? Or are there untapped territories ripe for exploration? Sadly, Doc Brown can’t pull up in his fusion-powered DeLorean to tell us what cancer-curing medicines await us in fifty years, so we went to the literature instead.
As authors of a recent J. Med. Chem ASAP (DOI: 10.1021/jm200187y) muse:
…discussions with other chemists have revealed that many of our drug discovery colleagues outside
the synthetic community perceive our syntheses…[are] predominantly composed of amine deprotections to facilitate amide formation reactions, and Suzuki couplings to produce biaryl derivatives. These “typical” syntheses invariably result in large, flat, achiral derivatives, destined for screening cascades.
Our intrepid British authors Roughley and Jordan - his full name, Allen Michael Jordan, was enough to entice this reader - make a case for a wide variety of possible molecular manipulation. Focusing on the output of three med-chem titans: GSK, Pfizer, and AstraZeneca, and papers from three high-impact journals (JMC, BMC, BMCL), they crunch the numbers on drug discovery synthesis and reveal some expected—and unexpected—results. Given the limited scope, the authors admit right away that their findings might not be comprehensive. But hey, let’s dive into the data anyway!
As a bench chemist, I’ve run my fair share of metal-catalyzed couplings, salt formations, and acetylations, so I fully expected total domination of the list by these “workhorse” reactions.
Was I right? In all, the authors analyzed 7,000 different reactions, and considered some 3,600 final compounds. Perhaps most surprising, given the recent Nobel prize given out for the mighty palladium catalysts, the C-C bond-forming reactions uncovered represented only ~10% of the total! Roughly 20% of the reactions covered are “tricks of the trade” - protecting group strategies used by chemists to deactivate or cap an otherwise reactive group of atoms. Further, half of the reactions analyzed actually couple carbon to other elements such as nitrogen, sulfur, or oxygen, so-called “C-X formation”.
Only 1.5% of these reactions were oxidations, while four times as many reductions were reported. The authors speculate that heavy metals used in oxidation may discourage its use in day-to-day med chem. Even less surprising,the authors observe several ether formations (for solubilizing molecules in water) and lactam / amide formation (polar atomic groups that help solubilize and keep the molecule rigid), which grants these reactions their rightful workhorse status.
OK, if we accept that alkylations and acylations are the Fords and Packards of the work-a-day atomic architect, where are the sleek, futuristic DeLoreans? Roughley and Jordan suggest a few types of structures or molecular features that fit that bill. For example, although the always trusty benzene is found in almost 99% of the entries, the structurally intriguing fused-ring heterocycles – 2 or more aromatics linked together like those found in antimalarial medicines – are much less so. Much of medicinal chemistry revolves around making compounds that mimic those found in our cells; thus, the lack of heterocycles, which are building blocks of DNA, proteins, and signaling molecules, is perhaps even more shocking…Great Scott!
One time capsule candidate: the lonesome aliphatic fluorine. A biochemistry darling, found as a functional group in many enzyme probes, it has since fallen to a mere 0.2% of the pool. The core sulfoxide group of Prilosec, a common treatment for upset stomachs, was almost completely off the radar. The authors also dispelled a chemical Old Wives’ Tale: chirality, the inclusion of mirror-image forms of molecules, was surprisingly not avoided, but welcomed: almost 35% of the compounds contained at least one asymmetric center. This flies in the face of the stereotype that the molecules that come out of med-chem labs are always flat like pancakes, devoid of chiral centers.
Perhaps we’ll leave the final word to the guys who did the research:
Not an author. (Wikimedia Commons/Steve Lipofsky)
[A] large diversity of reaction types and functionalities are reported by medicinal chemists, although the
favoring of a small number of reliable performers is understandable given the pressures of compound delivery… clearly, medicinal chemists must continue to investigate and embrace new enabling technologies.
The authors’ scientific call to arms for those sick of the status quo? - Go try a new reaction today!