These are my notes from lecture 13 of Harvard’s Chemistry 101: Chemical Biology Towards Precision Medicine course, taught by Dr. Stuart Schreiber on October 20, 2015.
Years ago, it was thought that just creating one skeleton with three R groups and combinatorically adding in every possible R group at each position would give you all the diversity you need. We now recognize that a truly diverse chemical library must also contain skeletal diversity (topic of today’s lecture) and stereochemical diversity (topic of next lecture).
Today we’ll cover:
- Several-step build/couple/pair (B/C/P) syntheses of skeletally diverse small molecule collections.
- Cheminformatic methods to compare the 3D shapes of resulting compounds to support strategic decision-making.
- Outlines of a few particular B/C/P pathways that achieve skeletal diversity.
A review of skeletal diversity B/C/P strategies can be found in [Nielsen & Schreiber 2008]. One interesting example relying on transition metals can be found in [Kumagai 2006]. Nature produces terpenes through different cyclizations of farnesyl pyrophosphate, and this same idea has been used to generate a variety of skeletons through B/C/P [Brummond & Mitasev 2004].
For years, Mycobacterium tuberculosis was thought to be insensitive to penicillin and other β-lactam antibiotics. However, decades ago it was reported that amoxicillin, another β-lactam, is effective against TB if used in conjunction with clavulanate, a drug which inhibits the bacterial enzyme that inactivates β-lactams [Cynamon & Palmer 1983]. This combination has recently gained new attention for its ability to treat multi-drug resistant (MDR) tuberculosis as well as extensively drug-resistant (XDR) tuberculosis [reviewed in Keener 2014]. This motivates further efforts to use diversity strategies to synthesize new analogues of old antibiotics in search for ways to treat drug-resistant bacterial infections.
Principal moments of inertia (PMI) is a way to plot the 3D shapes of molecules on a triangular continiuum with vertices representing rods, spheres, and discs [Sauer & Schwarz 2003]. Further study is still needed to figure out whether 3D diversity calculated by such metrics is predictive of performance diversity.
A group at the Broad was interested in generating a diversity library optimized for probability of being blood brain barrier-permeable [Lowe 2012]. They wanted looked at already-known molecules that known to cross the BBB, as inspiration for how to design a good BBB-permeable drug. A good starting point was neurotransmitters, which of course cross the BBB. Many of these contain an azetidine, so they set about creating a whole library based on an azetidine skeleton.