Our latest paper, which arose from a collaboration with the radiochemistry team at BMS, describes a new approach to carbon-14 radiolabeling and is now published in JACS (after previous submission to the ChemRxiv). I began working on this project as a visiting student in October 2017 and it has been an amazing learning experience with some tough challenges and valuable lessons. In this blog post I want to discuss some of the interesting problems we encountered upon diving into the field of radiochemistry.
To summarize, our synthetic approach consists of the formation of a redox-active ester followed by a chemoselective nickel-catalyzed decarboxylative carboxylation to reintroduce the (now radiolabeled) carboxylic acid. That carbon-14 label is a useful tool for radiochemists to perform the absorption, distribution, metabolism, excretion and toxicity studies required for drug development. Our method has some nice advantages including the two-step degradation-reconstruction synthetic approach (starting from the unlabeled product which is usually available in plentiful supply) and the use of [14C]-CO2, the cheapest and most common carbon-14 reagent.
During the initial reaction development, we successfully performed the desired decarboxylative carboxylation sequence using 50 psi of [13C]-CO2. Although this procedure allowed us to test the scope of the reaction (scheme 2 in the paper), our collaborators let us know the pressure and use of excess labeled reagent were big problems when it came to translating to radiolabeled [14C]-CO2.
We first considered limiting the stoichiometry of the [14C]-CO2 to about 10 equivalents while maintaining the pressure in the system. Several options were considered included pressurizing with an inert gas (discounted due to the dependency of the concentration of CO2 on its partial pressure, rather than the overall pressure), pressurizing with unlabeled CO2 (avoided because it would dilute the radiolabel too much) or using amine additives to sequester the CO2 into solution and promote reactivity (no reaction observed presumably due to coordination to/deactivation of the catalyst). Thinking further about the factors which influence the concentration of a gas in solution we realized it may be possible to accomplish our goal by simply cooling the reaction down. Gases are more soluble at colder temperatures (when most gases dissolve in solution the process is exothermic, an increased temperature causes an increase in kinetic energy which promotes gas escape from solution). Through use of Henry’s law, the concentration of CO2 in solution under the pressurized [13C]-CO2 conditions was estimated. With this figure as a target, the temperature required to obtain a similar concentration at atmospheric pressure of CO2 was calculated to be about -25 °C.
Running the reaction at -25 °C gave, perhaps unsurprisingly, only partial conversion of the starting material to the product. Using 10 equivalents of CO2 at atmospheric pressure and maintaining the cryogenic period for just one hour then sealing the vessel, warming to room temperature and stirring overnight gave full conversion of the sm with efficient isotopic incorporation (see the SI S12-S14 for further details)! The BMS team were now satisfied with this procedure and after making a few minor adjustments, they produced a working carbon-14 setup using standard radiochemical glassware and techniques. Upon testing the method with [14C]-CO2 we observed isotope incorporations suitable for use in all preclinical and clinical ADME studies and the project could be called a success. All in all, application of the method to the strict requirements of carbon-14 setting led us to some fun and interesting challenges and provided the answer to the age-old question…
Although we think Breeder Steve may have been considering one particular decarboxylation process, we can conclusively say the answer is yes #chemistrygoals! I would like to thank everyone in the lab for all their help on this project and please check out the complementary carbon isotope exchange methods just published by the Shultz group at Merck and the Audisio group at Université Paris-Saclay.
Cian Kingston
La Jolla, CA
January, 2019
Cool stuff. Curious to see how else a radical thus generated could be used!
ReplyDelete