Phil Baran on the Quest to Make Chemistry Boring Again

By Kyle Proffitt  

September 3, 2020 | Phil Baran of Scripps Research delivered the keynote address at the Drug Discovery Chemistry Virtual meeting last Thursday, providing a captivating mixture of hardcore synthetic chemistry and his underlying philosophy. The focus of his talk was translational chemistry, the synthetic techniques that move beyond esoteric academic pursuits to find mainstream use in creating successful pharmaceuticals.  

Baran began by reviewing the most commonly used reactions in modern drug synthesis, finding that most are “boring”, not specific to any synthetic stage, and fall primarily into the four categories of amide bond formation, reductive amination, Suzuki coupling, and Buchwald-Hartwig Ullman Goldberg reactions. Baran says this is not because medicinal chemists lack creativity but is instead an indicator that a truly “useful” synthetic reaction implies a wide scope, is simple to run, and is an essential transformation of readily available starting materials, paraphrasing Barry Sharpless. Thus, “If you can invent a reaction that becomes boring, that’s a great compliment,” according to Baran. “Some of the reactions we thought of as being exciting 20 years ago, we take for granted today”. 

Baran’s group is focused on the synthesis of marine natural products, such as palau’amine, which often requires the development of new synthetic techniques. Baran quickly recounted how a Minisci-type radical-based reaction using silver to accomplish direct arylation of electron-deficient heterocycles started a new direction for the lab that culminated in the development of zinc sulfinate salts (now known as diversinates, available from Sigma-Aldrich) that easily form radicals in the presence of Tert-butyl hydroperoxide and add to electron-deficient heterocycles. The facile nature of the reaction was demonstrated unconventionally by the difluoromethylation of caffeine in oolong tea performed in a paper cup. Baran believes industry players saw this and said to themselves “if this is working in this rather cowboy way on late-stage tea-derived material, it’s probably going to work at the end of my analog campaign to explore late-stage diversification”. 

Baran repeated at several points throughout his talk that he wants his efforts to simplify synthesis. “We don’t want to do chemistry that is gimmicky; we want to enable a transformation that couldn’t be easily or safely done before”. He showed the example of a pharma-useful, 5-step, weeklong synthetic effort to add a trifluoromethylcyclopropane group to a heterocycle with 3% yield. Working in one-electron chemistry using a sulfinate salt, the same product could be obtained in 2 steps, effectively removing the concept of stages from this synthesis.  

He spoke about radical chemistry with heterocycles and reported collaborative work with Donna Blackmond’s group to show that, contrary to some popular opinion that radical reactions are poorly controlled, substituted heterocycles have predictable and pH-modulable reactivity, work that was published in JACS (https://dx.doi.org/10.1021%2Fja406223k). He showed an example from a Gilead patent where a trifluoromethylation site could be easily predicted using Chemdraw 13C NMR shift predictions—the smallest shift indicating the most nucleophilic site is where substitution will occur. 

Reviewing the scope of these reactions seven years after their development illuminated their ubiquity in pharmaceutical patents (https://doi.org/10.1021/acs.jmedchem.8b01303). According to Baran, “What was remarkable to us, was that although the chemistry was developed from the vantage point of late-stage functionalization… people were using this at any stage.” He showed additional examples of similar chemistry to add olefins and to form carbon-nitrogen bonds, again greatly reducing steps and improving yields in select examples. 

Baran discussed his lab’s development in collaboration with Pfizer to develop strain-release amination reagents, “spring-loaded”, highly strained bridged building blocks that can be used to quickly append bioisosteres such as propellanes, azetidines, and cyclobutanes. Again, he discussed how this chemistry, using what were once considered molecular oddities and developed to work in late-stage modification, is amenable to all synthetic stages and can even be used in the extreme for bio-modification, such as selectively reacting with cysteine in a larger peptide molecule. This work found publication in Science (https://doi.org/10.1126/science.aad6252). He highlighted that the bicyclobutanes accessible with strain-release reagents are bioisosteres for acrylamide and have very similar GSH reactivity. According to Baran, “The strain-release reagents are fast ways of decorating your compounds kind of like you put ornaments on your Christmas tree.” 

The strain-release reagents were extended to the use of “housanes”, strained bridged five-membered rings that can be opened stereospecifically by amines, alcohols, thiols, acids, selenides, amides, and even carbon nucleophiles to obtain useful products. This work appeared in JACS (https://doi.org/10.1021/jacs.6b13229). Once again he shared a great example of a simple-looking aminofluro compound that previously required an 8-step synthesis but could be prepared rapidly in 2 steps using a housane.  

He highlighted other recent work involving the activation of carboxylic acids to redox-active esters for synthesizing ketones, alcohols, and amines through one-electron chemistry, work that was reported in JACS (https://doi.org/10.1021/jacs.9b02238). 

Baran spent a few minutes discussing the hot new (but incredibly old) field of electrochemistry. Baran’s group worked closely with IKA to help develop the ElectraSyn, a small benchtop device with a potentiostat for accurately applying current to enable redox chemistry that he says basically every pharma company now has on hand. “You can’t get cheaper than an electron”, but according to Baran, cost is not a sufficiently powerful motivator. “You need to tell a medicinal chemist about enablement.” An example of the utility of this setup was shown with the synthesis of a hindered ether, which Baran said could only previously be achieved after completing a miserable campaign of SN1 addition reactions over 6 days to obtain ~6% yield. Baran’s group expanded on an interrupted Kolbe-type reaction in which electrochemical oxidation frees carbocations from carboxylic acids to accomplish a direct decarboxylative etherification, and his group optimized reaction conditions to greatly improve yield and tolerate a vast range of substituents. The hindered ether shown prior could now be obtained in just 2 steps, in 15 hours, and with 51% yield. An additional example molecule’s synthesis was cut from 14 steps over 9 days with 5% yield to 3 steps, 21 hours, and 22% yield. That work was published recently in Nature (https://doi.org/10.1038/s41586-019-1539-y).  

Baran showed an expansion of this technique to create C-N bonds, trimming the synthesis of a cereblon binder compound from 6 steps in 6% yield to one step and 52% yield. Again, the method tolerates all sorts of different functional groups, and this work is not yet published. 

Finally, he shared new unpublished work in which olefins can add to ketones using electrochemical nucleophilic addition and showed a 7-step synthesis of a steroidal hedgehog signaling inhibitor being shortened to a 2-step procedure (though the Boc deprotection step is being optimized). The mechanism is still being elucidated, but they believe a ketal-type radical is forming on the carbonyl, which adds to the olefin. 

Baran summarized the synthetic technique developments of his lab and supplemented more philosophy. “It started off fundamental, we start with natural products, and it ended up being something that translates to useful chemistry” that medicinal chemists use to make useful molecules. “Radicals can be really, really useful vehicles to get you into late-stage or early-stage functionalization,” he adds. 

“Will these reactions become boring one day? I hope!” he exclaims. Maybe, “fifty years from now, we’ll look back in curiosity at chemists of our age who had stages in their syntheses instead of making everything in 1-3 steps.”