MOST people do their best to avoid the manchineel tree. Even sheltering under it during a rain shower could give you a nasty surprise. Its sap is toxic and water-soluble, so any drips bouncing off the trunk and on to your skin will yield blisters. And don’t even think about eating its small round fruit – in central America, where it grows, it is sometimes known as the “little apple of death”.
Phil Baran isn’t like most people. For him, the toxin the tree secretes, called phorbol, is an intriguing curiosity. It exemplifies the elaborate chemical structures and interesting effects that naturally produced molecules can have. Give phorbol a few tweaks and it morphs from a toxin into a potential anticancer agent. Other natural products might be the basis of the next great pill, perfume, plastic or pigment.
Advertisement
To study them we must first make and manipulate these natural products in the lab. And that’s where synthetic chemists such as Baran run up against a gnarly problem. Natural chemicals put up a barrier in the form of a particular atomic bond, one that crops up in all of them and is incredibly tough to break. Until now, that is: Baran and a few others think they have a way to break the unbreakable and usher in a new era of chemical alchemy.
Synthetic chemistry is all about breaking old bonds and making new ones: it is responsible for most of the human-made stuff we see around us. And the bonds we most want to make are between one carbon atom and another. That’s because carbon chemistry is the molecular language of living things, and so it is also the language medicines must understand.
To bond carbon to something, you must first remove one of the atoms already attached to it. Chances are that it will be a hydrogen atom. Ever since carbon was forged within the first stars, it has been pairing up with hydrogen, the most abundant element in the universe. Wherever you find carbon in the natural world, it will probably be nestled inside a hydrogen blanket.
Take crude oil, our favourite raw material for making drugs and much more. It is based almost exclusively on carbon and hydrogen. The clue is in the name we use for many of its ingredients: hydrocarbons. Their long chains of atoms can be chopped into smaller, workable fragments using an industrial process called catalytic cracking, which involves heating them up beyond 500°C. But this can’t break carbon-hydrogen bonds in the controlled way needed for useful chemistry.
Molecular rubble
The problem is that hydrogen, once attached, is supremely difficult to dislodge. Chemical bonds are often the result of two atoms sharing a pair of electrons, and because hydrogen is the smallest atom, it gets particularly close to carbon and forms a particularly strong bond.
“If you look in a textbook, you’ll find it’s a very thin chapter on CH bonds because there’s very little chemistry that can be done on them,” says , a chemist at Emory College in Atlanta, Georgia. Until recently, one of the few strategies we had for breaking CH bonds was combustion, which leaves molecular rubble that’s no good for making anything.
There is a more subtle option. Under normal circumstances, breaking a bond means shifting its two electrons to one of the atoms, while the other gets nothing. This creates two electronically charged fragments. But there is a fundamentally different sort of split where each atom gets one electron, generating uncharged fragments called radicals. In the case of the CH bond, you can do this by adding chlorine gas and shining particular frequencies of light on the chlorine. It will then rip away the hydrogen along with one of its electrons, leaving a carbon radical. This is then primed to react to form a fresh bond. The trouble is, radicals are flightly critters that will react quickly with almost anything. Chemical chaos is the result.
That leaves chemists in an sticky situation. It’s a bit like plotting a route across a sprawling city without being able to make any right turns. Whole blocks of the metropolis seem impossible to reach unless you execute an elaborate series of left turns to loop around to the waypoint a single right turn would have got you to. Not being able to manipulate CH bonds is, for chemists, akin to not being able to steer one way.
It means making a natural product in the lab is a lengthy task. A typical synthesis of phorbol, that toxin from the manchineel tree, takes 52 separate reactions. Each might take a day or so, with another day on top of that to purify the desired compounds from the reaction broth. Add in dead-end attempts, and perfecting the synthesis can take years.
That said, chemists have got extremely good at the turning-left-type reactions that break carbon’s bonds with atoms like nitrogen, oxygen, chlorine, sulphur and more. Since the 1970s, they have developed various metal catalysts that act like matchmakers. These will grab first one carbon and then another, snapping off the bonds to other atoms in the process and bringing the two carbons smoothly together, creating a new bond. These “cross-coupling” reactions have paved the way to many of the natural-product-inspired drugs in use today (see “Natural wonders“). No wonder they won the 2010 Nobel prize in chemistry.
Cross-couplings are, however, far from perfect. One niggle is that the catalysts will only do their grabbing if the carbon is first modified with something like a boron or zinc atom – not usually difficult, but another time-consuming step.
, at the Scripps Institute in La Jolla, California, has little patience with these meanders. He advocates a concept called ideal synthesis. “All it says is that you should do things in the laziest way possible,” he says. This involves “right turn” chemistry, if you will, and eschewing any unnecessary reactions like adding those boron and zinc atoms. Instead, Baran is taking aim at “native groups” of atoms that are already commonplace. His hit list includes arrangements of carbon and oxygen atoms called carboxylic acids, carbon-carbon double bonds – and the CH bond.
One way to dismantle them is to turn them into those supposedly skittish radicals. But rather than adopt the old approach of using chlorine, in 2014 Baran designed an iron-based catalyst that would grab a carbon-carbon double bond and . That radical can then react with a different carbon atom to form a new single bond, with no elaborate pre-activation required.
Baran and others have made similar progress using carboxylic acids. And his eye for the ideal is paying dividends: last year he brought phorbol synthesis down from 52 steps to just 19.
But although double bonds and carboxylic acids are common, they have nothing on the ubiquitous CH bond. Until we crack that, we can’t say the right turn reactions are nailed.
Nice, then, that serious progress is being made here too, including off the back of an accidental discovery by at Princeton University. He had been working on a reaction triggered by high-energy ultraviolet light, and designed to form a different type of bond. It worked, but its cost, complexity and safety considerations around powerful UV lamps meant it wasn’t practical.
“The idea of ideal synthesis is to do chemistry in the laziest way possible”
So he asked a member of his team to find an alternative. The researcher happened to talk to a friend working on catalysts used to capture sunlight for artificial photosynthesis systems. When the researcher gave them a go in his own reaction, it worked almost immediately using low-energy light.
Although chemists have worked on these “photoredox” catalysts for decades, MacMillan’s group was the first to show that their ability to both donate and receive single electrons – most catalysts can only do one or the other – makes them perfect for powering the breaking and making of carbon bonds via radical chemistry. They soon had one catalyst working in a , generating carbon radicals that can then be coupled to another carbon of their choice.
“Photoredox is very exciting; it leads to unusual ways of putting molecules together,” says Davies. Pharmaceutical companies are already considering how to adopt these catalysts in their drug discovery and manufacturing, says MacMillan. Merck, for example, has developed a custom photoredox reactor that uses blue LEDs and is shaped to maximise the uptake of the light. “We have reactions that are done now in 1 or 2 seconds. It’s wild,” says MacMillan.
Yet conquering the CH bond is a double-edged sword. The feat is useful because they are everywhere, but that ubiquity is also problematic. It’s all very well gaining the power to slice the bond open. But molecules with more than five or six carbon atoms – which is to say nearly all medicinally interesting natural products – can easily have 10 CH bonds. How to sever the correct one?
MacMillan has a plan on this front too. His catalysts can discern the dance of electrons around CH bonds, a dance that differs narrowly depending on the arrangement and type of the other atoms the carbon is connected to. He hopes to create versions of his catalysts that act selectively.
Davies’s lab has demonstrated an impressive example how this could work. Davies took a molecule of pentane, a simple compound consisting of five carbon atoms linked in a chain and covered with hydrogen atoms. Chemically, there’s virtually nothing to distinguish any one of its CH bonds from another. Yet by synthesising and screening a series of catalysts, the team found one that breaks with 95 per cent selectivity. “Pentane was a showcase challenge,” Davies says. Now he is moving to try the same trick on molecules that are useful waypoints en route to natural products.
None of this, of course, will replace the classic left hand chemical reactions like cross-couplings. The challenge is to come up with equally powerful right hand reactions, so that we can turn both ways. “I hope,” says Baran, “that some of the reactions we develop end up being as good as the old ones.”
Natural wonders
Chemicals produced by plants and animals are the basis of many successful drugs
Morphine
This painkiller is found in a number of plants, including poppies. It was isolated in the early 19th century and first used during the American civil war.
Aspirin
It has been used to treat pain and inflammation for thousands of years, but in pill form only since 1899. The active ingredient comes from the leaves of the willow tree.
Paclitaxel
This compound was isolated from the Pacific yew tree and approved to treat a range of cancers in 1993. Achieving the first chemical synthesis took 12 years.
Artemisinin
Part of a combination of drugs now used to treat malaria, artemisinin was isolated from the sweet wormwood herb by Tu Youyou, who shared a 2015 Nobel prize for the discovery.
This article appeared in print under the headline “Unbreakable”
