If you drink much from a bottle marked “poison,” it is almost certain to disagree with you, sooner or later.
—Lewis Carroll,
Alice’s Adventures in Wonderland (1865)
In novels, thick fog always envelops the city of London just when something dramatic is about to happen. It hides the robberies and the kidnapping in Oliver Twist. It shrouds the arrival of Dracula when he comes for Mina Harker, and Sherlock Holmes watches it swirl down the street before the fateful events in The Sign of Four. But on September 7, 1978, light morning showers had given way to sunshine by the time Georgi Markov parked his car and began walking toward the Waterloo Bridge. Had it been foggy, Markov might have left his windbreaker in the closet and donned an overcoat, or at least a pair of heavier trousers. Either one could have saved his life.
At home in Bulgaria, Markov’s novels and plays had made him a famous literary star, someone who mingled with the social and political elite. He’d even gone on hunting trips with the president. Since defecting to the West, that insider knowledge had helped him craft accurate and scathing commentaries on repression behind the Iron Curtain. He hosted a weekly show on Radio Free Europe and also worked at the BBC, where he was heading on that fateful afternoon. Markov knew that speaking out put him at risk, and he’d even received the occasional death threat. But he was a relatively minor figure—no one expected him to be the target of a plot, let alone one that would become the Cold War’s most infamous assassination. And no one could have predicted the murder weapon, something so preposterous even his widow found it hard to believe.
Walking past a bus stop on the south side of the bridge, Markov felt a sudden jab in his right thigh and turned to see a man bending over to pick up an umbrella. The stranger mumbled an apology, hailed a nearby cab, and disappeared. When he reached his office, Markov noticed a spot of blood and a tiny wound on his leg. He mentioned it to a colleague, but then dismissed the incident from his mind. Late that night, however, his wife found him stricken with a sudden and violent fever. He told her about the stranger at the bus stop, and they began to wonder—could he possibly have been stabbed by a poison umbrella? What had really happened was even more bizarre.
“The umbrella gun was invented by the KGB equivalent of Q’s lab,” Mark Stout told me, alluding to the fictional workshop for spy gadgetry made famous in James Bond movies. But while exploding toothpaste and flame-throwing bagpipes play well in Hollywood, exotic weapons are a rarity in real-life espionage. “It’s almost always low-tech,” Stout went on. “Somebody shoots somebody, or a bomb goes off. At the time, the umbrella gun, and the tiny pellets it fired, were a substantial feat of engineering.”
I called Mark Stout about the Markov case because for three years he had held the position of chief historian at the International Spy Museum. That job title must have looked great on a business card, but it also gave him access to a working copy of the umbrella gun, built by a veteran of the same KGB lab that created the original. The replica features prominently in a section of the museum called “School for Spies,” where it’s displayed alongside another KGB invention, the single-shot lipstick pistol. By the time I spoke with him, Stout had moved on to a more traditional academic post, but he still showed an obvious enthusiasm for the world of secret agents. “The umbrella used compressed air, exactly like a BB gun,” he explained eagerly. I could hear the squeak of his desk chair over the phone, and pictured him rolling around his office, pausing to lean back in thought. “But it was designed for extremely short range, an inch or two maximum. In Markov’s case, they literally pressed the tip against his leg before firing.”
For pathologists working in 1978, however, there were no spy museums or historians to turn to. Their patient soon died in a London hospital from what appeared to be acute blood poisoning, but they had no logical explanation for his symptoms. The autopsy did note an inflamed pinprick in his thigh, but it looked like an insect sting, not a stab wound. And the mysterious pellet lodged inside was so minute that technicians dismissed it as a blemish on the X-ray film. The investigation might have stopped there if another Bulgarian dissident hadn’t come forward with a similar story. He’d been attacked near the Arc de Triomphe in Paris, but had recovered after a short illness. This time, doctors paid attention to his account of a painful jab, and they soon recovered a tiny platinum sphere from the small of his back. Because he’d been wearing a heavy sweater, the pellet hadn’t penetrated beyond a layer of connective tissue surrounding the muscle, and most of its poison had failed to disperse. London’s coroner immediately reexamined Markov’s body, recovered an identical pellet from the wound in his leg, and came to a famously circumspect conclusion of foul play: “I cannot see any likelihood of this being an accident.”
To the public, Markov’s murder made the fantasy world of James Bond a sudden reality—in the same year, The Spy Who Loved Me became one of the highest-grossing British films of all time. To investigators, the case left two glaring questions unresolved: Who was the man with the umbrella, and—something British Intelligence and the CIA were keen to find out—what kind of poison could kill someone with such a tiny dose? The first question remains unanswered. Soviet defectors later confirmed that the KGB provided the umbrella and pellets to the Bulgarian government, but critical details remain hazy, and no one has ever been arrested for the crime. Resolving the poison puzzle, however, drew a unanimous opinion from an international team of pathologists and intelligence experts. They arrived at their conclusion after weeks of careful forensic analysis, with contributions from pharmacologists, organic chemists, and one 200-pound (90 kilogram) pig.
The first challenge lay in determining exactly how much poison had entered Markov’s body. Measuring less than a twentieth of an inch (1.5 millimeters) in diameter, the pellet removed from his thigh contained two carefully drilled holes with a total capacity estimated at sixteen millionths of an ounce (450 micrograms). (To put that into perspective, press a ballpoint pen lightly onto a piece of paper. The tiny ink-speck it leaves behind is the size of the pellet—seeing the holes would require a microscope.) Simply knowing that dosage narrowed the possibilities to a handful of the world’s deadliest compounds. The team immediately ruled out bacterial agents like botulin, diphtheria, or tetanus, all of which would have triggered telltale symptoms or immune reactions. Radioactive isotopes of plutonium and polonium didn’t fit the bill, either—they can be fatal, but their victims take a much longer time to die. Arsenic, thallium, and the nerve gas sarin weren’t nearly powerful enough, and while cobra venom might have produced a similar reaction, it would have required at least twice the dose. Only one group of poisons could have caused Markov’s deadly combination of symptoms so quickly: the poisons found in seeds.
For thousands of years, executioners and assassins have turned to seeds in search of ways to do in their victims. The plant kingdom in general offers a vast selection of toxins, but seeds offer the advantages of easy storage and high potency. They’re the most poisonous part of the hemlock plant that killed Socrates, as well as the white hellebore suspected of knocking off Alexander the Great. Strychnine trees bear seeds nasty enough to earn the nickname “vomit buttons,” and their poison has figured in the murders of everyone from a Turkish president to the young women targeted by Victorian serial killer Dr. Thomas Cream. In Madagascar and Southeast Asia, hundreds of deaths every year are attributed to the nuts of a salt-marsh species known simply as the “suicide tree.” The murderous potential of seeds was not lost on William Shakespeare when he needed a convincing concoction to pour into the ear of Hamlet’s father. Most scholars agree that his “leperous distilment” must have been an extract of henbane seeds, just as mystery fans know that Arthur Conan Doyle modeled the “devil’s foot” that nearly killed Holmes and Watson on the deadly calabar bean of West Africa. These plants all rely on alkaloids to provide their poisons, but investigators in the Markov case quickly narrowed in on a toxin more unusual, more deadly, and more difficult to trace. It’s something the Castrol Motor Oil Corporation inadvertently hit on the head with their company motto: “It’s More Than Just Oil.”
Castrol got its start, and its name, by formulating engine oils from the seeds of the castor bean plant, a shrubby African perennial related to spurges. Castor beans store most of their energy as a thick oil that boasts a rare ability to maintain viscosity at extreme temperatures. (Although Castrol now makes a range of petroleum-based products, bean oil remains the lubricant of choice for high-performance racecars.) But the beans contain something more—a peculiar storage protein called ricin. Chemists know ricin for the odd, double-chain structure of its molecules. In a germinating seed, those molecules break down like any other storage protein, providing nitrogen, carbon, and sulfur to fuel rapid growth. But inside an animal—or a Bulgarian dissident—their odd structure gives them the ability to penetrate and destroy living cells. One chain pierces the surface while the other detaches inside and wreaks havoc on the ribosomes—small particles essential for translating the cell’s genetic code into action. (In biochemistry, this puts ricin into a group called “ribosome inactivating proteins,” known by the fitting abbreviation RIPs.) Dispersed through the bloodstream, ricin sets off a wave of cell death so unstoppable that even scientific journals describe it with something like awe: “one of the most lethal substances known,” “one of the most fascinating poisons,” or simply, “exquisitely toxic.” As if to add insult to injury, castor beans also contain a potent allergen, so that, while dying, one can reasonably expect the further indignities of violent sneezing, a runny nose, and a painful rash.
FIGURE 11.1. Castor bean (Ricinus communis). Beautiful enough to be sought after by jewelry makers, the mottled seeds of the castor plant contain a valuable oil as well as ricin, one of the world’s deadliest poisons. The spiny, protective capsule bursts upon drying, hurling individual beans as far as thirty-five feet (eleven meters) from the mother plant. ILLUSTRATION © 2014 BY SUZANNE OLIVE.
Theoretically, the pellet from Markov’s leg could have held enough ricin to kill every cell in his body many times over. But investigators had precious little evidence to go on. He died too quickly for any recognizable antibodies to develop, and even though ricin was known to be deadly, documented poisonings were extremely rare, and there was no clinical description of the symptoms. So the pathologists decided to stage a test. They obtained their own batch of castor beans, refined a dose of ricin, and injected it into an unsuspecting pig. Within twenty-six hours the pig died in the same horrible manner as Markov. “The . . . animal defense people would be horrified,” a doctor on the case observed, but it emerged later that Bulgarian scientists had been even more brutal. They had fine-tuned the dose destined for Markov after testing a smaller amount on a prison inmate, who survived. When they worked out a quantity that would reliably kill a full-grown horse, they put the plan into action.
Georgi Markov’s murder shone a bright media spotlight on the homicidal potential of seeds. Criminal elements took note, and ricin continues to surface as a bioterror weapon of choice. Anonymous letters tainted with it have been sent to the White House, the US Congress, the mayor of New York, and various other government offices in recent years, sometimes closing down mail-processing facilities for weeks. When London police raided a suspected Al Qaeda cell in 2003, they confiscated twenty-two castor beans, a coffee grinder, and enough chemistry equipment to perform a simple extraction. (Their haul also included quantities of apple seeds and ground cherry pits, both of which contain traces of cyanide.) Seed poisons retain their appeal because they are not only potent, but also readily available. When I wanted some castor beans of my own, searching the Internet quickly revealed dozens of varieties openly and legally for sale. People still grow them for their oil and as an ornamental, and the plant has become a common roadside weed throughout the tropics. With a few clicks and a credit card, I had a batch delivered to my door—beautiful, glossy things the size of a thumbnail, their smooth coats mottled with burgundy swirls. They come in shades from umber to pink and often show up in beaded necklaces, earrings, and bracelets. In fact, bright “warning” colors make a number of toxic seeds fashionable in the bead industry, from rosary pea to coral bean, horse-eyes, and various cycads. But castor beans and other poisonous seeds remain common for another reason. It’s a principle that underlies the modern pharmaceutical industry, but it was also expressed perfectly in the nineteenth century by philosopher Friedrich Nietzsche and by children’s author Lewis Carroll.
People remember Nietzsche primarily for his views on religion and morality, but he also coined the maxim, “What does not kill me, makes me stronger.” He meant it as a general comment on life, but this phrase also describes a truth about seed poisons. Lewis Carroll made the same point when his most famous character, Alice, cautioned against “drinking much” from a bottle marked “poison.” By including the word “much,” Carroll implied that drinking “a little” from such a bottle wouldn’t be disagreeable at all, and might even do a person some good. Time and again, that is exactly the case with poisonous seeds. In doses short of deadly, many of those same toxins can be used medicinally—vital treatments for some of the world’s most serious diseases. For Alice, the bottle in question contained not poison but a shrinking potion, preparation for the next escapade in her continuing adventures in Wonderland. Nietzsche’s case seems more significant. He wrote his famed dictum shortly before suffering a mental collapse that scholars now interpret as the onset of brain cancer, one of the illnesses now being treated with seed extracts.
In the language of poisons, ricin is known as a cytotoxin—a cell killer. Along with similar compounds from the seeds of mistletoe, soapwort, and rosary pea, it shows great promise for assassinations at a much smaller scale: the targeted killing of cancer cells. By attaching these “RIP” proteins to the antibodies fighting a tumor, researchers have successfully attacked cancer in laboratory tests, clinical trials, and, in the case of mistletoe extracts, tens of thousands of patients. The challenge, of course, is twofold: finding the right dosage, and making sure the poisons don’t diffuse to other parts of the body.
Whether or not ricin will become a widespread cancer treatment remains to be seen. If it does, it will join a long list of other seed and plant-based curatives dating back to the origins of medicine itself. Wild primates from chimpanzees to capuchin monkeys regularly treat themselves with botanicals, choosing specific seeds, leaves, and bark known to have healing properties. When researchers in the Central African Republic observed a gorilla plucking junglesop seeds from the dung of elephants, no one was surprised to learn those seeds contained potent alkaloids, and that local healers prescribed them (as well as the plant’s leaves and bark) as a treatment for everything from sore feet to stomach problems. This pattern repeats itself throughout the tropics: primates shopping around the apothecary of the rainforest to help rid themselves of parasites, or relieve the pain of injury and disease. Few anthropologists doubt that our own ancestors did the same thing, and in fact a study in the Amazon found that hunter-gatherers used a list of plants that closely mirrored those preferred by monkeys. These ancient habits not only lie at the heart of traditional medicines, they continue to spur the development of new drugs.*
To gauge the importance of seeds in modern medicine, I contacted David Newman, an expert on drug development at the National Institutes of Health. He told me that until the mid-twentieth century, a huge proportion of medications came from plants, many of them from compounds found in seeds. Even today, in an era of synthetics, antibiotics, and gene therapy, nearly 5 percent of all new drugs approved for use in the United States come directly from botanical extractions. In Europe, the number is higher. A recent attempt to summarize medicinal research on seeds quickly ran to more than 1,200 pages, with contributions from 300 scientists working in labs around the world. Seed extracts play a role in treatments for everything from Parkinson’s disease (vetch and velvet bean) to HIV (blackbean, pokeweed), Alzheimer’s disease (calabar bean), hepatitis (milk thistle), varicose veins (horse chestnut), psoriasis (bishop’s flower), and cardiac arrest (climbing oleander). Like ricin, many of these compounds serve double duty as both poison and cure, and it turns out that another well-known example happens to come from the seeds of the almendro tree.
Fresh from their shells, almendro seeds look quite a bit like the almonds that give them their Spanish name, but stretched thin and polished to a dark sheen. The first time I roasted a batch, I immediately noticed the sweet, spicy smell that had brought them to the attention of perfumers in the nineteenth century. Known by the trade name “tonka beans,” the fragrant seeds also became popular as a vanilla substitute and as a flavoring for pipe tobacco and spiced rum. Commercial varieties came from an Amazonian almendro closely related to the ones I studied in Central America. They spawned an industry that was briefly lucrative enough to warrant huge tonka bean plantations in Nigeria and the West Indies. A French chemist isolated the active ingredient and called it coumarin in honor of an Indian name for the tree, cumarú. Things went along cheerfully for tonka bean farmers until the 1940s, when researchers discovered that coumarin was toxic to liver cells. Regulators warned that even small amounts could be harmful, and soon banned it altogether as a food additive. Needless to say, consumption of tonka beans has since plummeted, though daring chefs continue to add a few shavings to specialty chocolates, ice creams, and other desserts.
I knew this history when I sat down to sample a few of the roasted seeds with my dissertation adviser, and coauthor of all my almendro papers, Steve Brunsfeld. The fact that he was a liver cancer survivor didn’t faze us. In botany, tasting weird things goes with the job description—flavors and smells are often valuable tools in identifying plants. Still, we limited ourselves to a few nibbles, just enough to appreciate what struck me as a combination of vanilla and cinnamon, with a citrus finish. Steve wrinkled his mustache and described the flavor more simply: “These things taste like furniture polish.” The comment was typical Steve: sharp, funny, and straight to the point. But our almendro-tasting moment was also deeply ironic. At the time, neither of us knew that Steve’s cancer had reawakened and spread to other parts of his body, and that within a few months his doctors would probably start prescribing a variation of the very compound we’d been joking about.
Since the heyday of tonka beans, scientists have found traces of coumarin in a wide range of plants. It adds to the cinnamon fragrance of cassia bark, and it freshens the smell of cut hay from any field containing vernal grass or sweet clover. But scientists also noticed that something strange happens when plants containing coumarin start to rot. The presence of blue mold and other common fungi changes coumarin from a moderate liver toxin into a blood thinner potent enough to kill a full-grown cow. This discovery solved the riddle of why spoiled fodder sometimes wipes out a farmer’s stock. But once researchers mastered that small chemical tweak, it led to billion-dollar advances in two industries: pest control and pharmaceuticals.
Named warfarin after the group that funded the research (the Wisconsin Alumni Research Foundation), this modified coumarin quickly became the most widely used rat poison in the world. Mixed in with a tempting food bait, it kills rodents by causing anemia, hemorrhaging, and uncontrollable internal bleeding. But in people, a small dose thins the blood just enough to prevent dangerous clots inside the veins, one of the most common and deadly side effects of cancer and its treatment. Sold under the trade name Coumadin, a warfarin prescription often goes hand in hand with chemotherapy, particularly when the cancer spreads widely, like Steve’s. It’s also commonly used by stroke and heart patients, and remains one of the world’s top-selling drugs over half a century after its discovery.
All the while Steve and I worked on the almendro project, his body was fighting cancer. It was a situation that sick botanists must face all the time: struggling against diseases whose treatment may come from the very plants on their herbarium sheets and microscope slides. Steve never told me whether he was taking warfarin, but it wouldn’t have been the first time his research had overlapped with the medicine cabinet. He spent much of his career studying willow trees, the original source of aspirin, and once helped a biotechnology company find a good natural supply of false hellebore, a member of the lily family whose toxic seeds, leaves, and roots contain promising anticancer alkaloids.
In the end, no prescription was enough—Steve died a few short weeks before I defended my dissertation. It bothered him to leave things unfinished, in the lab as well as in his personal life, and he kept working long past the point where most people would have thrown in the towel. But while nothing could buy him more time on earth, he did live long enough to get some answers, to know what the research meant. And, for a curious mind like his, that was at least some reward. In the years since, I’ve often missed not only Steve’s friendship, company, and wicked sense of humor, but also his keen intellect. He had that rare ability to cut right through extraneous information, what he would have called “the bullshit,” and get at the heart of things. That’s a valuable skill both for conversation and for science, because, in nature, even straightforward ideas are rarely as simple as they seem.
On the surface, the notion of lethal seed poisons seems to make perfect sense. It’s a natural extension of the same adaptations that led to spices, caffeine, and other defensive compounds. After all, what better way to protect your seeds than to kill anything that tries to eat them? But in practice, taking that evolutionary step from disagreeable to deadly is more complicated. When a seed is attacked, the plant’s first imperative is to make the attacker stop, which is why bitterness, pungency, and burning sensations are so common. Immediate physical discomfort drives seed-eaters away and teaches them not to try again, a lesson they can even pass along to others. Poisons, on the other hand, may take hours or days to have an effect, which does nothing to stop a seed attack in progress. A flavorless toxin like ricin makes it theoretically possible for an animal to consume and destroy every seed on a castor bean plant, then wander off and die without even knowing the cause. (And certainly without developing and passing on a “castor-bean-avoidance” behavior!) So while chemicals that cause unpleasantness can discourage whole groups of seed predators, deadly poisons eliminate only individuals, a battle that must be fought again and again. This raises the question of what evolutionary incentive causes some toxins to keep getting stronger, to reach the almost absurd potency of compounds like ricin.
“There don’t seem to be any obvious answers,” Derek Bewley told me when I posed this question. I hadn’t contacted him in a while, but this “god” of seed research was always a generous resource when I ran into puzzles I couldn’t solve. He explained that seed poisons often affect different attackers in different ways. Something that evolved to give one animal a modest stomachache (and teach it not to eat that seed again) might prove utterly deadly to another. Or a poison that took days to dispatch large creatures might kill insects in seconds, thereby stopping an attack just as quickly as a foul taste. “Or, the whole thing might be a fortuitous accident,” he mused, and pointed again to the castor bean example. “Ricin is an easily and early mobilized storage protein, and its toxic properties might be just a useful side-effect.”
When Noelle Machnicki investigated the capsaicin in chili peppers, she learned that what began as an antifungal compound ended up influencing everything from insects and birds to the taste buds of mammals, including people. The same complexity applies to seed poisons, and it would probably take a determined doctoral student like Noelle to unravel the story behind any one of them. But there is something certain about all poisonous seeds: no matter how toxic they may have become, the plant must have also invented some way to disperse them. Because there’s no use in keeping your seeds safe if you can’t move them around. In the case of castor beans, the solution is twofold: an explosive pod that hurls the ripe seeds up to thirty-five feet (eleven meters) away from the mother plant, and a nutritious little package attached to the outside of the seed coat, which makes the seeds attractive to ants. Anywhere in the world, the scene near mature castor bean plants is pretty much the same—pods popping open, seeds flying, and thousands of ants busily dragging them home to their underground nests. Once there, they chew off the food packet and leave the seed untouched, safely buried and ready to sprout. Surprisingly, no one has yet looked into whether the food packets are harmless, or if the ants have developed an immunity to ricin. Either way, this clever system allows castor beans to become exceedingly deadly with no risk of compromising their ability to disperse. For almendro, on the other hand, the presence of coumarin in seeds is a little harder to explain.
Though it’s not technically rat poison without modification by mold or chemistry, coumarin still seems an unlikely compound for a seed dispersed by rodents. Even in its unadulterated form, it wreaks havoc on their livers. The toxicity that got it banned as a food additive was first noticed in an experiment on laboratory rats. Fed a diet supplemented with coumarin, the rats systematically lost weight, developed liver tumors, and died young. No one has studied this dynamic in the wild, but it’s hard to imagine a diet more rich with coumarin than that of agoutis, squirrels, and spiny rats living beneath an almendro tree. Yet these rodents continue feasting on the seeds—and occasionally dispersing them—with no apparent ill effects. Have they developed an immunity? Do their livers recover during other seasons, when almendro seeds aren’t available? Or perhaps they are, indeed, dying young, unnoticed in their nests and burrows. Nobody knows the answer, but there is another possibility that is even more intriguing.
Coumarin occurs in a wide variety of plants, but nowhere is the concentration higher than in almendro seeds. (That’s why European perfumers continue to get theirs from tonka beans rather than trying to squeeze it out of the vernal grass in their backyards.) Is it possible that the coumarin in almendro is on the rise? Could we be witnessing the early stages of a new chemical defense strategy? At the moment, rodents do indeed disperse almendro seeds, but that situation is only a snapshot in evolutionary time. From the plant’s perspective, it’s a messy and chancy business. Agoutis and squirrels will eat and destroy every seed they can, only dispersing the ones they happen to forget about. If almendro does develop a coumarin potency strong enough to keep them away, it wouldn’t be the first time a seed defense targeted rodents. Remember that capsaicin, to name but one example, burns the mouths of seed-eating rats and mice, but has no effect on the beaks of the seed-dispersing birds. But the almendro could only afford to deter rodents if, like chili peppers, it had an ace in the hole, another option for dispersing its seeds. And after walking hundreds of transects in the jungle, and analyzing thousands of samples in the lab, we realized that is exactly what is going on.
*Self-medication may be common practice for wild primates, and it probably helped inspire many traditional remedies, but it’s not something to trifle with. Like the ricin in castor beans, many compounds from seeds and other plant parts are highly toxic at the wrong dosage.
