by John McQuaid
He put his rats in Skinner boxes designed to let them sample dozens of liquid flavor formulations out of small, dimpled trays. Pleasure was measured by the number of times they licked from each mixture. They’d typically lick a sugar solution thirty times; water, about twenty; a bitter mix, only once or twice. With this baseline, any flavor compound could be rated by the number of licks. The rats were also trained to evaluate. If they were testing would-be sweeteners, they pressed one lever if a sample tasted like sugar, a second if it did not. After each sampling, they got a—tasteless—food pellet as a reward. Opertech rats spend their whole lives this way, living out a normal span of about three years; their minders get to know their personalities, and their tastes.
This may seem like a crude way to evaluate flavors, but mindlessness is its greatest virtue. To develop new ingredients, food companies sift through endless lists of chemicals, hoping to find even one with potential. By combining tasting with machine-like repetition, Palmer’s system generated impressive amounts of data in a short time, substituting brute force for nuanced observation. Four trained rats could test thousands of candidate compounds over a few days, pinpointing the tastiest for further study.
A competing biotech firm, Senomyx, took this approach a step further, using human tissue in place of lab rats. Senomyx scientists decoded, and patented, the DNA for sweet, umami, and some bitter human taste receptors. (Yes, taste genes and other parts of the human genome can be patented.) They infused these DNA strands into a line of kidney cells used in cancer research. The DNA went to work making taste receptors, giving Senomyx an unlimited supply of taste cells in petri dishes. These can be dosed with new flavors, the subtleties and flaws of each gauged down to the molecular level. “We can identify when we have a flavor agent that has a bitterness component to it, determine what bitter receptor it’s acting on, and basically dial out the bitter off-taste,” said David Linemeyer, a vice president at Senomyx. This system makes Opertech’s rats look lazy; it can do in hours what takes them days.
The Senomyx approach had one drawback. The host cells used were descended from stem cells taken from a human fetus that was aborted in the early 1970s. Since then, these cells have been a fixture in medical and biotech research. But Senomyx’s job was to develop flavors. However remote, an association with abortion would be devastating to a food or drink brand. Anti-abortion groups found out about this in 2011 and began to protest; a bill was introduced in the Oklahoma state legislature to ban any food developed using this technology. Senomyx, then working with Pepsi, promised not to use that cell line in its soft-drink research.
Both Senomyx and Opertech trained their technology on the hardest taste problem of them all, finding a truly sugary sugar substitute, and both hit on a similar potential solution. Opertech’s rats liked a compound called Rebaudioside C, or Reb C for short, a derivative of the stevia leaf. (The stevia extract already used in many products is a related compound called Reb A.) Reb C itself wasn’t sweet; however, it made sugar taste sweeter. With such a sweetness enhancer, soft drink makers could reduce the amount of sugar in a drink while maintaining the authentic taste. In 2013, Senomyx and Pepsi announced they had found a similar compound. However revolutionary, though, it was not clear the public would embrace this approach: “Americans Will Be Drugged to Believe Their Soda Is Sweeter,” read a headline on the website Gawker.
Pepsi researchers devised another way to enliven the mundane experience of sipping a soda. It was based on a straightforward premise: people form judgments about food by sniffing it first. Experience teaches people that delicious smells precede delicious food: the warm, bracing aroma of coffee, the smell of bacon crackling in a frying pan, the scent of chocolate chip cookies fresh out of the oven. Pepsi’s “aroma delivery system,” patented in 2013, was a gelatin capsule, less than half a millimeter across, containing an aroma designed to create a cola or citrus imprint on the senses and brain before the first swig. Twisting open the cap would break the capsules, releasing the pleasing smell. A signature fragrance might evolve into something like the sound of a pop-top can, the signal of a thirst-quenching drink. This is not the only recently discovered way to exploit the associative powers of aromas; in 2013 a Japanese company began selling a smartphone accessory and app that released pleasing smells into the air, including coffee, curry, strawberry, and Korean barbecue.
Food technologies have begun to dispense with nature itself, and with it, food and flavor traditions that reach back thousands of years or more. At an event in London in 2013, Dutch scientists staged a tasting of the world’s first hamburgers made from meat grown in a lab. The ecological costs of raising cattle are high; making meat without them might one day free up land for other purposes, reduce the beef industry’s environmental impacts, and feed millions. The research was funded by Google cofounder Sergey Brin, who provided a grant of $330,000; the project’s creators hoped to scale up and grow meat for the market within a decade or two.
The scientists, led by Mark Post, took adult stem cells from cattle, then grew them in a culture of antibiotics to prevent microbes in the surroundings from infecting them. They used serums derived from calf and horse fetuses to spur growth and make the stem cells develop into the right kind of muscle tissue. After a few weeks, the small clumps of cells were put into petri dishes. They grew into fibers that knotted together to form small strips of muscle, each about a centimeter long. To add bulk to the muscle tissue, the scientists stretched the meat over a scaffold made of soluble sugar. The stringy strips were then compacted into pellets and packed en masse to make a burger. The finished product comprised 20,000 strands of meat, each containing 40 billion cow muscle cells, along with bread crumbs and a binder added to hold them together. The burgers were cooked in a pan with sunflower oil and butter.
Real beef is red, fatty, and delicious. It grows marinating in a mixture of blood, natural hormones, and amino acids, and carries the imprint of an animal’s diet and experience. The lab burgers were white, and had to be colored with a mix of saffron and beet juice. They didn’t taste like meat—or like much of anything. Hanni Rützler, a food scientist, described it in pointedly non-meaty terms: “crunchy and hot” and “a bit like cake.” Post planned to add lab-grown fat (which can also be grown from stem cells) to later versions. The burgers may one day approach edibility, or taste good. But they are unlikely to rival the taste of real meat anytime soon.
Sometimes, food and flavor part ways completely. In the early 2010s, Rob Rhinehart, a Silicon Valley software engineer, became fed up with eating. Whatever pleasure he received from savoring tasty foods, or just filling his stomach, was outweighed by the constant inconvenience. He resented having to shop, cook, and wash dishes. He didn’t want to go to a restaurant or wait for takeout to arrive. Rhinehart was also suspicious of most of the food he ate. He knew only that it tasted good, and in truth was probably unhealthy.
The straightforward way to rebel against the tyranny of the food system is to return to natural ingredients and fresh, simple flavors. As author Michael Pollan put it: “Eat food. Not too much. Mostly plants.” There are many ways to do this, some not especially tasty. They include the Paleolithic diet, based on the idea that our genes and bodies are better suited to eating food that could have been acquired by hunter-gatherers, such as lean, grass-fed meats, milk, eggs, fruits, and nuts.
But Rhinehart was a tech guy, not a foodie or hipster, and he decided to use technology to create the perfect food, building it from first principles. He researched the human body’s nutritional needs and collected the most basic chemical ingredients available—the only recognizable elements were olive oil, fish oil, and salt. The contents were carbohydrates, proteins, fats, cholesterol, sodium, potassium, chloride, fiber, calcium, and iron, and a long list of vitamins and other nutrients. When blended together, they looked something like a milkshake, with a faint coffee-ish tint. To some, it resembled vomit. “At the time I didn’t know if it was going to kill me o
r give me superpowers,” Rhinehart wrote. “I held my nose and tentatively lifted it to my mouth, expecting an awful taste. It was delicious! I felt like I’d just had the best breakfast of my life. It tasted like a sweet, succulent, hearty meal in a glass.”
He called it Soylent, after the 1973 movie Soylent Green, set in a dystopian future New York City where the only food available is the eponymous green wafer, supposedly made from processed plankton. The lie is exposed in the movie’s final words: “Soylent Green is people!”
Rhinehart ingested nothing but Soylent and water for a month, making himself the guinea pig in an ongoing scientific experiment. He monitored his weight and drew blood daily to test for several important nutritional markers, fine-tuning the Soylent recipe as he went along to ensure he was getting the right balance of nutrients. At one point, his potassium level rose and his heart rate with it; he felt faint, so he reduced the concentration of potassium in the shake. When he began to lose weight, he drank more of it. The formulation cost $154.82 a month, plus shipping for the ingredients; previously, grocery shopping and eating out had cost him close to $500 per month. Pitching this combination of economy, nutrition, and time-saving helped him raise $1.5 million in a Kickstarter fund-raising appeal. He won another $1.5-million infusion of venture capital from Silicon Valley entrepreneurs to help bring Soylent to market.
Chemically, Soylent was not much different from the nutrient concoctions used in feeding tubes for decades. Nor did Rhinehart’s self-experimentation show much about how other people, let alone masses of consumers, would respond to a Soylent diet. But its effects on his food and flavor experiences were revealing. He felt sharper. He was never hungry, and no longer craved the junk food he had occasionally binged on before. Meeting the body’s nutritional needs so precisely, Rhinehart thought, had addressed the central problem of the junk food age, offering relief from the body’s constantly over-revved cycles of craving and gratification—a kind of biological reset. But then the monotony problem struck; bland shakes became a chore to drink. After his initial experiment, Rhinehart continued drinking them, but also allowed himself to indulge in old-fashioned eating and drinking a couple of times a week. He spiked his shakes with vodka. He ate sushi regularly, and came to appreciate its delicate flavors and the craft of the sushi chef, which he now took the time to observe. Only by giving up food was he able to appreciate it.
The apotheosis of this food trend, perhaps decades away or longer, is virtual flavor. When Nimesha Ranasinghe, a computer scientist in Singapore, studied virtual realities, he noticed something missing. Sophisticated head- and handsets could trick the eyes, ears, and even the skin into a feeling of immersion in a fabricated digital space: a spaceship, an alien world, ancient Rome. But without flavors, virtual reality would always be an incomplete and impoverished experience. Ranasinghe took a variant of the tongue electrodes used by Nestlé’s scientists and experimented with them to see if he could create tastes out of nothing but a mild electric current. He made a device he named the “digital lollipop”: a small sphere containing one electrode rests on the tongue; a second electrode is in contact with the tongue’s underside.
By subtly adjusting the current’s magnitude and frequency, along with temperature, he was able to induce sweet, salty, sour, and bitter sensations directly on the tongue (though not umami). These were crude, but Ranasinghe hoped to refine them, and to develop the means to simulate aromas in hopes of one day creating fully realized virtual flavors. He made digital records of the “tastes,” turning them into sequences of ones and zeroes that could be stored on a computer and transmitted over the Internet. Anyone with a digital lollipop device could download the file and “taste” it himself. As the technology improves, a chef may one day be able to create an entire meal, write its flavors to a digital format like a song or a movie, and share it with the world.
Soylent and virtual tastes separate flavor from food. Tasting and savoring become ends in themselves; a form of recreation, play, and art, a feast for the brain and mind without any negative consequences for the body. But can flavor ever truly be liberated from its connection to the body? It draws its power from the gut’s metabolic furnace, the call of ancient, implacable drives. Make those redundant, and flavor loses its essence. Taste has wreaked a lot of havoc in the industrial age. But every attempt to undo the damage—to water tastes down or to otherwise trick the senses—has produced unsatisfying results. This is the great catch-22 of flavor. Its great reward is the suggestion of self-indulgence, the whiff of danger from going too far.
CHAPTER 9
The DNA of Deliciousness
One day in 2010, chef David Chang decided to make some katsuobushi, the cured fish flakes that are a staple of traditional Japanese cuisine. Katsuo means “bonito,” a type of tuna, and bushi “pieces” or “shavings.” When placed in broth made with miso, kelp, and tofu, katsuobushi turns into delicate, rippling cellophane ribbons. Its complex flavors emerge in a process similar to that used for Icelandic hákarl, though more intricate. Thick cuts of bonito are smoked, then inoculated with molds and packed in dry rice. Over a period of months, the molds grow, then desiccate. They are scraped off, only to grow back again. Mold microbes infuse and dine on the fish flesh, making a suite of aromatic molecules that rival those of the subtlest cheeses. Amino acids with a strong umami signature also flourish, harmonizing these diverse flavors. The final product is a solid block, its surface mottled in greens and blues, ready to be grated.
Chang had already mastered this temperamental fermentation process, having made katsuobushi for his five Momofuku restaurants in New York City. But now he was in the Momofuku Culinary Lab, a cramped, 250-square-foot working kitchen whose sole purpose was experimenting—playfully—with culinary tradition.
Chang and his two partners, chefs Dan Felder and Daniel Burns, debated how to alter this particular recipe. They could fiddle around the margins—adjusting the heat applied during the drying and aging process, perhaps—or they could do something truly subversive and substitute pork for fish. In traditional Japanese kitchens, where katsuobushi has been made for three hundred years, this idea would be a culinary oxymoron. They chose to go with it.
As a practical matter, swapping out fish for pork made sense. Bonito and bluefin tuna, which is also used to make katsuobushi, are expensive and overfished. In Japan, bluefin tuna are so prized and in such short supply that some have been auctioned for over $1 million apiece. Pork is cheap and plentiful, and pigs can be raised organically. If the partners’ experiment worked, they would have a provocative take on a Japanese standard, save money, and minimize environmental harm.
They steamed, smoked, and dried a pork tenderloin, then packed it in raw sushi rice to age. No molds were added; they let the microorganisms on the meat grow instead. Six months later, the aged and petrified pork looked like a Jackson Pollock painting, layered with greens, whites, and coppers—in regular bushi preparation, a sign of success. They dubbed their concoction butabushi, substituting the Japanese word for pork. But as they were about to taste the finished product, they realized they had a serious problem.
Using pork had seemed like a straightforward ingredient swap, but the process of making katsuobushi is rigorously managed, all variables accounted for, the result of centuries of trial and error. By introducing an unknown element, they had profoundly altered the microbiology of that process. The chefs did not know the identity of the molds growing on their cured pork block. They could be toxic, a threat to public health. Even if benign, they could infect ingredients they came in contact with in the kitchen. Even the best-case scenario—that the butabushi was safe and its flavor good—was a chef’s worst nightmare: they wouldn’t know how to re-create it. Natural microbial communities are like snowflakes: each is unique. A different piece of pork would have its own distinctive population of microbes. Even if the molds were identical, small changes in temperature or humidity could influence them in unknown ways, producing wi
ldly different flavors at the end of the aging process. The chefs could duplicate every variable exactly, but the result might never be the same.
• • •
Humans mastered fermentation thousands of years ago, but the scientific understanding of it is still in its infancy. The origins can be traced back to 1856, when thirty-four-year-old Louis Pasteur was dean of the Faculty of Sciences at the University of Lille, in France’s northern industrial heartland. The father of one of his students, a local distiller named Bigo, approached him with a problem: the spirits he was making from sugar beets were mysteriously turning sour. He invited Pasteur to come and inspect the vats, drawing him into one of the major scientific debates of the era. Some doubted that the yeast that powered alcoholic fermentation was alive, maintaining the process was purely a chemical reaction. Others believed that yeast consisted of tiny organisms that sprang to life fully formed from rotting food and corpses, a process called “spontaneous generation.”
Pasteur literally plunged into the challenge that Bigo had presented to him. “Louis . . . is now up to his neck in beet juice. He spends all his days at the distillery,” his wife and lab assistant, Marie Pasteur, wrote to her father-in-law. He chemically analyzed the sour gunk from the vats. It contained lactic acid, the chemical that gives spoiled milk its unpleasant taste. Through a microscope, he examined samples taken from both good and bad batches. The good ones were swarming with yeasts. In the bad ones, no yeasts were present, but a smaller, rod-shaped organism was multiplying. Pasteur prepared a solution mixing the two. The rod microbes made more acid, which killed off the yeasts. Pasteur had discovered two distinct processes under way in the vats. The first was the intended one: yeasts making alcohol. The second was basically an infection. Bacteria were producing lactic acid—key in techniques for making cheese and yogurt, but inimical to brewing. Fermentation, it seemed, consisted of organisms living, digesting, and reproducing. The two major theories had both been wrong.