Kevin Verstrepen (centre) leads a beer-tasting session in his lab at the University of Leuven in Belgium. Credit: Justin Jin for <i>Nature</i>

Kevin Verstrepen’s lab meetings can be pretty boozy affairs. Twice a week, several members of his group at Belgium’s University of Leuven and the Flanders Institute for Biotechnology gather around a table loaded with black, tulip-shaped beer glasses, together with spit buckets and crackers.

Verstrepen holds a glass and takes a whiff of its contents. “For me, this was an ethyl acetate bomb,” he pronounces, referring to a chemical found in pear-flavoured sweets that, at high concentrations, reeks of nail polish.

Brigida Gallone, a graduate student in the lab, detects a second aroma: “ethyl acetate and 4-VG”, she says. That’s 4-vinyl-guaiacol, which smells of smoke, cloves and — according to a tasting sheet in front of her — a dentist’s surgery. “I like 4-VG and this was too much for me.” Another student, Stijn Mertens, catches a smell of wet cardboard, which is common to stale beers. “I got some trans-2-nonenal,” he says. With that, the group finishes its analysis of this brew and moves onto the ninth and final glass. It is not even 11 a.m..

“There’s only so many you can do before you lose focus,” says postdoc Miguel Roncoroni, who has been hosting these tastings for more than 4 months. They are part of a project to characterize some 200 commercially produced Belgian beers. Their assessments, alongside precise measurements of the dozens of chemicals that produce the flavours and aromas, could help consumers to identify new beers to try, by comparing the lab’s profiles to ones they like.

But Verstrepen has loftier ambitions than helping beer lovers to select their next bottle. He wants to build the perfect yeast. His lab is deploying what it is learning about the chemical and genetic basis of beer flavour to breed yeast strains that generate unique flavours and other qualities coveted by brewers and drinkers.

The beer geeks in his lab straddle the worlds of curiosity-driven science and industrial brewing. They study evolution, biochemistry, and even neuroscience through yeast. But they also have contracts with beer makers worldwide, from multinational conglomerates to small trend-setting craft breweries. In an upcoming Cell paper, the lab will report the genomes of some 150 yeast strains used to make beer, sake and other fermented products, a project done in collaboration with a leading supplier of yeast to brewers and a synthetic-biology firm.

### LISTEN

Ewen Callaway visits the beer lab and joins the researchers for a tasting session.

For a US$500-billion industry whose products depend on complex interactions between chemistry and microbiology, sophisticated yeast strains are hot commodities. “You always want to know what’s new in Kevin’s lab,” says Peter Bouckaert, the brewmaster at New Belgium, a leading craft brewery in Fort Collins, Colorado. “People are watching what he does.” Beer gets its flavour from just a few ingredients (see ‘Better brewing through biology’). Grains — mainly malted barley — provide sugar and body, but can also imbue flavour, such as the chocolate notes common to dark stouts. Hop flowers bring bitterness as well as the tropical fruit notes in some craft beers. Dissolved minerals in the water influence the flavours that come through from grain and hops. And brewer’s yeast, Saccharomyces cerevisiae, provides alcohol, bubbles and hundreds of aroma compounds. Fermentation produces everything from isoamyl acetate, which makes German Hefeweizens taste of banana, to the clove notes of 4-VG. Yeast science was spearheaded by beer makers. Denmark’s Carlsberg brewery established one of the world’s first yeast-biology labs in 1875, and it was there that Emil Christian Hansen isolated the first pure culture of a brewing yeast in 1883. In the 1930s and 40s, another Carlsberg scientist, Øjvind Winge, discovered that yeast reproduces both sexually and asexually and used this insight to breed new strains with useful brewing traits. Winge’s work moved yeast from the brewery to biology labs, and many scientists now use brewer’s yeast as a model to probe the inner workings of complex cells. But despite a long and fruitful marriage between yeast and bioscience, Verstrepen argues that many brewers are still stuck in the nineteenth century when it comes to the yeast they use. “Brewers, especially traditional brewers, are often not using the optimal yeast.” Most use just one strain — isolated from their brewery or borrowed from another decades ago — in all their beers. Verstrepen wants to change that. He started out working in a South African wine yeast lab, then joined a University of Leuven beer lab for his PhD in 1999. But he was disappointed to find that most of the research involved trouble-shooting brewers’ problems. “Nobody was doing any biology, really,” he says. Disillusioned, he moved to the Whitehead Institute for Biomedical Research in Cambridge, Massachusetts, to pursue a postdoc with Gerald Fink, who pioneered genetic engineering in yeast in the 1970s. But although the scientists there liked yeast, nobody was interested in beer — at least not by day — Verstrepen says. So, his work centred on the proteins that pathogenic yeast deploy to stick to human tissue. Verstrepen discovered that the level of stickiness is dictated by the number of repeating DNA sequences in a certain gene (K. J. Verstrepen et al. Nature Genet. 37, 986–990; 2005 ). “It’s like having longer Velcro loops that stick more easily,” he explains. The proteins are also responsible for flocculation, the process by which yeast cells clump together in a beer and fall out of solution. Flocculation varies between brewing strains and influences the flavour, clarity and alcohol content of a beer. In 2005, Verstrepen moved across town to open a lab at Harvard University, focusing on the roles of various repetitive DNA sequences in generating diversity. He taught Harvard undergraduates biology in a course that incorporated brewing — “It was a pretty tough course,” he says — but beer didn’t feature in his research until he returned to Leuven in 2009. Verstrepen hoped to combine the kind of research he was doing with his interests in beer and wine. But it was a phone call from a Swiss chocolate company that jump-started his work with industry. Zurich-based Barry Callebaut, one of the world’s largest cocoa makers, needed help transforming bitter cacao beans into cocoa powder (which is traditionally done by yeasts present in the environment). “My response was ‘Is chocolate fermented?’ That wasn’t the smartest thing to say,” says Verstrepen. Nevertheless, the company became the first client of Verstrepen’s lab consultancy, which now earns around €500,000 ($555,000) per year. Half of the 25 or so scientists in his lab do applied research on beer, biofuels and other fermented products; the rest pursue epigenetics, molecular evolution and other basic research.

Lager lab

At first glance, the lab looks like any other. Centrifuges, petri dishes and pipettes cover benches. There is also an incubator full of small glass bottles. It would look at home in any microbiology lab, were the bottles not filled with a strong broth of malted barley, sugar and hops. The cold room offers more clues. “There’s not too much beer in here now,” Verstrepen says, pointing to a few full crates. And a stockroom contains hundreds of the black glasses used for blind tastings; they are opaque so tasters cannot see what they are sipping.

His lab’s freezers house about 30,000 types of yeast, including 1,000 strains used by brewers, bakers and others worldwide, and another 1,000 wild isolates from fruit, flowers, insects and even people. Many of these have been characterized for the genes known to influence taste and other traits that brewers care about. The lab is collaborating with yeast supplier White Labs in San Diego, California, and with Craig Venter’s Synthetic Genomics in nearby La Jolla to build a family tree of industrial yeast.

The balance in the freezers is made up by the lab’s creations — completely new strains that have unique combinations of traits. The team makes them by mating different strains and then screening the offspring for the aromas they produce and, more recently, for genes that underlie these traits. This latter approach, known as genetic-marker-assisted breeding, is common in agriculture, and Verstrepen thinks that it will transform brewing.

A Canadian brewery commissioned one of these custom strains when it wanted a full-flavoured Belgian ‘tripel’ that has less alcohol than is typical for the style. Another brewery has asked for yeast that makes chocolatey aromas, a request that has stumped the lab so far.

Mass mating

Using a robot that can accomplish hundreds of matings in a day, the lab generates many more strains than it can analyse or use in its taste tests. To cope with the glut, the researchers are developing microfluidic chips that can brew with 2,000 different yeasts at a time in 20-picolitre batches, each of which contains a single yeast cell. They can automatically test the resulting ‘picobrews’ for alcohol content and hope eventually to be able to measure the aroma compounds produced, too. At the other end of the spectrum, the lab will soon take delivery of a kit to make 500-litre batches of beer to better appreciate the challenges of industrial-scale brewing.

Verstrepen’s archive makes the lab a one-stop shop for brewers looking for unique flavours. For example, when Bouckaert wanted “funky” aromas, such as the smell of barnyards or horse blankets, but without the usual accompanying fruity notes, he tried four of the lab’s strains that showed some promise. None was going to work for a New Belgium beer — at least not yet. “Kevin’s research is a little bit out there and on the edges of brewer’s applications,” Bouckaert says. “But that doesn’t mean it will not translate to something that could be huge in the future.”

Natural variation in brewer’s yeast offers much leeway to dial up and down flavours and other traits, but the approach can only go so far, Verstrepen says. Genetic-modification tools could improve on that. “With breeding we can increase flavour 10-fold, with genetic modification we can increase it 100-fold or 1,000-fold, and we’ve done this,” Verstrepen says. “The beers you make are more like banana milkshakes. Is this something we really want to do?” Brewers are excited about the work, but the stigma that surrounds genetically modified (GM) foods means that the lab always uses more conventional techniques such as breeding and directed evolution to make strains destined for industry.

The beers you make are more like banana milkshakes. Is this something we really want to do?

And gene-editing techniques such as CRISPR could be used to introduce naturally occurring variants linked to flavour production into yeast strains that perform well but don’t impart much flavour. This would accomplish the same goals as conventional breeding much more quickly. Food regulators are already grappling with the question of whether plants and livestock made this way count as GM foods, so it’s not inconceivable that CRISPR beers could force the issue.

A few craft breweries have asked the lab for GM yeast (they were turned down), but most of the brewing industry has little appetite for it, says Bouckaert. “American craft is pushing the boundaries, but on genetic modification, it’s a no-no,” he says. Multinational brewers are even more skittish about being linked to GM beer.

Leuven-based AB InBev, the world’s biggest brewing conglomerate and the maker of Budweiser and Stella Artois, confirmed that it works with Verstrepen. Philippe Malcorps, a senior scientist in the company’s yeast and fermentation division, says that AB InBev is interested in tapping the diversity of wild yeasts to get flavours and other traits that would be impossible to find in classic brewing strains.

Much of Verstrepen’s work with such companies is covered by strict non-disclosure agreements. “Some of our best applied research, we cannot publish, we cannot talk about,” he says, “That is frustrating as a scientist, because you want to share.”

But smaller breweries tend to be more liberal. In fact, Orval, one of the six Belgian breweries run by Trappist monks, declined to sign a non-disclosure agreement or even a contract when it asked Verstrepen to sequence its house strain. “If you cannot trust monks any more, what are you going to do?” says Verstrepen (who keeps his personal refrigerator stocked with Orval, one of his favourite beers). Bouckaert says that he was offered exclusive access to the yeast strains his company was trying, but he balked. “I said, ‘I’m a craft brewer. I don’t want exclusivity. I want you to send them to more people!’”

The applied work has led to scientific insights. Verstrepen’s team is still grappling with a project that originated from trouble-shooting a problem it heard about from brewers. Often, brewers reuse yeast to cut costs. But periodically, they start new cultures, growing them first on glucose (yeast’s preferred food), then adding them to unfermented beer, which contains mainly maltose. Yeast sometimes responds sluggishly to the new food source, and it can take days before the beer is fermenting at full clip (and that leaves it prey to contaminants). “Yeast remember not only what they were eating, but also what their grand-grand-grandmothers were eating, up to five or six generations,” Verstrepen says. Working out the mechanism, which seems to involve the inheritance of epigenetic modifications to DNA or proteins, is now a focus of the lab.

Often, the lab has found, good beer and good science go hand in hand. For his PhD project, Mertens was tasked with creating new strains of the yeast used to make lagers. Lager is brewed at cooler temperatures than other beers, using a yeast that emerged several centuries ago when S. cerevisiae hybridized with a related cold-tolerant species called Saccharomyces eubayanus.

Although responsible for the vast majority of beer sales, lagers tend to have a limited flavour range. To make yeasts that might expand this palette, Mertens coaxed various strains of the two species into mating. Some beers made with these yeasts tasted of onion, and many fermented poorly, but one combined the bracing crispness of a good pilsner (a pale lager) with hints of fruity aromas that are untypical of the style (S. Mertens et al. Appl. Environ. Microbiol. 81, 8202–8214; 2015 ). The beer was so popular among lab members that bottles vanished from the cold room. Several breweries are now testing that strain, known as H29.

Mertens would love to see a commercial beer made from his creation, but he also hopes to sequence the genomes of the other yeasts he has created to work out how species hybridize — and maybe even unpick the conditions that spawned the first lager yeast. “We have new yeast and the brewers love it,” he says. “But we’re looking at the fundamentals of how these hybrids work. That goes beyond beer science.”

By the end of the morning’s beer tasting, the spit buckets are filling up. Verstrepen has a meeting with a DNA-sequencing company, and Mertens and the other students have their research projects to get to. The lab might be a magnet for beer geeks, but it is no keg party.

“Yes, you’re working with a very fun product, but at the end of the day, it’s genetics work,” Verstrepen says. “We’re not drinking for fun.” At least, not until after work.