Engineered Yeast for Custom Wine Flavours? Manage Your Expectations

Yeast can help give wines flavours such as vanilla and raspberry, but more complex flavours are still trickier, and making Trebbiano taste like Sauvignon Blanc is still some way off, as Erika Szymanski of Colorado State University explains.

Reading time: 5m 15s

Wine flavours are difficult to create (Photo: ipopba/AdobeStock)
Wine flavours are difficult to create (Photo: ipopba/AdobeStock)


  • Yeast may change the flavour of wine, but three scientific caveats apply: the complexities of flavour, wine character, and yeast metabolism.
  • Yeast-induced flavour profiles are mostly the result of selective breeding and copying genes from other species.
  • Flavours such as raspberry and vanilla are easy to create. Wine flavours are  much more complex and harder to replicate.
  • The solution may lie in a mix of yeasts, but balancing the flavours will be difficult.
  • However, yeast-induced flavours might provide vanilla flavour at much lower cost than oak.
  • The challenges of climate change may lead to a relaxation in EU and OIV bans on the use of yeasts with extra or modified DNA.


Anyone with some simple laboratory equipment and an eye for following instructions can make Saccharomyces cerevisiae, standard winemaking yeast, glow under UV light or produce orange pigment. Scientists with a bit more expertise under their belts have engineered yeast to make an improbably wide range of useful molecules, from vanilla flavouring to milk proteins to spider silk. Like genetically modified plants, you’ve probably consumed products ‘fermented’ in genetically modified yeast: the heme that makes Impossible Meats ‘bleed,’ for example, or the limonene that gives some cleaning products their citrusy scent. A team at the Australian Wine Research Institute (AWRI) have even engineered a wine yeast to produce raspberry ketone, that fruit’s signature aroma compound, and churned out a batch of experimental raspberry-flavoured Chardonnay.

What to Expect from Engineered Yeast?

All of which makes it reasonable to ask: are flavoured wines made with engineered yeast the future face of big wine brands? Can climate- (and cash-) challenged winemakers ferment their way to making Trebbiano taste like Sauvignon Blanc?

Maybe, but the route to custom-fermented flavours is all about managing expectations. Three big scientific caveats apply—the complexity of flavour, the complexity of wine character, and the complexity of yeast metabolism—and that’s all in addition to regulatory concerns. Wine-lovers may live for complexity, but it’s bioengineering’s bane.

Defining ‘Engineered’

First, we need to clarify what the ‘engineered’ in engineered yeast means. Winemakers already routinely employ custom yeast to help build flavour profiles. With one exception, these are the product of selective breeding, little different than how agriculturists have been breeding and improving crops for centuries.

That exception is ML01, a yeast to which extra genes from malolactic acid bacteria have been added so that it conducts alcoholic and malolactic fermentation simultaneously. ML01 isn’t approved for use in countries that subscribe to the International Organization of Vine and Wine (OIV; the relevant exceptions are the United States, Canada, and China), which draws a firm line between organisms that result from selective breeding and those with extra or modified DNA courtesy of molecular biology tools. The latter are labelled engineered; the former are not.

Copying and Inserting Genes from Other Species

Engineered yeast can incorporate molecule-producing pathways from other species. Or ‘editing’ tools such as CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats – a genetic engineering technique modifying the genomes of living organisms) can be used to increase, decrease, or disable something yeast already does. Scientists don’t understand DNA well enough to ‘write’ genes from scratch, and that’s very unlikely to change any time soon, so giving yeast entirely new abilities requires copying and inserting genes from other species.

Easy to Engineer: Raspberry Flavour

The recipe for the AWRI’s raspberry ketone-producing yeast called for cobbling together four additional genes originally taken from parsley, rhubarb, another yeast species called Rhodosporidium toruloides, and a small plant called Arabidopsis thaliana. Conveniently, Saccharomyces cerevisiae happens to be the easiest organism in the world to genetically engineer because it’s better than any other species on the planet at incorporating foreign DNA. 

That provides a world of options, but yeast still can’t be made to do everything. 

Limits of Complexity

Raspberry wasn’t an incidental choice for the AWRI’s experiment, and it points to the first scientific caveat with dreaming up new yeast-driven wine sensations—flavour complexity. Few flavours are convincingly conveyed by a single molecule. Vanilla (from vanillin), mint (menthol), and generic citrusy-ness (limonene) are. Others such as coffee, chocolate, or banana, are quite more complicated, unless you’re satisfied with the artificial banana flavour found in cheap candies.

From Trebbiano to Sauvignon Blanc?

Every wine style is on the second list. Convincingly replicating the varietal distinctiveness of Sauvignon Blanc or Pinot Noir might require dozens of individual molecules in the proper proportions. While raspberry-flavoured chardonnay involves engineering yeast to make only one new molecule, making Trebbiano taste like Sauvignon Blanc is a much tougher proposition.

“While scientists are busy engineering them, yeast are busy doing their own thing.”

One yeast strain would need to be engineered with multiple flavour molecule-producing pathways, each comprising multiple genes. Alternatively, it would require a mixed fermentation by a group of yeasts engineered to each produce one molecule. Neither is trivial, because while scientists are busy engineering them, yeast are busy doing their own thing.

Yeast Tolerate Foreign DNA – but New Molecules Must be Beneficial

Yeast are unusually laid back about foreign DNA, but they’re still trying to maintain their own metabolisms at the end of the day. Every new molecule-producing pathway inserted into a yeast cell shunts resources away from that cell’s ordinary business of being alive. Cells include lots of redundant and highly interconnected pathways. They’re adept at redirecting their metabolisms to avoid making new molecules that don’t benefit them, especially when they’re asked to make several of them at once, or to grow so slowly as to make the whole exercise pointless. That makes using a mix of yeasts, each engineered to produce one molecule, a more likely option. But, then, we have the difficulty of balancing each of those flavours by convincing each yeast to grow in the desired proportions, without one outgrowing the rest.

Replication of Grape Varieties: Sauvignons Blanc and Vanilla Chardonnay are within Reach

Some varietal characteristics would be easier to replicate than others. Indeed, Sauvignon Blanc might be one of the easier targets, as we know from work led by Wendy Parr, a New Zealand-based sensory scientist, that a relatively few signature aroma compounds do the heavy stylistic lifting. Engineering yeast to produce them is within our current technical reach and could conceivably enable less popular varieties to mimic this market-leader. Vanilla-flavoured chardonnay, driven by cheap fermentation of vanillin instead of the more expensive oak products typically used for a similar purpose, is another easy target.

Hampering Infelicitous By-Products

CRISPR might also be used to hamper yeast’s production of infelicitous fermentation by-products such as stinky sulphor compounds, or to bolster desirable ones such as glycerol. Again, the complexity of yeast’s metabolism—its ability to subvert human engineering, and the need to keep it alive and growing—is a technical limitation that makes some changes easier than others.

Replacing Alcohol?

In theory, wine cooler-like products with simple fruity flavours are in reach. These won’t deliver complexity or nuance, but might appeal in the no- and low-alcohol niches where grape-derived characteristics are an inevitable casualty of alcohol-removal processing. The dream of engineering a yeast that ferments much less alcohol than usual, however, is remarkably difficult, as yeast really like making alcohol. But even for easy edits, selective breeding remains a more practical way to achieve these kinds of aims so long as CRISPR’s creatures are legislatively classed as genetically modified (GM) organisms.

Strict Regulation vs. Needs of Climate Change

In 2018, EU legislation declared gene-edited organisms to be GM for regulatory purposes, even when they don’t incorporate new genetic material from other species. That raspberry-flavoured chardonnay couldn’t be let out of the Australian Wine Research Institute labs because genetically modified yeast aren’t permitted for commercial fermentations under OIV regulations, and they still need regulatory approval for commercial use in the United States and Canada where OIV rules don’t apply. Those regulations will likely be revisited, however, as heat, drought, vine disease, and other climate-related threats press the industry to consider new options.

Genetically engineered vines have yet to receive approval for winemaking use, but they exist in test vineyards and are at least part of the conversation about how the wine industry will cope with the years ahead. Meanwhile, we’ve already seen new hybrids of European and other grape varieties, constructed through selective breeding for fungal disease tolerance, classified as Vitis vinifera for regulatory and marketing purposes and approved for use in OIV countries. Similar latitude will likely apply to engineered yeast, and even a bit more, since whether to consider yeast an ingredient or a processing agent is itself a matter of some debate.  

A Way to Preserve Varietal Flavours?

If and when that happens, gently edited yeast strains that merely reshape what yeast already do won’t be a game-changer; for many practical purposes, they’re already in use. Regulatory approval for GM wine yeast that ferment new flavours seems unlikely if we’re only talking about another variation on the wine cooler. When we reach a point at which climate seriously threatens varietal flavours, engineered yeast might be part of recreating them, though engineering the vines themselves may do more to preserve what we value most about them. And as for turning Trebbiano into Sauvignon Blanc? For now, that’s raspberry-flavoured pie in the sky.



Latest Articles