Inquiring Winemaker


Yeast Genetics Without GMOs

April 2009
by Tim Patterson

  • Two yeast producers, Phyterra and Maurivin, are poised to bring low-H2S-producing yeast strains to market, the result of different research and development methodologies that both make use of modern genetic analysis techniques.
  • Phyterra's potential offerings, based on research by Linda Bisson and her lab at UC Davis, benefited from laboratory-only genetic alteration to prove a concept, but rely on conventional breeding to yield the commercial yeast strain.
  • Maurivin's researchers, working in collaboration with the AWRI, used a chemical process to induce mutation and then employed genetic analysis to identify promising isolates.
  • In both cases, the final products are non-GMO strains.

What better time than the 200th anniversary of Darwin's birth to speculate about the evolution of yeast? For almost a decade, a debate has alternately aged and simmered, in the wine world and more broadly, about the pros and cons of genetic engineering, especially when applied to things people put in their mouths. By now, the warring sides are pretty well fossilized: You're likely to be for it or agin' it. One position says the technology holds incredible promise, an unprecedented capacity to confront vineyard diseases and make wines in a whole new way, and in any case, breeding has been part of agriculture forever. The other team draws a sharp line between classical breeding and GMOs (genetically modified organisms): Nature never tried inserting a flounder gene in a tomato to help shelf life, as one now-defunct vegetable producer once attempted. Down that road, critics assert, lie Frankenfoods and Frankenwines.

What gets counted as "natural" and what as "manipulation" drives many researchers to distraction. Molecular biologist Anthony Heinrich of Australia's Maurivin Yeast says, "It would take me four pages to even get started about the debate on GMO versus non-GMO." He observes that perfectly "natural" sexual hybrids can be quite unstable and unpredictable, while controlled tweaks of a single, known gene are verboten for commercial products. "I and many of my science colleagues find this hard to fathom, but we have to stick with the current rules and regulations."

Meantime, some of the most interesting research and development, at least in the realm of yeast, has been going on in the space between the two polarized positions: using the techniques of genetic analysis and engineering in the lab to chart a path, but basing potential commercial products on good old-fashioned methods. Two examples were showcased on the trade floor at the recent Unified Wine and Grape Symposium in Sacramento, with the announcement by two different suppliers of the impending release of yeast strains that produce little or no hydrogen sulfide (H2S) during fermentation, even under stress. One development track has been pursued by researcher Linda Bisson at the University of California, Davis, and is now being commercialized through Phyterra Yeast; the other represents a collaboration between Maurivin and the Australian Wine Research Institute (AWRI). Both are non-GMO products that owe their existence to modern genomics.

H2S formation and prevention

The formation of H2S and its many descendants (mercaptans, etc.) is surely one of the most irritating occurrences in winemaking. First of all, the stuff smells god-awful enough to ruin your day, just at a sensory level. It can mean a stuck fermentation in progress--rather, not in progress--which presents another whole set of problems.

Even if the fermentation finishes, there's stink to be removed. Aeration and splashing may or may not work, and may result in too much oxygen uptake in the process. Copper fining means putting poison in your wine, however briefly, which nobody likes to do, and the process takes a lot more than just the H2S out of the wine.

Even at lower, less immediately offensive levels, H2S can mask fruit and make a good wine boring. Since the most common cause of this fermented marsh gas is insufficient yeast nutrition, winemakers often resort to pre-emptive additions of DAP (diammonium phosphate, essentially nitrogen) to make sure the yeasts stay buzzed--which not only costs money but can have an adverse impact on wine aromatics and complexity (see Inquiring Winemaker, February 2009).

To summarize: H2S, not good. Most any winemaker would love to wake up in the morning knowing that the fungi at work in his or her cellar could handle the task of fermentation and possible nutrient stress without stinking up the place. And if you ask a roomful of winemakers for their three fondest wishes in this regard, the answer would surely be Montrachet, Montrachet and Montrachet--a workhorse strain in the industry with a well-known penchant for aromatic mischief.

Indeed, when Linda Bisson polled some winemakers about what yeast strain they'd like her to de-gas, Montrachet was the universal nomination. After several years of work with the grad students in her lab, she thinks she has found the solution through a combination of genetic wizardry and conventional breeding.

High-tech detective work

Bisson and her crew began by combing through 300 or so commercial yeast strains in the UC Davis collection, looking for low-H2S producers, and found several candidates. Best of all was UCD 932, which seemed to produce no H2S at all. Next, they tried to determine if this low-H2S behavior was the result of simple or complex genetics, by encouraging the strains to reproduce by sporulation. Sure enough, all the offspring from UCD 932 had the low H2S trait, while the offspring of the other promising strains were all over the map. The implication was that if they foun d the gene in 932 that was regulating H2S production, it should also do the same thing in other strains. The 932 strain by itself wasn't all that promising as a stand-alone yeast; the plan was to export its most endearing characteristic.

The next phase was a wave of crossing and re-crossing 932 with other strains, seeing what produced what, and sequencing the crossed strains to see what genetic material was controlling the low-H2S performance. The result was identification of the MET10 gene in the 932 as the key player. And here's where the genetic manipulation part comes in: to demonstrate that MET10 was the real deal, it was inserted into garden-variety Montrachet (UCD 522), and sure enough, no H2S. Bingo: proof of concept.

The MET10 gene is part of the complex that governs how sulfites are reduced by yeast, and in particular, whether sulfites get transformed into sulfides--the stinky stuff--or something else. Bisson thinks the MET10 somehow keeps the sulfides that are produced from being released into the wine, allowing them to be transformed into more benign compounds. The engineered strain was just as tidy in this regard as the donor strain had been.

Releasing that tweaked yeast would, however, be commercially unpalatable. So knowing what was needed, the Bisson lab went back into the breeding business, crossing 932 with 522 in conventional fashion until they found offspring that fit the genetic profile they wanted. (Fortunately, yeast can be bred and rebred with a short turn-around, unlike the years of waiting for a payoff from grapevine crossings.)

In partnership with UC Davis, Phyterra Yeast is taking this knowledge and technology and laying the groundwork for commercial release. Such new-wave yeasts are Phyterra's specialty; Unified also was the backdrop for its low-urea strain (see "New Methods to Limit Urea," February 2009). Phyterra's own enology staff will continue with quality-assurance testing, doing commercial-scale fermentations and further sensory validation. So far, this Montrachet relative seems just like good old Montrachet, without the embarrassing lapses. If all goes well, Phyterra should have the low-H2S strain on the market in time for the 2009 Northern Hemisphere harvest.

Random mutagenesis down under

Half a world away, Maurivin Yeast and AWRI have produced a low-H2S strain of their own--actually, three of them--billed as Advantage, Distinction and Platinum. All three are variations on another tried-and-true strain, Pris de Mousse, sometimes known as EC-1118. The three variations on this theme show slightly different performance characteristics--alcohol tolerance, etc.--but closely resemble Maurivin's own PDM strain in most respects. While PDM does not have Montrachet's reputation for H2S emissions, its huge role in the industry, especially in an era of high-sugar/high-alcohol fermentations often prone to sticking, makes it a good candidate for lowered hydrogen sulfide.

Yeast Strands
These yeast strains have reduced capacity to produce the amino acids methionine and cysteine. Grape juice contains sufficient quantities for normal growth and metabolism in fermentation.
PHOTO: Maurivin

The Maurivin/AWRI approach was somewhat different. The starting point was the encouragement of "random mutagenesis" with the base PDM strain, creating a pool of 30,000 isolates with slightly varying characteristics. The catalyst for this explosion of mutated yeast was exposure to a chemical process, the details of which are involved in currently pending patent applications. Anthony Heinrich likens the process to the human organism "walking out into the sun and being exposed to ultraviolet light," which can induce subtle variations. And as in the human example, the level of exposure to the chemical process sometimes meant too much change, sometimes too little--a bit like shades of sunburn.

The bumper crop of mildly mutant yeast was screened through a combination of methods, including plating some strains on agar. The low-H2S strains built white cultures, the medium-H2S strains tan, and the high-H2S variants grew black colonies. (Heinrich says that if anyone would like a truly high-H2S-yielding yeast, he's got some available.)

The researchers knew from the literature that the MET family of genes (MET10, but also MET5 and some others), named for their role in regulating the production of the enzyme methionine, were involved in H2S control; genetic sequencing was therefore employed as an analytical tool to keep track of what made the various isolates tick and help weed out the least promising.

The three survivors of this elaborate casting call all yield tiny amounts of H2S, but at levels 10 to 50 times lower than what is detectable in white wine. Better yet, according to Heinrich, there is evidence that they have the ability to take H2S out of the surrounding juice and turn it into amino acids--"a pretty useful property to have," he says, "if your juice is already stinky prior to inoculation."

Like the Phyterra offerings, the Maurivin strains are undergoing final testing and are slated, if all goes well, for release in time for the 2009 harvest. Also like Phyterra, Maurivin will charge premium prices for these strains to offset the costs of research--though, as Heinrich points out, even expensive yeast is but a small part of the cost of producing wine.

Both sets of producers expect to use the methods that worked for these initial strains on other base yeasts as well--crossing other strains with 932 in Phyterra's case, inducing mutations on other popular strains for Maurivin. Chances are good that within a few years, multiple low-H2S strains will be potential parts of the winemaking toolbox. And there's no reason the same approaches couldn't be applied to other aspects of yeast performance--temperature tolerance, alcohol tolerance, and so on.

Magic bullet? Hardly. Useful additions to the fermentation arsenal? Definitely. And most of all, a powerful demonstration of what modern genetic methods can do without crossing the GMO line.

Read an interview with Tim Patterson by Jo Diaz here:

Tim Patterson writes about wine and makes his own in Berkeley, Calif. Years of experience as a journalist, combined with a contrarian streak, make him interested in getting to the bottom of wine stories, casting a critical eye on conventional wisdom in the process. Contact him through

Currently no comments posted for this article.