The winemaker's bug: From ancient wisdom to opening new vistas with frontier yeast science.

Abstract

The past three decades have seen a global wine glut. So far, well-intended but wasteful and expensive market-intervention has failed to drag the wine industry out of a chronic annual oversupply of roughly 15%. Can yeast research succeed where these approaches have failed by providing a means of improving wine quality, thereby making wine more appealing to consumers? To molecular biologists Saccharomyces cerevisiae is as intriguing as it is tractable. A simple unicellular eukaryote, it is an ideal model organism, enabling scientists to shed new light on some of the biggest scientific challenges such as the biology of cancer and aging. It is amenable to almost any modification that modern biology can throw at a cell, making it an ideal host for genetic manipulation, whether by the application of traditional or modern genetic techniques. To the winemaker, this yeast is integral to crafting wonderful, complex wines from simple, sugar-rich grape juice. Thus any improvements that we can make to wine, yeast fermentation performance or the sensory properties it imparts to wine will benefit winemakers and consumers. With this in mind, the application of frontier technologies, particularly the burgeoning fields of systems and synthetic biology, have much to offer in their pursuit of "novel" yeast strains to produce high quality wine. This paper discusses the nexus between yeast research and winemaking. It also addresses how winemakers and scientists face up to the challenges of consumer perceptions and opinions regarding the intervention of science and technology; the greater this intervention, the stronger the criticism that wine is no longer "natural." How can wine researchers respond to the growing number of wine commentators and consumers who feel that scientific endeavors favor wine quantity over quality and "technical sophistication, fermentation reliability and product consistency" over "artisanal variation"? This paper seeks to present yeast research in a new light and a new context, and it raises important questions about the direction of yeast research, its contribution to science and the future of winemaking.

Full text

Page 1

© 2012 Landes Bioscience.
Do not distribute.
complex wines from simple, sugar-rich grape juice. Thus any
improvements that we can make to wine, yeast fermentation
performance or the sensory properties it imparts to wine will
benefit winemakers and consumers. With this in mind, the
application of frontier technologies, particularly the burgeon-
ing fields of systems and synthetic biology, have much to offer
in their pursuit of “novel” yeast strains to produce high quality
wine. This paper discusses the nexus between yeast research
and winemaking. It also addresses how winemakers and
scientists face up to the challenges of consumer perceptions
and opinions regarding the intervention of science and
technology; the greater this intervention, the stronger the
criticism that wine is no longer “natural.” How can wine
researchers respond to the growing number of wine com-
mentators and consumers who feel that scientific endeavors
favor wine quantity over quality and “technical sophistication,
fermentation reliability and product consistency” over “artisa-
nal variation”? This paper seeks to present yeast research in a
new light and a new context, and it raises important questions
about the direction of yeast research, its contribution to
science and the future of winemaking.
The winemaker’s bug
From ancient wisdom to opening new vistas
with frontier yeast science
Isak S. Pretorius,1,* Christopher D. Curtin2and Paul J. Chambers2
1University of South Australia; Adelaide, Australia;2The Australian Wine Research Institute; Adelaide, Australia
Keywords: Saccharomyces cerevisiae, wine, yeast
The past three decades have seen a global wine glut. So far,
well-intended but wasteful and expensive market-intervention
has failed to drag the wine industry out of a chronic annual
oversupply of roughly 15%. Can yeast research succeed where
these approaches have failed by providing a means of
improving wine quality, thereby making wine more appealing
to consumers? To molecular biologists Saccharomyces cerevi-
siae is as intriguing as it is tractable. A simple unicellular
eukaryote, it is an ideal model organism, enabling scientists to
shed new light on some of the biggest scientific challenges
such as the biology of cancer and aging. It is amenable to
almost any modification that modern biology can throw at a
cell, making it an ideal host for genetic manipulation, whether
by the application of traditional or modern genetic techniques.
To the winemaker, this yeast is integral to crafting wonderful,
Accidental Beginnings
The winemaker’s bug, Saccharomyces cerevisiae, is so closely asso-
ciated with humans it is rarely found in environs removed from
human habitation.1In fact, its evolutionary success can probably
be explained by its relationship with humans, particularly in the
production of alcoholic beverages, an activity that has been with
us for at least 7,000 years.2Because of us, S. cerevisiae enjoys
phenomenal reproductive success with, for example, an estimated
600,000 tons of baker’s yeast being produced every year.3But
how did this close relationship get started?
It is likely that the first alcoholic fermentations were “happy
accidents”: harvested grapes were not eaten quickly enough and
began to rot, Saccharomyces spp “moved in” and took advantage
of the free sugary meal and the first wines were made (Fig.1).
These early wines presumably tasted good and had an interesting,
pleasing, psychotropic effect. One can only assume that early
farmers learned from this experience and repeated the “accidents”
of previous “vintages.” Winemaking was born and wine yeast had
a secure future in the hands of its human guardians.
Modern Winemaking:
To Intervene or Let Nature Take its Course
It has probably been known since the earliest of times that wine is
susceptible to spoilage. The discovery that it is largely microbes
that are responsible for this has led, in more recent times, to a
debate on whether wine should be “natural” or whether we should
protect it from undesirable microorganisms by “pasteurization” or
the addition of sulfur dioxide. And this debate has broadened in
the past decade to include questions on whether wine should be
the fermentation product of its “natural” microflora or of a
controlled inoculated wine yeast (Fig.1).
Some winemakers and commentators believe that the ambient
yeast population in the vineyard and winery constitute part of the
characteristics of a “natural” wine. They believe that the unique
contributions of diverse yeast species confer a complexity upon
wine not seen in inoculated ferments. This might be true,4but it
comes with the risk of spoilage. There’s also an increased risk that
the fermentation will become “stuck”1—i.e., the ferment will stop
and be difficult to restart. In addition, spontaneous—“natural,”
“wild” or “feral”—ferments also tend to take longer.
*Correspondence to: Isak S. Pretorius; Email: sakkie.pretorius@unisa.edu.au
Submitted: 12/13/11; Revised: 02/09/12; Accepted: 02/13/12
http://dx.doi.org/10.4161/bbug.19687
REVIEW
Bioengineered Bugs 3:3, 147–156; May/June 2012;G2012 Landes Bioscience
www.landesbioscience.comBioengineered Bugs147

Page 2

© 2012 Landes Bioscience.
Do not distribute.
inocula at the start of fermentation. This is a very fertile research
field where advances in wine yeast strain development and
fundamental yeast science have leveraged from one another.
This new era in wine yeast research, embracing cross-
disciplinary expertise, is worthy of review. It started following
the revelation that genes are made of DNA; the stage was set
for an explosive growth of knowledge, driven by a convergence of
genetics, biochemistry, cell biology, microbiology and computing.
And work on yeast was often at the forefront of developments.
There were compelling reasons for molecular biologists from
all fields to look on this simple single-celled fungus as the ideal
“guinea pig” for fundamental research. Our close relationship
with S. cerevisiae in food and beverage production over millennia
tells us that it is safe to work with; for example, it is designated
“Generally Recognized as Safe” (GRAS) by the United States’
Food and Drug Administration.5In addition, it is inexpensive
and easy to grow and can be stored for long periods in sus-
pended animation. But, perhaps, its best asset is an accessible
genetic system that can be followed through asexual and sexual
cycles. The three basic cell types—a, a and a/a cells—can
undergo mitosis and reproduce through an asexual budding
process. The a and a haploid cells are also able to undergo mating,
a sexual process that culminates in nuclear fusion and creation of
a/a diploid cells, which can be induced to undergo meiosis to
produce asci carrying four haploid spores.
Since the mid 1970s, when recombinant DNA technologies
revolutionized the way research in biological sciences is con-
ducted, S. cerevisiae has been one of the most important model
The debate continues and sets the backdrop for wine yeast
research, but over the past two decades, active yeast strain
development programs have been launched the world over to
generate strains that can improve wine quality when used as
organisms in molecular biology and emerging fields. For exam-
ple, a haploid laboratory strain (S288c) was the first eukaryote
to have its genome sequenced, a feat achieved through a
collaborative international effort involving more than 600
scientists under the able leadership of André Goffeau and
Stephen Oliver.6This paved the way for the first chip-based
gene array experiments.7
S. cerevisiae was also the first organism to be used to build a
systematic collection of bar-coded gene deletion mutants enabling
high throughput functional genomics experiments.8. But the most
important resource available to the yeast scientific community is
the Saccharomyces Genome Database (SGD; www.yeastgenome.
org), which provides, free of charge, access or links to the most
comprehensive data sets (genomic, transcriptomic, proteomic,
metabolomics, etc.) available to a molecular biologist. All of this
has been achieved by international collaborations on a grand scale.
What does all of this mean for wine research? The above
international efforts have put the winemaker’s bug center stage
in thousands of laboratories worldwide. And our knowledge is no
longer limited to the S288c laboratory version of S. cerevisiae
whose genome sequence was announced in 1996. Recently, the
genomes of several industrial yeast strains were sequenced,
including two ale strains (Foster’s O and Foster’s B) and five
wine yeast, AWRI 696, QA23, VIN7, VIN13 and VL3,9,10
enabling us to conduct comparative genomic analyses. This means
we can better understand what makes wine yeasts “tick” and why
there is such variation in S. cerevisiae “winemaking phenotypes”.
Engineering Yeast to Make Better Wine
Wine research and wine yeast strain development are certainly
well placed to benefit from the privileged place that S. cerevisiae
Figure1. To the winemaker, yeast is integral to crafting wonderful, complex wines from simple, sugar-rich grape juice. Grape juice is converted into wine
by the action of wine yeast. Some wine components are wholly generated by yeast as part of metabolism while others are essentially created by
the grapevine. The large number of compounds synthesized or modified by wine yeast have a major impact on wine quality and style.
148Bioengineered Bugs Volume 3 Issue 3

Page 3

© 2012 Landes Bioscience.
Do not distribute.
glycerol transporters, mostly leading to increased glycerol yields
and accompanying reduced ethanol production. Alternative
approaches have included expression of the Aspergillus niger
occupies in the life sciences. The following includes some
examples that demonstrate this.
Getting control of alcohol levels in wine. Without question
the greatest challenge faced by the wine industry is rapidly
mounting concerns over alcohol consumption; excess consump-
tion creates problems for society and human health. In addition,
too much alcohol in wine can overwhelm flavor and make the
wine “hot” on the palate. The technical challenges associated with
reducing the alcohol content of wine, however, are substantial.11
Several genetic modification (GM) based metabolic engineer-
ing strategies have been explored to generate wine yeasts that
partially divert carbon metabolism away from ethanol production,
with the aim of decreasing ethanol yields during vinification.
Two glycerol-3-phosphate dehydrogenase isozymes, GPD1 and
GPD2, which divert carbon from glycolysis to glycerol produc-
tion, have proven to be the best candidates to date (Fig.2).
Enhanced expression of either GPD paralog achieved the
desired outcome with regard to ethanol yields;12-14however,
increased glycerol production was accompanied by undesirable
increased concentrations of acetic acid. This was probably due
to a perturbation in redox balance in the engineered strain,
requiring the action of one or more of the five aldehyde
dehydrogenase (Ald) isozymes; these enzymes help maintain
redox balance by reducing coenzymes NAD+or NADP+, when
they oxidize acetaldehyde to acetic acid. The problem, however,
was alleviated quite simply by knocking out ALD6.15,16
Similar approaches11have targeted S. cerevisiae pyruvate
decarboxylase isozymes, alcohol dehydrogenase isozymes and
glucose oxidase encoding gene (GOX) in S. cerevisiae, which
redirects glucose to gluconic acid, and extensive modification of
S. cerevisiae hexose transporters, which forces the yeast to respire
rather than ferment, regardless of the concentration of glucose
and fructose it encounters. Arising from these research efforts
are several promising candidate “low-ethanol” wine yeast strains
awaiting widespread acceptance of their use in commercial
winemaking. These strains could immediately enable production
of wines that contain 12% alcohol instead of 15%, from optimally
ripened grapes.
Enhancing varietal wine flavor during fermentation. To casual
wine drinkers it may seem fanciful, even pretentious, when a wine
enthusiast states that Shiraz offers impressions of “black pepper;”
Pinot Noir displays overtones of “earthy strawberries;” or that
Sauvignon Blanc is characterized by traces of “asparagus” and
“passionfruit.” But these descriptors of wine flavors may become
clearer to casual wine drinkers if they are informed that, for
example, strawberry flavor is considered to be relatively complex
among fruits and contributed to by a large number of aroma
compounds;17and that wine flavor is really the sum of complex
interactions between more than a thousand volatile compounds,
many of which overlap with those found in strawberry, and some
impact compounds that are, in fact, found in black pepper
(e.g., rotundone18) and passionfruit (e.g., polyfunctional thiols19).
Therefore, perceiving the aromas of these aforementioned fruits,
vegetables and spices in wine is not surprising. The relative
amounts of each compound, and the resultant flavor profile,
ultimately define differences among the vast array of wines and
wine styles produced throughout the world.
Grape variety is the starting point for differentiation—many
volatile compounds provide varietal distinction in addition to
Figure2. Reducing alcohol levels in wine: several GM-based strategies have been explored to generate wine yeasts that partially divert sugar metabolism
away from ethanol production. (A) Two glycerol-3-phosphate dehydrogenase isozymes, GPD1 and GPD2, can be harnessed to divert carbon from
glycolysis to glycerol production. However, increased glycerol production was accompanied by undesirable increased concentrations of acetic acid.
This problem was alleviated by knocking out ALD6. (B) Wild-type yeast convert most of the sugar they consume into ethanol and CO2.
www.landesbioscience.comBioengineered Bugs149

Page 4

© 2012 Landes Bioscience.
Do not distribute.
By mining the S. cerevisiae genome for putative carbon-sulfur-
lyase encoding genes, we identified four candidates (BNA3, CYS3,
giving wine its basic structure. The concentrations of many
volatile compounds are, however, dependent upon an almost
infinite number of variations in production, whether in the
vineyard or the winery. It is known, for example, that commercial
yeast strains possess different abilities to form and modulate
volatile compounds during alcoholic fermentation (Fig.3) that
significantly affect the flavor and overall quality of wines.20
Therefore, while the proportion of wine volatiles modulated by
yeast may be relatively low,21the choice of yeast strain controlling
fermentation is an effective method for shaping wine aroma
according to the preferences of consumers in target markets.22
A case in point is the incidence of powerful synergies between
Sauvignon Blanc grapes and yeast strains in formation of the
compounds responsible for “box-hedge” and “tropical fruit”
flavors—the polyfunctional thiols 4-mercapto-4-methylpentan-2-
one (4MMP), 3-mercaptohexan-1-ol (3MH) and 3-mercaptohexyl
acetate (3MHA). Odorless cysteine and glutathione conjugates
of 3MH and 4MMP form in the grape berry and during crush-
ing and can be found at higher concentrations in Sauvignon
Blanc juice in comparison to other white varieties.23Due to the
potency of the free thiols, with perception thresholds in the ng/l
range, only a fraction of the available conjugated precursors need
to be released to impart strong “passionfruit,” “grapefruit,”
“gooseberry” and “guava” flavors to wine. Yeast carbon-sulfur-
lyase enzymes are responsible for the release of 4MMP and
3MH from their cysteine conjugates, while 3MHA is produced
by yeast metabolism through the esterification of 3MH during
fermentation (Fig.4).
GLO1 and IRC7) that when deleted decreased the ability of yeast
to release 4MMP.24Subsequent studies have narrowed this list to
one gene, the β-lyase encoding IRC7,25and in fact, a particular
allele of this gene26as the main determinant of 4MMP formation
during winemaking. This line of research has, therefore, provided
a clear quantitative trait locus (QTL) for molecular breeding of
wine yeast. 3MH release, on the other hand, is not mono-
genetically determined;25,26therefore, optimization of its release
through non-GM approaches remains an empirical exercise.
Early research into precursors for the polyfunctional thiols
utilized a column-immobilized Escherichia coli carbon-sulfur
lyase enzyme, apo-tryptophanase, in a method designed to
measure aromatic potential.27We engineered a wine yeast,
VIN13, to constitutively express the gene encoding this enzyme,
tnaA.28Wine made from warm-climate Sauvignon Blanc grapes
with this yeast exhibited intense tropical characters, while in
model ferments the VIN13-tnaA strain released up to 20-fold
more 3MH and 4MMP.
The same wine yeast, VIN13, engineered to overexpress
S. cerevisiae alcohol-acetyltransferase encoding genes (ATF1 and
ATF2), was able to produce high concentrations of acetate
esters.29In neutral-tasting grape varieties, such as Colombard,
this results in lifted “banana” characters. Overexpression of ATF1
in VIN13 increased conversion of 3MH into its acetate ester,30
3MHA, which, due to its lower perception threshold, is signi-
ficantly more potent. Taken together, it is therefore possible to
specifically manipulate 4MMP production by controlling the
expression of S. cerevisiae’s own IRC7; and to release high
concentrations of both 3MH and 4MMP through the expres-
sion of E. coli’s tnaA gene in yeast and increase the potency of
Figure3. Commercial yeast strains possess different abilities to form and modulate compounds that impact on wine sensory properties.
These compounds are produced as a result of yeast metabolic processes.
150Bioengineered Bugs Volume 3 Issue 3

Page 5

© 2012 Landes Bioscience.
Do not distribute.
“tropical fruit” aromas by boosting the conversion of 3MH into
3MHA by overexpression of S. cerevisiae’s own ATF1 gene. While
significantly less powerful than the tnaA gene product, modest
enhancement of 3MH release is also possible through expression
of the S. cerevisiae cystathionine β-lyase encoding gene, STR3,31
thereby paving the way for a range of “self-cloned” thiol-
modulating wine yeast.
Polyfunctional thiols are not the only sulfur-containing aroma
compounds that contribute to wine style. One need not be a wine
expert to know that “reductive” aromas with descriptors such
as “rotten egg,” “burnt rubber” and “sewage” are not going to
appeal to wine consumers. While there are several chemical and
biological mechanisms that contribute to “reductive” aromas in
wine, “rotten egg” gas, also known as hydrogen sulfide (H2S), is
largely a by-product of yeast metabolism. Under certain ferment-
ation conditions, most wine strains of S. cerevisiae produce
H2S hile incorporating inorganic sulfur into the amino acids
methionine and cysteine—a process known as the sulfate
reduction sequence (SRS) pathway.
Several GM strategies in the laboratory have been successful in
limiting H2S production by S. cerevisiae and these are generally
based on overexpression or inactivation of one or more genes
involved in the SRS pathway.32,33One of the targets has been
sulfite reductase, which comprises two a- and two β-subunits
(a2β2) encoded by yeast’s MET10 and MET5 genes, respectively.
This knowledge informed a classical, non-GM, mutagenesis
approach to develop three “low-H2S” strains, derived from the
widely-used commercial wine yeast Maurivin PDM (Fig.5).34
These strains, commercialized under the names Maurivin
Advantage, Platinum and Distinction, provide winemakers with
new strategies to manage “reductive” aromas, especially in grape
musts low in assimilable nitrogen.
Making the first modest moves with GM yeast strains.
Australia, New Zealand and many European countries have
effectively banned the use of genetically-modified organisms
(GMOs) in commercial wine production. A multitude of inter-
connected agronomic, business, regulatory, cultural and social
factors have led to these bans, but consumer sentiment is clearly
one of the main drivers. While it is unlikely that the situation
will change in the near future, the first modest move to release
a commercialized GM yeast to market was made in 2005; a
transgenic wine yeast, ML01, was given the green light from
regulatory authorities in the USA, Canada and Moldova.
ML01 carries genes that enable it to perform malolactic
fermentation (MLF), in which grape-derived malic acid is
deacidified (decarboxylated) to lactic acid. MLF is performed
by lactic acid bacteria, particularly Oenococcus oeni, following
alcoholic fermentation. However, O. oeni is rather fastidious,
being inhibited by a range of conditions typical of fermented
grape juice (e.g., low pH, high alcohol content and poor nutrient
availability) and can become “stuck” or sluggish.35In addition,
some lactic acid bacteria produce biogenic amines that impose
Figure4. There are powerful synergies between Sauvignon Blanc grapes and yeast strains in formation of the compounds responsible for tropical fruit
flavors: 4-mercapto-4-methylpentan-2-one (4MMP), 3-mercaptohexan-1-ol (3MH) and 3-mercaptohexyl acetate (3MHA). Odorless cysteine and
glutathione conjugates are converted to aromatic thiols by carbon-sulfur-lyase enzymes. Alcohol acetyl transferase further modifies 3MH, converting it
to the more potent 3MHA.
www.landesbioscience.com Bioengineered Bugs151