Use of wild yeast and bacteria
in wine fermentation
Brewing Short Course
&
REG Meeting
Diversity of yeast and bacteria
Hanseniaspora
Saccharomyces
Oenococcus
Pichia
Candida
Pediococcus
Saccharomycodes
Leuconostoc
Dekkera
Lactobacillus
Debaromyces
Metschnikowia
… and many more
•
•
•
Acetobacter
… and many
more
11 to 20 different yeasts, every genus with sometimes more than 5 species
5 to 7 relevant bacteria including several species
Survival factors in the vineyard are completely different from the cellar
Survival factors of microorganisms
In the vineyard
• High oxygen
• No alcohol
• Low sugar levels
• High temperatures
In the cellar
• Very little oxygen
• Increasing ethanol level
• High sugar levels
• Low temperatures
Spontanous fermentation without SO2
Candida
stellata
Hanseniaspora
uvarum
„wild yeast“
Metschnikowia
pulcherrima
Typical Flavor caused by wild yeast strains
positive
negative
•
•
•
•
•
•
•
band aid
oxidized (acetaldehyde)
medically
volatile acidity and ethylacetate
hydrogensulfide
„Mousy-taint“ impressions
Lower extract and lactic flavors due
to flor yeast
• sweaty, fleshly
• Petroleum and kerosene aromas
•
•
•
•
•
•
Creamy character
Complex acidity composure
Intense mouth-feel
Exotic fruit
diverse fruity esters
Sulfur containing aroma
compounds
fruity, highly complex wines due
to the contribution of different
micro-organisms
Does spontaneous fermentation lead to more
interesting wines?
Parallel spontaneous and inocculated fermentation
of Riesling in wineries (yeast: R-HST, Lallemand)
Aroma profile with fast fermentation
Bitterness
Lemon
6
Green apple
5
Body
Apple
4
3
2
Fruity
Peach
1
0
Acidity
Quince
Green bean
Pineapple
Honey
Strawberry
R-HST
Mango/Passionfruit
Elder flower
Spontaneous
• Spontaneous
fermentation
considered more
fruit and body
• Positive
compared to
control
 Fermentation
time ~ 30 days
Aroma profile with slow fermentation
Bitterness
Lemon
6
Green apple
5
Body
Apple
4
3
Peach
2
Fruity
1
0
Quince
Acidity
• Pure culture
yeast judged
more fruity with
similar body
and bitterness
• Reduced effect
of wild yeast
flora
Pineapple
Green bean
Honey
Strawberry
R-HST
Mango/Passionfruit
Elder flower
Spontaneous
 Fermentation
time ~ 60 days
Cell count and sugar degradation
250
40
•
•
Sugar [g/L]
150
•
The higher the cell count,
the faster the fermentation
The higher the temperature,
the faster the proliferation
The slower the proliferation,
the lower the final cell count
30
20
100
10
50
0
0
0
5
10
15
20
25
30
35
Sugar degradation 21°C
Sugar degradation 18°C
Sugar degradation 15°C
Cell count 21°C
Cell count 18°C
Cell count 15°C
cells [Mio./mL] ]
200
Danger zone malolactic
fermentation
How great is the risk of unwanted bacterial
activity in spontaneous fermentations?
Comparison of 2 wines from the same must
Mineralic
Bitterness
Mouthfeel
7
6
5
Lemon
Apple
4
3
Body
2
Elder flower
1
0
Acidity
Peach
Buttery/Cheesy
Green bean
Green grass
Small Scale
Passionfruit
Honeymelon
Smokey
Winery
• Spontaneous malolactic fermentation can lead to buttery and
Sauerkraut-like characters
Effects of spontaneous MLF
Mineralic
Bitterness
400
Lemon
300
Mouthfeel
Apple
200
Body
Elder flower
100
0
Acidity
Peach
Buttery/Cheesy
Passionfruit
Green bean
Green grass
Control (Siha 7)
Honeymelon
Smokey
Small Scale 1
Small Scale 2
• Increase in
smokey, green,
and buttery
attributes
• Increase in
bitterness
• Decrease in
perceived fruity
aromas
MLF during spontaneous fermentation
3.0
Malic acid [g/L]
2.5
2.0
15°C a
15°C b
1.5
18°C a
18°C b
21°C a
1.0
21°C b
0.5
0.0
0
5
10
15
20
25
30
35
40
Days
• Speed of bacterial growth is depending on temperature
• Only delay in MLF, no inhibition by lower temperatures
Wild flora towards the end of fermentation
Lactobacillus brevis
Lactobacillus hilgardii
Decreasing yaest activity
Risk of unwanted MLF
Conclusion I
• With fast metabolic rates spontaneous fermentation can increase the
complexity of wines
• With slow metabolic rates (>60 days) spontaneous fermentations tend to
mask the fruity character of wine
• Cooler temperatures support growth of non-Saccharomyces yeasts and
increase the risk of spoilage or stuck fermentation
• Higher temperatures in the beginning increase the chance of early
fermentation start and high Saccharomyces cell counts
• The risk of unwanted MLF should be controlled with lysozyme; addition of
SO2 would inhibit non-Saccharomyces flora and change the composition
• Strategies to conduct spontaneous fermentation must be adapted to the
desired style of wine!
Does spontaneous fermentation enhance or mask the
Terroir character of wine?
• Wild yeast strains may contribute to complex aroma impressions,
but also to off-flavor formation
• Hypothesis: Terroir is influencing the composition of the wild yeast flora
Why connecting Terroir and Yeast?
Slope
Sunshine
Terroir character
Wind
Rain/Water
Soil
Cover crop
Why connecting Terroir and Yeast?
Almost the same factors
have an impact on the yeast
flora in the vineyard !
Wind
Sunshine
Rain/Water
Soil
Cover crop
Experimental setup
Normal handharvesting
Vinification under
common conditions
in the wineries
Bottling
„sterile“ handharvesting
Small scale
vinification under
„sterile“ conditions
„sterile“ handharvesting
Bottling
Vinification under
„sterile“ conditions
up to 2 %vol. alcohol
3 different wines from one vineyard
– Spontaneous fermentation in the winery incl.
Cellar-yeast (referred to as winery)
– Spontaneous fermentation small scale only with
yeast from the vineyard (small scale)
– Inoculation of pasteurized must (inoculation)
Inoculation of a
pasteurized
must
Bottling
Sensory evaluation of all experiments
4
138
Winery wines
3
132WGT
F2 (16,68 %)
2
Dried active yeast
Except for a
few wines,
grouping is
possible due
to sensory
properties
139
134WGT
1
132A
Small scale fermentations
130
134
133
0
132
133A
-1
131
134A
131A
137A
Inoculations
130A
-2
137
133WGT
-3
-6
-5
-4
-3
-2
-1
0
F1 (46,57 %)
1
2
3
4
5
Biplot (Axes F1 and F2: 63.24 %)
4
Dried active yeast
Winery wines
3
apple
132WGT
F2 (16.68 %)
2
134WGT
peach
138
139
lemon
Sauerkraut
130
cantaloup
1
acid intensity
acidity
mouth
feel
134
hay
133
132A
body
0
smoky
honey
-1
mineralic
133A
134A
131A
Inoculations
bitter
137A
Small scale fermentations
-2
131
132
130A
133WGT
137
-3
-6
-5
-4
-3
-2
-1
0
F1 (46.57 %)
1
2
3
4
5
Small scale spontaneous fermentation
Differences due to
terroir and site specific
yeasts (?)
mineralic
bitter
7
lemon
6
body
5
peach
4
3
2
mouthfeel
honey melon
1
Sonnenschein
0
Reiterpfad
Kastanienbusch
acidity
apple
acid intensity
honey
Sauerkraut
smoky
hay
•
Only little differences in fruity characters
Spontaneous fermentation in the wineries
Differences by including
the „cellar yeast“
mineralic
bitter
7
lemon
6
body
5
peach
4
3
2
mouthfeel
honey melon
1
Sonnenschein
0
Reiterpfad
Kastanienbusch
acidity
apple
acid intensity
honey
Sauerkraut
smoky
hay
•
•
Differences between fruity characters
No differences in negative parameters
Why do winery wines produce other profiles than „sterile“
small scale fermentation?
• Similar sensory results in four different vintages
• Hypothesis: yeast flora from the cellar dominates sensory profile.
• Proof: Identification and quantification of the yeast species involved
Identification of species with DGGE
Extract and
wash DNA
DNA Mix of
different species
Harvested yeast
Quantify by
evaluating
fluorescence
intensity
•
•
Differentiation on species level
No absolute quantification; educated guess
Polymerase Chain
Reaction
Compare with
Single cultures
DGGE procedure according to
Mills et al., 2000
Yeast dynamics in spontaneous fermentation
100
9
90
8
80
7
6
60
5
50
4
40
3
30
20
2
10
1
0
0
day 2
Pichia
•
•
•
•
%vol.
genera [%]
70
day 4
Hanseniaspora
day 6
day 8
Hansenula
day 10
Saccharomyces
day 13
Alcohol
At 18°C fermentation starts rapidly
Saccharomyces takes over after 6 days
Part of Hanseniaspora remains active for a longer period
Delayed onset of fermentation doubles growth time for „wild yeast“ !
Spontaneous flora without SO2 addition
•
•
Composition of yeast population changes
with increasing ethanol concentration
Above 4 %vol. ethanol only Saccharomyces
active
?
Alkoholgehalt
4 % vol.
Hansenula /
Pichia
Candida /
Metschnikowia
Kloeckera /
Hanseniaspora
Saccharomyces
Modified from Großmann et al. 2005
Conclusions II
• Spontaneous fermentation show higher batch-to-batch variation
• Cooler temperatures favor growth of wild yeast and yields a
different aroma profile
• Composition of the wild yeast flora varies only slightly between
sites
• Spontaneous fermentations in the wineries had more more
Saccharomyces yeasts than those fermented under sterile
conditions (cellar yeast flora)
• The individual cellar yeast flora seems to dominate the final
sensory properties of wines from spontaneous fermentations
and not the flora of individual vineyard sites
Impact of yeast composition on the
flavor profile of wine?
• Hypothesis: Variation among volatiles in finished wines is related to
yeast composition during fermentation.
• Use of analytical fingerprints to display possible differences
Analytical „Fingerprints“
• Comprehensive two dimensional gas chromatography yields high
separation of complex mixture of volatiles
• Fingerprinting method to display differences between wines
Volatiles explaining the most variation
Variables (Axes F1 and F2: 79.09 %)
74
89 19
53
1
0.75
160
43
F2 (24.21 %)
0.5
50 250
208
169
0.25
186
124
65
185
116
• Selection of 23
compounds out of 283
based on explained
variability
0
55
-0.25
Mainly:
Alcohols
Esters
Hydrocarbons
-0.5
-0.75
-1
-1
-0.75
-0.5
-0.25
0
F1 (54.88 %)
0.25
0.5
0.75
1
Variation among wines due to volatiles
•
132WGT
Winery wines
131
Dried active yeast
•
134WGT
139
134
132A
•
130
138
133
137
131A
133WGT
132
134A
130A
Dried active yeast and
inoculations together
– Influence of wild
yeast strains less
important
– Dominance of
Saccharomyces
„pure“ spontaneous
fermentation batches on
one side
More variation among
winery wines than
among „sterile“ batches
– Impact of cellar flora
increases
differences between
wines
133A
6
4
Inoculations
2
137A
0
2
4
Small6 Scale spontaneous
fermentation
How to make it work in your cellar?
What to recommend to the
winemakers?
Ways to conduct spontaneous fermentation
„Real“
spontaneous
fermentation
random
population of
microorganisms
from vineyard and
cellar
„Russian Roulette“
Ways to conduct spontaneous fermentation
„Real“ spontaneous
fermentation
random population of
microorganisms from
vineyard and cellar
Conducted
spontaneous
fermentation I
Yeast from successful
spontaneous
fermentation used to
inoculate other
batches
Ways to conduct spontaneous fermentation
„Real“ spontaneous
fermentation
random population of
microorganisms from
vineyard and cellar
Conducted spontaneous
fermentation I
Yeast from successful
spontaneous
fermentation used to
inoculate other batches
Conducted spontaneous
fermentation I
(Pied de Cuve)
Early harvesting of
healthy grapes to
establish a starter culture
for further inoculation
Ways to conduct spontaneous fermentation
„Real“ spontaneous
fermentation
random population of
microorganisms from
vineyard and cellar
Conducted spontaneous
fermentation I
Yeast from successful
spontaneous
fermentation used to
inoculate other batches
Conducted
spontaneous
fermentation I (Pied de
Cuve)
Early harvesting of
healthy grapes to
establish a starter
culture for further
inoculation
„optimized“
spontaneous
fermentation
spontaneous start
followed by
inoculation with
dried active yeast
„Diversity due to
wild yeast, safety
by using
Saccharomyces“
Be on the wild or safe side of fermentation
„Real“ spontaneous
fermentation
random population of
microorganisms from
vineyard and cellar
High
•
•
Conducted spontanous
Conducted spontanous
fermentation I (Pied de
fermentation I
Cuve)
Yeast from
successful
No
significant difference
Early harvesting of
spontanous
compared
to
dried
active
healthy grapes to
fermentation used to
yeast
fermentation
establish a starter
inocculate other
batches
culture for further
inocculation
„optimized“ spontaneous
fermentation
spontaneous start followed by
inoculation with dried active
yeast
Risk of stuck fermentation
„optimized“ spontaneous fermentation including established cellar yeast
recommendable
=> influence of cellar yeast seems to be much more important for the final
flavor composition than wild yeast from the vineyards
Low
Yeast and bacteria coinoculation
Focusing on diacetyl
Diacetyl in wine
• Dicarbonyl compound with distinct buttery flavor
• Sensory threshold 0.2 mg/L – 2.8 mg/L depending on wine
matrix
• Higher concentrations are considered as off-flavors
especially in white wines
• Formation and degradation by yeast and bacteria possible
 final concentration as a result of complex metabolic
activity
• Literature relates diacetyl formation to citrate degradation
during MLF
 Use of bacteria strains that lack the responsible enzyme
(citrate lyase negative, CLneg)
Solution so far…
250
14
8
150
Lalvin VP41 (Diacetyl)
6
100
Diacetyl [mg/L]
12
10
4
50
2
0
0
0
5
10
15
Citrat [mg/L]
200
VINIFLORA CiNe
(Diacetyl)
VINIFLORA CiNe
(Citrat)
Lalvin VP41 (Citrat)
20
Time [days]
•  CiNe (Eaton Begerow, Germany) leaves citrate untouched
•  VP 41 (Lallemand, Canada) is supposed to metabolize citrate
without diacetyl formation
• … but some of the wines contained diacetyl well above the sensory
threshold!
Citric acid degradation
O
O
HO
Uptake towards
the end of MLF
Bacteria
Constant release
throughout growth
OH
O
OH
OH
From yeast sugar
metabolism
Release
during amino
acid synthesis
(valine)
Yeast
O
O
Release of reduced diacetyl
strain depending
Uptake and
reduction to
2,3-butanediol
adapted from Quintans et al., 2008
Experimental setup
 Fermented with yeast
CY 3079 (Lallemand)
Pinot blanc
2011 and 2012
 Divided in simultaneous
and consecutive MLF
 Trials run in duplicates
CiNe (Eaton‘s
Begerow)
Control
no MLF
CLNX
(FermControl)
Citrate lyase negative
VP 41
(Lallemand)
Medium
production
ALPHA
(Lallemand)
BETA
(Lallemand)
Strong production
4
1,00E+08
4
1,00E+07
3
1,00E+06
3
1,00E+05
2
1,00E+04
Malic acid
2
1,00E+03
CFU
1
1,00E+02
1
1,00E+01
0
1,00E+00
0
5
10
15
20
25
30
CFU/mL
Malic acid [g/L]
Influence of cell concentration
35
5
4
4
3
3
2
2
1
1
0
1,00E+09
1,00E+08
1,00E+07
1,00E+06
1,00E+05
1,00E+04
1,00E+03
1,00E+02
1,00E+01
1,00E+00
Malic acid
CFU
0
5
10
15
20
25
vinification time [days]
30
35
• some commercial
products contain
less viable cells
• delayed MLF
CFU/mL
Malic acid [g/L]
vinification time [days]
• no MLF with less
than 106 CFU/mL
• risk of undesired
flavor development
Accumulation of diacetyl
Simultaneous Inoculation
O. oeni
Diacetyl [mg/L]
5
4.73
4
3
Start citrate
degradation
Yeast
2
2.97
1.37
1
0.30
0
-1
1
3
5
7
9
11 13 15 17 19 21 23 25 27 29 31 33 35
vinification [days]
ALPHA
BETA
CREAM
control
• No reduction of the diacetyl formed by bacteria  accumulation
• Reason: low pH prior to fermentation (2.9) delays MLF and citrate
degradation  higher final level of diacetyl
Diacetyl concentration during vinification
1.00E+08
0.7
1.00E+07
0.5
O. oeni
Diacetyl O. oeni
Diacetyl control
CFU O.oeni
0.4
0.3
1.00E+06
0.2
1.00E+05
O. oeni and S.
cerevisiae
0.1
0.0
1.00E+04
0
2
4
6
8
10
12
14
vinification [days]
• First increase induced by both yeast and bacteria
• Second increase mainly caused by O. oeni
• Correlation between biomass production and diacetyl accumulation
CFU/mL
Diacetyl [mg/L]
0.6
Gene expression analysis (qRT-PCR)
Citrate transporter (maeP)
Citrate lyase (citE)
Lactate dehydrogenase (ldh)
 House-keeping gene
Acetolactate synthase (alsS)
Acetolactate decarboxylase (alsD)
Acetoin reductase (butA2)
adapted from
Quintans et al., 2008
Gene expression analysis (qRT-PCR)
alsS
Sugar/pyruvate
induced
Citrate induced
• first alsS over-expression induced by sugar metabolism
• second alsS over-expression induced by citrate degradation
The central role of pyruvate
Pyruvate addition
The central role of pyruvate
 Degradation von citrate also leads to
an increase in pyruvate
Diacetyl degradation by yeast
60
Diacetyl [mg/L]
50
40
Diacetyl 0 mg/l
Diacetyl 10 mg/L
Diacetyl 20 mg/L
Diacetyl 50 mg/L
control
30
20
10
0
0
2
4
6
8
10
vinification [days]
• Diacetyl addition does not affect growth or MLF rate, even in high
concentrations
• The dried active yeast strain CY 3079 (Lallemand) is able to reduce 50 mg/L
diacetyl in the first two days of fermentation
Diacetyl degradation by yeast
 Independent of the start concentration 2 mg/L diacetyl remain due to bound
diacetyl (to sulfur dioxide, amino acids and others)
What does that mean for the winemaker?
• Only a rapid MLF can inhibit diacetyl accumulation
• At least 106 CFU/mL are required to start MLF  cell count
after inoculation if commercial preparation is unknown
• Initial medium must be suitable for MLF (pH at least 3.1-3.3) 
higher pH might favor growth of other bacteria (e.g. Pediococcus
damnosus) and therefore diacetyl formation  lower pH delays
MLF and therefore favors diacetyl formation
• Once diacetyl is formed curative measures like fresh yeast
fining may be applied
Summary and conclusion
 Diacetyl accumulation shows two characteristic peaks: first increase caused by
yeast and bacteria, second increase due to bacteria only
O. oeni
 Correlation between biomass increase and
diacetyl accumulation observed (more
bacteria produce more diacetyl)
 Over-expression of alsS gene due to high
concentrations of pyruvate as well as citrate
 Formation of diacetyl by citrate
degradation only can be excluded
S. cerevisiae
 Dried active yeast (CY 3079) is able to
reduce 50 mg/L diacetyl in two days
 Independent from the starting level,
residual diacetyl of 2 mg/L always remains;
reason possibly bound diacetyl