Journal of Applied Microbiology ISSN 1364-5072

ORIGINAL ARTICLE

The timing of diammonium phosphate supplementation

of wine must affects subsequent H2S release during
fermentation
A. Mendes-Ferreira, C. Barbosa, A. Inês and A. Mendes-Faia
Institute for Biotechnology and Bioengineering, Centre of Genetics and Biotechnology, (IBB ⁄ CGB-UTAD), Universidade de Trás-os-Montes e Alto
Douro, Vila Real, Portugal

Keywords Abstract
DAP supplementation, fermentation,
hydrogen sulfide, wine, yeast. Aims: The aim of this study was to evaluate the impact of supplementation by
diammonium phosphate (DAP) on hydrogen sulfide (H2S) production, when
Correspondence DAP given either prior to fermentation or during the early stationary growth
Ana Mendes-Ferreira, Institute for phase of yeast.
Biotechnology and Bioengineering, Centre of
Methods and Results: Three contrasting Saccharomyces cerevisiae wine strains
Genetics and Biotechnology, (IBB ⁄ CGB-
UTAD), Universidade de Trás-os-Montes e
were used to ferment synthetic grape juice (GJ) containing 67 mg l)1 of initial
Alto Douro, Vila Real, Portugal. yeast assimilable nitrogen (YAN), supplied either as DAP or as mixture of
E-mail: anamf@[Link] amino acids. Sufficient DAP was added either prior to or 72 h after the initia-
tion of fermentation to achieve a final YAN concentration of 267 mg l)1. Sup-
2009 ⁄ 0692: received 17 April 2009, revised plementation prior to fermentation stimulated H2S production. The results
26 May 2009 and accepted 4 June 2009 obtained in model solutions were validated using natural GJ.
Conclusion: The timing of DAP supplementation is critical for ensuring that
doi:10.1111/j.1365-2672.2009.04457.x
fermentation proceeds without excessive release of H2S.
Significance and Impact of the Study: This result has important implications
for the wine-making industry, because it highlights the value of determining
the initial nitrogen level of a GJ. It raises awareness of the dependence of wine
quality on the correct timing of DAP supplementation.

compounds could arise from reactions of H2S with other

Introduction
wine components (Rauhut 1993). In Saccharomyces cerevi-
One of the most important requirements for the yeast in siae, sulfide is a metabolic intermediate of the tightly reg-
wine-making is to be able to complete the fermentation ulated sulfate reduction sequence (SRS), which represents
process, reducing the residual fermentable sugar concen- a part of the biosynthesis of sulfur-containing AA path-
tration to below 2 g l)1. Grape juice (GJ) contains suffi- way (Thomas and Surdin-Kerjan 1997). The relationship
cient sugar to meet the yeast’s nutritional needs, but between nitrogen deficiency and H2S formation is
frequently a low content of assimilable nitrogen [amino thought to reflect the combined activation of the SRS and
acids (AA) and ammonia], which has been associated the depletion of intracellular nitrogenous precursors of
with poor yeast growth (Varela et al. 2004), stuck ⁄ slug- sulfur AA, which would normally act to incorporate sul-
gish fermentation (Salmon 1989; Bely et al.1990; Bisson fide (Jiranek et al. 1995a). An ineffective incorporation of
1999; Mendes-Ferreira et al. 2004) and hydrogen sulfide sulfide can therefore lead to the formation, and ultimately
(H2S) release (Vos and Gray 1979; Henschke and Jiranek the release of H2S from the cell (Henschke and Jiranek
1993; Jiranek et al. 1995a; Gardner et al. 2002). The 1993). Under given growing conditions, the extent of H2S
release of H2S is of major concern to the wine-maker production is dependent on yeast strain (Spiropoulos
because it imparts an unpleasant rotting odour to the et al. 2000; Gardner et al. 2002), and some attempts have
wine even when present at extremely low concentrations. been made to predict the tendency of a yeast strain to
In addition, other persistent and disagreeable volatile S produce off-flavours during fermentation (Jiranek et al.

ª 2009 The Authors

540 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549
A. Mendes-Ferreira et al. Effect of DAP supplementation on H2S release

1995b; Mendes-Ferreira et al. 2002). A recent analysis has nineÆHCl, 1Æ70 g l)1 asparagineÆH2O, 3Æ50 g l)1 aspartic
shown that the negative correlation between the initial acid, 6Æ22 g l)1 glutamic acidÆHCl, 2 g l)1 glutamine,
nitrogen content in the medium and the amount of H2S 0Æ74 g l)1 glycineÆHCl, 2Æ04 g l)1 histidineÆHClÆH2O,
generated is strain dependent (Mendes-Ferreira et al. 2 g l)1 isoleucine, 3 g l)1 leucine, 3Æ12 g l)1 lysineÆHCl,
2009). 1Æ5 g l)1 phenylalanine, 5 g l)1 proline, 4 g l)1 serine,
Nitrogen supplementation, in particular by diammo- 3Æ50 g l)1 threonine, 1 g l)1 tryptophan, 0Æ20 g l)1 tyrosine
nium phosphate (DAP), is widely used not only to inhibit and 2 g l)1 valine. The given solution contains 7Æ20 g l)1 of
H2S release, but more importantly to prevent sluggish YAN (sum of AA without proline). The pH was adjusted to
and stuck fermentations. Optimizing the timing of DAP 3Æ7 with NaOH prior to sterilization by filtration. Control
provision to improve fermentation kinetics has been fermentations (DAP67 and AA67) were conducted without
attempted by a number of researchers (see review by Bell any supplementation. DAP supplementation (200 mg YAN
and Henschke 2005). The supplementation of inorganic per litre) was given either prior to fermentation (DAP0 h,
forms of nitrogen throughout fermentation can be effec- AA0 h) or at 72 h after the beginning of fermentation
tive in assuring rapid sugar degradation, particularly in (DAP72 h, AA72 h).
low nitrogen musts (Bely et al. 1990). However, the effect Natural GJ was obtained by pressing grapes of the cvs.
of variation in the timing of such supplementation on Touriga Nacional and Tinta Roriz, clarified by centrifuga-
H2S formation to nitrogen-deficient GJs has not been tion at 12 734 g for 10 min and held at 65C for 10 min.
fully investigated. Thus, here we set out to evaluate the The must was stored at )20C. Yeast strain UCD522 was
effect of adding DAP both prior to and during fermenta- used to ferment this GJ either nonsupplemented or with
tion to determine how best to minimize H2S release. DAP supplementation at either 0 h (GJ0 h) or 72 h (GJ72 h).
Each set of fermentations was performed in triplicate.
Materials and methods
Inocula and fermentation conditions
Strains and maintenance conditions
Starter cultures were prepared by growing the yeast over-
The three S. cerevisiae strains used were UCD522 (kindly night in 100-ml flasks, containing 70-ml synthetic GJM
supplied by the Enology Culture Collection, Department with 267 mg l)1 YAN supplied as DAP. The flasks were
of Viticulture and Enology, University of California, incubated at 25C in an orbital shaker set at
Davis, USA); PYCC4072 (obtained from the Portuguese 150 rev min)1. The experimental cultures were inoculated
Yeast Culture Collection (PYCC), New University of Lis- with 5 · 105 CFU ml)1 of starter culture. Fermentation
bon, Portugal and originally isolated from a sample of was carried out in 500-ml flasks filled to 2 ⁄ 3 of their vol-
Fermivin, an industrial wine yeast distributed by Rapi- ume. The flasks were fitted with a side-arm port sealed
dase; and EC1118 (obtained from a commercial source as with a rubber septum to allow anaerobic sampling and
dry yeast). UCD522 is a high producer of H2S when were maintained at 20C in an orbital shaker set at
grown on a solid medium, PYCC4072 produces an 120 rev min)1. Weight loss was used to estimate the vol-
intermediate amount, and EC1118 is a poor producer ume of CO2 produced. Aseptic sampling was accom-
(Mendes-Ferreira et al. 2002). All yeast cultures were plished with a syringe-type system, in which a stylet was
maintained at 4C on yeast peptone dextrose agar (YPD) inserted in the needle holder to avoid medium accumu-
slants, containing 20 g l)1 glucose, 10 g l)1 peptone, lating in the system. Each flask was sealed with a rubber
5 g l)1 yeast extract and 20 g l)1 agar. Immediately before stopper, allowing the fermentation gases to escape
use, the yeasts were transferred to a new slant of YPD through a glass tube connected to a two-way valve by tef-
and cultured for 24–48 h at 25C. lon tubing, which was connected to a fermentation lock
and an H2S trap. Each 24 h, all flasks were removed from
the shaker, weighed, disconnected from the trap and con-
Culture media
nected to a new trap. Samples were collected daily for
Model solutions based on the chemically defined grape assessing fermentation and growth parameters. Prior to
juice medium (GJM) formulated by Henschke and Jiranek sampling, the flasks were stirred to ensure homogeneity.
(1993) were used, with some modifications. Glucose Clinitest tablets (Roche) were used to determine whether
(200 g l)1) was used as the sole carbon and energy source, alcoholic fermentation had been completed, at which time
and the initial concentration of yeast assimilable nitrogen the viable cell number, culture dry weight and residual
(YAN) was 67 mg l)1 supplied either as DAP or as a mix- nitrogen levels were measured. All experiments were
ture of AA prepared from a stock solution containing repeated at least three times, and all reported data repre-
7Æ20 g l)1 YAN, comprising 1 g l)1 alanine, 9Æ10 g l)1 argi- sent mean values of these replicated experiments.

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549 541
Effect of DAP supplementation on H2S release A. Mendes-Ferreira et al.

(30 m · 0Æ25 mm) with a 0Æ5 lm film thickness (Agilent,

Determination of nitrogen content
Santa Clara, CA, USA). The column was maintained at
Ammonium content was determined by a continuous- 40C for 5 min after desorption, ramped at 4C min)1 up
flow analysis system, as described by Mendes-Ferreira to 200C and then at 10C min)1 up to 240C, where it
et al. (2004). YAN content in natural GJ was determined was held for 15 min. The carrier gas was helium, supplied
by the formol method (Aerny 1996). at 34 cm s)1 average linear velocity. A 0Æ75-mm liner was
used, and analysis performed in the splitless mode. All
mass spectra were acquired in electron impact mode at
Determination of cell growth and culture dry weight
70 eV using full scan with a scan range of 26–250 atomic
The optical density measured at 660 nm of appropriately mass units at a rate of 6Æ12 scans per seconds. The identi-
diluted culture samples was used to assess yeast cell fication of all compounds was confirmed by comparing
growth. Samples (either 2 · 50 ml and ⁄ or 3 · 15 ml) mass spectra and retention indices with those of authentic
were centrifuged in weighed tubes for 10 min at 2300 g, standards. The compounds were quantified in selected
washed twice with deionized water, dried by holding at ion monitoring mode by selecting the most characteristic
100C for 24 h and stored in a desiccator before weighing. ion for each compound: to eliminate variation in extrac-
tion efficiency caused by small differences in the sample
matrix (particularly ethanol content), 2-octanol was used
H2S determination
as an internal standard to quantify the analytes. SPME
The quantity of H2S released by the cultures was deter- extraction procedure, preparation of standards and SPME
mined colorimetrically following the selective collection of fibre calibration were performed according to Martorell
fermentation gases with a modified fermentation lock and et al. (2002), with minor modifications (Mendes-Ferreira
sulfide-trapping system, as described by Jiranek et al. et al. 2009).
(1995a). A cadmium hydroxide suspension (5Æ59 mmol l)1
CdSO4Æ8H2O, 15 mmol l)1 NaOH) was used as a trapping
Statistical analysis
solution, and the amine reagent contained 37Æ5 mmol l)1
N,N-diethyl-p-phenylene diamineÆHCl (Sigma, MO, USA) JMP [Link] (SAS Institute Inc., Cary, NC, USA) was
in 9Æ12 mmol l)1 H2SO4, with the FeCl3 solution (60% used to generate an analysis of variance. The Tukey test
w ⁄ w) being added immediately prior to use. This mixture was used for paired comparisons. Where significant
(325 ll) was added to 10 ml trapping solution in a 30-ml differences were found, the least significant difference test
screw-cap bottle, which was immediately closed and shaken was used to determine the differences between treat-
vigorously for 30 s. After 1 h of incubation at room ments. A value of P < 0Æ05 was considered statistically
temperature, the absorbance was measured at 672 nm. The significant.
H2S content was calculated using a calibration curve
prepared with 0–12 lg sulfide, following Acree et al.
Results
(1971) and Rees et al. (1971). H2S released was
quantitatively measured during fermentation as well as in
The effect of the timing of DAP supplementation on
the final wines.
H2S production, using DAP as initial nitrogen source
The timing of DAP supplementation had no effect on the
Determination of oenological parameters
specific growth rate of any of the strains. However, cell
Total SO2, volatile acidity, total acidity, pH and ethanol growth stopped later in the DAP0 h treatment, probably
production were determined by standard methods (Office because of the different time periods required to deplete
International de la Vigne et du Vin 1990). the available YAN, which was directly correlated with its
initial content (Fig. 1, Table 1). On the other hand, the
final biomass was significantly higher in the DAP72 h
Solid-phase microextraction (SPME) and GC–MS
treatment (Table 1). For all three strains, the length of
analysis
the fermentation period was reduced by 2 days in the
Volatile compounds released during fermentation were DAP0 h treatment in line with maximum fermentation
quantified using SPME coupled to GC–MS as described rates obtained (Fig. 1a–c, Table 1).
elsewhere (Mendes-Ferreira et al. 2009). Briefly, an For strain UCD522, independent of the onset of
Agilent 6890N gas chromatograph equipped with a ammonium exhaustion (24 h in DAP72 h and 48 h in
5973N mass spectrometer was used. The target analytes DAP0 h, see Fig. 1a,d), H2S production peaked at 48 h, at
were separated using an Innowax capillary column a time when the cells entered the stationary growth phase.

ª 2009 The Authors

542 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549
A. Mendes-Ferreira et al. Effect of DAP supplementation on H2S release

(a) 35 (d)

Nitrogen concentration (mg l–1)

350 10 300
30 300 250
Weight loss (g)

Log OD660 nm
25 250

H2S (µg l–1)

200
20 200
1 150
15 150
100
10 100
5 50 50

0 0 0 0
0 48 96 144 192 240 0 48 96 144 192 240
Time (h) Time (h)

(b) 35 350 (e)

Nitrogen concentration (mg l–1)

10 300
30 300 250
Weight loss (g)

25

Log OD660 nm
250

H2S (µg l–1)

200
20 200
1 150
15 150
100
10 100
5 50 50

0 0 0 0
0 48 96 144 192 240 0 48 96 144 192 240
Time (h) Time (h)
(c) 35 (f)
350 10 300

Nitrogen concentration (mg l–1)

30 300 250
Weight loss (g)

25
Log OD660 nm

250
H2S (µg l–1)

200
20 200
1 150
15 150
100
10 100
5 50 50

0 0 0 0
0 48 96 144 192 240 0 48 96 144 192 240
Time (h) Time (h)

Figure 1 Fermentation kinetics (d,s), hydrogen sulfide release (histograms), yeast growth (4, ) and nitrogen consumption profiles ( ,h) by
UCD522 (a and d), PYCC4072 (b and e) and EC1118 (c and f). Experimental cultures were grown in chemically-defined synthetic grape-juice
media, containing 67 mg l)1 yeast assimilable nitrogen, supplied as diammonium phosphate (DAP), and supplemented by 200 mg l)1 of DAP prior
to fermentation (open symbols) or 72 h after inoculation (filled symbols). Values represent the mean of triplicated measurements.

In the DAP72 h treatment, both cell number and fermen- supplementation appeared to have a stimulatory effect on
tation rate increased, but H2S release was not completely H2S formation, because the total H2S produced was
suppressed (Fig. 1a). By 96 h, the yeast had depleted the higher in the DAP72 h than in the DAP0 h treatment
medium of ammonium (Fig. 1d), and a further peak in (Table 1). Strain EC1118 consumed ammonium and
H2S release occurred at 120 h, with release continuing sugar more slowly than did the other two strains
until the end of the fermentation (Fig. 1a). The total (Fig. 1c,f). The level of sulfide detected, irrespective of the
amount of H2S produced by UCD522 was highest in the timing of DAP supplementation, was significantly lower
DAP0 h treatment (Table 1). The profile of ammonium than that in UCD522 or PYCC4072 (Table 1). In the
consumption by strain PYCC4072 was similar to that of DAP0 h treatment, little H2S was released, while in
UCD522 (Fig. 1e). The DAP72 h treatment also induced DAP72 h, release was detected immediately after ammo-
an increase in cell number and fermentation rate, but nium depletion and continued until the end of fermenta-
unlike for UCD522, H2S release was suppressed until the tion (Fig. 1c) at concentrations above the odour
added ammonium was depleted (Fig. 1b,e). However, the threshold (Table 1).

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549 543
Effect of DAP supplementation on H2S release A. Mendes-Ferreira et al.

Table 1 Maximum fermentation rate,

Yeast Max fermentation Dry weight Total H2S biomass yield and total H2S release from the
strain Experiments rate*(g h)1) (g l)1) (lg l)1) fermentation of model synthetic GJ by yeast
UCD522 DAP67 0Æ139 ± 0Æ005 3Æ46 ± 0Æ29 281Æ3 ± 2Æ8 strains UCD522, PYCC4072 and EC1118.
DAP67 ± DAP200 (0 h) 0Æ374 ± 0Æ004 5Æ81 ± 0Æ02 805Æ6 ± 6Æ3 Results represent mean values with their
DAP67 ± DAP200 (72 h) 0Æ253 ± 0Æ009 6Æ52 ± 0Æ22 612Æ5 ± 9Æ7 standard deviation
AA67 0Æ133 ± 0Æ006 2Æ36 ± 0Æ19 18Æ3 ± 5Æ7
AA67 ± DAP200 (0 h) 0Æ393 ± 0Æ006 5Æ55 ± 0Æ13 508Æ1 ± 32Æ6
AA67 ± DAP200 (72 h) 0Æ238 ± 0Æ001 6Æ44 ± 0Æ15 205Æ7 ± 34Æ1

PYCC4072 DAP67 0Æ102 ± 0Æ010 3Æ17 ± 0Æ16 103Æ1 ± 3Æ6

DAP67 ± DAP200 (0 h) 0Æ358 ± 0Æ005 4Æ95 ± 0Æ16 222Æ7 ± 43Æ7
DAP67 ± DAP200 (72 h) 0Æ285 ± 0Æ005 5Æ52 ± 0Æ09 339Æ8 ± 24Æ0
AA67 0Æ108 ± 0Æ000 2Æ55 ± 0Æ08 4Æ0 ± 3Æ1
AA67 ± DAP200 (0 h) 0Æ396 ± 0Æ012 5Æ40 ± 0Æ07 240Æ8 ± 42Æ9
AA67 ± DAP200 (72 h) 0Æ247 ± 0Æ011 6Æ25 ± 0Æ06 8Æ6 ± 1Æ6

EC1118 DAP67 0Æ078 ± 0Æ001 2Æ38 ± 0Æ03 11Æ2 ± 4Æ1

DAP67 ± DAP200 (0 h) 0Æ336 ± 0Æ000 4Æ47 ± 0Æ09 6Æ4 ± 2Æ5
DAP67 ± DAP200 (72 h) 0Æ293 ± 0Æ018 4Æ71 ± 0Æ09 42Æ0 ± 5Æ2
AA67 0Æ117 ± 0Æ000 2Æ32 ± 0Æ03 3Æ1 ± 0Æ4
AA67 ± DAP200 (0 h) 0Æ280 ± 0Æ003 6Æ19 ± 0Æ07 38Æ6 ± 9Æ6
AA67 ± DAP200 (72 h) 0Æ281 ± 0Æ003 6Æ49 ± 0Æ11 0Æ0 ± 0Æ0

Subscript values indicate the YAN (mg l)1).

DAP, diammonium phosphate; AA, amino acids; YAN, yeast assimilable nitrogen; H2S, hydrogen
sulfide.
*The maximum fermentation rate was determined using time points corresponding to the
steepest decline in weight.

ner as was observed in DAP as sole nitrogen source.

The effect of the timing of DAP supplementation on
Thus, in the AA0 h treatment, the sugar consumption was
H2S production, using amino acid as initial nitrogen
reduced in 2 days (Fig. 2a,b); but for strain EC1118, the
source
timing of supplementation had no influence over the time
As the timing of nitrogen supplementation had such an needed to completely degrade the sugar (Fig. 2c). The lat-
impact on H2S production when DAP was used as the ter strain achieved a similar peak fermentation rate in
sole nitrogen source, a second set of experiments were both AA0 h and AA72 h, treatments. Only UCD522
performed under conditions chosen as closely as possible released remarkable amounts of H2S, irrespective of the
to mimic normal wine-making practice. The initial nitro- timing of the supplementation (Table 1). However, the
gen source was a mixture of AA, equivalent to 67 mg l)1 total H2S released by this strain was less than half in
YAN. AA72 h fermentation than in AA0 h fermented medium.
As given in Table 1, for UCD522 and PYCC4072, DAP This reduction in H2S release was more evident for
supplementation remarkably increased the amount of H2S strains PYCC4072 and EC1118, which released, respec-
produced compared to the control fermentation, tively, little or no H2S during AA72 h fermentation
although, in this case, fermentation extent was apprecia- (Table 1).
bly longer (results not shown). The timing of DAP sup-
plementation did not influence growth rate (Fig. 2), but a
The effect of the timing of DAP supplementation to
higher biomass was obtained in the AA72 h than in the
natural GJ on the release of H2S and other volatile
AA0 h treatment (Table 1). The release of H2S over time
compounds
by the three strains is shown in Fig. 2a–c. Based on its
analysis, we found a strain-independent trend of higher The high H2S producing strain UCD522 was exposed to a
total H2S formation when DAP was supplied at the low nitrogen (122 mg YAN per litre) natural GJ to study
beginning of fermentation (AA0 h) (Table 1). Yet impor- the impact of the timing of DAP supplementation on
tant differences in the time of release and total H2S pro- H2S release. Three fermentation conditions were carried
duction caused by yeast strain were found (Fig. 2a–c). out (i) without DAP supplementation (GJ); (ii) with
For strains UCD522 and PYCC4072, the timing of DAP DAP supplementation prior to fermentation (GJ0 h) and
supplementation affected fermentation in a similar man- (iii) with DAP supplementation 72 h after inoculation

ª 2009 The Authors

544 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549
A. Mendes-Ferreira et al. Effect of DAP supplementation on H2S release

(a) 10 35 350 10 35 350

30 300 30 300
Weight loss (g)

250

Weight loss (g)

Log OD660 nm

25 250

Log OD660 nm
25

H2S (µg l )

H2S (µg l )
20 200 20 200
1 1
15 150 15 150

–1

–1
100 100
10 10
50 50
5 5
0 0 0
0 48 96 144 192 240 0 0 0
0 48 96 144 192 240 288
Time (h)
Time (h)
35 350
(b) 10 Figure 3 Fermentative activity (d, ,s) and total hydrogen sulfide
30 300
release (histograms) by strain UCD522, in a natural grape-juice having
Weight loss (g)

250 an initial yeast assimilable nitrogen (YAN) of 122Æ5 mg l)1, supple-

Log OD660 nm

25

H2S (µg l )
20 200 mented by 144Æ5 mg l)1 YAN as diammonium phosphate, prior to
1 fermentation ( ) or at 72 h after inoculation (d). A control experi-
15 150
ment lacked any supplementation (s). Values represent the mean of
10 100 –1
triplicated measurements.
5 50
0 0 0 variation was detected for any of the oenological
0 48 96 144 192 240
parameters (ethanol, volatile acidity, pH and total SO2,
Time (h)
see Table 2) assessed in these experiments.
35 350 To evaluate whether DAP supplementation at different
(c) 10
30 300 fermentation stages could have any detrimental effect on
other aroma compounds, the wines obtained were analy-
Weight loss (g)

250
Log OD660 nm

25
H2S (µg l–1)

200
sed by GC–MS. At the end of the fermentation, 21 vola-
20
1 tile by-products could be identified and quantified
15 150
(Table 3). The levels of hexanoic and octanoic acids were
10 100
significantly higher in the DAP-supplemented wines, par-
5 50
ticularly when the supplementation was supplied prior to
0 0 0 fermentation. DAP supplementation significantly reduced
0 48 96 144 192 240
Time (h) the amounts of 2-phenylethanol and isoamyl alcohol
although this decrease was more accentuated in the GJ0 h
Figure 2 Fermentation (d,s) and yeast growth profiles (4, ) and treatment than in GJ72 h. On the other hand, the levels of
hydrogen sulfide release (histograms) by strains UCD522 (a), the acetate esters (2-phenylethyl acetate and isoamyl ace-
PYCC4072 (b) and EC1118 (c). Experimental cultures were grown in
tate) and total ethyl esters, which are responsible for the
chemically-defined synthetic grape-juice media, containing 67 mg l)1
yeast assimilable nitrogen, supplied as a mixture of amino acids, and
pleasant fruity and floral bouquet of wine, were signifi-
supplemented by 200 mg l)1 diammonium phosphate prior to cantly higher in the wines obtained from both GJ0 h and
fermentation (open symbols) or 72 h after inoculation (filled symbols). GJ72 h treatments than from untreated GJ. Some concen-
Values represent the mean of triplicated measurements. tration differences in individual ethyl esters were detected
(Table 3).
(GJ72 h). The amount of DAP added was equivalent to
145 mg l)1 YAN, so that the same final YAN concentra-
Discussion
tion as was used in the synthetic juice media applied. The
timing of DAP supplementation strongly influenced the Nitrogenous compounds are important components of
release of H2S (Fig. 3), consistent with the outcome from the wine must. Where there is an imbalance between the
the synthetic media. Addition of DAP prior to fermenta- levels of sugar and nitrogen, the growth and fermentation
tion more than doubled the amount of H2S released com- activity of the yeast is restricted, and the release of H2S is
pared to the untreated GJ (Table 2) although the accentuated (Vos and Gray 1979; Salmon 1989; Bely
duration of alcoholic fermentation was appreciably et al.1990; Henschke and Jiranek 1993; Jiranek et al.
reduced by 4 days (Fig. 3). Note that the GJ72 h treatment 1995a; Bisson 1999; Gardner et al. 2002; Mendes-Ferreira
was just as effective in reducing the duration of fermenta- et al. 2004; Varela et al. 2004). DAP supplementation is a
tion, without inducing any considerable increase in H2S common wine-making practice, as it provides an effective
released compared to the untreated GJ. No important means of avoiding sluggish and stuck fermentations (Bely

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549 545
Effect of DAP supplementation on H2S release A. Mendes-Ferreira et al.

Table 2 Maximum fermentation rate, specific

Yeast strain UCD522 growth rate, biomass yield and oenological
Experiment GJ GJ0 h GJ72 h parameters of the wines obtained from
Time to reach dryness (h) 288 192 192 fermentation of natural GJ by yeast strain
Maximum fermentation rate* (g h)1) 0Æ198 ± 0Æ007 0Æ283 ± 0Æ004 0Æ184 ± 0Æ004 UCD522. Results are shown as mean values
Specific growth rate (h)1) 0Æ062 ± 0Æ002 0Æ085 ± 0Æ001 0Æ064 ± 0Æ001 with their standard deviation
Dry weight (g l)1) 4Æ01 ± 0Æ09 4Æ79 ± 0Æ06 5Æ63 ± 0Æ21
pH 3Æ27 ± 0Æ01 3Æ37 ± 0Æ01 3Æ23 ± 0Æ01
Total SO2 (mg l)1) 20Æ1 ± 0Æ6 20Æ5 ± 0Æ0 16Æ4 ± 0Æ0
Final ethanol (% v ⁄ v) 12Æ6 ± 0Æ1 12Æ4 ± 0Æ2 12Æ4 ± 0Æ1
Volatile acidity (g l)1) 0Æ1 ± 0Æ0 0Æ1 ± 0Æ0 0Æ1 ± 0Æ0
Total acidity (g l)1) 7Æ6 ± 0Æ1 7Æ6 ± 0Æ1 7Æ8 ± 0Æ1
Total H2S (lg l)1) 403Æ8 ± 34Æ6 824Æ4 ± 88Æ9 537Æ7 ± 72Æ1

GJ, Grape Juice; GJ0 h, Grape Juice supplemented with DAP before fermentation was initiated;
GJ72 h, Grape Juice supplemented with DAP 72 h after inoculation; DAP, diammonium
phosphate; H2S, hydrogen sulfide.
*Maximum fermentation rate was determined using time points corresponding to the steepest
decline in weight.
As g l)1 acetic acid.
As g l)1 tartaric acid.

et al.1990; Mendes-Ferreira et al. 2004). Our experiments less H2S. Jiranek et al. (1995a) have suggested that the
based on model synthetic GJ media have shown that the later the onset of nitrogen depletion, the longer the time
fermentation kinetics of all three yeast strains were lag before H2S is released, and the lower the peak rate
strongly affected by the initial YAN concentration, and in and total amount of H2S released. As UCD522, even in
particular, that a low initial YAN concentration was asso- the presence of ammonium, produced sufficient sulfide to
ciated with a decline in activity during the early stages of be detectable (Mendes-Ferreira et al. 2009), it appears
fermentation. However, the initial YAN concentration that for this strain, the availability of nitrogen is not the
hardly affected the growth rate of the yeast. The timing of only factor involved in the formation of H2S. H2S release
DAP supplementation had a major effect on yeast bio- is clearly influenced by whether DAP supplementation is
mass, especially when the DAP was added at 72 h after given as a single dose at the start of the fermentation, or
the fermentation was initiated. In contrast, Hernández- whether the dosage is staggered. PYCC4072 and EC1118
Orte et al. (2006) were not able to observe any growth released little H2S when the DAP was supplied prior to
stimulation as a result of nitrogen supplementation late fermentation, whereas in UCD522, the release was
in the stationary phase. This lack of agreement could be decreased when supplementation was given after 72 h.
explained by the timing of the supplementation, which Importantly, the replacement of DAP by AA, as initial
was when half of the sugar had been consumed. By this nitrogen source, suppressed these genotypic differences.
stage, the level of ethanol in the culture was likely rather That is, for the three strains, higher H2S release was
higher than that in our experiments. Because it has been detected when DAP supplementations were given prior to
established that ethanol compromises the ability of the fermentation. The presence of sufficient methionine has
yeast cell to uptake ammonium (Cardoso and Leão been reported to lead to the repression of the SRS path-
1992), the higher ethanol concentrations in the Hernán- way (Thomas and Surdin-Kerjan 1997) and therefore to
dez-Orte et al. (2006) experiments would have been suffi- inhibit H2S release. As the methionine level in natural GJ
cient to have inhibited the yeast growth and thus reduced is typically below that shown to be repressive (Henschke
the accumulation of yeast biomass. and Jiranek 1993), the model synthetic GJM used here
In the present work, we found that unlike growth and did not include any of this amino acid. Thus, the distinct
fermentation rate, both the yeast strain and the timing of behaviour of EC1118 and PYCC4072 corroborates the
DAP supplementation strongly affected H2S release. In suggestion of Wainwright (1971) that AA other than
the experiments where ammonium was used as the sole methionine can also influence H2S release.
nitrogen source, a correlation between H2S release and In very low nitrogen media (67 mg l)1), yeast strains
nitrogen depletion was found although the extent of the produced remarkably less sulfide irrespective of the nitro-
release was strain dependent, as previously reported by gen source. It is worthwhile pointing out that while this
Jiranek et al. (1995a). The EC1118 strain differed mark- is an interesting observation it might have the disadvan-
edly from the other two in that it released considerably tage of leading to sluggish or stuck fermentation. The

ª 2009 The Authors

546 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549
A. Mendes-Ferreira et al. Effect of DAP supplementation on H2S release

Table 3 Concentration of aroma compounds in natural GJ fermented with yeast strain UCD522

Volatile compound Odour descriptor GJ GJ0 h GJ72 h

)1
Acids (lg l )
Butyric acid Rancid, cheese 393Æ2 ± 39a 381Æ8 ± 33Æ9a 406Æ2 ± 4Æ6a
Isovaleric acid Sweat, cheese 1204Æ6 ± 62Æ9c 357Æ4 ± 11Æ8a 853Æ3 ± 56Æ7b
Hexanoic acid Rancid, cheese 1138Æ3 ± 29Æ4a 2164Æ5 ± 238Æ5b 1350Æ4 ± 133Æ5a
Octanoic acid Sweat, cheese 1976Æ4 ± 42Æ3a 4407Æ9 ± 277Æ9c 2861Æ4 ± 466Æ3b
Decanoic acid Rancid, fat 3511Æ8 ± 280Æ4a 3937Æ8 ± 507Æ5a 3473Æ0 ± 445Æ8a
Alcohols (mg l)1)
2-Phenylethanol Honey, rose 139Æ4 ± 1Æ4c 90Æ8 ± 3Æ1a 103Æ2 ± 1Æ4b
1-octanol Jasmine, lemon 6Æ4E-3 ± 0Æ2E-3a 6Æ5E-3 ± 0Æ8E-3a 6Æ1E-3 ± 0Æ2E-3a
Isoamyl alcohol Alcohol, nail polish 360Æ8 ± 5Æ1c 240Æ4 ± 0Æ7a 332Æ0 ± 12Æ6b
Acetates (lg l)1)
Ethyl acetate Acetic acid 21 800 ± 500a 33 600 ± 1400b 29 800 ± 4800b
2-Phenylethyl acetate Fruity 744Æ8 ± 32Æ4a 1602Æ1 ± 118Æ2b 1378Æ6 ± 288Æ6b
Isoamyl acetate Banana, fruity 1036Æ1 ± 51Æ1a 2840Æ8 ± 114Æ4b 2923Æ6 ± 772Æ3b
Ethys esters (lg l)1)
Ethyl propionate Banana, apple 126Æ2 ± 13Æ9b 94Æ0 ± 2Æ2a 144Æ7 ± 17Æ7b
Ethyl isobutyrate Sweet, rubber 11Æ3 ± 0Æ9a 8Æ5 ± 1Æ6a 10Æ1 ± 2Æ3a
Ethyl butyrate Apple 115Æ1 ± 2Æ6a 214Æ8 ± 9Æ6b 210Æ6 ± 27Æ8b
Ethyl 2-methylbutyrate Apple 1Æ70 ± 0Æ12c 0Æ66 ± 0Æ08a 1Æ28 ± 0Æ01b
Ethyl isovalerate Fruit 2Æ54 ± 0Æ05c 1Æ05 ± 0Æ08a 2Æ16 ± 0Æ10b
Ethyl hexanoate Banana, apple 171Æ9 ± 5Æ0a 261Æ4 ± 25Æ9b 255Æ2 ± 10Æ2b
Ethyl octanoate Banana, pear, floral 120Æ8 ± 12Æ7a 247Æ2 ± 58Æ4c 176Æ7 ± 10Æ9b
Ethyl decanoate Grape 79Æ9 ± 10Æ7a 134Æ5 ± 12Æ2b 88Æ4 ± 20Æ1a
Diethyl succinate Fruity, wine 20Æ4 ± 0Æ3c 11Æ2 ± 0Æ3b 9Æ6 ± 0Æ2a

Values in same row shown with different letters differ significantly from one another (P < 0Æ05).
GJ, Grape Juice; GJ0 h, Grape Juice supplemented with DAP before fermentation was initiated; GJ72 h, Grape Juice supplemented with DAP 72 h
after inoculation; DAP, diammonium phosphate.

inferior amounts of H2S detected could not be attribut- when DAP was added at the beginning of fermentation.
able to the reduced fermentation rate observed since every This could be because those compounds are produced
24 h the CO2 was forced to be released into sulfide trap very early in fermentation, as reported by Stashenko et al.
by vigorous agitation. Probably by the same reason, we (1992) and Vianna and Ebeler (2001), and therefore the
could not find sulfide in the finished wines even in those stimulatory effect of DAP might be partially lost.
with late sulfide production. In sum, the quantity of YAN is closely associated with
The timing effect of DAP supplementation seen in the aroma development in wine. Yeast genotype is a signifi-
model synthetic GJ was reproduced in a low nitrogen cant variable in this context, preventing the estab-
natural GJ using strain UCD522. Variation in the concen- lishment of any general relationship between YAN
tration of particular aroma compounds was present composition or concentration and H2S release. However,
within wines obtained from both untreated and DAP- this study shows that independent of yeast strain, con-
supplemented natural GJs, in agreement with previous sidering the two parameters, fermentation length and
studies (Hernández-Orte et al. 2005; Ugliano et al. 2008). H2S release, supplementation of nitrogen-deficient musts
In addition to the quantitative variation for each aroma should be carried out 72 h after the initiation of fermen-
compound (for which the yeast genotype may be respon- tation, corresponding to early stationary growth phase.
sible), the composition of the GJ and the fermentation To our knowledge, this is the first time such relationship
conditions used, most of them displayed the same trend between H2S formation and the timing of DAP supple-
in response to DAP supplementation at the beginning of mentation has been reported. Because late nitrogen
fermentation. In the present study, it was clear that a addition was also associated with late H2S release and
given yeast strain can produce a range of volatile profiles, given that sulfide reacts with other wine components to
depending on timing of the DAP supplementation. From form mercaptans, the possible sensory modifications
the data presented here, addition of DAP later in arising from this issue should be further investigated
fermentation had a more moderate effect on some of the under industrial conditions. Overall, the importance of
acetate esters as well as on the fatty acid ethyl esters than an accurate estimate of the initial YAN concentration

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549 547
Effect of DAP supplementation on H2S release A. Mendes-Ferreira et al.

has become clear, questioning the wisdom of prophylac- producing S. cerevisiae strains by assimilable nitrogen. Appl
tic DAP supplementation prior to fermentation. When Environ Microbiol 61, 461–467.
supplementation is needed, attention needs to be paid to Jiranek, V., Langridge, P. and Henschke, P.A. (1995b) Valida-
its timing. tion of bismuth-containing indicator media for predicting
H2S-producing potential of Saccharomyces cerevisiae wine
yeasts under enological conditions. Am J Enol Vitic 46,
Acknowledgements 269–273.
This work was carried out in the framework of project Martorell, N., Martı́, M.P., Mestres, M., Busto, O. and Guasch,
J. (2002) Determination of 4-ethylguaiacol and 4-ethylphe-
PTDC ⁄ AGR-ALI ⁄ 71460 ⁄ 2006. The authors thank J. Cout-
nol in red wines using headspace-solid-phase microe-
inho for the ammonium determination and V. Falco for
xtraction-gas chromatography. J Chromatogr A 975,
support on GC–MS analysis of wines.
349–354.
Mendes-Ferreira, A., Mendes-Faia, A. and Leão, C. (2002)
References Survey of hydrogen sulphide production by wine yeasts.
J Food Prot 65, 1033–1037.
Acree, T.E., Sonoff, E.P. and Splittstoesser, D.F. (1971) Deter- Mendes-Ferreira, A., Mendes-Faia, A. and Leão, C. (2004)
mination of hydrogen sulfide in fermentation broths con-
Growth and fermentation patterns of Saccharomyces
taining SO2. Appl Microbiol 22, 110–112.
cerevisiae under different ammonium concentrations and
Aerny, J. (1996) Composés azotes des moûts et des vins. Revue
its implications in winemaking industry. J Appl Microbiol
Suisse Vitic Arboric Hortic 28, 161–165.
97, 540–545.
Bell, S.J. and Henschke, P.A. (2005) Implications of nitrogen
Mendes-Ferreira, A., Barbosa, C., Falco, V., Leão, C. and
nutrition for grapes, fermentation and wine. Aust J Grape
Mendes-Faia, A. (2009) Hydrogen sulphide and other
Wine Res 11, 242–295.
aroma compounds production by wine strains of Saccharo-
Bely, M., Sablayrolles, J.M. and Barre, P. (1990) Automatic
myces cerevisiae in synthetic media with different nitrogen
detection of assimilable nitrogen deficiencies during concentrations. J Ind Microbiol Biotechnol 36, 571–
alcoholic fermentation in enological conditions. J Ferment 583.
Bioeng 70, 246–252. Office International de la Vigne et du Vin (1990) Recueil des
Bisson, L.F. (1999) Stuck and sluggish fermentations. Am J méthodes internationales d’analyse des vins et des moûts.
Enol Vitic 50, 107–119.
Paris: OIV.
Cardoso, H. and Leão, C. (1992) Sequential inactivation of
Rauhut, D. (1993) Yeasts – production of sulfur compounds.
ammonium and glucose transport in Saccharomyces
In Wine Microbiology and Biotechnology ed. Fleet, G.H. pp.
cerevisiae during fermentation. FEMS Microbiol Lett 73,
183–223. Chur, Switzerland: Harwood Academic.
155–159.
Rees, T.D., Gyllenspetz, A.B. and Docherty, A.C. (1971) The
Gardner, J.M., Poole, K. and Jiranek, V. (2002) Practical
determination of trace amounts of sulphide in condensed
significance of relative assimilable nitrogen require-
steam with N,N-diethyl-p-phenylenediamine. Analyst 96,
ments of yeast: a preliminary study of fermentation
201–208.
performance and liberation of H2S. Aust J Grape Wine Res Salmon, J.M. (1989) Effect of sugar transportation inactivation
8, 175–179. in Saccharomyces cerevisiae on sluggish and stuck enologi-
Henschke, P.A. and Jiranek, V. (1993) Yeasts-metabolism of cal fermentations. Appl Environ Microbiol 55, 953–958.
nitrogen compounds. In Wine Microbiology and Biotechno-
Spiropoulos, A., Tanaka, J., Flerianos, Y. and Bisson, L.F.
logy ed. Fleet, G.H. pp. 77–164. Chur, Switzerland:
(2000) Characterization of hydrogen sulfide in commercial
Harwood Academic.
and natural wine isolates of Saccharomyces. Am J Enol Vitic
Hernández-Orte, P., Ibarz, M.J., Cacho, J. and Ferreira, V.
51, 233–248.
(2005) Effect of the addition of ammonium and amino
Stashenko, H., Macku, C. and Shibamato, T. (1992) Monitor-
acids to musts of Airen variety on aromatic composition
ing volatile chemicals formed from must during yeast fer-
and sensory properties of the obtained wine. Food Chem
mentation. J Agric Food Chem 40, 2257–2259.
89, 163–174.
Thomas, D. and Surdin-Kerjan, Y. (1997) Metabolism of
Hernández-Orte, P., Ibarz, M.J., Cacho, J. and Ferreira, V. sulfur amino acids in Saccharomyces cerevisiae. Microbiol
(2006) Impact of ammonium additions on volatile acidity, Mol Biol Rev 61, 503–532.
ethanol, and aromatic compound production by different Ugliano, M., Siebert, T., Mercurio, M., Capone, D. and
Saccharomyces cerevisiae strains during fermentation in
Henschke, P.A. (2008) Volatile and color composition of
controlled synthetic media. Aust J Grape Wine Res 12,
young and model-aged Shiraz wines as affected by
150–160.
diammonium phosphate supplementation before alcoholic
Jiranek, V., Langridge, P. and Henschke, P.A. (1995a)
fermentation. J Agric Food Chem 56, 9175–9182.
Regulation of hydrogen sulphide liberation in wine

ª 2009 The Authors

548 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549
A. Mendes-Ferreira et al. Effect of DAP supplementation on H2S release

Varela, C., Pizarro, F. and Agosin, E. (2004) Biomass content Vos, P.J.A. and Gray, R.S. (1979) The origin and control of
governs fermentation rate in nitrogen-deficient wine H2S during fermentation of grape must. Am J Enol Vitic
musts. Appl Environ Microbiol 70, 3392–3400. 30, 187–197.
Vianna, E. and Ebeler, S.E. (2001) Monitoring ester formation Wainwright, T. (1971) Production if H2S by yeasts: role of
in grape juice fermentations using solid phase microextrac- nutrients. J Appl Bact 34, 161–171.
tion coupled with gas chromatography–mass spectrometry.
J Agric Food Chem 49, 589–595.

ª 2009 The Authors

Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 108 (2010) 540–549 549