Comparison between Malolactic Fermentation Container and Barrel

Subscriber access provided by HACETTEPE UNIVERSITESI KUTUPHANESI

Article

Comparison Between Malolactic Fermentation Container and Barrel Toasting Effects on Phenolic, Volatile and Sensory Profile of Red Wines María Reyes González-Centeno, Kleopatra Chira, and Pierre-Louis Teissedre J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.6b05497 • Publication Date (Web): 01 Apr 2017 Downloaded from http://pubs.acs.org on April 3, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 36

Journal of Agricultural and Food Chemistry

Comparison Between Malolactic Fermentation Container and Barrel Toasting Effects on Phenolic, Volatile and Sensory Profile of Red Wines

María Reyes González-Centeno,a,b,c Kleopatra Chira,a,b,c,* Pierre-Louis Teissedre,a,b a

Univ. Bordeaux, ISVV, EA 4577, Œnologie, 210 Chemin de Leysotte, 33140 Villenave d’Ornon, France

b

INRA, ISVV, USC 1366 Œnologie, 210 Chemin de Leysotte, 33140 Villenave d’Ornon, France c

Tonnellerie Nadalié, 99 Rue Lafont, 33290 Ludon-Médoc, France * Tel : + 33 (0)5 57 58 20 51, Fax : + 33 (0)5 57 57 58 13 e-mail: [email protected]

1 ACS Paragon Plus Environment


Page 2 of 36

1

Abstract

2

Ellagitannin and anthocyanin profiles, woody volatile composition and sensory

3

properties of wines in which malolactic fermentation (MLF) took place in barrels or

4

stainless steel tanks, have been compared after 12 months of barrel ageing. Three

5

different barrel toastings were evaluated.

6

Barrel-fermented wines generally presented 1.2-fold higher total phenolics, whereas

7

tank-fermented wines exhibited 1.1 and 1.2-fold greater total proanthocyanidin and

8

anthocyanin contents, respectively. Concerning ellagitannin composition, barrel toasting

9

effect seemed to be more important than differences due to MLF-container. Certain

10

woody and fruity volatiles varied significantly (p < 0.05) depending on whether MLF

11

occurred in barrels or tanks. Barrel-fermented wines were preferred in mouth, while

12

olfactory preference depended on barrel toasting.

13

This is the first study that evaluates the impact of oak wood during MLF on ellagitannin

14

composition of wine, as well as the barrel toasting effect on wine attributes during

15

ageing when MLF occurred whether in barrels or tanks.

16 17

Keywords malolactic fermentation, oak wood barrel, toasting, ellagitannins, volatile

18

composition, sensory analysis


Page 3 of 36


19

INTRODUCTION

20

Chemical composition, aromatic profile and sensory attributes of wines are bind to grape

21

variety, winemaking procedure, maturation and ageing. Traditional red wine production

22

usually consists of performing both alcoholic and malolactic fermentations in the same

23

tank, and then, ageing in oak barrels for a variable period of time, ranging from a few

24

months to over a year.1 In an attempt to obtain particular quality wines with their own

25

personality and a higher organoleptic complexity, alternative production technologies, such

26

as carrying out the malolactic fermentation (MLF) in the same oak barrels where ageing

27

will take place, are being introduced in the wineries in an increasingly widespread way.2,3

28

MLF is a biochemical stage, supported by lactic acid bacteria (LAB), typically occurring

29

upon the winemaking process.4 This bacterial activity ensures wine acidity reduction, by

30

transforming the malic acid into lactic acid, while also contributes to wine stabilization and

31

enrichment of its aroma and flavor complexity, by the production of odor-active

32

compounds and the transformation of both grape and yeast derived volatiles and flavor

33

precursors.3,5

34

Within this context, because of the own micro-oxygenation properties of oak barrels, which

35

may help to stabilize wine color, as well as the supply of volatile and non-volatile

36

compounds to the wine, these wooden containers may also offer good fermentation

37

conditions.6,7 Transfer of those compounds (mainly ellagitannins and woody volatiles) from

38

oak wood to wine will depend on the composition of the wine matured in barrels, on the

39

contact time wood-wine, as well as on the available amount of compounds potentially

40

extractable. In addition, the latter is especially conditioned by the barrel toasting, a heating

41

treatment that induces severe modifications on wood chemical composition and, in turn,

42

may influence wine composition during barrel storage.8

43

At present, despite the increasing use of barrels during MLF, there is still scarce scientific

44

investigation focused on the effect that oak wood has on the phenolic, aromatic and sensory

45

properties of red wines during this new winemaking procedure. Specifically, the effect of 3 ACS Paragon Plus Environment


Page 4 of 36

46

barrel wood during MLF has been reported on color and oenological parameters (pH,

47

alcoholic degree, volatile and total acidity),4 total phenolic contents,3,4 phenolic profile,2,9

48

some fruity and/or woody aromatic compounds,3,10-12 N compounds (amino acids and/or

49

biogenic amines),1,11 and sensory analysis.3,10,13 The above works evidenced that the nature

50

of the MLF-container may lead to variations on the phenolic and/or aromatic composition

51

of wines, changing significantly their organoleptic perception. From the above-mentioned

52

studies, only Moreno-Arribas et al.2, Izquierdo-Cañas et al.3 and Hernández-Orte et al.10

53

evaluated wine evolution during ageing after using barrels as MLF-containers. And none of

54

those studies investigated either the impact of barrel wood during MLF on ellagitannin

55

composition of wine, or the effect of barrel toasting on wine attributes during ageing when

56

MLF occurred in barrels or in tanks.

57

To learn more about what differences are due to the MLF-container and/or the barrel

58

toasting, and to enable the wine industry to exert greater control over the use of oak wood

59

during MLF, more in-depth research is required. Thus, the present study aimed to evaluate

60

the impact of both MLF-container and barrel toasting on the phenolic composition

61

(including ellagitannins), the aromatic profile and the sensory attributes of two sets of 12-

62

months aged wines: one in which MLF and ageing were carried out in oak barrels, and

63

another were MLF was performed in tanks previous to barrel ageing.

64

MATERIALS AND METHODS

65

Chemicals. Vanillin, eugenol, isoeugenol, guaiacol, β-methyl-γ-octalactone, furfural, 5-

66

methylfurfural, dodecan-1-ol, ethyl propanoate, ethyl 2-methylbutanoate, ethyl butanoate,

67

ethyl hexanoate, ethyl octanoate, ethyl 3-methylbutanoate, isobutyl acetate, butyl acetate,

68

hexyl acetate were all purchased from Sigma-Aldrich (Saint-Quentin-Fallavier, France) at

69

the highest purity available. Chlorogenic acid (≥95%), gallic acid monohydrate (≥98%),

70

Folin Ciocalteu reagent, formic acid (HPLC grade), sodium carbonate (≥99%) and sodium

71

bisulfate (95%) were also supplied from Sigma-Aldrich (Saint-Quentin-Fallavier, France).

72

Acetone, ethanol 96%, acetonitrile (HPLC grade), methanol (HPLC grade), sodium sulfate 4 ACS Paragon Plus Environment

Page 5 of 36


73

anhydrous (≥99%), sodium chloride (99%) and isoamyl acetate were obtained from VWR-

74

Prolabo (Fontenay-sous-Bois, France). Hydrochloric acid 37% and dichloromethane were

75

from Fisher Scientific (Illkirch, France), and malvidin-3-O-monoglucoside (≥97%) was

76

from Extrasynthese (Lyon, France).

77

Oak wood origin and drying conditions. All barrels used were made up of French oak

78

from two species (Quercus robur and Quercus petraea) from the same forest located in the

79

Center region of France. The raw staves (100 cm x 11 cm x 2.2 cm) were naturally

80

seasoned for 24 months in the Tonnellerie Nadalié (Ludon-Médoc, France) wood yard.

81

Once assembled, barrels (225 L) were submitted to different toasting procedures using the

82

traditional way over an oak wood fire. Specifically, three different toasting levels were

83

provided: Noisette, 62 min at 52±3 ºC; MT (medium toast) and MTAA (medium toast with

84

watering), 68 min at 57±3 ºC. In the case of MTAA toasting, the watering process took

85

place twice (20 L of water in both cases): first, before starting the toasting procedure, and

86

secondly, after 53 min of toasting; then, the barrel drum was placed again above the fire in

87

order to be heated for 15 min more. The barrel heads were not toasted. For the purpose of

88

the study, five barrels of each toasting procedure were provided to the wine cellar.

89

Red wine vinification and sample collection. Cabernet Sauvignon grapes (Vitis vinifera

90

L.) were manually harvested at maturity in Domaine Costa Lazaridi winery (Adriani,

91

Drama, Greece) during the 2013 vintage. The same day, grapes were crushed and some SO2

92

(5 g/hL) was added during the transfer of must to a stainless steel tank (50 hL).

93

Saccharomyces cerevisiae was included to perform alcoholic fermentation at 25-30 ºC.

94

Two sets of experiments were performed depending on the container where malolactic

95

fermentation (MLF) took place (Figure S1). Thus, when alcoholic fermentation concluded,

96

one part of the wine was used to fill new oak barrels (modality -B) presenting MTAA, MT

97

and Noisette toastings, whereas the other part was kept in the stainless steel tank (modality

98

-T). In both types of container, MLF extended nearly for one month at a maintained



99

Page 6 of 36

temperature of 20 ºC. The L-malic acid content of the wines was monitored by thin-layer

100

chromatography to control the development and end point of MLF.

101

Once the MLF was finished (malic acid content ≤ 0.2 g/L), wines that carried it out in

102

barrels were immediately sulfitated (3 – 3.5 g SO2/hL), stayed with the lees for a month,

103

and then, were racked and after barrels cleaning and sanitization, were returned to the same

104

ones. Meanwhile, wines that carried out MLF in stainless steel tanks, were directly racked,

105

additionally sulfitated (3 – 3.5 g SO2/hL) and transferred to new oak barrels presenting the

106

same three toastings than the barrel-fermented modality.

107

From the five barrels of each toasting procedure provided to the wine cellar, two were used

108

for the wines of the MLF-tank modality and other two for the wines of the MLF-barrel

109

modality. The fifth one was used for the ullage or fill level of the barrels during ageing.

110

Wines were kept in oak barrels for ageing during 12 months at a controlled temperature of

111

15-16 ºC. All three toasting methods (MTAA, MT and Noisette) described above were

112

tested in each set of experiments. After 12 months of wood contact, wine was sampled from

113

oak barrels of each MLF-modality and toasting method, and then, bottled and stored at 16

114

ºC until further analysis.

115

Oenological parameters in wines. Conventional oenological parameters of wines, i.e., pH,

116

alcoholic degree (%), titratable acidity (g/L tartaric acid) and total polyphenol index (TPI),

117

were determined by Infrared Spectrometry with Fourier Transformation (IRTF) with a

118

WineScanTM Flex (FOSS Analytical, Denmark), which was previously calibrated with wine

119

samples analyzed in accordance with official OIV methods.14 A duplicate per barrel was

120

performed (n=4 per each MLF-container x toasting).

121

Total phenolics, proanthocyanidins and anthocyanins analysis. Total phenolic,

122

proanthocyanidin and anthocyanin contents of wines were spectrophotometrically

123

determined. An automated microplate reader (FLUOstar Optima, BMG LabTech, France)

124

was used for the first analysis, and a V-630 UV-VIS spectrophotometer (JASCO, Japan), 6 ACS Paragon Plus Environment

Page 7 of 36


125

for the other ones. Experimental procedures and wine dilution conditions were as

126

previously described by González-Centeno et al.8 These spectrophotometric analyses were

127

all performed in triplicate (n=6 per each MLF-container x toasting).

128

HPLC analysis of anthocyanins. Anthocyanin separation was performed on an Agilent

129

Nucleosil 100-5C18 (250 mm × 4.0 mm, 5 µm) column by using a Thermo-Accela HPLC

130

instrument including a UV−vis detector (Accela PDA Detector), an autosampler (Accela

131

autosampler), and a quaternary pump (Accela 600 − pump), controlled by Xcalibur data

132

treatment software. Methodology, adapted from Lorrain et al.,15 consisted of a flow rate set

133

at 1 mL/min and UV-Vis detection monitored at 520 nm. The mobile phases were 5% (v/v)

134

aqueous formic acid (solvent A) and 5% (v/v) formic acid in acetonitrile (solvent B). The

135

initial elution conditions were set at 10% B followed by a binary mobile phase gradient:

136

35% B at 25 min, 100% B at 35 min, 100% B from 35 min to 40 min, 10% B at 41 min and

137

re-equilibration of the column with 10% B for 4 min before next injection.

138

Anthocyanin 3-O-monoglucosides (delphinidin, Dp; cyanidin, Cy; petunidin, Pt; peonidin,

139

Pn; and malvidin, Mlv), as well as the acetylated and p-coumaroylated forms of Pn and

140

Mlv, were identified by comparison to injected external standards and previous results.

141

Two replicates were performed per barrel (n=4 per each MLF-container x toasting). Results

142

were expressed in mg of Mlv-3-O-monoglucoside/L wine.

143

Oak ellagitannins of wines: extraction and identification by HPLC. Ellagitannin

144

fraction was obtained after column fractionation as described by González-Centeno et al.8

145

Prior to HPLC−UV/MS analysis, the solid residue obtained was dissolved in H2O/HCOOH

146

(996/4, 1 mL) and filtered (0.45 µm). Ellagitannin identification was performed on a

147

reversed-phase LiChrospher 100 RP18 (250 mm × 4.6 mm, 5 µm) column by using a

148

Thermo-Finnigan Surveyor HPLC system. Elution conditions, flow rate, and composition

149

of the mobile phases were fixed as previously described by González-Centeno et al.8 Each

150

target compound (castalagin, vescalagin, grandinin and roburins A-E) was identified by

151

using their molecular ion. Chlorogenic acid (20 mg/L) was used as an internal standard. For 7 ACS Paragon Plus Environment


Page 8 of 36

152

calibration, the ratio of the chromatographic peak areas between chlorogenic acid and

153

castalagin was plotted against the ratio of their concentrations. The response factor was

154

then calculated as the slope of the plot. Thus, ellagitannin concentrations, expressed as

155

equivalents of castalagin (mg castalagin/L wine), were calculated relative to the

156

chromatographic peak area of the chlorogenic acid. All ellagitannin analyses were

157

performed in a duplicate per barrel (n=4 per each MLF-container x toasting).

158

Volatile composition of wines: extraction and gas chromatography analysis. Woody

159

and fruity aroma profiles were established by adapting the gas chromatography methods

160

reported by Barbe and Bertrand,16 and Antalick et al.17 respectively. Procedures of the

161

volatile extraction prior to gas-chromatographic analyses, equipment and calibration

162

conditions were considered as specified by González-Centeno et al.8 Target compounds

163

were identified by comparing their retention times and mass spectra with those of the pure

164

reference standards. All samples were analyzed in duplicate (n=4 per each MLF-container x

165

toasting). Results were calculated from calibration curves previously established using pure

166

reference standards analyzed under the same conditions than wine samples.

167

Woody aroma. The following ions were used to identify the target compounds: vanillin, m/z

168

151; eugenol and isoeugenol, m/z 164; guaiacol, m/z 124; β-methyl-γ-octalactone, m/z 99;

169

furfural, m/z 96; 5-methylfurfural, m/z 110; and m/z 83 for the internal standard (dodecan-1-

170

ol).

171

Fruity aroma. For the identification of the target compounds, selected ions were m/z 102

172

for ethyl propanoate and ethyl 2-methylbutanoate, m/z 88 for ethyl butanoate, ethyl

173

hexanoate, ethyl octanoate and ethyl 3-methylbutanoate, m/z 70 for isoamyl acetate, and

174

m/z 56 for isobutyl acetate, butyl acetate and hexyl acetate.

175

Sensory analysis. To assess the influence of the MLF-container on the organoleptic quality

176

of wines and to elucidate the potential differences according to the barrel toasting method,

177

sensory analysis was performed by a panel of 20 expert judges (fifteen women and five

178

men) from the Institute of Vine and Wine Sciences of the University of Bordeaux. 8 ACS Paragon Plus Environment

Page 9 of 36


179

To familiarize the subjects with wine taste and aroma recognition, judges were trained over

180

a period of two months as previously described by Chira et al.18 After training sessions, the

181

judges were homogenized and became familiar with the olfactory (vanilla, fruity, spicy,

182

overall woody) and gustative (sweetness, astringency, bitterness) descriptors considered in

183

this research.

184

All evaluations were conducted in a standard sensory-analysis chamber,19 equipped with

185

separate booths, where an uniform source of lighting, absence of noise and distracting

186

stimuli were guaranteed, and the ambient temperature was maintained at 19-22 ºC. Wines

187

(30 mL) were presented in standard black wine glasses,20 covered with a Petri dish to

188

minimize the escape of volatile components and randomly coded with three-digit numbers.

189

The position of the samples was balanced in all sensory tests.

190

Descriptive sensory analysis was first performed to assess the sensory profile of wines

191

matured in barrels with different toasting methods, for each MLF-modality separately.

192

Judges were asked to rate all descriptors on a 7-point scale (0 = ‘absence’, 7 = ‘at maximum

193

intensity’), starting with evaluation of the orthonasal odor (vanilla, fruity, spicy, overall

194

woody), and then, after a short break, the taste (sweetness, bitterness) and tactile sensation

195

(astringency). The intensity level of each descriptor was then expressed as the mean value

196

of all the judges from two formal tasting sessions. The olfactory and gustative preference

197

among the three toasting methods was also requested for each MLF-modality.

198

Discriminatory tests including triangle test21 and bilateral paired comparison test22 were

199

also performed to determine whether the panel was able to distinguish between wines

200

which carried out MLF in barrels or in stainless steel tanks, and if yes, to elucidate to what

201

sensory attributes they relate that difference. Trained judges attended two formal tasting

202

sessions per each barrel toasting considered. The olfactory and gustative preference

203

between both MLF-modalities was also requested for each barrel toasting method.

204

For the triangle test, three glasses were presented and judges were asked to indicate the one

205

olfactory perceived as different from the others. For the bilateral paired comparison test, 9 ACS Paragon Plus Environment


Page 10 of 36

206

judges started with evaluation of the orthonasal odor (vanilla, fruity, spicy, overall woody),

207

and then, after a short break, the taste (sweetness, bitterness) and tactile sensation

208

(astringency).

209

Statistical analysis. All experimental results (n=4 or 6 per each MLF-container x toasting

210

and analysis) were reported as mean values with their corresponding standard deviations.

211

Statistical analysis was performed by the statistical package R version 3.1.1 (R Foundation

212

for Statistical Computing, Wien, Austria). Normality and homocedasticity of the residuals

213

were evaluated for all parameters, by using the Shapiro−Wilk test and Levene’s test,

214

respectively. As populations were normally distributed and presented homogeneity in

215

variance, parametric tests were used. Thus, data was submitted to two-way ANOVA (MLF

216

container × toasting method). If the interaction p-value was statistically significant, Tukey

217

test was directly applied to evaluate the degree of the significant differences. If not,

218

previously to the Tukey evaluation, a one-way ANOVA was run individually for data

219

corresponding to each significant factor. Differences at p ≤ 0.05 were considered to be

220

statistically significant.

221

The results of the sensory tests were analyzed by the probability theory that the number of

222

right answers follows a binomial distribution (n, p = 1/3 for triangle test, and p = 1/2 for

223

paired comparison test), where n is the panel size (n = 20). Wines were considered as

224

differently perceived for a probability lower than 5%.

225

RESULTS AND DISCUSSION

226

Oenological parameters and phenolic attributes of wines. Regardless of the MLF-

227

container and the barrel toasting, all aged wines presented a pH of 4.0, an alcohol

228

concentration of 14.5%, a titratable acidity of 3.6 g tartaric acid/L wine and a total

229

polyphenol index of 76 ± 1. The research of Gockowiak et al.23 demonstrated that wine

230

matrix, together with pH and alcohol content, may significantly impact on the rate of MLF.

231

Taking into account that the same initial wine was considered for both sets of experiments,

232

and that pH and alcohol concentration were the same for all final wines, the significant 10 ACS Paragon Plus Environment

Page 11 of 36


233

differences observed among their phenolic, aromatic and sensory attributes may be entirely

234

attributed to the different wine evolution according to the MLF-container and the barrel

235

toasting method.

236

Results of total phenolics, total proanthocyanidins and total anthocyanins for all the wines

237

in study are depicted in Table 1. Regardless of the toasting method, total phenolic content

238

of barrel–fermented wines (3903 – 4000 mg GA/L wine) was significantly higher (p
0.05) were found between

285

both MLF-container modalities for some of the barrel toastings. That is the case of MTAA

286

toasting for cyanidin-3-O-glucoside.

malvidin-3-O-glucoside,

malvidin-3-(6”-acetylglucoside),


Page 13 of 36


287

The effect of the toasting method, previously observed for the total anthocyanin content,

288

was especially evident for barrel-fermented wines, with Noisette toasting leading to wines

289

with the highest content of most individual anthocyanins (2.3-14.6% greater than MT and

290

MTAA toastings).

291

Ellagitannin composition of wines. Individual quantitation of the main wood ellagitannins

292

extracted from oak barrels is depicted in Figure 1. It is worth noting that this is the first

293

study in the literature presenting the comparison of the oak wood ellagitannins content in

294

wines whether MLF is done in tanks and then aged in barrels or whether both processes,

295

MLF and ageing, are done in barrels.

296

Total ellagitannin content of wines after 12-months ageing, calculated by adding up the

297

individual concentration of each ellagitannin compound, ranged from 8.2 ± 0.1 mg

298

castalagin/L wine to 23.9 ± 1.0 mg castalagin/L wine, for MTAA-T and MT-T wines,

299

respectively. A general ellagitannin trend persisted regardless of the MLF-container and/or

300

the barrel toasting, castalagin being the most abundant component and accounting for 77 –

301

86% of the total ellagitannin fraction. Vescalagin was the second main ellagitannin, whose

302

contribution to the total content varied between 8 and 10%. This distribution may be

303

attributed to the higher chemical reactivity of vescalagin, due to the β-position of the

304

hydroxyl group compared to its α-position in castalagin.26 The glucosidic monomers

305

grandinin and roburin E, the dimers roburin A and D, and the glucosidic dimers roburin B

306

and C were present as minor constituents (≤ 5% in all cases). These results showed the

307

same order of magnitude and ranking order as those previously reported in the literature for

308

the major ellagitannin compounds.8,27-31

309

According to the two-way ANOVA results, the effect of the barrel toasting seemed to be

310

more important than differences due to MLF-container. Indeed, a particular behavior was

311

observed for each toasting method when comparing both fermentation vessel modalities. In

312

the case of MT toasting, individual ellagitannin content was from 1.6 to 2.5-fold greater in

313

MLF-tank wines than in wines with oak wood contact during MLF (p < 0.05). A similar 13 ACS Paragon Plus Environment


Page 14 of 36

314

trend was observed for Noisette toasting (higher ellagitannin values for Noisette-T wines),

315

even if no significant differences (p > 0.05) according to the MLF-container were detected

316

for some of the ellagitannin compounds. Conversely, for the MTAA toasting, wines having

317

undergone MLF in barrels presented between 1.5 and 2.2-fold higher content of individual

318

ellagitannins than wines resulting from MLF in tank (p < 0.05), except for roburins D and E

319

which denoted similar values for both MLF-container modalities. These observations

320

suggest that the watering process during barrel toasting might play an important role in the

321

extractability of ellagitannins during MLF in barrels, whose effect is maintained through

322

the 12-months ageing period.

323

Furthermore, a particular ellagitannin profile was noted in terms of quantitation of the

324

individual compounds for each toasting method, regardless of the MLF-container.

325

Specifically, MT-T and MTAA-B wines stood out clearly from the other toastings for their

326

greatest ellagitannin content within MLF-tank and MLF-barrel modalities, respectively.

327

Overall, the barrel toasting has an important impact on individual ellagitannin composition,

328

which in turn may have sensory consequences on the final investigated wines.

329

Volatile composition of wines

330

Woody aroma. The impact of the MLF-container and the barrel toasting on the main

331

contributors to overall oak wood aroma of wines was also studied. Concentrations of the

332

furanic aldehydes furfural and 5-methylfurfural, the two isomers of β-methyl-γ-octalactone

333

(cis- and trans-whiskey lactones), the phenols guaiacol, eugenol and isoeugenol, and the

334

phenolic aldehyde vanillin, are depicted in Figure 2 for wines of both MLF-modalities after

335

12-months ageing in barrels representing three different toasting methods. Except for

336

guaiacol, the two-way ANOVA of the raw experimental data revealed that the MLF-

337

container and the toasting method, as well as the interaction between both factors, had a

338

significant impact (p < 0.05) on the oak wood volatiles concentration of wines. In general,

339

barrel toasting showed a greater effect on woody aroma, except for the furanic compounds,

340

which were further influenced by MLF-container. 14 ACS Paragon Plus Environment

Page 15 of 36


341

It is noteworthy to point out that, in the case of MT and Noisette toastings, MLF-barrel

342

wines presented, respectively, 9.9% and 22.0% greater total content of woody volatiles

343

(calculated by adding up the individual concentration of the most abundant ones:

344

whiskeylactones, vanillin and eugenol) (p < 0.05) than the corresponding wine of the MLF-

345

tank modality. This phenomenon suggests that when MLF is carried out in oak wood

346

contact, lactic acid bacteria (LAB) may also interact with wood components and increase

347

the content of certain woody volatiles in wine, as previously stated by Bloem et al.32 In

348

contrast to the two above-mentioned toasting methods, no significant differences (p > 0.05)

349

were found between MTAA-T and MTAA-B wines with regard to the total concentration

350

of woody aromatic compounds. As for ellagitannins, the watering process during barrel

351

toasting (MTAA) led anew to a different behavior than that observed for MT and Noisette

352

toastings.

353

A general woody aroma trend persisted throughout all of the wines considered. The major

354

oak wood volatile constituent of the 12-months aged wines was cis-whiskey lactone (179–

355

393 µg/L wine), whereas vanillin (146–228 µg/L wine) and trans-whiskey lactone (133–

356

238 µg/L wine) displayed moderate values. The rest of woody volatiles analyzed were

357

present as minor constituents, with concentrations lower than 32 µg/L wine. As observed in

358

Figure 2, significant differences (p < 0.05) could be noted among wines with regard to the

359

amount of each woody volatile, leading to a particular woody aroma profile for wines of

360

each MLF-modality and/or barrel toasting method.

361

Due to the high perception thresholds of furfural (20 mg/L) and 5-methylfurfural (45

362

mg/L), these volatiles have low direct impact on the organoleptic attributes of aged wines.

363

Nevertheless, it is well known that they may enhance the oaky aroma of wines by acting as

364

precursors of some other potent odorant compounds or as masking agents for the fruity

365

notes.33 In the present research, wines undergoing MLF in stainless steel tanks presented

366

higher content of both furanic compounds than the corresponding barrel-fermented wines

367

(Figure 2A, B). Izquierdo-Cañas et al.3 reported the same trend when comparing wines



Page 16 of 36

368

undergoing tank or barrel MLF, followed by just 45-days ageing in barrels. With regard to

369

the barrel toasting method, no significant differences (p > 0.05) were observed among

370

wines of the MLF-barrel modality. Meanwhile, Noisette-T wine presented significantly

371

lower amounts of furfural and 5-methylfurfural (p < 0.05) than MTAA-T and MT-T wines.

372

Concerning the oak whiskey lactones, different conclusions were drawn when comparing

373

both MLF-modalities, depending on the toasting method (Figure 2C, D). In the case of MT

374

toasting, no significant differences (p > 0.05) were detected between wines undergoing tank

375

or barrel MLF, according to the cis- and trans-whiskey lactone contents. The same results

376

were reported by Izquierdo-Cañas et al.3 for the cis-whiskey lactone concentration of

377

Cabernet Franc wines. Nevertheless, the MLF-container led to significant differences (p
0.05) between both MLF-modalities (Figure

419

2H). Thus, the barrel toasting method may play an important role on the availability of

420

vanillin precursors in oak wood.

34,35



Page 18 of 36

421

Fruity aroma. Esters are enzymatically synthesized by yeast during alcoholic fermentation

422

and may be modulated by LAB during MLF. These compounds are known to play an

423

important role in the fruity character of wines. For instance, both ethyl butyrate and ethyl

424

hexanoate are reported to provide the wine with strawberry-like aromas, and ethyl

425

octanoate, with hints of ripe fruit; whereas isoamyl acetate, isobutyl acetate, and hexyl

426

acetate are described by banana, apple and pineapple flavors, respectively. Furthermore,

427

among the ethyl esters branched acids, ethyl 2-methylbutanoate is characterized by apple,

428

strawberry and anise odors, and ethyl 3-methylbutanoate, by pineapple, floral and lemon

429

notes. 37

430

The concentrations of ethyl esters of straight-chain fatty acids (ethyl propionate, ethyl

431

butyrate, ethyl hexanoate, ethyl octanoate), higher alcohol acetates (isoamyl acetate,

432

isobutyl acetate, butyl acetate, hexyl acetate) and ethyl esters branched acids (ethyl 2-

433

methylbutanoate, ethyl 3-methylbutanoate) in the wines investigated, are given in Table 3.

434

Among the ethyl esters of straight-chain fatty acids, the most abundant ones were ethyl

435

hexanoate (237 to 248 µg/L wine) and ethyl octanoate (221 to 245 µg/L wine). Ethyl

436

propanoate and ethyl butyrate presented concentrations 1.7 and 1.9-fold lower, respectively.

437

Within the higher alcohol acetates, isoamyl acetate was found at concentrations above its

438

perception threshold (160 µg/L wine)37 as the major component in all cases, with values

439

ranging from 360 to 424 µg/L wine depending on MLF-container and barrel toasting. The

440

second main component in this family of esters was isobutyl acetate, contributing with

441

values of 48 – 56 µg/L wine, whereas butyl and hexyl acetates were minor components

442

with concentrations lower than 7 µg/L wine. With regard to the ethyl esters branched acids,

443

the ethyl 3-methylbutanoate showed 1.7 times higher values than the ethyl 2-

444

methylbutanoate (~ 26 µg/L wine). Ester concentrations obtained in the present research

445

showed the same order of magnitude as those previously described in the literature for red

446

wines aged in barrels.8,17,38,39


Page 19 of 36


447

In general, wines did not contain significantly different ester concentrations in relation to

448

the MLF-container (p > 0.05), except for isobutyl, isoamyl and hexyl acetates in the case of

449

Noisette wines (14.6%, 15.3% and 20.7% greater for the MLF-barrel modality,

450

respectively). This general trend has been previously described in the literature for wines

451

undergoing tank/barrel MLF and ageing in barrels for 8 months10 or 45 days,3 and for wines

452

with/without oak chips during MLF,34 but never for a longer contact time as in the present

453

research (12 months). The effect of MLF on the esters content of wine seems to be linked to

454

the esterase activity of the LAB strains.3 The fact that, in the present research as well as in

455

the above-mentioned bibliographic studies, the same LAB strain has been used in both

456

MLF-modalities, may justify that their ester concentration did not differ significantly from

457

one to another.

458

Concerning the effect of barrel toasting on the fruity volatile content, no important

459

differences were found between MTAA, MT and/or Noisette wines of each MLF-modality

460

(Table 3).

461

Sensory evaluation of wines. Spider web diagrams obtained from average values of

462

olfactory and gustative descriptors from sensory analysis of wines aged 12 months in

463

barrels, presenting MTAA, MT and Noisette toastings, are reported in Figure 3 for both

464

MLF-modalities. The toasting level did not lead to significant differences (p > 0.05) in

465

wines from both MLF-modalities for any of the attributes considered, apart from the fruity

466

descriptor in barrel-fermented wines. In this case, MT barrels led to wines significantly

467

perceived as the less fruity (p < 0.05), which may be associated to their highest toasting

468

intensity compared to MTAA and Noisette barrels. Since this difference among the three

469

toasting methods was not observed in wines from the tank modality (Figure 3A), these

470

results suggest that the sensory response of wine to the barrel toasting effect is influenced

471

by the MLF-container.



Page 20 of 36

472

From both olfactory and gustative points of view, results from the preference test revealed

473

that, in general, Noisette wines were the least preferred (39 – 47%), whereas MTAA wines

474

were usually the most appreciated (39 – 50%) for both MLF-modalities.

475

The triangle test was carried out on both barrel and tank-fermented wines for each toasting

476

method separately. For MT and MTAA toastings, the triangle test showed no significant

477

differences (p > 0.05) between wines of both MLF-modalities. Nevertheless, in the case of

478

MT toasting, the number of right answers was so close to the minimum number of correct

479

answers for differentiation at a 95% confidence level, that the bilateral paired comparison

480

test was also performed. Tasters detected statistically significant differences (p < 0.05) in

481

the olfactory descriptor vanilla, being the barrel-fermented wine the MLF-modality

482

characterized with the highest vanilla flavor. This sensory observation is in accordance with

483

the greater vanillin content of the barrel-fermented wine (p < 0.05) for the MT toasting.

484

Finally, for the Noisette toasting, the triangle test exhibited significant differences (p

Comparison between Malolactic Fermentation Container and Barrel

Recommend Documents