Année 2012

Thèse n°2019

THÈSE

pour le

DOCTORAT EN CO-TUTELLE DE

L’UNIVERSITÉ BORDEAUX 2 ET DE

L’UNIVERSITÉ DE GIESSEN

Ecole doctorale des Sciences de la vie et de la santé

Mention : Sciences, Technologie, Santé

Option : Œnologie

Présentée et soutenue publiquement à Villenave d’Ornon

Le 18 décembre 2013

Par Armin SCHÜTTLER

Né le 31 octobre 1977 à Nordhorn, Allemagne

Influencing factors on aromatic typicality of wines from Vitis vinifera L. cv. Riesling

- sensory, chemical and viticultural insights -

Membres du Jury

M. Juan CACHO Professeur de l’Université de Saragosse, Espagne ...................... Président et Rapporteur M. Peter WINTERHALTER Professeur de l’Université Braunschweig, Allemagne ................ Rapporteur M. Jordi BALLESTER Maître des conférences à l’Université de Bourgogne, France ..... Examinateur Mme Doris RAUHUT Professeure de l’Université des Sciences Appliquées RheinMain/ Forschungsanstalt Geisenheim, Allemagne ................................ Co-directrice de thèse M. Holger ZORN Professeur de l’Université de Gießen, Allemagne ...................... Co-directeur de thèse M. Philippe DARRIET Professeur de l’Université Bordeaux Segalen, France ................ Directeur de thèse

„Wege entstehen dadurch, dass man sie geht“ Franz Kafka (1893 – 1924)

… für Eva und Magdalena

Acknowledgements First of all I would like to thank my supervisor Prof. Dr. Philippe Darriet, Professor at the Faculté d’Œnologie and director of the Unité de rechérche Œnologie, EA 4577 – Université Bordeaux Segalen – INRA Œnologie for his supervision, helpful advice and support during the whole time of my thesis. I highly appreciated to get introduced into the universe of Œnologie and aroma chemistry during my time as his doctoral student. I want to thank for his patience with and his confidence in me and his sympathetic guidance not only in scientific details.

Further I would like to thank the co-supervisors Prof. Dr. Doris Rauhut, Professor at the Hochschule RheinMain and the Forschungsanstalt Geisenheim and Prof. Dr. Holger Zorn, Professor and Dean of the Fachbereich 08 – Biologie and Chemie at the Justus-Liebig-University Gießen:

I thank Prof. Dr. Doris Rauhut for the opportunity doing this thesis. I want to thank for all her support, her valuable advice and supervision during my thesis. I am grateful that I had the possibility to develop not only the project, but also my professional skills during this time.

I thank Prof. Dr. Holger Zorn for his helpful advice, his support of this project, and his efforts to realise the co-tutelle.

Also I wish to express my sincere gratitude and appreciation to the members of the thesis committee:

Prof. Dr. Juan Cacho, Professor at the University of Saragossa for giving me the honour to accept presiding the committee and to evaluate this thesis as a reviewer.

Prof. Dr. Peter Winterhalter, Professor at the University of Braunschweig for giving me the honour to review and evaluate this thesis.

Prof. Dr. Jordi Ballester, Associate Professor at the University of Burgundy for giving me the honour to evaluate this thesis as an examiner.

Prof. Dr. Aline Lonvaud-Funel, the former director of the UMR 1219 Œnologie Generale at the Faculté d’Œnologie and Prof. Dr. Manfred Grossmann, head of the Fachgebiet Mikrobiologie and Biochemie at the Forschungsanstalt Geisenheim for accepting me to work in their laboratories.

Prof. Dr. Hans Reiner Schultz, Director of the Forschungsanstalt Geisenheim, for his valuable implications into the project. Without his enthusiasm and his efforts to realise this co-tutelle this project would not have been started.

Prof. Dr. Otmar Löhnertz, Dean of the Hochshule RheinMain – Fachbereich Geisenheim, for his support and the financial covering by the Hochschule RheinMain.

The Conseil Régional d’Aquitaine and the Ministerium für Wissenschaft und Kunst of the Land of Hesse for financing the project through the Hesse-Aquitaine program.

Dr. Manfred Stoll, head of the Fachgebiet Weinbau at the Forschungsanstalt Geisenheim, for helpful advice, reviewing the manuscript and supporting the viticultural experimental trials.

Prof. Dr. Gilles de Revel, Dean of the Faculté d’Œnologie and head of the group Chimie Analytique, for allowing me to analyse the trans-ethyl cinnamate in his laboratory.

Prof. Dr. Helmut Dietrich, the head of the Fachgebiet Weinanalytik und Getränkeforschung at Forschungsanstalt Geisenheim, for supporting analysis of several parameters in his laboratories and a special thank to Anja Giehl, Anja Rheinberger and Petra Kürbel for their assistance with FTIR and HPLC analysis.

Further I would like to thank:

The whole Equipe of the laboratory Œnologie of the Faculté d’Œnologie situated in the Institut des Sciences de la Vigne et du Vin for being warmthly welcomed, especially Monique Pons for explaining me the French language and the advantages of strolling at Carrefour, Dr. Alexandre Pons and Dr. Cécile Thibon for all: from A like accommodation to Z in m/z. Merci beaucoup pour tout!

The whole staff of the Fachgebiet Mikrobiologie und Biochemie at the Forschungsanstalt Geisenheim, in particular Beata Beisert, Stefanie Fritsch and Heike Semmler for lots of analysis and their moral support. Also a great ‘Thank you!’ to Anja Abd Elrehim for the hours spent correcting incomprehensible languages. I also thank Helmut Kürbel for his support and will remember him as a great person.

The staff of the Fachgebiet Weinbau at the Forschungsanstalt Geisenheim especially Dr. Bernd Gruber for introducing me into the magic of water deficit, his support and helpful discussions, Magalie Lafontaine for her valuable input and practical advice, and Marco Hofmann for computing water potential values.

The staff of the Fachgebiet Kellerwirtschaft at the Forschungsanstalt Geisenheim especially Prof. Dr. Rainer Jung and Christoph Schüssler for helpful advice and assistance within sensory analysis.

The staff of the Fachgebiet Bodenkunde und Pflanzenernährung at the Forschungsanstalt Geisenheim, especially Ralph Lehnart for assistance with analysis and Marco Pfliehinger for sharing his viticultural knowledge.

The wineries for all the excellent wines they donated for this project and especially the “Hessischen Staatsweingüter Kloster Eberbach” for providing the experimental vineyard.

All the people, who assisted in sensory tastings and especially Dr. Maria-José Ruiz Moreno, Johanna Hoppe and Karl Rainer Schneider (‘Danke Stampfi’) for their excellent work within the project.

Finally there is a whole bunch of people to thank. Without having met them the last years would have been a lot more complicated.

I would like to thank Andrea Romano and Katharina Zott for giving me shelter during several stays in Bordeaux, Leila Falcao for organising my bank account in France, Gregory Schmauch for his jokes and his support during the writing, Daniel Gerhards for spending his time at the ‘Kellerbesprechung’, Maria Silva and Gregory Pasquier for spending their weekends with me in the lab and their help, Guillaume Antallick and Philippe Renault for starting the weeks with a short chat about soccer, Thomas Vidalies for taking me to the line up, Maria Nikolantonaki, Takis Stamodopoulos, Kentaro Shinoda, Davide Slaghenaufi, Georgia Lytra and Guilherme Martins, Bernd Lochbühler, Sophie Tempere... to name only a few. Thank you all for making the time such a good time.

My parents, Diana, Werner and Götz, Inge and Elmar

Anyway,...

...the most important gratitude is reserved to my family – my wife Eva and my daughter Magdalena. I want to thank you for your love, your patience and the energy you offered me every day.

Table of Contents

I

Table of Contents Acknowledgements ..................................................................................................... 4 

Table of Contents ........................................................................................................... I 

Abbreviations ............................................................................................................... X 

General Introduction ...................................................................................................... 1 

Chapter 1 ...................................................................................................................... 3 

1  Bibliography ............................................................................................................ 3 1.1  The grape variety Riesling ............................................................................ 4 1.1.1  History of the grape variety Riesling ............................................................. 4 1.1.2  Ampelographic and viticultural characteristics of Riesling vines ................... 5 1.1.3  Economical importance of Riesling in viticulture ........................................... 5 1.1.4  Organoleptic attributes of Riesling wines ...................................................... 6 1.2  The grape variety Sauvignon blanc .............................................................. 7 1.2.1  History of the grape variety Sauvignon blanc ............................................... 7 1.2.2  Ampelographic and viticultural characteristics of Sauvignon blanc vines ..... 7 1.2.3  Economical importance of Sauvignon blanc in viticulture ............................. 8 1.2.4  Organoleptic attributes of Sauvignon blanc wines ........................................ 8 1.3  Possible impact of ‘Climate Change’ on viticulture ....................................... 9 1.4  Effects of climate-dependent environmental parameters on vines and fruits10 1.4.1  Effect of temperature .................................................................................. 10 1.4.2  Effect of sun exposure – temperature and radiation ................................... 10 1.4.3  Effect of water deficit and nitrogen uptake .................................................. 11 1.5  Aroma compounds in Riesling wines’ aroma and their genesis ................. 11 1.5.1  Alcohols and esters .................................................................................... 11 1.5.2  Terpenes ..................................................................................................... 13 1.5.2.1  Monoterpenes ............................................................................................. 14 1.5.2.2  Norisoprenoids ............................................................................................ 16 1.5.3  Thiols .......................................................................................................... 20 1.6  ‘Typicité’ and ‘typicalité’ .............................................................................. 23 1.7  Measurement of typicality - sensory methodical aspects............................ 29 1.8  ‘A Not-A’ discrimination methodology ......................................................... 31 

Chapter 2 .................................................................................................................... 32 

2  Concept of Riesling wines’ typicality ..................................................................... 32 2.1  Introduction ................................................................................................. 33 2.2  Material and methods ................................................................................. 35 

Table of Contents

II

2.2.1  Chemicals and consumables ...................................................................... 35 2.2.2  Standard compounds and other chemical products ................................... 36 2.2.3  Wine samples ............................................................................................. 39 2.3  Chemical analysis ....................................................................................... 42 2.3.1  General wine composition .......................................................................... 42 2.3.2  Analysis of aroma compounds (‘Kaltron-method’): higher alcohols, esters

and terpenols .............................................................................................. 42 2.3.2.1  Sample preparation for ‘Kaltron’ extracts .................................................... 42 2.3.2.2  Instrumental settings for ‘Kaltron’ extracts analysis .................................... 43 2.3.3  Analysis of C13–norisoprenoids ................................................................... 43 2.3.3.1  Sample preparation for C13–norisoprenoid analysis .................................... 43 2.3.3.2  Chromatographic conditions for C13-norisoprenoid analysis ....................... 44 2.3.4  Analysis of low boiling sulfur compounds ................................................... 44 2.3.4.1  Sample preparation for low boiling sulfur compound analysis .................... 44 2.3.4.2  Chromatographic conditions for low boiling sulfur compound analysis ....... 44 2.3.5  Analysis of varietal thiols ............................................................................ 44 2.3.5.1  Sample preparation for thiol extract analysis .............................................. 44 2.3.5.2  Solutions for sample preparation ................................................................ 45 2.3.5.2.1  Solution for pH adjustment ......................................................................... 45 2.3.5.2.2  Solution for complex formation ................................................................... 45 2.3.5.2.3  Acetate buffer solution ................................................................................ 45 2.3.5.2.4  Solution for analyte elution ......................................................................... 45 2.3.5.3  Preparing anion exchange resin column .................................................... 45 2.3.5.4  Selective extraction of volatile thiols ........................................................... 45 2.3.5.5  Chromatographic conditions for thiol extract analysis ................................ 46 2.3.6  Analysis of free and bound terpenes and C13-norisoprenoids - potentially

volatile terpenens and norisoprenoids (PVTN) ........................................... 46 2.3.6.1  Sample preparation of must and berry extracts for potentially volatile

terpenes and C13-norisoprenoids (PVTN) analysis ..................................... 47 2.3.6.2  Chromatographic conditions for potentially volatile terpenes and C13-

norisoprenoids (PVTN) analysis ................................................................. 48 2.4  Sensory analysis ......................................................................................... 49 2.4.1  Sensory panel ............................................................................................. 49 2.4.2  Methodology ............................................................................................... 49 2.4.2.1  A Not-A test ................................................................................................ 49 2.4.2.2  Measurement of typicality and free choice descriptor frequency ................ 49 2.4.3  Statistical methodology ............................................................................... 50 

Table of Contents

III

2.4.3.1  Data transformation - descriptor families .................................................... 50 2.4.3.2  Analysis of panel performance ................................................................... 50 2.4.3.2.1  A Not-A test – Chi-square (χ2) and Fishers exact test ................................ 50 2.4.3.2.2  Typicality rating - data normalisation .......................................................... 50 2.4.3.2.3  Principal Component Analysis – panels’ homogeneity ............................... 50 2.4.3.2.4  Repeatability of typicality ratings ................................................................ 51 2.4.3.2.5  Descriptor family reproducibility .................................................................. 51 2.4.3.3  Panel comparison ....................................................................................... 51 2.4.3.3.1  Generalised Procrustes Analysis (GPA) ..................................................... 51 2.4.3.4  Product characterisation ............................................................................. 52 2.4.3.4.1  Typicality ratings - Analysis of Variance (ANOVA) ..................................... 52 2.4.3.4.2  Descriptor family citation frequencies – Chi-square (χ2) proportions .......... 52 2.4.3.5  Correlation of sensory and analytical data .................................................. 52 2.4.3.5.1  Univariate Analysis - Pearson product moment correlation ........................ 52 2.4.3.5.2  Multivariate Analysis ................................................................................... 52 2.4.3.5.2.1  Partial Least Square Regression (PLS) .................................................................. 52 2.4.3.5.2.2  Canonical Correspondence Analysis (CCA) (ter Braak and Verdonschot, 1995) .. 53 

2.5  Results and discussion ............................................................................... 53 2.5.1  Panel Performance – measurement of typicality ........................................ 53 2.5.1.1  Repeatability of measures .......................................................................... 53 2.5.1.2  Panels’ homogeneity - Principle Component Analysis (PCA) and Analysis

of Variance (ANOVA) .................................................................................. 54 2.5.1.2.1  Typicality ratings - Principle Component Analysis (PCA) ........................... 54 2.5.1.2.2  Typicality ratings - Analysis of Variance (ANOVA) for factor judge ............ 55 2.5.1.3  Panel performance on distinguishing the category Riesling ....................... 55 2.5.1.3.1  Typicality ratings - Anaysis of Variance (ANOVA) for factor wine ............... 55 2.5.1.3.2  A Not-A test derived Riesling wines recognition frequencies ..................... 56 2.5.1.3.3  Comparison of the two applied methods: Pearson product-moment

correlation - Typicality ratings vs. A Not-A test derived dry Riesling wines’ recognition frequencies ............................................................................... 57 

2.5.2  Discussion on panel performance and methodology comparison concerning the measurement of typicality ..................................................................... 58 

2.5.3  Panel performance on descriptive task ....................................................... 59 2.5.3.1  Descriptor families - choice of descriptors .................................................. 59 2.5.3.2  Descriptor families – average reproducibility index (Ri) .............................. 59 2.5.4  Discussion on panel performance for the descriptive task.......................... 60 2.5.5  Essential outcomes regarding panel performance and methodology ......... 62 

Table of Contents

IV

2.5.6  General sensory concept of typical Riesling wines ..................................... 63 2.5.6.1  German panels’ and French panels’ concept of typical Riesling wines ...... 63 2.5.6.2  Univariate Pearson product-moment correlation - typicality ratings vs.

descriptor family frequencies ...................................................................... 64 2.5.6.3  Multivariate methodology - Generalised Procrustes Analysis (GPA) .......... 65 2.5.7  Essential outcomes regarding universality of Riesling wines’ typicality ...... 69 2.5.8  Chemical base of Riesling wines’ typicality ................................................ 70 2.5.8.1  Univariate methodology .............................................................................. 70 2.5.8.1.1  Pearson product-moment correlation - typicality ratings vs. aroma

compound concentrations ........................................................................... 70 2.5.8.1.2  Pearson product-moment correlation - Riesling wines’ typicality correlated

aroma compounds vs. total frequency of descriptor families ...................... 72 2.5.8.1.3  Pearson product-moment correlation - typicality correlated total descriptor

family frequencies vs. aroma compound concentrations ............................ 73 2.5.8.2  Multivariate methodology ............................................................................ 74 2.5.8.2.1  Partial Least Square Regression (PLS) ...................................................... 74 2.5.8.2.2  Canonical Correspondence Analysis (CCA) on common variables from both

panels ......................................................................................................... 75 2.5.8.2.2.1  General properties of the CCA ................................................................................ 76 2.5.8.2.2.2  Factor composition .................................................................................................. 76 2.5.8.2.2.3  Explanation of sensory descriptor families by factors ............................................. 76 2.5.8.2.2.4  Interpretation analogous to ecological data – the sensorial niche .......................... 78 2.5.8.2.2.5  Interpretation of wine samples ................................................................................ 78 2.5.8.2.2.6  Interpretation for descriptor families ....................................................................... 79 2.5.8.2.2.7  Co-correlations of metric variables ......................................................................... 79 

2.5.9  Discussion on aroma compounds contribution to Riesling wines typicality 80 2.5.9.1  Univariate methodology .............................................................................. 80 2.5.9.2  Multivariate methodology ............................................................................ 80 2.5.9.3  Observations for selected aroma compounds’ concentrations regarding

typicality ...................................................................................................... 82 2.5.10  Essential outcomes regarding aroma compounds and Riesling wines’

typicality ...................................................................................................... 84 2.5.11  Consideration of an alternative approach: taking into account dissimilarities85 2.5.11.1  Differentiation into factors of similarity and factors of dissimilarity of a

category ...................................................................................................... 85 2.5.11.1.1  Principle ...................................................................................................... 85 2.5.11.1.2  Chi-Square (χ2) - descriptor family frequencies .......................................... 88 

Table of Contents

V

2.5.11.1.3  Pearson product-moment correlation – total frequency of descriptor families vs. aroma compounds’ concentrations ....................................................... 88 

2.5.12  Discussion on the Riesling wines’ typicality related descriptor families ...... 88 2.5.13  Essential outcomes regarding an alternative approach to typicality ........... 90 2.6  Conclusion .................................................................................................. 91 

Chapter 3 .................................................................................................................... 92 

3  Viticultural studies ................................................................................................. 92 3.1  Introduction ................................................................................................. 93 3.2  Material and methods ................................................................................. 94 3.2.1  Vineyard ...................................................................................................... 94 3.2.2  Viticultural treatments ................................................................................. 95 3.2.3  Viticultural measurements .......................................................................... 95 3.2.3.1  Grape vines’ water status ........................................................................... 95 3.2.3.1.1  Pre-dawn leaf water potential Ψ(PD) ............................................................. 95 3.2.3.1.2  Fraction of transpirable soil water (FTSW) derived pre-dawn leaf water

potential Ψ(PD) .............................................................................................. 96 3.2.3.1.3  Carbon isotope ratio – δ13C......................................................................... 96 3.2.4  Temperature - calculation of growing degree days (GDD) ......................... 96 3.2.5  Canopy micro-climate - ‘Point-quadrat’ measures ...................................... 96 3.2.6  Samples ...................................................................................................... 96 3.2.6.1  Berry samples for maturity measures ......................................................... 96 3.2.6.2  Berry samples for berry extracts ................................................................. 96 3.2.6.3  Grape must ................................................................................................. 97 3.2.7  Berry extract preparation ............................................................................ 97 3.2.8  Chemical analysis ....................................................................................... 97 3.2.8.1  Total soluble solids ..................................................................................... 97 3.2.8.2  Reducing sugars ......................................................................................... 97 3.2.8.3  pH ............................................................................................................... 97 3.2.8.4  Total acidity ................................................................................................. 97 3.2.8.5  Primary bound α-amino nitrogen as o-phthaldialdehyde - N-actetyl-L-

cysteine derivates (N-OPA) ........................................................................ 97 3.2.8.5.1  Buffer solution ............................................................................................. 98 3.2.8.5.2  N-actetyl-L-cysteine (NAC) solution ........................................................... 98 3.2.8.5.3  o-phthaldialdehyde (OPA) solution ............................................................. 98 3.2.8.5.4  Sample preparation and analysis ............................................................... 98 3.2.8.6  Analysis by means of Fourier transform infrared spectroscopy (FTIR) ....... 98 

Table of Contents

VI

3.2.8.7  Total nitrogen concentration after KJELDAHL - decomposition detected as ammonia ..................................................................................................... 99 

3.2.8.8  Free ammonia in grape juice and grape must ............................................ 99 3.2.8.9  Reduced glutathione ................................................................................... 99 3.2.8.10  Phenolic compounds ................................................................................ 100 3.2.8.11  Cysteinylated and glutathionylated precurser of 3-sulfanylhexan-1-ol ..... 101 3.2.9  Analysis of aroma compounds .................................................................. 102 3.2.10  Harvest and wine making ......................................................................... 102 3.2.11  Sensory analysis ....................................................................................... 103 3.3  Results and Discussion ............................................................................ 105 3.3.1  General climatic conditions ....................................................................... 105 3.3.1.1  Temperature and precipitation .................................................................. 105 3.3.1.2  Water status in three years ....................................................................... 106 3.3.1.2.1  Pre-dawn leaf water potential Ψ(PD) ........................................................... 107 3.3.1.2.2  Fractionated Transpirable Soil Water (FTSW) derived Ψ(PD) .................... 107 3.3.1.2.3  δ 13C – isotope Ratio ................................................................................. 109 3.3.1.3  Canopy micro-climate - effects of defoliation ............................................ 111 3.3.1.4  Infection by Botrytis cinerea at harvest ..................................................... 112 3.3.2  Development of grapes’ maturity .............................................................. 112 3.3.3  Monoterpenol and C13-norisoprenoid concentrations ............................... 115 3.3.3.1  Monoterpenols in grape must ................................................................... 115 3.3.3.2  Monoterpenol composition of wines after 12 months ............................... 115 3.3.3.3  C13-norisoprenoids in grape must ............................................................. 116 3.3.3.4  C13-norisoprenoid composition in wines after 12 months .......................... 117 3.3.3.5  C13-norisoprenoid composition in wines after 22 months .......................... 119 3.3.4  Discussion on monoterpenols in grape must ............................................ 120 3.3.5  Discussion monoterpenol composition of wines after 12 months ............. 122 3.3.6  Essential outcomes for monoterpenols and their precursor molecules .... 123 3.3.7  Discussion on C13-norisoprenoids ............................................................. 123 3.3.7.1  Discussion on C13-norisoprenoids in grape must ...................................... 124 3.3.7.2  Discussion C13-norisoprenoid composition in wines after 12 months and 22

months ...................................................................................................... 126 3.3.8  Essential outcomes for C13-norisoprenoids and their precursor molecules128 3.3.9  Thiol concentrations in wines and thiol precursor concentrations in berry

extracts and grape must ........................................................................... 128 3.3.9.1  Observations on thiol precursor molecules and other related metabolites in

berry extracts ............................................................................................ 128 

Table of Contents

VII

3.3.9.2  Thiol precursor molecules and related metabolites in grape musts .......... 130 3.3.9.3  Comparison of berry extracts and grape musts composition .................... 132 3.3.9.4  Free thiol concentrations in young wines .................................................. 133 3.3.10  Discussion on observations on thiol precursor molecules and other related

metabolites in berry extracts ..................................................................... 134 3.3.10.1  Discussion on comparison of berry extracts and grape musts composition136 3.3.10.2  Discussion on thiol precursor molecules and related metabolites in grape

musts ........................................................................................................ 137 3.3.10.3  Discussion on free thiol concentrations in young wines ........................... 138 3.3.11  Essential outcomes regarding thiols and their precursor molecules ......... 140 3.3.12  Sensory aspects ....................................................................................... 141 3.3.13  Discussion on sensory aspects ................................................................ 142 3.4  Conclusion on viticultural influence on Riesling wines’ aromatic aromatic

expression ................................................................................................ 143 

Chapter 4 .................................................................................................................. 144 

4  Aroma compound identification........................................................................... 144 4.1  Introduction ............................................................................................... 145 4.2  Material and methods ............................................................................... 146 4.2.1  Global aroma extracts ............................................................................... 146 4.2.2  Thiol selective aroma extracts .................................................................. 146 4.2.3  Fractioning global aroma extracts by means of high pressure liquid

chromatography (HPLC) according to Ferreira et al. (1999) and Pineau et al. (2009) ................................................................................................... 147 

4.2.4  Sensory analysis of aroma extract fractions ............................................. 147 4.2.5  Triangle tests of aroma extract fractions and reconstituted aroma extracts148 4.2.6  Quantitative Descriptive Analysis (QDA) of aroma extract fractions and

reconstituted aroma extracts .................................................................... 148 4.2.7  Liquid-liquid micro extraction of aroma extract fractions ........................... 148 4.2.8  Gas chromatography coupled to olfactometric detection (GC-O) ............. 149 4.2.9  Two-dimensional heart-cut gas chromatography coupled to olfactometric

detection and high-resolution mass-spectrometry (GC-GC-O-HRMS) ..... 150 4.2.10  Calculation of lineary retention indices (LRI) ............................................ 150 4.2.11  Quantification of trans-ethyl cinnamate by means of HS-SPME-GC-MS . 151 4.2.12  Quantification of trans-ethyl cinnamate by means of SBSE-GC-MS ........ 151 4.2.13  Determination of the odour - detection threshold of trans-ethyl cinnamate151 

Table of Contents

VIII

4.2.14  Effect of trans-ethyl cinnamate concentrations on Riesling wines’ typicality using the A Not-A method ......................................................................... 152 

4.3  Results and discussion ............................................................................. 153 4.3.1  Comparison of the odoriferous zones in aroma extracts of typical and not

typical Riesling wines by means of GC-O ................................................. 153 4.3.2  Comparison of the odoriferous zones in thiol selective aroma extracts

(TSAE) of typical Riesling and typical Sauvignon blanc style wines by means of GC-O ......................................................................................... 154 

4.3.3  Comparison of the aroma extract fractions (AEF) of typical Riesling wines and typical Sauvignon blanc style wines .................................................. 156 

4.3.4  Comparison of the aroma extract fractions (AEF) of typical Riesling wines and typical Sauvignon blanc style wines by means of triangle tests and Quantitative Descriptive Analysis (QDA) – Fractions 54 and 55 ............... 158 

4.3.5  Comparison of the aroma fraction reconstitution for typical Riesling wines and typical Sauvignon blanc style wines by means of triangle tests and Quantitative Descriptive Analysis – Fractions 54 and 55 .......................... 158 

4.3.6  Comparison of the odoriferous zones (OZ) in aroma extract fractions (AEF) of typical Riesling wines and typical Sauvignon blanc style wines using GC-O – Fraction 55 ......................................................................................... 159 

4.3.7  Identification of trans-ethyl cinnamate ...................................................... 160 4.3.7.1  One-dimensional GC-O/HR-TOF-MS ....................................................... 160 4.3.7.2  Two-dimensional heart-cut GC-GC-O/HR-TOF-MS ................................. 161 4.3.8  Impact of trans-ethyl cinnamate on Riesling wines’ aromatic expression . 162 4.3.8.1  Quantification of trans-ethyl cinnamate in Riesling wines and Sauvignon

blanc style wines ....................................................................................... 162 4.3.8.2  Determination of the odour detection threshold of trans-ethyl cinnamate . 163 4.3.8.3  Correlation of trans-ethyl cinnamate concentrations with Riesling wines’

typicality .................................................................................................... 163 4.3.8.4  Effect of trans-ethyl cinnamate additions on Riesling wines’ typicality ..... 165 4.3.9  Discussion on the role of trans-ethyl cinnamate in wine aroma ................ 166 4.3.10  Effect of viticultural factors on trans-ethyl cinnamate concentrations in

finished wines ........................................................................................... 167 4.4  Conclusion ................................................................................................ 169 

General conclusion and perspectives ........................................................................ 170 

Cited Literature .......................................................................................................... 175 

Appendix ........................................................................................................................ i 

Table of Contents

IX

Publications and Conference communications ................................................................  

Declaration of academical honesty ..................................................................................  

Abbreviations

X

Abbreviations Abbreviation used for chemical compounds 2M3SF 2-methyl-3-sulfanylfuran 2FM 2-furanmethanethiol 2M3SB 2-methyl-3-sulfanylbutanol 3M3SB 3-methyl-3-sulfanylbutanol 3SH 3-sulfanylhexan-1-ol 3SHP 3-sulfanylheptanol 3SP 3-sulfanylpropanol 3SPOH 3-sulfanylpentanol 4MSP 4-methyl-4-sulfanylpentan-2-one 4MSPOH 4-methyl-4-sulfanylpentan-2-ol A3SH 3-sulfanylhexyl acetate A3SP 3-sulfanylpropylacetate AE2S 2-sulfanylethylacetate BMT benzenemethanethiol Cn chain of carbon atoms with n number of carbon atoms CA caftaric acid CO2 carbon dioxide cys-S-3-SH 3-S-L-cysteinylhexan-1-ol cys-S-MSP 4-S-L-cysteinyl-4-methylpentan-2-one cys-S-prc sulfur bound L-cysteinylated precursor of thiols cysgly-S-3SH 3-S-L-cysteinyl-L-glycinylhexan-1-ol DMH DMS

2,6-dimethylhept-5-en-2-ol dimethylsulfide

E2SP ethyl-2-sulfanylpropionate E3SP ethyl-3-sulfanylpropionate glucys-S-3SH 3-S-L-glutamyl-L-cysteinylhexan-1-ol GRP 2-S-glutathionyl caftaric acid GSH glutathione (γ-glutamyl-cysteinyl-glycine) GSH-S-3SH 3-S-glutathionyl-hexan-1-ol GSH-S-Prc sulfur bound glutathionylated precursor of thiols GSSG glutathione-glutathiondisulfide (oxidised form of glutathione) H2O water H2O2 hydrogen peroxide H2S hydrogen sulfide H2SO4 sulfuric acid

Abbreviations

XI

homo GSH homoglutathione (γ-L-glutamyl-L-cysteinyl-β-alanine) IPP isopentenyldiphosphate K2SO5 potassium disulfite MBB monobromobimane NAC N-acetyl-L-cysteine NaCl sodium chloride

NH4+ ammonia N-OPA primary α-amino nitrogen as o-phthaldialdehyde-N-actetyl-L-cysteine OPA ortho-phthaldialdehyde p-HMB para-hydroxymercuribenzoate SO2 sulfur dioxide TDN 1,1,6-trimethyl-1,2-dihydronaphthalene TFA trifluoroacetic acid TPB (E)-1-(2,3,6-trimethylphenyl)-buta-1,3-diene TRIZMA 2-amino-2-(hydroxymethyl)-1,3-propanediol Units % percent °Brix degree Brix °C degree Celsius °Oe degree Oechsle GDD growing degree days h, min, s hour, minute, second ha hectare hl hectolitre kg, g, mg, µg, ng Kilogramme, gramme, milligramme, nanogramme KJoule kilojoule kPA kilopascal L, mL, µL litre, millilitre, microlitre m, mm, cm, µm, nm metre, millimetre, centimetre, nanometre M, mM, nM mol, millimol, nanomol m/z mass-to-charge ratio MPa Megapascal ppm parts per million ppt parts per trillion rpm revolutions (rotations) per minute %vol percent by volume

Abbreviations

XII

Abbreviation used in statistical analysis

α alpha, significance level χ2 Chi-square test p probability ANOVA analysis of variance CCA canonical correspondence analysis CF citation frequency ci constrained inertia contr contributions cos2 cosinus square CV coefficient of variation Fischer LSD method Fischer Least Square Differences method GPA Generalised Procrustes Analysis nc no correlation ns not significant PCA principle components analysis PLS Partial Least Square Regression r correlation factors RF recognition frequencies Ri average reproducibility index rim repeatability index for single subjects rin repeatability index for the whole panel Ris average repeatability index

σ standard deviation SEM standard error on means T relative typicality ratings Ti total inertia VIP variable importance in the projection

Abbreviations

XIII

Others AD Anno Domini (counting years after the birth of Jesus) AOC Appelation d'Origine Controlé δ 13C Isotope ratio of 13C/12C-atoms CIS cold injection system cv. cultivar HPLC high pressure liquid chromatography

HS-SPME-GC-MS head space - solid phase micro extraction - gas chromatography - mass spectrometry

I irrigated treatment IdB irrigated treatment defoliated at flowering and veraison IdV irrigated treatment defoliated at veraison INAO Institute national d'origine et de la qualité ISO International Organisation for Standardisation IUPAC International Union of Pure and Applied Chemistry JAR just about right analysis LoD limit of detection LoQ limit of quantification LRI lineary retention index max maximum min minimum MLF malolactic fermentation n number of samples NIST National Institute of Standards and Technology, USA nt no treatment OAV odour activity value OZ odoriferous zone P prototype p.a. pro analysi pH negative decadic logarithm of H+-ion concentration PDO protected designation of origin PRD partial root zone drying PVTN potentially volatile terpene and norisoprenoid concentration QDA quantitative descriptive analysis RI Riesling RIonly Data set containing only Riesling wines RT retention time

Abbreviations

XIV

SB Sauvignon blanc

SBSE-GC-MS stirr-bar-sorptive-extraction – gas chromatography – mass spectrometry

SH functional group: sulfhydryl; thiol SIM selective ion monitoring SPE solid-phase extraction TA total acidity TDU thermo desorption unit TIC total ion chromatogram TSAE thiol selective aroma extract TSS total soluble solids UV/VIS ultraviolet / visible UV-B ultraviolet-B UV-C ultraviolet-C vs. versus Ψ(PD) pre-dawn leaf water potential

General Introduction

1

General Introduction The grape varieties Vitis vinifera L. cv. Riesling and Vitis vinifera L. cv. Sauvignon blanc are considered to be two noble grape varieties. White wines made from these varieties are highly appreciated upon their distinctive aromatic bouquet. Since about 200 years Riesling grapes are grown traditionally in the vine growing region of the Rheingau Valley in Germany, producing wines of superior quality. It is considered that the famous wine style Spätlese, which is produced from over ripe and partially botrytised grapes originally comes from this region or rather from the domain of Schloss Johannisberg. Sometimes fully botrytised wines, labelled as Auslese and Beerenauslese, or the world-famous icewines are produced from this variety as well. Today about 79% of the grape vines planted in the Rheingau Valley are Riesling vines. To a similar extent the variety of Sauvignon blanc represents the traditional white vine variety of the Bordelais. It is used in combination with the grape varieties of Sémillon and Muscadelle to produce the recognised dry white wines from Bordeaux, but it is also elaborated as the famous sweet wines from botyrised grapes of Sauternes, Loupiac or Sainte-Croix du Mont. Riesling wines’ aromatic expression is very versatile and manifold fruity and floral odours, as well as vegetative and mineral aromatic nuances are associated with its varietal characteristic (Fischer and Swoboda, 2005). Despite several works were conducted on Riesling wines’ aroma, there is still a lack of knowledge towards the impact of the aroma compounds considered to be imparted in Riesling wines’ aromatic expression. Riesling is a variety, which is known to be highly capable to reflect the climatic conditions, the soil, and viticultural practices applied during a year in its wines’ aroma and taste (Fischer and Swoboda, 2005). Different projects showed the influence of terroir, in its geological and geographical factors, on aromatic expression and some volatile compounds of Riesling wines (Böhm et al., 2008; Bauer, 2008). But none of these works considered the concept of typicality in the notion of the French term typicité. Typicality [la typicité] includes the sensorial, technical and environmental dimensions and can be defined as a set of properties of distinction as well as properties of affiliation (Casabianca and Sylvander et al. (2005), cited by Cadot et al., 2010). Dubourdieu (2012) defines typicality [la typicité] as reflecting originality, being identifiable, being associated to a terroir, and reflecting a contemporary image. It also integrates all human choices at all stages of vine growing and winemaking. For Salette (1997) it is also linked very close to the notion of terroir – especially for single varietal vines [les vins de cépage].

General Introduction

2

As typicality is linked closely to terroir and terroir was shown to influence Riesling wines’ aromatic expression, it is likely, that Riesling wines’ aromatic typicality is impacted when single or multiple factors of terroir change. Since a decade or two, the global warming scenario has become more and more evident (ICCP, 2007) and the regional consequences of probable regional climate effects are studied since then (ICCP, 2012). For the wine growing region of the Rheingau, regional models predicted a higher probability for extreme drought events, which could lead to an increased number of water stress situations for grape vines grown in vineyards showing low water holding capacity (Hofmann and Schultz, 2008). Grape vines’ water status was identified to be an important factor of terroir (van Leeuwen et al., 2004; van Leeuwen and Seguin, 2006) and therefore Riesling wines’ aromatic typicality probably gets impacted directly by the predicted climate effects. Only little is studied about the impact of water deficit on biosynthesis of aroma compounds and their precursor molecules in grape berries and finished wines, especially regarding the grape variety Riesling. Therefore, in order to get deeper knowledge on those possible consequences, the opportunity of the collaboration between two research institutions in the Federal State of Hesse, namely the ‘Forschungsanstalt Geisenheim’ in cooperation with the ‘Justus Liebig Universität Gießen’, and in the Region of Aquitaine, namely the ‘Institute des Sciences de la Vigne et du Vin’ associated with the ‘Université Bordeaux Segalen’, was taken for this present study. After a first Chapter of bibliography, in the Chapter 2, the concept of Riesling wines’ typicality was studied using sensory and instrumental analytical methods by confrontation to Bordeaux white, Sauvignon blanc-style wines. In the Chapter 3, consequences of water deficit and micro climate manipulation on precursor molecules and aroma compounds considered to be imparted in Riesling wines’ typicality perception were studied. Finally, the Chapter 4 describes the identification of a new marker molecule considered to impact Riesling wines’ typicality using a comparative study with the better understood varietal aroma of wines from Vitis vinifera L. cv. Sauvignon blanc.

3

Chapter 1

1 Bibliography

Bibliography

4

1.1 The grape variety Riesling

1.1.1 History of the grape variety Riesling The origin of the variety Riesling is discussed to be situated in Germany. Suggestions are made, Riesling being a descendent of a variety described by the roman writer Plinius (23 - 79 AD). Others suggest King Ludwig the German (843 - 876 AD) to be responsible for the first plantings on river banks of the Rhine. Also discussed is its development in the Rhine region from the wild variety Vitis vinifera L. cv. Silvestris, being as speculative as former suggestions (Anonymous, 1982; Lott et al., 2010). Genetic analysis conducted by the group of F. Regner implicates the origin of Riesling as a interbreeding of a small berry providing and frost resistant autochthonic variety, which was cultivated by German tribes and Traminer, brought by the Romans to the Rhine region. Interbreeding with Heunisch is discussed as a further step of development to bring consistency, vitality and acid potential into Riesling (Lott et al., 2010). In the German Viticulture lexicon of 1930 Vitis vinifera L. cv. Riesling is clearly originated from Germany, probably in the Rheingau valley. First white grape varieties’ plantings are documented for 1392 AD at Geisenheim, replacing dominant red grape varieties’ plantings. In that period of time especially the Cistercian abbey Eberbach pushed white wines’ plantings to face the competition of French red wines on the Cologne wine market through typical regional white wines of high quality. Despite there is no documentation concerning the exact specification of the grape variety Riesling, one can consider these wines were produced from a close Riesling progenitor, due to first documentation findings for rießlingen in wine cellar accounting documents from March 13th, 1435 addressed to the Earl of Katzenelnbogen in ‘Rüsselsheim am Main’. Probably not coincidently, the members of the house of Katzenelnbogen were usually buried in the Eberbach abbey (Staab, 1986). Generally clerical structures led to spreading of the variety Riesling through ordinances, as for vineyards of Schloss Johannisberg since 1716 AD and at the Mosel region since 1787 AD. By the end of the 19th century, German Riesling wines were highly appreciated and valued with extremely high prices (Anonymous, 1982; Lott et al., 2010).

Bibliography

5

1.1.2 Ampelographic and viticultural characteristics of Riesling vines Vitis vinifera L. cv. Riesling grape vines’ shoots are characterised by long distanced internodes. They show medium secondary shoot growth. Grape cluster are small, dense and generally cylindrical (shouldered). Riesling produces small to medium sized, round, greenish-yellow (golden yellow in the case of sun exposure), black spotted berries. Phenological stages are characterised mostly as being late to very late during growing season (Galet, 1998, Lott et al., 2010). Cultivation of the variety Riesling is favoured by moderate or cool climate. Despite being favoured through cool climate, Riesling needs a sufficient quantity of warmth during the ripening season, especially in September (for Germany), to achieve full ripeness. In warmer wine growing regions, hot weather causes problems in production of high quality wines, often leading to non-fruity wines of low quality (Lott et al., 2010). Being frugal concerning its demands from soil, Riesling can be grown on a broad range of soils, from flat grounded, weathered rocky soils to deep grounded, loamy soils. In dry, flat grounded, rocky south exposed steep slope vineyards, robustness distinguishes Riesling vines (Kerridge and Antcliff, 2004; Lott et al., 2010). Riesling generally is not highly endangered by pests and diseases, like Peronospora, Oidium, or Esca. Concerning noble rot (Botrytis cinerea), grapes are not super sensible in comparison to other varieties, whereas pedicel rot as primary or secondary infection seems to be more frequent. In years of early maturity and humid and warm climatic conditions, Riesling berries are highly susceptible to bursting berries, which favours putrefaction (Galet, 1998, Lott et al., 2010). Clone selection led to good yield quantity potential. Selected Riesling clones today permit constant yields from 60 and 110 hl/ha, depending on viticultural practices. Grape maturity is generally estimated regarding sugar concentration in grape juice measured in °Brix (in Germany in °Oe). Riesling grapes show good maturity from 15.9 °Brix to 17.1 °Brix for simple wine qualities (Kabinett). Late Vintages (Spätlese, Vendange tardive) and sweet dessert wines are made from >19 °Brix and >21.3 °Brix respectively (Lott et al., 2010).

1.1.3 Economical importance of Riesling in viticulture World-wide an estimated area of 34 209 ha are planted with Riesling. Major Riesling producing countries are Germany, Australia and France (Table 1-1). In Germany an area of 22 601 ha was planted with Riesling in 2010, representing 22.1% of the total vineyard surface in Germany (DWI, 2011). As the major variety in six of the 13 wine growing areas in Germany, with a share of up to 79 % (Rheingau Valley) (3125 ha), Riesling is an important factor in German viticulture in general and especially for the regional Hessian viticulture (DWI, 2011b).

Bibliography

6

Table 1-1 Riesling vineyard surface world-wide (Data from DWI, 2011a)

Country total Riesling vineyard area in

[ha] [%]

Germany 22601 61.4 Australia 4256 12.1 France 3407 9.9 USA 1700 4.9

Austria 1643 4.8 New Zealand 636 2.5

Canada 440 1.3 Chile 293 0.8

South Africa 276 0.8

others 193 1.5

1.1.4 Organoleptic attributes of Riesling wines Riesling wines’ aromatic quality is highly dependent on vineyard characteristics and yield. Riesling wines show high capability to reflect the concept of terroir, including soil, climatic conditions, viticulture and oenological practises (Lott et al., 2010; Kerridge and Antcliff, 2004; Fischer and Swoboda, 2005). A Riesling wine is always considered to have pronounced acidity. It is designated to be the Riesling wines spine, bearing its fruity and fresh character and permitting long storage and bottle aging (Lott et al., 2010; Fischer and Swoboda, 2005). Taste and aromatic properties of young Riesling wines can include versatile fruitiness, including grapefruit, lemon, lime, apple, peach, passion fruit, pineapple, mango, melon and quince. In addition to this, they can also show distinctive aromatic properties like flowers as rose or vegetative as fresh cut grass, hey, straw and herbs, but also mineral aromatic nuances like the smell of wet rocks, smoke, empyromatic, crustacean shell or iodine aroma. It is evident that fermentation derived aromas, such as esters or higher alcohols, show responsibility for some fruity or even cooling sensory expressions (Kerridge and Antcliff, 2004; Lott et al., 2010; Fischer and Swoboda, 2005). During the aging of wines, the aromatic profile of Riesling wines changes towards less fruity to more mineral or ‘ripe’, aromatic properties like honey, orangepeel, walnut and the for aged Riesling wines well known ‘kerosene’ or ‘petrol’ aroma, which corresponds to the odour of linseed oil (Lott et al., 2010; Fischer and Swoboda, 2005).

Bibliography

7

1.2 The grape variety Sauvignon blanc

1.2.1 History of the grape variety Sauvignon blanc The cultivation of Sauvignon blanc was reported to happen simultaneously for the regions le Berry (today’s Departements Cher and Indre) and the Bordelais (today’s Departement Gironde) in the first century AD. This gets explained by kinship or trading relations between the Gallican people Bituriges Vivisques (Bourges) and Bituriges Cubi (Bordeaux) (Garrier, 2008). Ampelogists indicate strong morphological similarities to Chenin de Loire. The name is referred to originate from forest (silva). In the sixth century Sauvignon blanc was mentioned by Grégoire de Tours and in the twelveth century it is cited in La Bataille des vins of Henri d’Andéli (1225 AD) to produce the wine of Pouilly-sur-Loire and Sancerre. Victor Pulliat already indicated in 1879 the ‘finesse et [sa] saveur spéciale’ of Sauvignon blanc wines (Garrier, 2008).

1.2.2 Ampelographic and viticultural characteristics of Sauvignon blanc vines Vitis vinifera L. cv. Sauvignon blanc grape vines’ shoots are characterised by short to long distance internodes. They show strong vigour (Galet, 1998; Kerridge and Antcliff, 2004; Lott et al., 2010). Grape cluster are small, dense and generally conical (shouldered). Sauvignon blanc produces small to medium sized, ovoide, yellow (golden yellow at maturity), and thick-skinned berries. Phenological stages are characterised as being timed averagely during the growing season (Galet, 1998; Kerridge and Antcliff, 2004; Lott et al., 2010). Due to late lignification in the growing period, Sauvignon blanc vines can suffer from a lack of winter hardiness, if they are exposed to unfavourable climatic conditions during the growing season. Therefore, moderate or warmer climatic regions are favourable (Lott et al., 2010). Sauvignon blanc vines can be grown on a wide range of soil types. Generally poor soils suit more than heavy loamy grounds, although they achieve good performance on rich deep grounded soils. Vigour can be controlled by selection of a suitable rootstock for soil conditions. Same as for Riesling, Sauvignon blanc is generally endangered by pests and diseases like oidium, black rot, grey rot and eutypa (Galet, 1998). Concerning berry rot or noble rot (Botrytis cinerea), grapes are more sensible at the stage of inflorescence in comparison to other varieties, as it is to grape leaf hoppers (Lott et al., 2010). Like for Riesling, clone selection led to Sauvignon blanc clones permitting constant yields from 50 and 70 hl/ha (Galet, 1998).

Bibliography

8

1.2.3 Economical importance of Sauvignon blanc in viticulture World-wide an estimated area of 45 000 ha is planted with Sauvignon blanc. Major Sauvignon blanc producing countries are France, Chile, USA, South Africa, Italy and Australia (Table 1-2). Table 1-2 Sauvignon blanc vineyard area world-wide (Data from Galet, 1998; *Data from Lott et al., 2010)

Country total Sauvignon blanc vineyard area in [ha] [%]

France 19974 44.4 Chile 5572 12.4 USA 5181 11.5 South Africa 3900 8.7 Italy 2947 6.5 Australia 1725 3.8 Hungary 1200 2.7 New Zealand 857 1.9

Austria 582 1.3

Germany* 434 1.0 others 2609 5.8

France, with its close to 20 000 ha of Sauvignon blanc, grows 44.4% of the worlds’ Sauvignon blanc vines. From these stocks about 8082 ha are grown in the South West of France, in the Region of Aquitaine, and represent with the red Bordeaux varieties an enormous economical factor, especially for the biggest appellation in the Bordeaux region ‘Entre deux mers’ (Galet, 1998).

1.2.4 Organoleptic attributes of Sauvignon blanc wines Sauvignon blanc wines are generally assigned to wines of great aromatic diversity. Their varietal aroma is defined by spicy odours, with an aromatic expression of green capsicum, when grapes were not mature at harvest, and by fruity odours like blackcurrant and citrus fruit (Darriet, 1993; Lott et al., 2010). Especially the characteristic boxtree odour, reminding sun heated blackcurrant buds and grapefruit or passion fruit odours are appreciated in wines produced from fully ripened crop of grapes (Galet, 1998; Lott et al., 2010).

Bibliography

9

1.3 Possible impact of ‘Climate Change’ on viticulture Promotive climatic conditions are basic requirements for agricultural and viticultural production. Geographic suitability of grape varieties in regions is qualified according to climatic factors, e.g. mean temperature during the wine growing season (Huglin, 1986; Gladstone, 1992). This kind of categorisation always reflects certain socio-historically emerged ‘climatic optimum’ for grape varieties, which, in a way is also reflected in the French AOCs and the notion of terroir. The equilibrium of climatic conditions with other factors of ‘terroir’ and ‘typicité’ could get disequilibrated by the prognosted ‘climate change’, which is likely to impact general wine quality (Mira de Orduna, 2010). Meanwhile the global rise of mean temperature is undoubted (ICCP, 2007). Jones and co-workers predicted, on the basis of wine spectator ratings for wines from different regions, world-wide changes in quality ratings towards positive ratings for 25 of 30 grapevine growing regions (Jones et al., 2005). The applied model took only the effects of temperature into account, but not the effects of higher sugar accumulation, rising pest risks or drought events. Therefore such a model would be insufficient. Temperature changes also would impact the distribution of extreme weather situation during a year, like frequencies of extreme cold or extreme hot weather, which therefore is likely to interact with the distribution of precipitation during a year on global as well as regional level (IPCC, 2012). According to the results of a regionalised predicting model for the Hessian region, it is likely that the precipitation pattern will change during summertime, the growing period for grapevines, in a range up to 20 % less precipitation (Hofmann and Schultz, 2008). Less precipitation and increasing frequency of drought periods could impact vineyard sites showing low water holding capacity, which in Germany are commonly situated in areas of steep slope viticulture, like the upper Rheingau Valley and the Middle Rhine Valley. The probable consequences of these scenarios on grapevines’ water status were predicted on the base of well established grapevine water status and soil water status relations (Lebon et al., 2003; Schultz and Lebon, 2005; Gruber and Schultz, 2005; van Leeuwen et al, 2010) for two representative types of vineyards in the Rheingau Valley showing an increasing number of drought induced water deficit events until the year 2100 (Hofmann and Schultz, 2008). These results indicate that vineyard sites showing low water holding capacity, could have to deal every second to third year with 31 and more days of water deficit (Ψ(pd) = <-0.3 MPa) in the future. Therefore, the challenge for the future will be to handle these probable consequences in an economical way and to keep up a high quality level in wines.

Bibliography

10

1.4 Effects of climate-dependent environmental parameters on vines and fruits

1.4.1 Effect of temperature The effect of temperature has extensively been reviewed by Mira de Orduna (2010). Warmer temperatures extend the period of physiological activity of vines and increase metabolic rates and metabolites’ accumulation (Coombe, 1987; Mira de Orduna, 2010). Therefore, dates for budbreak, flowering and fruit maturity are affected through climatic conditions (Mira de Orduna, 2010). At temperatures of 30 °C and higher, it is reported that grapevines react with reducing berry size and weight and sugar accumulation may be stopped (Coombe, 1987; Kriedemann and Smart, 1971, cited by Mira de Orduna, 2010). Even if higher temperatures accelerate grape maturation, temperature effects on final sugar amounts are reported to be not very large (Coombe, 1987; Mira de Orduna, 2010). High sugar accumulation in berries, resulting in more than 24 – 25 °Brix, is considered to occur through evaporative loss at high temperatures (Keller, 2010; Mira de Orduna, 2010). Amino acid concentrations in berries were not considered to be influenced by temperature (Buttrose et al., 1971; Mira de Orduna, 2010), whereas total acidity shows lower levels with higher temperatures, especially through a decrease of malic acid (Buttrose et al. 1971; Mira de Orduna, 2010). Temperature as well, affects polyphenol synthesis, but in most studies it is considered to be difficult to separate light from temperature effect (Reynolds et al., 1995).

1.4.2 Effect of sun exposure – temperature and radiation The indirect effect of sun exposure – increasing temperature in exposed grapes – leads to lower total acidity in must (Reynolds et al., 1995). However, the quantity of radiation through sun exposure can be controlled through canopy management, whereas the quality of radiation is influenced by climatic factors such as atmospheric and stratospheric gas composition (Mira de Orduna, 2010). Therefore, the effects of different levels of e.g. UV-B radiation on plants and vines were studied. Plants responded with decreases in leaf expansions, fresh and dry weight, total biomass and photosynthetic capacity (Schultz et al., 1998; Schultz et al., 2000). A biochemical reaction to UV-B radiation is an upregulation of enzymes, which are involved in the metabolisation of UV absorbing compounds such as polyphenols as well as for antioxidants like Glutathione. Furthermore an influence on carotenoids’ degradation and amino acid formation was observed (Schultz, 1998; Schultz et al., 2000).

Bibliography

11

1.4.3 Effect of water deficit and nitrogen uptake Grape vines are often grown under sub-optimal conditions. Especially limited water supply is a key factor for vines’ physiological and biochemical responses, leading to enhanced winemaking potential of the grape (van Leeuwen et al., 2009). Water deficit was shown to reduce shoot growth, berry weight and yield, as well as it increases berry anthocyanin and tannin content (Hardie and Considine, 1976; Matthews and Anderson, 1988; Matthews and Anderson, 1989; van Leeuwen and Seguin, 1994; Koundouras et al., 2006; van Leeuwen et al., 2009). A mild to moderate deficit in the first phase of the rapid berry expansion led to limited berry growth and reduced canopy density (Dry and Loveys, 1998; Dry et al., 2001; Keller, 2005). Berries grown under mild water deficiency were smaller and often showed higher sugar concentration (Keller, 2005). A decrease in photosynthesis and sugar transport from the leaves under more severe water deficit could lead to a reduction in berry sugar accumulation (Quick et al., 1992; Rogiers et al., 2004; Keller, 2005). Pre-veraison as well as post-veraison water deficiency was shown to reduce berry size (Kennedy et al., 2002; Ojeda et al., 2002; Keller, 2005). Generally, water stress should be avoided before fruit set, whereas the time between fruit set and veraison is the period when shoot growth and berry size can be most effectively controlled by water deficit (Keller, 2005). Another factor coupled to the water status is the uptake and inner plant delocalisation of nitrogen compounds. Nitrogen uptake is realised through an active nitrate transporter, which is driven by boron cations, and therefore boron deficiency decreases nitrate uptake in roots (Camacho-Cristobal et al., 1999; Keller, 2005). When there is no sufficient soil moisture, nitrogen cannot be uptaken by the roots (Keller, 2005). Moreover, if vines suffered long-term water deficit, abscisic acid was increased and cytokinin contents were reduced. This accelerated leaf aging and finally led to senescence of older leaves (Jackson, 1997; Yang et al., 2002; Keller, 2005). ‘Senescence is accompanied by a decline in chlorophyll and remobilisation of carbon, proteins, and nutrients from these leaves and followed by leaf abscission’ (Keller, 2005). Some of the sugars, amino acids and mineral nutrients could be recycled to the fruit (Keller, 2005).

1.5 Aroma compounds in Riesling wines’ aroma and their genesis

1.5.1 Alcohols and esters Volatile higher alcohols and the corresponding acetic esters as well as ethyl acetates of fatty acids are considered to be the major volatile fraction to aroma of all kinds of fermented beverages (Nykänen, 1986). Among other volatile compounds, they were identified in

Bibliography

12

Riesling wines’ aroma extracts (Schreier and Drawert, 1974a; Schreier et al., 1976). Being omnipresent in all fermented wines, many of the esters and alcohols were not considered to make significant contributions to distinctive flavour and aroma perception and therefore also not in varietal aromatic expression (Ebeler and Thorngate, 2009). Considered as most significant for wine aroma are ethyl acetate (fruity, solvent like), 3-methylbutyl acetate (iso-amyl acetate, banana), 2-methylpropyl acetate (iso-butyl acetate; pear-drops), ethyl hexanoate (ethyl caproate, apple), and 2-phenylethyl acetate (honey, fruity, floral), which are produced in variable amounts by commercial yeast strains (Swiegers et al. 2005). Produced quantities are influenced by yeast selection, especially through mixed culture fermentations with wild yeast strains as Hanseniaspora guilliermondii or Pichia anomala and Saccharomyces cerevisiae, which led to higher acetate ester concentrations (Rojas et al, 2003). In spontaneous fermented Riesling wines, showing high share in Hanseniaspora uvarum during fermentation, ethyl acetate concentrations were elevated (Gerhards et al., 2012). Esters are considered to impact the typical fruity character of red wines, mainly through synergetic effects and perceptive interactions (Escudero et al., 2007; Pineau et al., 2009). Especially ethyl esters of fatty acids as butanoic acid, hexanoic acid, octanoic acid (caprylic acid), decanoic acid (capric acid), and to a lesser extent dodecanoic acid are implicated as perceptive interactors with other aromatic compounds as β-damascenone (Escudero et al, 2007; Pineau et al., 2007), β-ionone (Escudero et al., 2007) and dimethyl sulfide (Segurel et al., 2004; Escudero et al., 2007), even if these are present below their odour detection threshold (Escudero et al., 2007; Pineau et al., 2007). Ethyl butanoate, ethyl hexanoate and ethyl octanoate were shown to participate in the aromatic expression of ‘fresh fruit’ and ‘red fruit’, whereas ethyl propionate was shown to participate in the notion of ‘candied fruit’ and ‘black fruit’ (Pineau et al., 2009). Ethyl esters of fatty acids as well as acetic acid esters of higher alcohols, like 3-methylbutyl acetate (iso-amyl acetate; banana), hexyl acetate (pear) or 2-phenylethyl acetate (floral), are produced during fermentation and participate in the ‘fermentative’ aroma of young wines. Acetic acid esters are degraded through hydrolysis during bottle aging (Rapp et al., 1985), whereas ethyl esters of branched acids like ethyl 2-methylpropionate (ethyl iso-butyrate), ethyl 2-methylbutanoate, ethyl 3-methylbutanoate (ethyl iso-valeriate), and ethyl phenylacetate are formed through chemical esterfication (Ferreira et al., 2001; Diaz-Maroto et al., 2005). Same was observed for mono- and diethyl succinate, diethyl glutarate and diethyl malate (Rapp et al., 1985). Despite being strongly related to fermentation, some esters were identified and considered to impact varietal aroma of grape varieties. In wines of Pinot noir methyl anthranilate,

Bibliography

13

ethyl anthranilate (sweet-fruity, grape-like), ethyl cinnamate (cinnamon-like, sweet-balsamic, sweet-fruity, plum and cherry-like), and ethyl 2,3-dihydrocinnamate were identified exhibiting distinctive varietal aroma (Moio and Etiévant, 1995). Ethyl cinnamate is influenced by oenological practices such as ‘macération carbonique’, which favours the formation of this ester (Bitteur et al., 1992). As others, ethyl cinnamates’ aromatic properties are influenced by β-damascenone (Pineau et al., 2007) or show synergistic effects with those of other aroma compounds (Loscos et al., 2007). Another group of esters contributing to fruity odours, mostly in red wines, are hydroxylated ethylesters like ethyl 3-hydroxybutyrate (Pineau, 2007b and Pineau et al., 2009), ethyl 4-hydroxybutyrate (Ugliano and Moio, 2005), ethyl 2-hydroxyhexanoate, ethyl 6-hydroxyhexanoate (Pineau, 2007b), and ethyl 4-oxopentanoate (Pineau, 2007b). Recently 2-hydroxy-4-methylpentanoate was identified and demonstrated to be involved in blackberry aroma (Falcao et al., 2012). The organoleptic impact of its enantiomers and their distribution in wine subsequently was studied (Lytra et al., 2012). In general, fermentation derived esters have been classified as less important for aroma prediction of Riesling wines than grape derived aroma compounds (Smyth, 2005), whereas for Croatian Riesling, esters were found to be the most odoriferous aroma compounds (Komes et al., 2006).

1.5.2 Terpenes Until now more than 30 000 terpenes are known in literature, filling several books and reviews (Breitmeier, 2006). Since the finding of the isoprene rule by Otto Wallach in 1887,

natural compounds built up from isoprene subunits (C5)n are specified as terpenes (Ruzicka,

1963), but are nevertheless denoted as isoprenoids (Breitmeier, 2006). The terpenes get differentiated into subgroups of hemi- (C5), mono- (C10), sesqui- (C15), di- (C20), sester- (C25), tri- (C30), tetraterpenes (C40) and polyterpenes (C5)n with n>8. The isopropyl part of 2-methylbutane is defined as the head, the ethyl residue is defined as the tail. Whereas isoprene units in mono-, sesqui-, di- and sesterterpenes usually are exclusively linked from head-to-tail, the tri- and tetraterpenes do contain a centered tail-to-tail linkage (Breitmeier, 2006). The biosynthesis of activated isoprene (isopentenyldiphosphate (IPP)) can either occur via the acetate mevalonate pathway as well as via activated acetaldehyde and glyceraldehydes-3-phosphate forming 1-deoxypentulose-5-phosphate as precursor for IPP (Davis and Croteau, 2000; Breitmeier, 2006). Isomerisation of IPP produces γ,γ-dimethylallyl pyrophoshate as electrophilic reaction partner for the nucleophilic methylene group of IPP. Connection in this position leads to geranyl pyrophosphate as monoterpene. Subsequent head-to-tail reaction of geranyl diphosphate

Bibliography

14

with IPP yields farnesyl diphosphate as sesquiterpene, whereas reaction of farnesyl diphosphate with IPP results in geranylgeranyl pyrophosphate as diterpene. Tail-to-tail connection of either two units farnesyl pyrophosphate or either two units of geranylgeranyl pyrophosphate lead to squalene as triterpene or 16-trans-phytoene as tetraterpene (carotenoid). It is reported that terpenoid biosynthesis in plants is compartmentalised with the sesquiterpenes (and triterpenes) being produced in the cytosol via the acetate-mevalonate pathway, and the monoterpenes and diterpenes (and tetraterpenes) getting produced in the plastids via the still incompletely understood mevalonate independent pathway (Davis and Croteau, 2000). Terpenes of greatest interest regarding wines’ organoleptic properties are mainly monoterpenes and sesquiterpenes (extensively reviewed in Marais, 1983; Rapp and Mandery, 1986 and Mateo and Jiménez, 2000), recently elucidated triterpenes (Marchal, 2010) and, indirectly via degradation, tetraterpenes (carotenoids).

1.5.2.1 Monoterpenes Grape varieties are classified regarding their monoterpene concentrations into ‘muscat varieties’ (e.g. diverse Muscat varieties, Gewürztraminer) showing free monoterpene concentrations as high as 6 mg/L, ‘non-muscat aromatic varieties’ (e.g. Riesling, Scheurebe, Müller-Thurgau, Sylvaner, Kerner) showing a concentration range of 1 – 4 mg/L and ‘neutral varieties’ (e.g. Aryan, Sauvignon blanc, Semillon, Chardonnay, Viognier, Bacchus) with a concentration range between the first to classes (Rapp and Hastrich, 1976; Rapp and Güntert, 1985a; Rapp et al., 1985b; Günata et al., 1985a; Rapp et al., 1993; Mateo & Jiménez, 2000). Most abundant terpenes in grapes and wines are the monoterpene alcohols (monoterpenols) linalool, geraniol, nerol, α-terpineol (van Wijk, 1967a; van Wijk, 1967b; Schreier et al., 1974a; Ribéreau-Gayon et al., 1975; Schreier et al., 1976) and hotrienol (Schreier et al., 1974b). Some of them form isomeric oxides like linalool oxide (cis-, trans- in pyranoide and furanoide form), cis- and trans-rose-oxide and nerol oxide (Ribéreau-Gayon, 1975; Schreier and Drawert, 1974a; Simpson et al., 1983; Rapp et al., 1985a; Rapp et al., 1985b). In addition to these major terpenes also monoterpenediols were identified in wines (Rapp and Knipser, 1979; Rapp et al., 1984; Rapp and Mandery, 1986; Versini et al., 1991). For Vitis vinifera L. cv. Riesling the principal free monoterpenes in grapes and wine are linalool, α-terpineol, geraniol and nerol and the isomeric linalool oxides (Rapp et al., 1985a; Rapp et al., 1985b; Chisholm et al., 1994). However, other monoterpenes were considered to contribute to the distinctive Riesling aroma, such as nerol oxide (Simpson, 1979), hotrienol

Bibliography

15

(Rapp et al., 1985a; Rapp et al., 1985b), 2-carene and ocimene (Skinkis et al., 2008). Terpene esters are also considered to impact on wines’ aroma (Feddrizzi et al., 2011). Monoterpenes are shown to accumulate as glycosidic bound non-volatile conjugates in grape berries in form of β-D-glucopyranosides, 6-O-(α-L-rhamnopyranosyl)-β-D-glucopyranosides, 6-O-(α-L-arabinofuranosyl)-β-D-glucopyranosides and 6-O-(β-D-apiofuranosyl)-β-D-glucopyranosides (Williams et al., 1982; Schneider et al., 2001) as it was already suggested by Cordonnier and Bayonove earlier (Marais, 1983). This fact directly implicates the influence of winemaking practices as alcoholic fermentation by yeasts on free monoterpenes concentration in the wine, whereas an influence of yeast strains on the release potential by glycosidical enzymatic activity during the alcoholic fermentation was more and less significantly established (Grossmann et al., 1987; Delcroix et al., 1994; Zoecklein et al., 1997; Zoecklein et al., 1999). Free and bound monoterpenes are indifferently distributed in grape berries. For Vitis vinifera L. cv. Muscat d’Alexandria 94% of free geraniol and 96% of free nerol is located in the berry skin, whereas linalool is almost equally distributed between the juice (50%) and skin (26%) and cellular debris (24%), which was shown by Bayonove and co-workers in 1974 (Marais, 1983). For Vitis vinifera L. cv. Riesling, similar results were obtained, even though showing 60% of the linalool located in the grape skins. Linalool oxides showed equal distribution between mesocarp and berry skin (Versini et al., 1981). For total (free and bound) monoterpenols, these results were either confirmed by showing higher concentration in berry skins (Günata et al., 1985; Wilson et al., 1986) or not by showing higher values in the mesocarb (Park et al., 1991) of Vitis vinifera L. cv. Muscat of Alexandria (and Vitis lambrusca L. cv. Frontignac and Vitis vinifera L. cv. Traminer). Higher concentrations in berry skins are also reported in a Vitis vinifera L. cv. Chardonnay clone comparison study, which showed differing monoterpenol concentrations regarding the clones (Duchêne et al., 2008). Therefore pressing conditions, maceration and must heat treatment alter monoterpene concentrations in wine, leading to higher terpene concentrations or alternating concentration pattern of single terpenic compounds in grape must and wine (Williams et al., 1982; Versini et al., 1981; Marais, 1987). Monoterpenes are accumulated in the berry in the end of the ripening period and are not coupled to sugar accumulation, whereas in over-ripe berries the terpene concentration pattern alters towards oxidised forms (Hardy, 1970; Günata et al., 1985; Marais, 1987; Park et al., 1991; Kalua and Boss, 2010). Linalool addition to Botrytis cinerea cell cultures led to tertiary linalool diol isomers and was also found in grape must and wines (Rapp et al., 1986b). A study concerning ‘Amarone’ style wines’ aroma composition showed increased monoterpenediol and monoterpenol oxide concentrations with increasing percentage of botyrized berries (Fedrizzi et al., 2011).

Bibliography

16

Monoterpene’s composition of Riesling wines from different vine growing regions (South Africa, Germany, and Northern Italy) was shown to be a discriminating factor in a comparison study, implicating the influence of climatic conditions like temperature, quantity of sunlight and water availability (Marais et al., 1992a). Reynolds et al. (1995) showed that total terpene concentrations in grape juice were elevated under warmer conditions as well as for vines with applicated basal leaf removal. Same was shown for Chilean muscat grape cultivars, whereas free and bound terpenes’ concentrations were highest in juice from 50% sun exposed grape cluster in comparison to 20% and 100% sun exposure (Belancic et al., 1997). Similar results were obtained for Croatian Riesling, with the exception of contrary results for linalool oxides’ concentrations, which decreased with an increasing number of removed leaves (Kozina et al., 2008). The influence of water supply of grape vines during the growing period on the monoterpene concentration in must and wine is little studied. For table grapes of Vitis vinifera L. cv. Muscat of Alexandria in Japan, terpene concentrations increased with increasing post-veraison regulated water deficiency, showing highest concentrations for the highest water deficit level in that experiment (El-Ansary et al., 2005). In Vitis vinifera L. cv. Sauvignon blanc berry skins’ monoterpenes increased with a decrease of stem water potential value showing a minimum of Ψ(Stem) – 1.1 MPa (Giorgessi et al., 2007). Wines obtained from Vitis vinifera L. cv. Merlot grown under different water status (from fruit set to veraison) showed a higher monoterpenol level for deficit irrigated experimental plots

(Ou et al., 2010).

1.5.2.2 Norisoprenoids Cleavage products of tetraterpenic carotenoids, like the C20 carbon skeleton compound (retinoid) retinal, have important natural functions as they are acting as a photoreception chromophore in the process of vision in mammals and energy harvesting mechanisms in bacteria and algae. Other cleavage patterns, either enzymatically or non-enzymatically by oxidative conditions, lead to the group of norisoprenoids showing carbon skeletons from C9 up to C15 (Britton, 2008). ‘In most cases enzymatic processes co-exist with non-enzymatic chemical reactions that form the same compounds’ (Fleischmann and Zorn, 2008). There is a high structural diversity within this group of compounds and they show chemical and sensory properties which make them important contributors to perfume and aroma or other specific functions, e.g. signalling molecules in nature. Examples include sensory signalling of plants to attract insects and other animals to ensure pollination and seed dispersal, pheromone properties of insects, repellent effects against insects, like it is observed for grasshopper keton, and hormone-like growth regulation in plants through abscisic acid (Britton, 2008).

Bibliography

17

The most widespread – and most abundant in grapes and wine – volatile carotenoid metabolites show a C13-carbon skeleton – the so called C13-norisoprenoids – which are derived through oxidative cleavage between position C(9) and C(10). Among others, C13-norisoprenoids were isolated from a broad range of biological material as leaf products (tobacco, tea and mate), essential oils (rose), fruits (grapes, passion fruit, star fruit, quince, apple and nectarine), vegetables (tomato, melon), spices (saffron, red pepper) and other processed products, such as wine, rum, coffee, oak wood, honey and seaweeds (Winterhalter and Rouseff, 2001). C13-norisoprenoids can be divided chemically into megastigmanes and non-megastigmanes, referring to their megastigmane carbon skeleton, which can be oxidised in different carbon positions either resulting in the ionone (C9) or damascenone (C7) family (Ribéreau-Gayon et al., 2006). Traditionally, C13-norisoprenoids lacking an oxygen function in the side chain are referred to megastigmanes. Beside these families, many other structures are identified showing spiroether, acetal and benzofuran properties (Winterhalter and Rouseff, 2001). ‘Although present at only trace levels, sensory thresholds of most norisoprenoids are very low’ like 700 ng/L for β-ionone, 200 ng/L for β-damascenone (Ebeler and Thorngate, 2009), 2 µg/L for 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN) (Sacks et al., 2012) and 40 ng/L for 1,2,4-trimethylphenyl-3-butadiene (TPB) (Janusz et al., 2003). In the case of TDN, an odour threshold of 20 µg/L was considered to show an undesirable petroleum note in aged Riesling wines (Simpson and Miller, 1983). TDN concentration below 20 µg/L were suggested to have a positive contribution to varietal aroma of younger Riesling wines and to general Riesling wine quality up to a maximum concentration of 4 µg/L (Sponholz and Hühn, 1997). In grapes and wines a range of norisoprenoids have been identified, whereas some seem to be ubiquitous abundant in a wide range of varieties and others less abundant in just a few varieties. The potent odorous compounds β-damascenone and β-ionone are found in a wide range of white and red varieties, such as Vitis vinifera L. cv. Riesling, Müller-Thurgau and Scheurebe (Schreier and Drawert, 1974a), Vitis labruscana L. cv. Concord and Catawba (Acree et al., 1981), Vitis vinifera L. cv. Muscat (Etiévant et al., 1983), Chardonnay (Sefton et al., 1993), Melon de Bourgogne (Schneider et al., 2001), Fiano (Ugliano and Moio, 2008), Merlot and Cabernet Sauvignon (Sefton, 1998; Ou et al., 2010; Qian et al., 2009; Bindon et al., 2007; Lee et al., 2007) and ‘Amarone’ style wines (Fedrizzi et al., 2011), impacting these wines’ aroma. Despite TDN, vitispirane and Riesling acetal are found in several varieties e.g. Cabernet Sauvignon and Cabernet franc (Eggers et al., 2006; Lee et al., 2007; Bindon et al., 2007; Sacks et al., 2012), it is considered that they mainly impact Riesling wines’ aroma in young wines as well as in aged wines (Simpson, 1979; Simpson and Miller, 1983; Rapp et al., 1985; Marais et al.,1992; Silva Ferreira et al., 2004; Sacks et al., 2012), whereas the

Bibliography

18

direct contribution of the four vitispirane and the two Riesling acetal enantiomers has not been elucidated yet. Numerous research works have been carried out on the formation of these compounds. Carotenoids were suggested to be precursors and therefore research on this chemical class was done concerning the grape varieties, their localisation in berries and changes during ripening, showing β-carotene and lutein to be the principal carotenoids present in the berries’ skins. Decreasing concentrations during maturation have been effected to climatic conditions and shading (Razungles et al., 1987; Razungles et al., 1988; Marais et al., 1991; Kwasniewski et al., 2010). Lutein was shown to be a possible precursor for TDN via chemical thermal degradation in sulfuric acid media (Marais, 1992b). Vitispirane, TDN and β-damascenone precursors were shown to increase during maturation (Strauss et al., 1987). Among other C13-norisoprenoids, vitispirane isomers and TDN were generated by hydrolytical cleavage from megastigm-4-ene-3,6,9-triol and megastigm-4,7-diene-3,6,8-triol and their presence in grape must, possibly bound as glycosides, was suggested (Strauss et al., 1986), which was supported by droplet counter-current chromatography (DCCC) with subsequent GC-MS analysis after hydrolysis (Winterhalter et al., 1990a). Further, the presence of various precursors for isomeric vitispiranes, TDN and β-damascenone from the glycoside fraction were elucidated (Winterhalter et al., 1990b). With the identification of 2,6,10,10-tetramethyl-1-oxaspiro[4.5]dec-6-ene-2,8-diol, after enzymatic hydrolysis of a glycoside fraction from a Riesling grape must, a possible precursor for TDN was proposed (Winterhalter et al., 1991). In combination to previous work concerning quince fruit (Winterhalter and Schreier, 1988) and the elucidation of 1-(3-hydroxybutyl)-6,6-dimethyl-2-methylene-3-cyclohexen-1-ol as a possible precursor of vitispirane isolated from Riesling wine (Waldmann and Winterhalter, 1992), an integrated pathway was proposed for the genesis of vitispirane, Riesling acetal and TDN, with a hypothetic triol as key intermediate (Waldmann and Winterhalter, 1992) (Fig. 1-1). Other studies implicated at least one additional precursor (Versini et al., 1996) and the transformation of Riesling acetal into TDN (Daniel et al., 2009). Regarding the stereochemistry of enantiomeric Riesling acetals, nearly racemic distribution was shown for leafs of Vitis vinifera L. cv. Riesling and for Riesling wines, whereas enantiomers showed different sensory properties (Dollmann et al., 1995). Volatile vitispirane in grape juices showed differing proportions for its four enantiomers, but sensory properties were not reported in this work (Herion et al., 1993). However, C13-norisoprenoid levels in grapes and therefore in wines were clearly shown to be dependent on enzymatic activities during berry maturation (Baumes et al., 2002; Mathieu et al., 2005; Mathieu et al., 2009). Labelling experiments evidenced that carotenoids were

Bibliography

19

produced from berry set to veraison and that they degrade during the ripening period and are stored as glycosidical bound compounds (Baumes et al., 2002).

Fig. 1-1 Proposed pathway for vitispirane formation in Riesling wine and the potential role of hypothetic triol as key intermediate in vitispirane and TDN genesis (reproduced after Waldmann and Winterhalter et al., 1992) Furthermore it was shown, that Bortrytis cinerea bioconverted C13-norisoprenoids (α-damascenone) into several oxidation products (Schoch et al., 1991). In a study concerning different levels of botrytised grapes, an increase in 3-oxo-α-ionol was observed, regarding an increasing percentage of botrytis infection (Fedrizzi et al., 2011).

- H2O

- H2O

H+

H+

H+

H+

H+

- R

- R

Riesling acetal

TDN(1,1,6-trimethyl-1,2-

dihydronaphthalene)

Vitispirane

with - R: glycosidical bound sugar moiety

Bibliography

20

A comparative evaluation of C13-norisoprenoid concentrations in Riesling wines from different vine growing regions with different climatic conditions showed higher concentrations of TDN in wines from warmer regions (Marais et al., 1992; Marais et al., 1992a). Nitrogen fertilisation during the growing period was shown to decrease TDN concentrations, whereas it increased β-damascenone concentrations in Riesling wines (Linsenmeier and Löhnertz, 2007). Furthermore, grape vine clones as well as yeast strains during fermentation influenced the aging behaviour of Riesling wines regarding their TDN concentrations (Sponholz and Hühn, 1997). Another factor considered is the oxygen uptake during bottling and bottle storage, due to different closure systems (Silva, 2011). Few studies were carried out on micro climates’ influence on norisoprenoid concentrations. In Riesling and Chenin blanc an increase of vitispirane and TDN has been observed regarding grape maturity and natural sun exposure (without canopy manipulation) (Marais et al., 1992c). Another study showed increased levels of the C13-norisoprenoids TDN and Riesling acetal in white Riesling grapes, but recognised that leaf removal decreased TDN in the berries, albeit being extensively sunlight exposed (Gerdes et al., 2001). These results were confirmed in a study on Cabernet Sauvignon, which was defoliated at berry set, showing increasing TDN and vitispirane concentrations with increasing sunlight exposure but a decrease with defoliation (Lee et al., 2007). The influence of leaf removal timing on TDN and vitispirane concentrations in Riesling grapes and wines was elucidated, showing highest concentrations for the treatment defoliated after berry set in comparison to the treatment being defoliated at veraison (Kwasniewski et al., 2010). The observed effects of grape vines’ water status on C13-norisoprenoid concentration in berries of Cabernet Sauvignon, using partial root zone drying (PRD) to establish differences in water status, were observed higher TDN concentrations for PRD modalities, which were correlated with carotenoid concentrations for lutein and β-carotene (Bindon et al., 2007). Merlot wines from grapes grown under deficit irrigation showed higher vitispirane concentrations, whereas TDN was not detectable (Qian et al., 2009). Another study of the same group showed higher β-damascenone concentrations in Merlot wines from deficit irrigated grapes (Ou et al., 2010).

1.5.3 Thiols ‘Among chemicals, sulfur-containing molecules and especially thiols are probably the most famous key flavour compounds in many foods and beverages. They are often characterised by sensorial thresholds as low as the ppt [ng/L or ng/kg] level.’ (Vermeulen et al., 2005) Chemically, thiols are sulfur analogues of alcohols, characterised by their functional group, a -SH moiety, which is responsible for their name being derived from the Greek word for sulfur – theion (Vollhardt, 1990). Due to precipitating as salts with low solubility in presence of

Bibliography

21

mercury, thiols are traditionally named mercaptans, originating from the Latin expression corpus mercurium captans (Beyer and Walter, 1988). According to today’s IUPAC nomenclature rules the functional group should be named with the prefix sulfanyl- , even though the ancient prefix mercapto- is still widely used in actual literature. The role of thiols for the aroma of food and beverages is extensively reviewed in Vermeulen et al. (2005). A wide range of thiol compounds has been identified in these products being grouped into sulfanyl - acids, - alcohols, - ketones, - esters, - ethers, aliphatic and aromatic thiols, terpenic thiols, acyclic bisulfurous thiols or heterocyclic thiols. Thiols are reported to be abundant in all kinds of products from vegetables (e.g. onion, garlic, and asparagus), beverages (e.g. brewed coffee, beer, wine, orange juice) and fruits (grapefruit, blackcurrant) to other food, such as meat, brooth and dairy products like cheese (Vermeulen et al., 2005). In the aroma of wine, the group of thiols is subdivided into varietal thiols and non-varietal thiols (Ribéreau-Gayon et al., 2006; Roland et al., 2011). Varietal thiols have been suggested to participate in guava aroma in form of 4-methyl-4-sulfanylpentan-2-one (4MSP) of Chenin blanc and Colombard wines, without proof of its identity in wines from these cultivars (Du Plessis and Augustyn, 1981). Later, this compound was identified by Bordeaux researchers in wines from Sauvignon blanc contributing to characteristic box tree and broom like aroma of this variety (Darriet, 1993; Darriet et al., 1995). Additional thiol compounds as 3-sulfanylhexan-1-ol acetate (A3SH) (Tominaga et al., 1996), 4-methyl-4-sulfanylpentan-2-ol (4MSPOH) and 3-sulfanylhexan-1-ol (3SH) and 3-methyl-3-sulfanylbutan-1-ol (3M3SB) (Tominaga et al., 1998a; Tominaga et al., 1998b; Tominaga et al., 2000a) contributing to the characteristic aromatic expressions of Sauvignon blanc wines, were identified subsequently by this research group. Additionally 2-methyl-3-sulfanylbutan-1-ol (2M3SB), 3-sulfanylpentan-1-ol (3SPOH) and 3-sulfanylheptan-1-ol (3SHP) were characterised as impact compounds in sweet wines from botrytised Sauvignon blanc (Sarrazin et al., 2007). In addition to these compounds, 3-sulfanylpropan-1-ol (3SP), 2-sulfanylethyl acetate (AE2S) (Lavigne et al., 1998), ethyl 2-sulfanylpropionate (E2SP), ethyl 3-sulfanylpropionate (E3SP) (Tominaga et al., 2003b), 2-furanmethanethiol (2FM) (Tominaga et al., 2000b; Tominaga and Dubourdieu, 2006), 2-methyl-3-sulfanylfuranthiol (2M3SF) (Tominaga et al., 2006) and benzenemethanethiol (BMT) (Tominaga et al., 2003) were identified in wines, but not considered to be varietal impact compounds (Roland et al., 2011), Lately Nikolantonaki and Darriet (2011) identified ethyl 2-sulfanylacetate (E2SA) as compound to cause off-flavour in Sauvignon blanc wines. After being identified in Sauvignon blanc wines, varietal thiols have been identified in wines of other white varieties as Scheurebe (Guth, 1997), Gewürztraminer, Pinot Gris, Riesling, Muscat, Pinot blanc, Sylvaner (Tominaga et al., 2000a), Sémillon, Petit Manseng (Tominaga

Bibliography

22

et al., 2000a), Bacchus, Melon de Bourgogne (Schneider et al., 2003) Petite Arvine (Fretz et al., 2005), Albarino, Malvasia, Parellada, Verdejo (Campo et al., 2005), Maccabeu (Escudero et al., 2004), and Koshu (Kobayashi et al., 2010). They were also found in red variety wines from Merlot, Cabernet Sauvignon (Bouchilloux et al., 1998; Murat et al., 2001), Grenache (Ferreira et al., 2001) as well as in rosé wines from the Provence region (Masson and Schneider, 2009). In Alsatian Riesling wines the most abundant varietal thiol was shown to be 3SH in a concentration range of ~400 – ~1000 ng/L, exceeding its perception threshold by the factor of 7 to 16, (Tominaga et al., 2000a). In that work, two late harvested wines showed 4MSP concentrations, exceeding its perception threshold of 0.8 ng/L (determined in ethanolic model solution) (Tominaga et al., 2000a). Three biogenesis pathways are considered to participate in the formation of 3SH during fermentation, whereas each of them also can act as one step in a multistep pathway. These pathways also seem to be partly valid for 4MSP, while A3SH formation is automatically linked to these pathways, due to being an esterfication product of 3SH and acetate during fermentation (Roland et al., 2011). The first precursors identified, and therefore the first pathway, were cysteinylated precursors bound as thioether in γ-position to an alcohol-, keto- or aldehyde group, which can release free thiols 3SH, 4MSP and 4MSPOH by means of β-lyase activity (Tominaga et al., 1998c). As another type of precursors glutathionylated precursors, also bound as thioether γ-position to electrophilic alcohol (3SH) or keto group (4MSP), were identified (Peyrot des Gachons et al., 2002a; Fedrizzi et al., 2009). Furthermore, the degradation of 3-S-glutathionylhexan-1-ol (GSH-S-3SH) by γ-glutamyl transpeptidase showed increasing concentrations of 3-S-L-cysteinylhexan-1-ol (cys-S-3SH) in grape must samples, concluding that γ-glutamylcysteinylated and cysteinylglycinylated precursors for 3SH (glucys-S-3SH and cysgly-S-3SH) are present in grape must (Peyrot des Gachons et al., 2002b).This was confirmed by identification of 3-S-cysteinylglycinylhexan-1-ol in grape juice (Capone et al., 2011a; Capone et al., 2012). In addition to this, it was shown that cys-S-3SH was formed from GSH-S-3SH in grape vine cell cultures (Thibon et al., 2011). These authors hypothesised that GSH-S-3SH acts as a pro-precursor for cys-S-3SH, and therefore for 3SH by means of a plant tissue detoxification process for cell toxic substances (Peyrot des Gachons et al., 2002b; Thibon et al., 2011), whereas others indicated GSH-S-3SH as a direct precursor of 3SH, which corresponds to the second biogenesis pathway (Kobayashi et al., 2010; Roland et al., 2010; Grant-Preece et al., 2010). The role of stereochemistry of GSH-S-3SH has not been elucidated yet, despite some studies did monitor proportions of (S)- and (R)-configured GSH-S-3SH and did show the (S)-configured stereoisomer as the most abundant in grape must, but a nearly racemic distribution of free 3SH in wines (Capone

Bibliography

23

et al., 2010a). The racemic distribution of 3SH was observed also in Sauvignon blanc wines and Semillon wines by Tominaga et al. (2006a). A third alternative pathway implicating direct thiol group transfer on C6 unsaturated compounds like trans-2-hexen-1-al during double binding hydration. Until now, no sulfhydryl donor could be pointed out, but H2S is supposed to be involved or any other compound showing a free thiol function (Schneider et al., 2006; Roland et al., 2011). Maturity was found to favour the precursor (cys-S-3SH) concentration in berries of Sauvignon blanc grapes from Bordeaux (Peyrot des Gachons et al., 2000). Similar results were observed for Sauvignon blanc grapes from Sancerre and Tours in a study also including GSH-S-3SH (Roland et al., 2010). Overripening of grape berries led to 10 fold higher cys-S-3SH concentrations in Sauvignon blanc grapes from Bordeaux (Thibon et al., 2009). The biogenesis of the thiol precursor is likely to be linked to nitrogen assimilation in the vineyard. Glutathione and cys-S-3SH increased with assimilable nitrogen content in the grape berries (Choné et al., 2006; Lacroux et al., 2008). Botrytis cinerea was shown to impact cys-S-3SH concentrations in grape berries, by increasing them in the beginning of the grape berries infection (Thibon et al., 2009). Similar results were obtained for Petite Arvine affected by rot (Luisier et al., 2008) and for Koshu, Chardonnay and Merlot vines under laboratory conditions (Kobayashi et al., 2011). In addition to this, it was shown that wines obtained from botrytised grapes showed a modified ratio on absolute configuration concentration of (R)- and (S)- forms of 3SH (Thibon et al., 2009). Only little data is available concerning the effect of microclimate on berries’ thiol precursor concentrations. The work of Japanese researchers on seedlings of Koshu, Chardonnay and Merlot grape vines under laboratory conditions, indicated an increasing effect of UV-C irradiation on precursor (GSH-S-3SH and cys-S-3SH) concentrations in grape vine berries as well as in grape vine leaves (Kobayashi et al., 2011). Moderate water deficiency is considered to favour thiol aromatic potential (Peyrot des Gachons et al., 2005). Water deficiency also was observed to increase GSH-S-3SH and cys-S-3SH in Koshu, Chardonnay and Merlot grape vines under laboratory conditions (Kobayashi et al., 2011).

1.6 ‘Typicité’ and ‘typicalité’ When it comes to the terminology of typicité, typicalité in the French language it becomes quite confusing, due to both get translated into English by the word typicality. One has to

Bibliography

24

consider that each expression is originating from different languages as well as from the different fields oenology and cognitive psychology. Development of the term typicité In the context of oenology the French term typicité was introduced in the years 1979/1980 by French researchers as a neologism for better characterisation of experimental vines derived from differrent terroirs by sensory analysis at the INRA research center in Angers (Salette, 1997). The proposition to introduce the concept into AOC regulations was effectuated in 2001 when typicité was introduced as a criterion for organoleptic testing for AOC wines (Ministère de l’Agriculture et de la Pêche, 2001). At the same time, the term Typizität was introduced into the German language in analogism to the French expression of typicité in the wine sector and found its way into the regulations based on the German wine law according to sensory examinations of wines and as a criterion for labelling the grape variety (BGBL, 1994). On the semantic level, typicité is defined according to the French dictionary ‘Le Petit Robert’ as the ‘totality of characteristics [of a wine], coming from the variety, the land, and the vinification practices’ for the domain of oenology (Robert, 1993). In the German equivalent ‘Duden’ the definition is made as – ‘characteristic originality; exemplary peculiarity’ (Duden, 2010), whereas it does not define the term according to its use in the domain of wine language. Meanwhile, the concept of typicité is no more limited to the wine sector; it got its universal importance in the whole field of food products, e. g. for the label ‘Protected Designation of Origin’ (Maitre et al., 2010). Whereas the French term is more related to terroir, including the grape variety, the German terms’ application is mostly limited to the grape varietal expression of a wine. But what kind of concept can be found behind this definition? The word typicité was formed as a neologism from the word ‘typique’ like the words ‘élasticité’ from ‘élastique’ or ‘spécifité’ from ‘spécifique’ (Salette, 1997). The same can be concluded analogeously for the German word Typizität from ‘typisch’ (‘Elastizität’ from ‘elastisch’, ‘Spezifität’ from ‘spezifisch’), which corresponds more or less to the French expression typicité. Being originated in the French language, the concept of typicité will be discussed focussing on its French interpretation. Definition of ‘typicité’

i) Nature of similarity The concept of typicité is the determination of whom or what corresponds to a (predefined) type (Salette, 1997). Salette defines the type as the model, the reference, the ideal form, which unifies its properties, its design, its essential characters of a class of being or things of the same nature, the same category on the highest level (‘le modèle, la référence, la forme idéale, qui réunit au plus haut degré les propriétés, les traits, les caractères essentiels d’une

Bibliography

25

classe d’être ou des choses de même nature, de même catégorie’). Further the type is some

synthetic form around which the individual variations can oscillate in a considered group. A very representative sample of a type is called idéo-type (‘[...] est en quelque sorte une forme synthétique autour de laquelle peuvent osciller des variations individuelles à l’intérieur du groupe considéré. L’échantillon particulièrement représentatif d’un type est appelé idéo-type.’) (Salette, 1997). This can be interpreted in accordance to a well known cognitive psychological concept for categorisation, based on the so called family resemblance, introduced by Rosch and Mervis (Rosch and Mervis, 1975; Mervis and Rosch, 1981). Salette’s idéo-type is in correspondence to Rosch’s prototype, and his oscillating individual variations can be seen as Rosch’s inner categorical gradients of representativeness (Mervis and Rosch, 1981). Salette’s definition of type is already the product of an abstraction process in the sense of Mervis and Rosch (Mervis and Rosch, 1981), which gets underlined by his remark that typicité can only be defined and recognised through memorised experiences and through precise references routed to the notion of the sensorial image (‘La typicité qui ne peut être définie et reconnue qu’à partir d’expériences mémorisées et de références précises conduit à la notion d’image sensorielle [...]’ (Salette, 1997)). ‘Typicality’ [...] is usually defined as the degree to which an item is perceived to represent a category’ and is positive correlated to family resemblance (Loken and Ward, 1990). According to these definitions, it was established that typicité, in the domain of oenology, and typicality, as it is used in cognitive psychology, they both refer to the idea of being representative or being typical for a category or a concept and can be used synonymously dependent on their domain of application (Ballester, 2004).

i) Nature of dissimilarity – role of ‘quality’ and ‘terroir’ In addition to this corresponding sense, Salette points out that typicité is a characteristic value of a product, which allows to differenciate, to identify and to recognise (‘La typicité, [...] est une donnée caractéristique du produit: elle permet de le différencier, de l’identifier et de le reconnaître’). This suggests an existing discriminating nature of the notion of typicité in addition to its assigning nature, which itself indicates the concept of typicité to be based on dissimilarities as well as it is based on similarities at the same time. Therefore, a wine presenting a certain typicité would have to to show differenciating properties (dissimilarities) and assigning properties (similarities) regarding other wines of the same category. In the concept of typicité, the role of sensory properties is highlighted upon their importance for type determination and their use in the process of attachment of a certain typicité to a certain sample (‘L’analyse sensorielle est un outil privilégié pour definir les types, pour déterminer le caractère typique de tel ou tel échantillon et, par conséquent, pour attributer la possession d’une typicité à un échantillon donné’) (Salette, 1997). According to the above

Bibliography

26

described process, the attachment of typicité would in fact be a measure of quality, due to evaluating a sensory profile of a sample. This interpretation is also used by other authors, which state, that the notion of typicité includes implicitly a judgement of value and that it is a synonym of quality for agricultural food products (Sauvageot, 1994). According to Sylvander and Lassaut (cited in Trognon, 2005), there are two types of quality for agricultural products:

• the quality, which is absolute and embodies the conformity of a product to an expected level, regulatory or not regulatory (social consensus; good manufacturing practices). The products can be rated and have to satisfy a minimum standard regarding these fixed parameters of this quality for all consumers.

• a quality, which is relative and expressed by quality-specification and product profile and embodies the presence of specifities or not and the conformity to a particular quality profile. A product profile is somehow its record of characteristic techniques used in elaboration and is characteristic for a certain product. The typicité of a product translates through its particular sensory profile, which distinguishes the product of another.

The latter, the relative, quality approach either embodies or is emodied in the concept of typicité and its distinguishing nature mentioned by Salette (1997), in which the product profile and quality-specifications function as objectifying tools (Trognon, 2005). Another dimension of typicité, reflected in that relative quality approach, is its relation to the concept of terroir, which can also be defined as a part of the product profile. As Trognon (2005) figures out, the association typicité-terroir is a postulate for the totality of authors (‘Pour l’ensemble des auteurs l’association typicité-terroir est un postulat’). In fact the terroir concept is essential, due to being the origin of the concept of typicité (Salette, 1997). A terroir itself is defined as a characterised agricultural system, able to produce particular products, to which it grants a certain originality and an own character (‘[...] un agrosystème caractérise, dote d’une capacité à donner des produits particuliers auxquels il confère une originalité et un caractère propre’) (Salette, 1998, cited in Trognon, 2005). The French ‘Institut National de l’Origine et de la Qualité’ (INAO) defines terroir as a defined geographical area, in which a human community gains over the time of its history a collective know how of production, based on the system of interactions between the physical and biological environment and a totality of the human factors. Thus applied, these socio-technological itineries create originality, confirm a typicité and end up in a reputation for being originated from that geographical area. (‘[...] un espace géographique délimité, dans lequel une communauté humaine, construit au cours de son histoire un savoir collectif de production fondé sur un système d’interactions entre un milieu physique et biologique, et un ensemble de facteurs humains. Les itinéraires socio-techniques ainsi mis en jeu, révèlent

Bibliography

27

une originalité, confirment une typicité, et aboutissent à une reputation, pour bien originaire de cet espace géographique’) (INAO, 2011). This leads to the situation, that terroir cannot exist without typicité and typicité cannot exist without terroir. Furthermore, within a category of products it implicates a distinguishing factor, a nature of dissimilarities, for both, terroir and typicité. In conclusion, the concept of typicité reflects the notion of wines regarding five parameters: the originality, the association with a difficult-to-copy terroir, the recognition for professionals (and amateurs) with reflection to a contemporary product image and the taste, and the ability of its conservation. It also integrates the human decisions for all levels of vine growing and vinification and elaboration (Dubourdieu, 2012). This conclusion merges the factors mentioned above and is consistent to other interpretations (Salette, 1997; Perrin, 2008; Cadot et al., 2010; Cadot et al., 2012a). Complementary concepts of ‘typicality’ and ‘typicité’ In international literature typicité is often translated as typicality, leading to a confusing interpretation being limited to its dimension of representativeness derived from the category structure theory, which in the French language is translated as typicalité. As Parr et al. (2007) mentioned ‘[...] the term typicality [is employed] as the best, albeit imperfect, translation of the French notion of typicité [.]’ and therefore it could be considered that describing its differentiating nature by the concept of a category’s gradient of representativeness maybe is not sufficient. This problem also got pointed out by several researchers from the marketing and consumer science and choices of pragmatical solutions for this problem were proposed for these sectors. For Giraud (cited in Perrin, 2008) typicité is a notion which persists between the world of production, where it is originated, and the world of consumption, where it operates. (‘[…] la typicité est une notion qui hésite entre le monde de la production, dont elle est originaire, et celui de la consommation, où elle opère’). This leads to a double positioning of typicité:

• from the point of view of the production, it is a feature of a distinctive character of a product, based on a link to a geographic place (typicité) (in the French context the regulations of the ‘appelations’ will be implicated; generally the concept of terroir would impact)

• from the point of view of the consumer, it can be considered as the degree of representation for an object of a category (typicalité)

In the marketing sector Passebois-Ducros et al. (2012) also distinguishes typicality (typicalité) and typicité from the point of view of either professionals or consumers. They conclude that the point of view is determined by either extrinsic or intrinsic factors. She states, that in any case the gustative qualities of a wine are determining factors of the perceived typicality [in the sense of cognitive psychology], but that experts (wine

Bibliography

28

professionals) do not talk about typicality but about typicité, due to taking into account, that a typical wine from a sensory point of view possesses several intrinsic characteristics, which are related to its terroir. Further she points out, that for consumers typicité is simply determined with the place of production (as an extrinsic information) and not to perceived sensory properties and, that the consumer’s prototype gets more precise (in intrinsic factors) the more a consumer gets into the complexity of the concepts of terroir and typicité. She forms the hypothesis, that a consumer estimates wine’s intrinsic typicité through extrinsic estimation of typicality (e. g. packaging, production place) for a purchase decision, and that perceived typicité then gets influenced by these (e.g. packaging).

Fig. 1-2 Position of products S1, S2, S3 regarding a categorical prototype P with vectors indicating inter product typicité (Δ) relations (black), product-prototype typicité (Δ) relations (orange), both based on dissimilarities, and product-prototype typicality (Θ) relations (green), based on similarities, in a categorical space defined on similarities (black ellipse) (modified after Trognon, 2005) Trognon (2005) presented a theoretical model of product positioning on a market. He distinguishes a product’s typicality (typicalité) and typicité in a certain product category based on:

• their similarities, a certain typicality is attached to a product, which, if present at a high degree, acts as a factor of security

• their differences, a certain typicité is attached to a product, which, if present at a higher degree, acts as a factor of originality

PP

S1

S2

S3

Δ(S1;S3)

Δ(S1;S2)

Δ(S2;S3)

Θ(S1;P)

Δ(S3;P)

Δ(S1;P)

P prototypeS product

product‐product typicité relation Δ

product‐prototype typicité relation Δ

product‐prototype typicality relation Θ

Legend :

Bibliography

29

These two measurements towards a prototype are interacting and lead to different positioning of similar categorical products in a categorical space (Fig. 1-2). Inspired by this approach, which was presented by Trognon for a categorical space of a product from the perspective of marketing relevant properties, an adaption for the categorical space of a product from the perspective of sensory properties could open a new dimension of information, which might be helpful to characterise the importance of dissimilarities within a sensory space of a product category. It could be considered that a measure focused on dissimilarities could deliver valuable information on a sensory space of a product’s category.

1.7 Measurement of typicality - sensory methodical aspects When it comes to the question how to determine typicality, it is important to figure out what aspect of this concept exactly should be determined and by whom (Salette, 1997). As outlined in the former paragraph, sensory properties are key aspects to the concept of typicality in the sense of typicité in the oenological sector. So how is sensory characterisation of categorical members of a group done so far? A quite simple approach of typicality measurement is the determination of distinctiveness of various members of various categories, as it can be carried out for wines obtained from different varieties by wine professionals (Winton et al., 1975). Such an experiment is, in fact, a measurement of typicality in the sense of similarities, which means a comparative measure towards different prototypes or exemplar models, on predefined categories - the varietal aromatic properties of wines of a certain variety. A direct measure of typicality was presented for Chardonnay wines from the Bourgogne area (Moio et al., 1993). Experienced tasters were asked to rate the presented wines as not typical, medium typical and very typical. In addition, wine tasters were asked to name a maximum of five free chosen descriptors linking citation frequency of descriptors to the degree of typicality. Intentional typicality rating, respecting cognitive psychological fundamental considerations, has been carried out by Ballester et al. (2005) as well as by Parr et al. (2007). Taking into account inner categorical gradients of representativeness with their work on Chardonnay wines from the Bourgogne region as well as for Sciaccarelo wines from Corsica, typicality was characterised (Candelon et al., 2004; Ballester et al., 2005; Ballester et al., 2008; Lorrain et al., 2006; Jaffré et al., 2011). Typicality was measured on an unscaled bar, which was anchored by ‘very poor example’ and ‘very good example’ on the left and the right end respectively. Panellists had to respond to the question if a presented sample is considered to be a good or a bad example for a wine of the category of Chardonnay de Bourgogne. All panellists were highly experienced wine tasters and no further information was asked regarding a sensory profile (Ballester, 2004; Ballester et al., 2005). Similar procedures were carried out to identify

Bibliography

30

representative botrytised wines from the Sauternes appelation (Sarrazin, 2007b), representative red wines from the Bordelais, which was carried out in combination with descriptive analysis (Pineau et al., 2010), Sauvignon blanc wines from New Zealand, France and Austria (Parr et al., 2007; Parr et al., 2010; Green et al., 2011; Pineau et al, 2011), Chenin varietal wines from the central Loire valley in France (Perrin and Pagès, 2009), Cabernet franc wines from the Loire region (Cadot et al., 2010; Cadot et al., 2012a; Cadot et al., 2012b), and also for Italian olive oil (Caporale et al., 2006). In most cases, the total strategy included sorting tasks, in order to verify global distinctiveness of the sample sets (Ballester et al., 2005; Parr et al. 2007; Parr et al., 2010; Green et al., 2011; Perrin and Pagès, 2009) and/or descriptive tasks for the characterisation of the most typical samples specifities (Moio et al., 1993; Parr et al., 2007; Parr et al., 2010; Green et al., 2011; Pineau et al., 2010; Pineau et al., 2011; Perrin and Pagès, 2009; Cadot et al., 2010; Cadot et al., 2012a; Cadot et al., 2012b; Caporale et al., 2006). Quantitative Descriptive Analysis (QDA) based approaches Since the introduction of quantitative descriptive analysis (QDA) into the field of food sensory (Stone et al., 1974), it was considered as the most appropriate tool for sensorial characterisation of wines’ sensory spaces also. This technique was often used for varietal characterisation or terroir and/or establishment of appellations, mostly in the new world, in combination with univariate and multivariate statistical analysis. For example it was applied to characterise these aspects for Cabernet Sauvignon (Heymann and Noble, 1987), Pinot noir (Guinard and Cliff, 1987), Chardonnay (McCloskey et al., 1996), Riesling (Fischer et al., 1999) and Sauvignon blanc (Lund et al, 2009). Anyway, with this approach, perceptual typicality is measured by forming cluster of similar aromatic profiles, and not by comparison to a categorical prototype. This means, that a degree of typicality is analysed into the data sets’ of the samples mostly including analytical data for aroma compounds also, but without any information wether the wines are good or bad examples of the concerning category. As an extension of the QDA, the ‘Just about Right Analysis’ (JAR) emerged, taking a prototype into account and measuring the deviation between perceived and conceptual prototypical ideal intensity for each chosen descriptor, positive as well as negative. This method was applied for Dornfelder and Beaujolais, towards a exemplary model, which was defined as the expectations of an ordinary consumer, supposing being represented by an average quality product found in the supermarket (Fischer et al., 2005). In another study Cabernet franc wines from the Loire Valley were evaluated towards wine professionals prototype from the Loire Valley region (Cadot et al., 2010; Cadot et al., 2012a; Cadot et al., 2012b). These researchers both applied a combination of direct typicality measurement according to Ballester et al. (2005) on an unscaled bar, JAR - analysis and QDA. Citation frequency as an alternative to QDA

Bibliography

31

Several characterising tasks have been carried out using citation frequency to describe sensory spaces of wines e.g. for Chardonnay wines from Bourgogne (Moio et al., 1993), Pinot noir and Aglianico wines (Piombino et al., 2004), sweet Fiano wines (Genovese et al., 2007), various Spanish white wine varieties (Campo et al., 2008) and Grechetto wines (Esti et al., 2010). They used either trained panels or oenology students and wine professionals. A citation frequency-based descriptive method, performed by a well trained panel, turned out to be a convenient alternative to conventional descriptive analysis for odour profiling of Burgundy Pinot noir wines (Campo et al., 2010). When comparing conventional profiling results of a trained panel with results, obtained using free profiling by wine experts, it was demonstrated, that free profiling with a free descriptor generation step can identify wine characteristics that were not considered using a conventional method (Perrin et al., 2007). The use of panellists having a specialised expertise on the product was shown to favour reliability of results. It was shown, that expert wine consumers matched wine descriptions better to a set of wines than non-expert wine consumers (Lawless, 1984) and that a panel of wine professionals performed better in replicating descriptive terms and distinguishing wine samples than a trained panel of non-wine professionals (Zamora ang Guirao, 2004). These authors also figured out, that experts perceived the samples with reference to their remembered specific features, or, as pointed out by Brochet and Dubourdieu (2001), towards a prototype, which is constructed on the experts’ personal experiences and preferences.

1.8 ‘A Not-A’ discrimination methodology Testing differences of general sensory properties in food, is mostly carried out by direct confrontation of confusable products or samples in various variations from triangle testing or duo - trio testing (Lawless and Heymann, 2010). As an alternative to these methods, the same different methodologies are considered (Lee et al., 2007a; Lee et al., 2007b; Hautus et al., 2009; Min-A Kim et al., 2012; Santosa et al., 2011). Representing one variation in this category of tests, the A Not-A testing methodology was introduced to food science by Pfaffmann et al. (1954). The test does not have a standard form and many different versions of the A Not-A test are reported (Lawless and Heymann, 2010; Lee et al., 2007b). Generally, the term A Not-A test applies to any procedure, where the judge undertakes a familiarisation phase to build a memory of the product A at the beginning of a testing session, and then a set of unknown products is given, of which some are A, and others are not. The task of the judge is to report whether each product presented was A or Not-A (Lee et al., 2007b).

32

Chapter 2

2 Concept of Riesling wines’ typicality

Concept of Riesling wines’ typicality

33

2.1 Introduction Tasted among other wines, Riesling wines are highly recognised by wine tasters regarding their aromatic expression. Multiple aroma active compounds were shown to be present in Riesling wines and were already correlated with their aromatic expression (Schreier and Drawert, 1974a; Rapp and Hastrich, 1976; Simpson et al., 1983; Rapp and Güntert, 1985a: Chisholm et al., 1994; Tominaga et al., 2000a; Smyth, 2005; Skinkis et al., 2008). In that way Riesling wines’ terpenes pattern was used for differentiation from other varieties (Rapp and Güntert, 1985a),C13-norisoprenoids were identified to be imparted in the bottle aging aroma of Riesling wines also known as ‘kerosene aroma’ (Simpson et al., 1979) and powerful thiols were identified in Riesling wines also (Tominaga et al., 2001a). Until now, no systematic study was carried out to relate ‘what is typical for a Riesling wine’ from a sensory point of view with analytical data considering such a broad range of aroma compounds like acetic acid esters, ethyl esters, terpenols, terpenol oxides, C13-norisoprenoids, low boiling sulfur compounds and varietal thiols. As a tool, the concept of typicality was chosen in order to select good examples of Riesling wines and to distinguish them from bad examples, using the methodology proposed by Ballester et al., (2005). Additionally, it is considered, that the A Not-A test procedure, even if not in accordance with the family resemblance theory of Rosch, could be used for sensory typicality determination, as a kind of sorting task. The question, a judge would have to respond to when a wine gets presented is: ‘Is the wine part of a certain conceptual category (in this case typical Riesling wine), or not?’ This test represents a categorical In/Out test, when the familiarisation procedure in the beginning of the session is replaced by sufficient experience of the panelists with knowledge of the product, like wine professionals have. The global objective of this survey is to study the universality of the Riesling wines’ aromatic concept in a German-French binational study, using two different panels of wine experts performing both two different methods and then link these results to instrumental analytical data regarding aroma active compounds. In the sensory part of this study:

i) both panels and their performances on the sensory task were monitored and were evaluated.

ii) an alternative testing method was applied and compared to the standard method for typicality testing.

iii) both panels’ results were confronted and the validity of an universal aromatic typical Riesling wine concept was validated by comparison of the data.

Concept of Riesling wines’ typicality

34

Results of the instrumental analytical part were linked to the results of the sensory part of this study by different univariate and multivariate statistical methods, in order to point out aroma compounds involved in perception of Riesling wines’ typicality.

Concept of Riesling wines’ typicality

35

2.2 Material and methods

2.2.1 Chemicals and consumables All solvents used for the purpose of extraction, were from HPLC graded purity (99.9%) or were freshly destilled. Acetonitrile for analysis of cysteinylated and glutathionylated precursors were suitable for high resolution mass spectrometry. They were purchased at different suppliers upon laboratories geographic situation (Carl Roth Laborbedarf, Karlsruhe, Germany; VWR, Darmstadt, Germany; VWR, Fonteney sur bois, France). Gases used for gas chromatography (Hydrogen; Helium, Nitrogen) have been purchased from Air Liquide (France) or Linde Gas (Germany) upon Laboratory’s situation. Water indicated as H2’O in the methods was from ultra-pure quality (Millipore®/ MilliQ® Quality).

Concept of Riesling wines’ typicality

36

2.2.2 Standard compounds and other chemical products Standard compounds and other chemical products are presented in the tables 2-1 to 2-6 sorted by analytical methods. Table 2-1 Applied standards in analysis of compounds by means of the ‘Kaltron’-method, purity and suppliers; (IS) indicates compounds used as internal standard Compound Purity Supplier

ethyl acetate >99% Sigma-Aldrich, Seelze, Germany 2-methylpropan-1-ol (iso-butanol) >99% Sigma-Aldrich, Seelze, Germany ethyl propanoate 99% Sigma-Aldrich, Seelze, Germany 3-methyl-1-butanol (iso-amyl alcohol) 99% Sigma-Aldrich, Seelze, Germany 2-methyl-2-butanol (tert-amyl alcohol) 99% Sigma-Aldrich, Seelze, Germany ethyl 2-methylpropanoate (ethyl iso-butyrate) 99% Sigma-Aldrich, Seelze, Germany ethyl butanoate (ethyl butyrate) >98% Sigma-Aldrich, Seelze, Germany ethyl 2-hydroxypropanoate (ethyl lactate) >99% Sigma-Aldrich, Seelze, Germany hexan-1-ol >99% Sigma-Aldrich, Seelze, Germany 3-methylbutyl acetate (iso-amylacetate) >98% Sigma-Aldrich, Seelze, Germany 2-methylbutyl acetate (amylacetate) 99% Sigma-Aldrich, Seelze, Germany caproic acid (hexanoic acid) >98% Sigma-Aldrich, Seelze, Germany ethyl hexanoate (ethyl caproate) >99% Sigma-Aldrich, Seelze, Germany hexyl acetate >98% Sigma-Aldrich, Seelze, Germany trans/cis-linalool oxide >99% Sigma-Aldrich, Seelze, Germany linalool 95% Sigma-Aldrich, Seelze, Germany 2-phenylethanol >99% Sigma-Aldrich, Seelze, Germany octanoic acid (caprylic acid) >99% Sigma-Aldrich, Seelze, Germany diethyl succinate (diethyl butanedioate) >99% Sigma-Aldrich, Seelze, Germany ethyl octanoate (ethyl capryloate) >99% Sigma-Aldrich, Seelze, Germany α-terpineol >99% Sigma-Aldrich, Seelze, Germany ethyl 2-phenylacetate (ethyl benzeneacetate) >98% Sigma-Aldrich, Seelze, Germany 2-phenylethylacetate (benzylethyl acetate) >99% Sigma-Aldrich, Seelze, Germany decanoic acid (capric acid) >98% Sigma-Aldrich, Seelze, Germany ethyl decanoate (ethyl capriate) >99% Sigma-Aldrich, Seelze, Germany DMH (2,6-dimethylhept-5-en-2-ol) (IS) 98% Carl Roth, Karlsruhe, Gemany cumene (IS) 99% Sigma-Aldrich, Seelze, Germany

Concept of Riesling wines’ typicality

37

Table 2-2 Applied standards in analysis of C13-norisoprenoids, purity and suppliers; (IS) indicates compounds used as internal standard Compound Purity Supplier

cumene (IS) >99% Sigma-Aldrich, Seelze, Germany DCA (dichloroaniline) (IS) >99% Sigma-Aldrich, Seelze, Germany vitispirane isomers 95%1 gifted by Dr. A. Rapp, Sieberdingen, Germany TDN (1,1,6-trimethyl-1,2-dihydronaphthalene) 98%1 Faculté d'Oenologie, Bordeaux, France β-damascenone >99.9% gifted by Firmenich, Geneva,Switzerland β-ionone >99% Sigma-Aldrich, Seelze, Germany TPB (1,2,4-trimethylphenyl-3-butadiene) 98%1 Faculté d'Oenologie, Bordeaux, France

trans-ethyl cinnamate 99.9% Sigma-Aldrich, Seelze, Germany 1 purity determined by means of total ion chromatogram peak areas oft the pure standard

Table 2-3 Applied standards in volatile sulfur low boiling sulfur compounds, purity and suppliers; [ ] indicates the compound used for standard curve; (IS) indicating compounds used as internal standard Compound Purity Supplier

[sodium sulfide] >99% Sigma-Aldrich, Seelze, Germany

methanethiol >99% Sigma-Aldrich, Seelze, Germany methyl sulfide >97% Alfa Aeser, Karlsruhe, Germany carbon disulfide >99% Acros Organics, Geel, Belgium methylthioacetate >99.9% Acros Organics, Geel, Belgium methyl disulfide >98% Alfa Aeser, Karlsruhe, Germany ethyl thioacetate >99.5% Alfa Aeser, Karlsruhe, Germany diethyl disulfide >99% Sigma-Aldrich, Seelze, Germany dimethyl trisulfide >98% Acros Organics, Geel, Belgium butylmethyl sulfide (IS) >98% Lancaster, Ward Hill, MA, USA methyl-iso-propyl sulfide (IS) >96% Alfa Aeser, Karlsruhe, Germany

Concept of Riesling wines’ typicality

38

Table 2-4 Applied standards in analysis of varietal and polyfunctional thiols, purity and suppliers; (IS) indicates compounds used as internal standard

Compound Purity Supplier

benzenemethanethiol >98% Sigma-Aldrich, Saint-Quentin Fallavier, France 2-furanylmethanethiol >98% Sigma-Aldrich, Saint-Quentin Fallavier, France 3-sulfanylhexan-1-ol >98% Lancaster, Bischheim, France 3-sulfanylhexyl acetate >98% Interchim, Montlucon, France 4-methyl-4-sulfanylpentan-2-one >98% Interchim, Montlucon, France ethyl-2-sulfanyl acetate >98% Alfa Aesar, Schiltigheim, France ethyl-2-sulfanylpropionate >98% Sigma-Aldrich, Saint-Quentin Fallavier, France ethyl-3-sulfanylpropionate >98% Sigma-Aldrich, Saint-Quentin Fallavier, France 1-methoxy-3-methylsulfanylbutane (IS) >98% Oxford Chemicals, Hartlepool, Great Britain 6-sulfanylhexan-1-ol (IS) >98% Interchim, Montlucon, France Table 2-5 Applied standards in analysis of bound and free terpenes and C13-norisoprenoids, purity and suppliers; (IS) indicates compounds used as internal standard Compound Purity Supplier

geraniol >98% Sigma-Aldrich, Seelze, Germany linalool 95% Sigma-Aldrich, Seelze, Germany nerol >98% Sigma-Aldrich, Seelze, Germany β-citronellol >95% Sigma-Aldrich, Seelze, Germany trans/cis-linalool oxide >99% Sigma-Aldrich, Seelze, Germany nerol oxide >95% Interchim, Montlucon, France α-terpineol >99% Sigma-Aldrich, Seelze, Germany β-damascenone >99% Firmenich, Geneva,Switzerland α-ionone >98.5% Sigma-Aldrich, Seelze, Germany β-ionone >98.5% Sigma-Aldrich, Seelze, Germany vitispirane isomers 95%1 gifted by Dr A. Rapp, Sieberdingen, Germany TDN (1,1,6-trimethyl-1,2-dihydro-naphthalene) 98%1 Faculté d'Oenologie, Bordeaux, France TPB (1,2,4-trimethylphenyl-3-butadiene) 98%1 Faculté d'Oenologie, Bordeaux, France octan-3-ol (IS) >98% Sigma-Aldrich, Seelze, Germany DMH (2,6-dimethylhept-5-en-2-ol) (IS) >98% Carl Roth, Karlsruhe, Gemany 1 purity determined by means of total ion chromatogram peak areas oft the pure standard

Concept of Riesling wines’ typicality

39

Table 2-6 Applied chemical products, others than volatile standard compounds for Gas Chromatography analysis Chemical Product Purity Supplier

resins and cartridges Dowex resin 1x2 Acros Organics, Noisy-le-Grand, France SPE - LC-18 (500 mg, 6 mL) Supelco France, Saint Germain-en-Laye Cedex, France SPE - SDBL (500 mg, 6 mL) Phenomenex, Darmstadt, Germany

other products n-alcanes mixture (C6-C23) Sigma Aldrich, Seelze, Germany caffeic acid >99% Sigma Aldrich, Seelze, Germany CHES >99% Sigma-Aldrich, Saint-Quentin Fallavier, France citric acid >98% Carl Roth, Karlsruhe, Germany L-cysteine chloride >98% Sigma-Aldrich, Saint-Quentin Fallavier, France formic acid >99% Sigma-Aldrich, Saint-Quentin Fallavier, France hydrochloric acid (conc.) 38% VWR, Fontenay sous Bois, France monobromobimane >93% Sigma-Aldrich, Saint-Quentin Fallavier, France p-hydroxymercuribenzoate sodium salt >99% Sigma-Aldrich, Saint-Quentin Fallavier, France S-3-(hexan-1-ol)-glutahione-d5 Faculté d'Oenologie, Bordeaux, France1

S-3-(hexan-1-ol)-glutathione Faculté d'Oenologie, Bordeaux, France1

S-3-(hexan-1-ol)-L-cysteine Faculté d'Oenologie, Bordeaux, France1

sodium acetate >99% Sigma-Aldrich, Saint-Quentin Fallavier, France sodium sulfate (anhydreous) >99,9% Sigma-Aldrich, Saint-Quentin Fallavier, France trans-ethylcinnamate-d5 Faculté d’Oenology, Bordeaux, France2

tartaric acid >99,5% VWR, Fontenay sous Bois, France trifluoroacetic acid >99,5% Sigma-Aldrich, Saint-Quentin Fallavier, France Trizma-base >99,8% Sigma-Aldrich, Saint-Quentin Fallavier, France 1synthesized by C. Thibon; 2 synthesized by G. Antalick

2.2.3 Wine samples In total 41 commercial mostly monovarietal white wines from different geographic origins and varieties were analysed regarding their sensory and chemical attributes (Table 2-7).

Concept of Riesling wines’ typicality

40

Table 2-7a Wines used in the present study; x indicates, which wine was tested in which location by which panel (GH: Geisenheim, Hesse, Germany; BX: Bordeaux, Aquitaine, France)

wine- no.

identifier variety origin winery - wine vintage panel

GH BX

1 RIRH1 Riesling D - Rheingau Stefan Breuer - Riesling trocken 2008 x x 2 RIRH2 Riesling D - Rheingau Schloss Vollrads - Kabinett trocken 2007 x x 3 RIRH3 Riesling D - Rheingau Weingut der Forschungsanstalt Geisenheim - Von Lade QbA trocken 2007 x 4 RIRH4 Riesling D - Rheingau Schloss Vollrads - Kabinett trocken 2008 x x 5 RIRH5 Riesling D - Rheingau Schloss Johannisberg - Rotlack Kabinett trocken 2007 x x 6 RIRH6 Riesling D - Rheingau Schloss Johannisberg - Rotlack Kabinett trocken 2008 x x 7 RIRH7 Riesling D - Rheingau Weingut Georg Breuer - Riesling Estate Rüdesheim 2008 x 8 RIRH8 Riesling D - Rheingau Schloss Johannisberg - Riesling Grünlack Spätlese 2008 x x 9 RIRH9 Riesling D - Rheingau Weingut der Forschungsanstalt Geisenheim - Von Lade QbA trocken 2008 x x 10 RIRH10 Riesling D - Rheingau Hessische Staatsweingüter - Kloster Eberbach - Berg Rottland 2008 x 11 RIRH11 Riesling D - Rheingau Schloss Johannisberg - Riesling Grünlack Spätlese 2007 x x 12 RIRH12 Riesling D - Rheingau Weingut Robert Weil, Rheingau - QbA trocken 2007 x 13 RIRH13 Riesling D - Rheingau Schloss Vollrads - Edition QbA 2007 x 14 RIRH14 Riesling D - Rheingau Schloss Vollrads - Edition QbA 2008 x 15 RIAL1 Riesling F - Alsace Domaine Bernhard Reibel - Hahnenberg 2007 x 16 RIAL2 Riesling F - Alsace Paul Ginglinger - Grand Cru Pfersichberg 2007 x 17 RIAL3 Riesling F - Alsace Domaine Frédéric Mochel - G.C. Altenberg de Bergbieten - Cuvée Henriette 2007 x x 18 RIAL4 Riesling F - Alsace Domaine Hugel 2007 x 19 RIAL5 Riesling F - Alsace Domaine Josmeyer - Grand Cru Hengst 2007 x 20 RIAL6 Riesling F - Alsace F.E. Trimbach - Riesling Reserve 2007 x 21 RIAL7 Riesling F - Alsace Domaine Albert Mann - Rosemberg 2007 x x

Concept of Riesling wines’ typicality

41

Table 2-7b Wines used in the present study; x indicates, which wine was tested in which location by which panel (GH: Geisenheim, Hesse, Germany; BX: Bordeaux, Aquitaine, France)

wine- no.

identifier variety origin winery - wine vintage panel

GH BX

22 RIAL8 Riesling F - Alsace Domaine André Kienzler - Cuvée Francois Alphonse 2007 x 23 RIMO1 Riesling D - Mosel Knebel-Lehnigk - Winninger Röttgen-Spätlese trocken 2008 x x 24 RIMO2 Riesling D - Mosel Weingut Heymann Löwenstein - Röttgen - QbA trocken 2007 x 25 RIMO3 Riesling D - Mosel Weingut Reichsgraf von Kesselstatt - Wiltinger Riesling trocken 2007 x 26 RINA1 Riesling D - Nahe Weingut Stein - Oberhäuser Leistenberg, Riesling Kabinett 2008 x x 27 RINA2 Riesling D - Nahe Weingut Dönnhoff - Oberhäuser Leistenberg, Riesling Kabinett 2008 x 28 RIPF Riesling D - Pfalz Weingut Müller-Catoir - Haardt, Kabinett trocken 2007 x 29 RIAUT1 Riesling AUT - Wachau Weingut Prager - Riesling Prager Smaragd 2008 x x 30 RIAUT2 Riesling AUT - Wachau Domäne Wachau - Burggarten Riesling Federspiel 2007 x x 31 RINZ Riesling NZ - Wairarapa Johner Estate - Riesling Wairarapa 2008 x x 32 SBBX Sauvignon blanc F - Bordeaux Vignobles Darriet -Sauvignon Blanc 2007 x x 33 SBBX2 Sauvignon blanc F - Bordeaux Vignobles Darriet - Sauvignon Blanc2 2007 x 34 SBSA1 Sauvignon blanc F - Sancerre Henri Bourgeois, Chavignon - Petit Bourgeois 2007 x 35 SBSA2 Sauvignon blanc F - Sancerre Henri Bourgeois, Chavignon - Le MD de Bourgeois 2007 x 36 SBNZ Sauvignon blanc NZ - Marlborough Montana Wines Ltd - Marlborough Sauvignon Blanc 2007 x x 37 SBAL Sauvignon blanc D - Pfalz Bergdoltd - Reif & Nett - Sauvignon Blanc trocken 2007 x 38 CHAUS Chardonnay AUS - Victoria Lindemans Wines - Bin 65 - Chardonnay 2007 x 39 CHRCH Chardonnay RCH – Valle Central A.L.C. Warenvertriebsgesellschaft mbH - Los Pagos 2009 x 40 PGAL Pinot gris F - Alsace Albert Boxler - Pinot Gris 2008 x 41 GTAL Gewürztraminer F - Alsace F.E. Trimbach - Gewürztraminer 2008 x

Concept of Riesling wines’ typicality

42

2.3 Chemical analysis

2.3.1 General wine composition A range of parameters in wines were measured by means of a FTIR-inferometer using an ‘OenoFossTM’ (Foss, Hilleroed, Denmark). Parameter included, were total soluble solids (TSS) in °Brix, reducing sugars in g/L, pH, total acidity as g tartaric acid/L for berry extracts and grape juice (must) and ethanol concentration in vol. %, reducing sugars in g/L and total acidity as g tartaric acid/L for wines. Samples were analysed at the Department Wine Chemistry and Beverage Research at the Geisenheim Research Center.

2.3.2 Analysis of aroma compounds (‘Kaltron-method’): higher alcohols, esters and terpenols

2.3.2.1 Sample preparation for ‘Kaltron’ extracts According to Rapp et al., 1994 modified by Fischer and Rauhut (2005) (not published): 10 mL of the sample (wine) were given into a 10 mL falcon tube (orange cap) using a piston-driven air displacement pipette (Eppendorff). 2 g of sodium chloride (NaCl) were added. 5 µL of the internal standard solution from 2,6-dimethylhept-5-en-2-ol (DMH) and cumene (c(DMH) = 1188 µg/L; c(cumene) = 107 µg/L) were added to each sample using a 10 µL micro syringe (Thermo-Fischer). 100 µL of 1,1,2-trichloro-1,2,2-trifluoroethan (Kaltron) was added using a piston-driven air displacement pipette and agitated for 20 min on an agitation device (Intelli-Mixer; NeoLab, Germany). After centrifugation (3000 rpm; 1700 x g; 8 min) (Megafuge 1.0, Heraeus Instruments) the organic phase was removed using a glass capillary pipette and dried by 50 mg sodium sulfate deposited on glass wool in an Eppendorff pipette tip. For gas chromatography 2 µL were injected. Ions (m/z) and equation for identification and quantification are indicated in Appendix 15.

Concept of Riesling wines’ typicality

43

2.3.2.2 Instrumental settings for ‘Kaltron’ extracts analysis Instrumental settings are given in Table 2-8: Table 2-8 Instrumental settings for ‘Kaltron’ extract analysis

Autosampler Injector CIS-Parameter Temperatures (CIS) Gas Chromatograph Capillary Column Carrier Gas Oven Program Detector Temperatures Aquisition Mode Ionisation SCAN parameters Data acquisition

MSP4 (Gerstel, Mühlheim a.d.R., Germany) Cooled Injection System CIS 3 (Gerstel, Germany) Splitless (1 min) 30 °C (0.01 min) – 230 °C at 12 °C/s; Hold 4 min GC 5890 Series II (Hewlett Packard, USA) Varian VF-5MS (5% phenyl methyl siloxane) (60m x 320 µm ID x 1 µm FT) Helium; constant flow (1 mL/min); Initial pressure 41.5 kPa; Average velocity 27 cm/s; Outlet pressure vacuum 40 °C (1 min) – 125 °C at 3 °C/min; 125 °C – 200 °C at 6 °C/min; Hold 14.2 min 5972 Mass Selective Detector (Hewlett Packard, USA) Transfer line 210 °C ; MS Quad 150 °C; MS Source 230 °C Scan 70 eV m/z 35 – 250; 3.43 Scans/sec Agilent Chemstation (Agilent Technologies, USA)

2.3.3 Analysis of C13–norisoprenoids

2.3.3.1 Sample preparation for C13–norisoprenoid analysis 10 mL of the wine sample were given into a 10 mL head space vial using a piston-driven air displacement pipette (Eppendorf). 5 µL internal standard solution of cumene and 2,6-dichloroaniline (DCA) (c(cumene) = 100.1 mg/L; c(DCA) = 100.0 mg/L) were added to each sample using a 10 µL micro syringe (Thermo-Fischer) and a polydimethylsiloxane coated stir bar (length 10 mm; film-thickness 0.5 mm) was added. The vial was crimped and stirred for 60 min at 1000 rpm for extraction on a multi agitation plate. Then the loaded stir bars were removed and rinsed properly with H2O (MilliQ), dried with lint free tissue and transferred into desorption tubes. Ions (m/z) and equation for identification and quantification are indicated in Appendix 16.

Concept of Riesling wines’ typicality

44

2.3.3.2 Chromatographic conditions for C13-norisoprenoid analysis Instrumental settings are given in Table 2-9: Table 2-9 Chromatographic conditions for C13-norisoprenoid analysis

Autosampler TDU Temperatures (TDU) Injector CIS-Paramter Temperatures (CIS) Gas Chromatograph Capillary Column Carrier Gas Oven Program Detector Temperatures Aquisition Mode Ionisation SCAN parameters SIM parameters Data aquisition

MSP4 (Gerstel, Mühlheim a.d.R., Germany) Transfer temperature fixed (300 °C); splitless; sample remove 20 °C (0.01 min) – 280 °C at 120 °C/min; Hold 4 min Cooled Injection System CIS 4 (Gerstel, Germany) solvent vent; cryo cooling used -150 °C (0.01 min) – 280 °C at 12 °C/s; Hold 5 min GC 6890 (Agilent Technologies, USA) J & W Scientific HP-5 capillary column (5% Phenyl Methyl Siloxane) (60m x 320 µm ID x 0.25 µm FT) Helium; constant flow (1.1 mL/min); Initial pressure 41.5 kPa; Average velocity 27 cm/s; Outlet pressure vacuum 60 °C (1 min) – 110 °C at 5 °C/min; 110 °C – 170 °C at 3 °C/min; 170 °C – 230 °C at 30 °C/min; Hold 9.33 min MSD 5973 N (Agilent Technologies, USA) Transfer line 280 °C ; MS Quad 150 °C; MS Source 230 °C Scan/SIM EM voltage 2141.2 eV m/z 35 – 250; Sample # 2 A/D Samples 4 listed in Appendix 16 Agilent Chemstation (Agilent Technologies, USA)

2.3.4 Analysis of low boiling sulfur compounds

2.3.4.1 Sample preparation for low boiling sulfur compound analysis The sample preparation was carried out according to Rauhut et al. (2005), modified by Irmler et al. (2008). Compounds analysed and calibration parameters for quantification are listed in Appendix 17.

2.3.4.2 Chromatographic conditions for low boiling sulfur compound analysis Chromatographic conditions were applied according to Rauhut et al. (2005), modified by Irmler et al. (2008).

2.3.5 Analysis of varietal thiols

2.3.5.1 Sample preparation for thiol extract analysis For selective extraction of volatile thiols from wine the method described by Tominaga and Duburdieu (2006) was applied.

Concept of Riesling wines’ typicality

45

2.3.5.2 Solutions for sample preparation

2.3.5.2.1 Solution for pH adjustment

For adjusting the pH of the solution and of the samples during sample preparation 10 M sodium hydroxide solution was prepared using H2O.

2.3.5.2.2 Solution for complex formation

2 mM of p-hydroxymercuribenzoate sodium salt (p-HMB) were disolved in a 0.1 M 2-amino-2-(hydroxymethyl)-propan-1,3-diol (TRIZMA®) aqueous buffer solution to result in 1 L solution, which then was adjusted to pH 10.

2.3.5.2.3 Acetate buffer solution

50 mL of 0.1 M sodium acetate buffer solution was prepared for each sample using H2O and was then adjusted to pH 7.

2.3.5.2.4 Solution for analyte elution

60 mL of 35 mM L-cysteine chloride in H2O was prepared for each sample and adjusted to pH 7.

2.3.5.3 Preparing anion exchange resin column To prepare the anion exchange column DOWEX® 1x2 (50 – 100 mesh; 1,5 x 3 cm) was reactivated with 0.1 M HCl and washed with H2O until the eluate showed pH 6. The activated resin was transferred to a flash chromatography glass column (20 cm x 1 cm ID; upside: NS 14; bottom: Teflon rotation tap) to build a 5 cm column bed. This column was rinsed two times with 10 mL H2O.

2.3.5.4 Selective extraction of volatile thiols To 50 mL of centrifuged wine (3000 rpm; 1600 x g; 10 min) containing 50 µL of internal standard solution 100 µg/L 1-methoxy-3-methyl-3-sulfanylbutane (MMB) and 1000 µg/L 6-sulfanylhexan-1-ol (6SH), 7.5 mL of the solution containing p-HMB (2.3.5.2.2) was added and stirred for 1 min using a magnetic stirr bar on a multi agitation plate (600 rpm; 1 min). After that, the pH was adjusted to pH 7 by drop wise addition of the pH adjusting solution (2.3.5.2.1) and the pH-meter. This mixture was agitated to form the thiol-p-HMB-complex at pH 7 for 10 min at 600 rpm. After agitation the sample was percolated through the anion exchange resin column (2.3.5.3) to fix the thiol-p-HMB-complex on the resin. Then the column was washed with 50 mL of an acetate buffer solution (2.3.5.2.3) before releasing the volatile thiols from the column by

Concept of Riesling wines’ typicality

46

replacing them in the thiol-p-HMB-complex with L-cysteine through percolating of 60 mL of an elution solution (2.3.5.2.4). This aqueous solution of selective enriched volatile thiols was extracted twice by liquid-liquid extraction using dichloromethane (4 mL, 3 mL) and using 500 µL of ethyl acetate as mediator during the first extraction. The organic phases were reunified and dried over anhydreous disodium sulfate (~ 1 g) and concentrated under a nitrogen gas stream to 50 µL. 2 µL of this extract were injected for gas chromatography. Ions (m/z) and equation for identification and quantification are indicated in Appendix 18.

2.3.5.5 Chromatographic conditions for thiol extract analysis Instrumental settings are given in Table 2-10: Table 2-10 Chromatographic conditions for thiol extract analysis

Autosampler Injector Mode Temperature Gas Chromatograph Capillary Column Carrier Gas Oven Program Detector Temperatures Aquisition Mode Ionisation SIM parameters Data aquisition

MSP4 (Gerstel, Mühlheim a.d.R., Germany) Split – splitless Injector (Agilent technologies, USA) Splitless (1 min); constant temperature 230 °C GC 5890 Series II (Hewlett Packard, USA) SGE – BP20 (SGE Europe Ltd, UK) (50m x 250 µm ID x 0.22 µm FT) Helium; constant flow (1 mL/min); Initial pressure 151 kPa; Average velocity 27 cm/s; Outlet pressure vacuum 45 °C (10 min) – 230 °C at 3 °C/min; Hold 15 min 5973 Mass Selective Detector (Hewlett Pacckard, USA) Transfer line 250 °C ; MS Quad 150 °C; MS Source 230 °C SIM 70 eV see Appendix 18 Agilent Chemstation (Agilent Technologies, USA)

2.3.6 Analysis of free and bound terpenes and C13-norisoprenoids - potentially volatile terpenens and norisoprenoids (PVTN) For the analysis of free and bound monoterpenes and C13-norisoprenids in berry extracts a modified solid-phase extraction (SPE) approach using XAD-2 resin equivalent SPE cartridges (Günata et al., 1985) and following gas chromatography coupled to mass spectrometry was adapted by replacing the enzymatical release of aglycones by acid hydrolysis using the procedure described by Kotseridis et al. (1999).

Concept of Riesling wines’ typicality

47

2.3.6.1 Sample preparation of must and berry extracts for potentially volatile terpenes and C13-norisoprenoids (PVTN) analysis Frozen must samples and berry extracts were defrosted gently at room temperature over night prior to analysis and were centrifuged (Megafuge – Hereaus, Germany) (3000 rpm; 1600 x g; 10 min) in 50 mL falcon tubes (VWR, Germany). 50 mL and 10 mL of musts and berry extracts respectively, were spiked with octan-3-ol and 2,6-dimethylhept-5-en-2-ol (DMH) as internal standards (c(octan-3-ol) = 50 µg/L and c(DMH)=25 µg/L), for free monoterpenes and C13-norisoprenoid analysis before being loaded on to solid phase extraction cartridges containing 500 mg styrene-divenylbenzene polymer. SPE cartridges (Phenomenex©, Strata SDB-L©) were preconditioned with 5 mL pentane/dichloromethane (2/1; v/v), 10 mL of methanol, 10 mL of methanol/H2O (1/1; v/v) and 10 mL of H2O. After washing with 10 mL of H2O the cartridge was dried under nitrogen flow (~100 ml/min) prior to the elution of free monoterpenes and C13-norisoprenoids using 5 mL of pentane/dichloromethane (2/1; v/v). The eluate was dried by means of anhydrous disodium sulfate and concentrated to the final volume of 50 µL under a gentle nitrogen flow. 2 µL of that extract were injected for gas chromatographic analysis. For bound monoterpenes and C13-norisoprenoids the cartridge was eluted using 10 mL of ethyl acetate. This fraction was evaporated to dryness under vacuum using a rotary evaporator (BÜCHI, Switzerland) (water bath temperature 40 °C; p = 138 kPa). The residue was resolved in 5 mL 0.2 M citric acid (pH=2.5), transferred into a 10 mL head space vial, sealed, and heated for 60 min at 100 °C in a laboratory oven. After cooling down in iced water, octan-3-ol and DMH were added resulting in concentrations of 50 µg/L and 25 µg/L related to the initial sample volume. Afterwards the reaction mixture was extracted 3 times by liquid-liquid extraction with pentane/dichloromethane (2/1; v/v) (2 mL, 1 mL, 1 mL). The recombined organic extracts were dried by means of anhydrous disodium sulfate and concentrated under nitrogen flow to a final volume of 50 µL. 2 µL of these extracts were injected for gas chromatographical analysis. Ions (m/z) and equation for identification and quantification are indicated in Appendix 19. Listed equations for the calibration curves were used for free volatile C13-norisoprenoids and terpenes as well as for acid hydrolised volatile C13-norisoprenoids and terpenes.

Concept of Riesling wines’ typicality

48

2.3.6.2 Chromatographic conditions for potentially volatile terpenes and C13-norisoprenoids (PVTN) analysis Instrumental settings are given in Table 2-11: Table 2-11 Chromatographic conditions for free and bound terpenes and C13-norisoprenoids extract analysis

Autosampler Injector Mode Temperatures Transfer temperature Transfer time Gas Chromatograph Capillary Column Carrier Gas Oven Program Detector Temperatures Aquisition Mode Ionisation SIM parameters Data aquisition

Tri Plus Autosampler (Thermo Fisher Scientific, USA) Programmable Temperature Vaporizing (PTV) Injector (Thermo Fisher Scientific, USA) PTV Splitless (splitless time 1.5 min) 35 °C – 180 °C at 14.5 °C/s 180 °C 1 min Trace GC Ultra (Thermo Fisher Scientific, USA) Agilent DB-Wax (Agilent Technologies, USA) (30 m x 250 µm ID x 0.5 µm FT) Helium; constant flow (1.4 mL/min); Initial pressure 151 kPa; Average velocity 27 cm/s; Outlet pressure vacuum 40 °C (1 min) – 60 °C at 10 °C/min; 60 °C – 200 °C at 3 °C/min; 200 °C – 220 °C at 10 °C/min; Hold 10 min ITQ 900 Ion Trap MS (Thermo Fisher Scientific, USA) Transfer line 240 °C; MS Source 200 °C SIM 70 eV Segment 1 (9 – 18 min): m/z 81, 91, 121, 136, 139, 154 Segment 2 (18 – 20.5 min): m/z 69, 139, 154 Segment 3 (20.5 – 22.5 min): m/z 83, 95, 101, 105, 112, 121, 136, 169 Segment 4 (22.5 – 26 min): m/z 79, 83, 93, 109, 124, 137, 152, 155 Segment 5 (26 – 34 min): m/z 93, 121, 136, 177, 192 Segment 6 (34 – 36.5 min): m/z 91, 93, 108, 109, 121, 123, 138, 142, 157, 172 Segment 7 (36.5 – 38 min): m/z 104, 108, 119, 121, 139 Segment 8 (38 – 39 min): m/z 93, 121, 136, 139, 142, 157, 172, 175, 190, 192 Segment 9 (39 – 41.5 min): m/z 91, 121, 135, 142, 149, 157, 164, 172, 177, 192 Segment 10 (41.5 – 46 min): m/z 91, 100, 121 135, 142, 149, 157, 172, 177 XCalibur (Thermo Fisher Scientific, USA)

Concept of Riesling wines’ typicality

49

2.4 Sensory analysis

2.4.1 Sensory panel The German panel comprised 11 wine professionals and researchers (5 female and 6 male), who were highly experienced in wine sensory analysis, due to their participation in weekly tastings with descriptive tasks at the Geisenheim Research Center, which are often related to research projects concerning white wines, especially Riesling wines. The French panel comprised 10 wine professionals and researchers (3 female; 7 male) who were highly experienced in wine sensory analysis also, due to regularly tastings with descriptive tasks for red and white wines, especially for Sauvignon blanc style wines, at the ISVV Bordeaux. No special training on descriptors was carried out prior tasting sessions for none of the both panels.

2.4.2 Methodology

2.4.2.1 A Not-A test A set comprising 30 wines (wine choice for each panel given in Table 2.7) was evaluated in a first session by the panelists in 5 randomised batches comprising 6 wines by each panel. In each batch of 6 wines a minimum of one non-Riesling wine was presented. Every wine was assessed one time. The wines were presented in dark tasting glasses (ISO, 1977) covered with petri dishes in an air-conditioned (20 °C) tasting room equipped with individual booths. 20 ml of each wine was served at 14 °C in randomised manner to obtain different sample order for each panelist to avoid order effects. The assessors were asked to answer the question, whether the sample tested orthonasally is a Riesling wine or not.

2.4.2.2 Measurement of typicality and free choice descriptor frequency According to a modified procedure of Ballester (2005) the same sets of wines tested in the first session (2.4.2.1) were presented the same way to the judges in 2 additional tasting sessions on two different days. The panellists therefore were asked to rate typicality regarding Riesling wines after othonasal testing on an unstructured 10 cm anchored scale with ‘not typical’ on the left end and ‘very typical’ on the right end. In addition to this, the panellists were asked to name descriptors for each wine. Though a list of Riesling descriptors was provided, the assessors were free in descriptor choice. Descriptors were given in German by the German panel and in French by the French panel.

Concept of Riesling wines’ typicality

50

2.4.3 Statistical methodology

2.4.3.1 Data transformation - descriptor families The free chosen descriptors from both panels were translated from German and French into English descriptors by the panel leader. 19 major descriptor families were formed deducted from the ‘aroma wheel for German white wines’ (DWI, 2008), Fisher and Swoboda (2005), Campo et al., (2008); Campo et al., (2010) and Jaffré et al. (2011) upon discussion with the panellists (Appendix 1). Citation frequencies were calculated on these descriptor families.

2.4.3.2 Analysis of panel performance

2.4.3.2.1 A Not-A test – Chi-square (χ2) and Fishers exact test Chi-Square (χ2) proportions and Fishers exact test were calculated to evaluate the panels’ and the single judges’ performances respectively. A recognition frequency was calculated for each sample over the two assessments.

2.4.3.2.2 Typicality rating - data normalisation

Relative typicality values (T) were computed by normalising each value on each judge’s personal scale with:

Mean values of T from each judge of both repetitions were used to compute the means and the standard deviation for each sample. This method of normalisation was used for further statistical calculations in order to respect the importance of normalisation, which is reported to be important for multivariate statistical methods (McEvan and Schlich, 1991).

2.4.3.2.3 Principal Component Analysis – panels’ homogeneity

Principal Component Analysis (PCA) was carried out on panelists’ typicality ratings for both panels including both repetitions using the R-software (Lê et al., 2008).

Concept of Riesling wines’ typicality

51

2.4.3.2.4 Repeatability of typicality ratings

Repeatability rim of the panel was calculated with:

1

using:

1/

1/

2.4.3.2.5 Descriptor family reproducibility

Descriptor family repeatability in both repetitions for each panellist in both panels was

calculated using the average reproducibility index (Ri) (Campo et al., 2008; Campo et al.,

2010):

12 /

2.4.3.3 Panel comparison

2.4.3.3.1 Generalised Procrustes Analysis (GPA)

In order to compare the sensory concepts of the two panels a Generalised Procrustes Analysis (GPA) was carried out with the ‘Commandeur Method’ using XLStat® (Addinsoft, Andernach, Germany) on the total citation frequencies of the two sessions, relative typicality ratings (T) ,and the A Not-A test derived recognition frequencies for the 18 intersecting wine samples from both panels’ data sets.

Concept of Riesling wines’ typicality

52

2.4.3.4 Product characterisation

2.4.3.4.1 Typicality ratings - Analysis of Variance (ANOVA)

A 3-factorial ANOVA was carried out on relative typicality ratings for each sample including both repetitions with Holm-Sidak post hoc significance test at a significance level of α = 0.05 using Sigma PlotTM (Systat Software GmbH, Erkrath, Germany). The factors wines, judges and repetition were considered.

2.4.3.4.2 Descriptor family citation frequencies – Chi-square (χ2) proportions

A chi-square (χ2) analysis on the citation frequencies (two repetitions) of each descriptor family was carried out for both sets of samples to evaluate significant descriptor families using the k proportions test with ‘Monte Carlo simulation’ (n=5000) at a significance level of α = 0.05 in XLStatTM (Addinsoft). The same was conducted on the data sets separated in the natural categories of Riesling wines and non-Riesling wines in order to distinguish wines of one natural category by single descriptor proportions.

2.4.3.5 Correlation of sensory and analytical data

2.4.3.5.1 Univariate Analysis - Pearson product moment correlation

Linear Pearson product-moment correlations were calculated between i) the relative typicality and the descriptor frequencies ii) the relative typicality and the volatile aroma compounds iii) the descriptor frequencies and the volatile compounds

using Sigma PlotTM (Systat Software GmbH, Erkrath, Germany).

2.4.3.5.2 Multivariate Analysis

2.4.3.5.2.1 Partial Least Square Regression (PLS)

PLS was carried out in XLStatTM (Addinsoft) on both sets of data, including Riesling and non-Riesling wines for the two panels using ‘automatic’ settings, in order to optimise the model. Relative typicality ratings were chosen as quantitative dependent variable (Y) and the aroma compounds’ quantitative data as explanatory variables (X), which were normalised, centered and reduced before. Components were added according to the Q2 criterium explained by Tenenhaus et al. (2005), while the confidence interval was set at 95%. Due to the model-fit expressed through Q2 (cum) = 0.398; R2Y (cum) = 0.636; R2X (cum) = 0.125) and Q2 (cum) = 0.411; R2Y (cum) = 0.737; R2X (cum) = 0.119 respetively, the model was not satisfying. Therefore indicated outliers were removed and the procedure was repeated on the

Concept of Riesling wines’ typicality

53

new set (Q2 criterium in results). Important variables showing values bigger than 0.8 in the ‘Value of Importance in Projection’ (VIP) tables were considered to be useful for interpretation (Addinsoft, Andernach, Germany; personal communication).

2.4.3.5.2.2 Canonical Correspondence Analysis (CCA) (ter Braak and Verdonschot, 1995)

CCA was applied on both total (Riesling and non-Riesling wines) as well as on both inner categorical (only Riesling wines) data sets from the 2 panels with user defined settings using XLStatTM (Addinsoft). Wine samples were defined as ‘sites’ and descriptor family citation frequency and Riesling wines recognition frequencies (from A Not-A test) were specified as ‘objects’, whereas volatile compounds’ concentrations and relative typicality ratings (T) were computed as quantitative data. Factors’ filter were set to a minimum of 80%. The variables used for CCA were chosen in the PLS section (2.4.3.5.2.1). For the permutation statistics, the number of permutations was set at 1000 with a significance level of 5% (α = 0.05). Interpretation of the results was done according to ter Braak and Vandenschot (1995).

2.5 Results and discussion To obtain reliable sensory product characterisation, two essential requirements are obvious – a good panel performance and an adequate methodology. In order to ensure these requirements in this study, an evaluation of both panels’ performances regarding the global aromatic typicality appreciation of dry Riesling wines and regarding the applied descriptive tasks was done. In order to verify the methodical approach, two methods of typicality determination, based on different principles, were compared by means of linear correlation.

2.5.1 Panel Performance – measurement of typicality

2.5.1.1 Repeatability of measures Regarding dry Riesling wines direct typicality rating, the panels’ repeatability was characterised by means of the average of all rin (rim), which itself represents the deviation of the mean from the mean normalised typicality rating of a sample. For the German panel rim  is 0.16, whereas for the French panel it is 0.37. A value of 0 is interpreted as a very good repeatability, whereas a value of 1 represents a very poor repeatability. Despite both panels show generally good repeatability for direct Riesling’ wines typicality ratings (Fig. 2-1.), single wines were rated differently in both panels. Whereas the German panel showed no good repeatability for the wines RIPF and RINZ the French panel showed

Concept of Riesling wines’ typicality

54

higher variation in the rating of the typicality of the wines GTAL, RIAL1, RIAL2, RIAUT1, RINA2, RIRH10. For the German panel RIAL7 was assessed only once, due to strong corky off-flavour for the bottle used in the second session. Problems like this could explain why some wines were rated not in the same way in two sessions. Being assessed from different bottles, differences apart from off-flavours can occur from one bottle to another, due to transport conditions or closure issues (e.g. cork quality or leak tightness of screw caps). Most of the wines showing great variability were purchased through retailers and therefore storage and transport conditions were not monitored personally.

Fig. 2-1 Riesling wines typicality ratings for total wine sets tested by the a) German (GH) and b) French (BX) panel in two sessions on two different days; error bars indicate standard errors; wine samples according to Table 2-7.

2.5.1.2 Panels’ homogeneity - Principle Component Analysis (PCA) and Analysis of Variance (ANOVA)

2.5.1.2.1 Typicality ratings - Principle Component Analysis (PCA)

To evaluate the two panels inter individual consistency for Riesling wines’ typicality principle component analysis (PCA) was carried out separately on the two panels’ data sets, containing relative typicality scores for the two panels for all wines. The first two principal components (PC) explained 65.87% and 43.53% for the German and the French panel respectively. This difference in explained variance suggests a greater agreement in the German panel than in the French panel, which is presented graphically in Fig. 2-2. However, in both cases, the panellists all rate the wines’ typicality not inverse along the first dimension.

samples

SBN

ZSB

BX2

SBSA

1SB

SA2

SBB

XC

HAU

SC

HR

CH

RIA

L8R

IAL3

RIM

O2

RIR

H12

RIN

A1R

IAL7

RIM

O3

RIR

H14

RIR

H13

RIR

H4

RIR

H5

RIP

FR

INZ

RIA

UT1

RIR

H1

RIR

H6

RIR

H2

RIM

O1

RIR

H11

RIA

UT2

RIR

H9

RIR

H3

RIR

H8

typi

calit

y

0,0

0,2

0,4

0,6

0,8

1,0

session 1session 2

samples

SBN

ZSB

BX

GTA

LR

IAL1

RIA

L2SB

ALPG

ALR

IAL3

RIA

L4R

IAL5

RIN

A1R

IAU

T1R

IRH

1R

IRH

3R

INA2

RIA

UT2

RIR

H2

RIR

H5

RIR

H4

RIM

O1

RIA

L6R

IRH

6R

IRH

8R

IAL7

RIR

H7

RIR

H9

RIN

ZR

IRH

11R

IRH

10

typi

calit

y

0,0

0,2

0,4

0,6

0,8

1,0

session 1session 2

a) b)

Concept of Riesling wines’ typicality

55

For the French panel Judges B and H rated typicality differently on the second and the third dimension of the PCA, whereas these two dimensions explained only 15.6 % and 12.08 % of the total variance).

Fig. 2-2 Correlation cycles of PCA for relative typicality scores (German panel (a); French panel (b))

2.5.1.2.2 Typicality ratings - Analysis of Variance (ANOVA) for factor judge

For the German panel significant differences were observed for judge G versus judges K (p<0.003) and A (p<0.016) at a significance level of α = 0.05. For the French panel significant differences at a significance level of α = 0.05 were observed between judge A and judges B (p < 0.002), C (p < 0.001), D (p < 0.008) and F (p < 0.018) as well as between judge B and judge G (p < 0.005), between judge C and judges E (p < 0.011), G (p < 0.001) and I (p < 0.039), between judge D and judge G (p < 0.024), between judge F and judge G (p < 0.049) and between judge J and judges A (p < 0.001) and G (p < 0.001). These results show that the German panel ratings were homogenous, whereas differences between the ratings of the French panel were more important, as already deduced from the PCA in 2.5.1.2.1.

2.5.1.3 Panel performance on distinguishing the category Riesling

2.5.1.3.1 Typicality ratings - Anaysis of Variance (ANOVA) for factor wine

In the German panels’ data set 233 pairs of wines were significantly different regarding rated relative typicality at a significance level of α = 0.05, whereas in the French’s data set only 49 pairs of wines were rated significantly different (Fig. 2-3).

a) b)

Concept of Riesling wines’ typicality

56

Fig. 2-3 Multiple dot-plot with dots indicating significant different typicality ratings for wines (ANOVA; Holm-Sidak post-hoc test; α = 0.05; p< 0.05) for a) the German panel (Geisenheim) and b) the French panel (Bordeaux) Despite showing much less significant different results for typicality ratings of wine pairs, the French panel rated non-Riesling wines significantly different from the most typical Riesling wines regarding Riesling wines’ typicality. This shows that both panels, the German and the French one, were able to discriminate typical dry Riesling wines from other wine styles regarding their aromatic expression.

2.5.1.3.2 A Not-A test derived Riesling wines recognition frequencies

Chi-square (χ2) statistics on contingency tables with the categories Riesling and non-Riesling were highly significant for both panels, the German and the French, showing χ2-values of 143.43 and 111.81 respectively at a critical χ2 -value of 3.84 (α = 0.05). For single judges mean χ2-values were 13.77 and 12.37, the lowest values were 9.36 and 6.32, the highest values 25.11 and 20.64 for the German and the French judges respectively (Appendix 2). In this alternative sensory task to direct typicality rating, both panels were able to distinguish Riesling wines from non-Riesling wines to a highly significant degree. The German panel members showed slightly higher agreement of ‘what’ is a Riesling wine like, than the French panel members. This verifies the results of the direct typicality ratings (2.5.1.3.1). Additionally χ2-values could function as statistical characteristic for the ability of a single subject or a whole panel to differentiate categories.

RIR

H1

RIN

A1

RIA

UT1

RIM

O1

RIR

H6

RIR

H4

RIA

L1R

INA

2R

IAL3

RIN

ZR

IAL5

RIR

H3

SB

BX

GTA

LR

IAL6

RIR

H2

RIR

H5

RIA

L2P

GA

LR

IAL7

RIA

L4R

IRH

7R

IRH

11R

IRH

10R

IRH

9S

BN

ZS

BA

LR

IRH

8R

IAU

T2

RIAUT2RIRH8SBALSBNZ

RIRH9RIRH10RIRH11RIRH7RIAL4RIAL7PGALRIAL2RIRH5RIRH2RIAL6GTALSBBX

RIRH3RIAL5RINZ

RIAL3RINA2RIAL1RIRH4RIRH6RIMO1

RIAUT1RINA1RIRH1

RIR

H1

RIN

A1

RIA

UT1

RIM

O1

RIR

H6

RIR

H4

RIR

H13

RIR

H12

RIA

L3S

BS

A2

RIM

O2

RIR

H3

SB

BX

CH

AU

SR

INZ

RIA

L8R

IRH

2R

IRH

5S

BB

X2R

IPF

RIA

L7S

BS

A1

RIM

O3

RIR

H11

RIR

H14

RIR

H9

SB

NZ

CH

RC

HR

IRH

8R

IAU

T2

RIAUT2RIRH8

CHRCHSBNZ

RIRH9RIRH14RIRH11RIMO3SBSA1RIAL7RIPF

SBBX2RIRH5RIRH2RIAL8RINZ

CHAUSSBBX

RIRH3RIMO2SBSA2RIAL3

RIRH12RIRH13RIRH4RIRH6RIMO1

RIAUT1RINA1RIRH1

a) b)

Concept of Riesling wines’ typicality

57

2.5.1.3.3 Comparison of the two applied methods: Pearson product-moment

correlation - Typicality ratings vs. A Not-A test derived dry Riesling wines’

recognition frequencies

Two methods of typicality rating were presented, a direct measurement by means of a scale and an indirect measurement by means of a frequency based approach. In order to compare both results, Pearson product-moment correlations were computed between recognition frequencies for dry Riesling wines and direct measured typicality ratings at a significance level of α = 0.05 for both panels’ total data sets (Appendix 5 and Appendix 6) as well as the data sets comprising only Riesling wines. Pearson product-moment correlation tables for the data sets consisting only Riesling wines are shown in Appendix 7 and Appendix 8 for the German and French panel respectively. For the German panel, direct relative typicality ratings showed good correlation with the A Not-A test derived recognition frequencies for Riesling wines (r = 0.738; p < 0.001) in a data set only containing Riesling wines. This was not verified by the French panels results (r = 0.338; p = 0.106), but in the total data sets, correlations were strong and highly significant for the German (r = 0.935; p < 0.0001) and the French panel (r = 0.671; p < 0.0001).

Fig. 2-4 Correlation of direct typicality ratings and A Not-A derived recognition frequencies for Riesling wines for the total data sets including non-Riesling wines for a) the German panel (Geisenheim) with r=0.935, p<0.0001 and b) the French panel (Bordeaux) with r=0.671, p<0.0001

0,00

0,10

0,20

0,30

0,40

0,50

0,60

0,70

0,80

0,90

1,00

0 20 40 60 80 100

typicality

frequency (A Not‐A)

Geisenheim

Geisenheim Linear (Geisenheim)

0,00

0,10

0,20

0,30

0,40

0,50

0,60

0,70

0,80

0,90

1,00

0 20 40 60 80 100

typicality

frequency (A Not‐A)

Bordeaux

Bordeaux Linear (Bordeaux)

non‐Riesling

non‐Riesling

a) b)

Concept of Riesling wines’ typicality

58

2.5.2 Discussion on panel performance and methodology comparison concerning the measurement of typicality Two different sensory methodogical approaches led to the deduction of a clear evidence of a varietal sensory space for Riesling wines at a very high level of significance for the German panel and a little less significant for the French panel. The observed higher agreement on Riesling wines typicality within the German panel in comparison to the French panel might be explained according to the theory of family resemblance (Rosch and Mervis, 1975). The German panels’ high agreement on Riesling wines’ typicality can result from the formation of a distinctive prototype (idéo-type). In comparison to that, the prototype for the French panellists was less pronounced and showed more fuzzy borders than the category, which was formed by the German panel, which is more often exposed to Riesling wines. This would be confirmed by hypothesis of refining a prototype and adjustment of category boundaries depending on the habitual exposure to a product (Brochet and Dubourdieu, 2001). Furthermore, this would be in accordance to the results from earlier studies, where panellists, which frequently were exposed to the concerned product, generally showed a better performance on typicality ratings on a certain product than those being not frequently exposed to that product (Ballester et al., 2008; Yusop et al., 2009). The German panellists are likely to be more frequently exposed to Riesling wines, than the panellists from Bordeaux, as they participate in far more sensory tasting sessions concerning Riesling wines than the French panellists do. Anyway, the introduction of Pinot gris and Gewürztraminer wines might influenced the total rating behaviour of the French panel regarding Riesling wines’ typicality rating, as the same Riesling wines tested orthonasally were presented in a different context, which can be seen as a different olfactory setting. The influence of the setting and the context on typicality rating of a product was already observed in another more general study on packaging (Blijlevens et al., 2012). The results of the Pearson product-moment correlations between the results obtained by the applied methods could be interpreted also in this way. For the German panel inner categorical differences are much more evident than for the French panel. The correlation for the total data set is stronger for the German panel and is still strong, after removing the non-Riesling wines, whereas it is not significant for the French panels’ data. For the French panel it can be imagined that the panellists associate Riesling wines’ typicality to those wines, whose sensory concept they are not that familiar with. For example, the Bordeaux panel will have more expertise on Sauvignon blanc style white wines and on red wines. Therefore it could be considered that the strategy of the French panellist was, to check first, if the wine to

Concept of Riesling wines’ typicality

59

be rated, is a member of those better known categories or not. And if it is not, it would probably be a Riesling wine, due to being in a sensory testing session, which is asking for Riesling wines’ typicality. This could explain the lack of correlation between direct and frequency based measurement results for the data sets only containing Riesling wines, but showing correlation if non-Riesling wines, especially Sauvignon blanc style wines are included.

2.5.3 Panel performance on descriptive task Regarding the descriptive task of describing aromatic expression of the wines by free chosen descriptors, it is essential to monitor the panels’ performances on this task for a reasonable interpretation of the obtained results. Therefore the inter panel semantic diversity and the inner panel repeatability of descriptor citation was evaluated.

2.5.3.1 Descriptor families - choice of descriptors The German panel used 154 different descriptors, whereas the French panel used 134 different descriptors. In total 173 single descriptors were applied by both panels, 111 of these were commonly applied. These descriptors were grouped into 19 families shown in Appendix 1. Slightly different descriptor numbers were used in some of the odour families, depending on the panel. Whereas in the German panel the descriptors blackcurrant, elderberry and gooseberry were named regarding the fruits of these plants, the French panel used the descriptor blackcurrant with respect to the vegetative parts of the plant (e.g. blackcurrant wood, blackcurrant bud) and the descriptors elderberry and gooseberry not at all. For the descriptor family citrus fruit, the German panel applied less different descriptors than the French panel, which used a broader range of descriptors. A similar effect was observed for the descriptor family dried / candied fruits, when the French panel used more different discriptors than the German panel. Inversely, the German panel used more terms for the descriptor family fermentative aroma. In conclusion, there was a high agreement among applied descriptors of both panels. However, some variability was observed for single descriptor families.

2.5.3.2 Descriptor families – average reproducibility index (Ri) To monitor individual panelists’ and total panel performance, applied descriptor families’ average reproducibility indices (Ri) were calculated for single judges representing single measurement units as well as for the total panel considered to be one unit of measurement (Table 2-12). In contrast to Campo et al. (2008), discrimination was undertaken neither for descriptors and descriptor families nor for single judges concerning further data exploitation.

Concept of Riesling wines’ typicality

60

Individual reproducibility for judges, expressed as Ri, ranged from 0.1 to 0.36 in the German panel and from 0.02 to 0.39 in the French panel, representing a mean of 0.2 for the German panel and a mean of 0.28 for the French panel. Therefore, the shown individual repeatability was relatively low for individual judges from both, the German and the French panel. Interestingly, the French panellists generally showed higher reproducibility, than the German panelists did. Regarding the two panels as one unit of measurement, Ri’s were 0.59 and 0.68 for the German and the French panel respectively, higher reproducibility is reflected for the performance of the French panel. Table 2-12 Average reproducibility indices (Ri) (n=2) for single judges and total panels of the German panel (Geisenheim) and French panel (Bordeaux) regarding descriptor family frequencies

German panel French panel judges Ri judges Ri

A 0.16 A 0.29 B 0.10 B 0.38 C 0.17 C 0.20 D 0.36 D 0.39 E 0.12 E 0.29 F 0.16 F 0.28 G 0.15 G 0.02 H 0.29 H 0.30 I 0.24 I 0.33 J 0.17 J 0.33 K 0.24

total panel 0.59 total panel 0.68

2.5.4 Discussion on panel performance for the descriptive task Good consistency between the two panels concerning odour vocabulary was expected. Due to working with similar products, wine professionals would be assumed to develop similar linguistic structures concerning the specific product facilitating within-group communication (O’Mahoney, 1991; cited by Parr et al., (2007). Generally, there was a good consistency between the two panels concerning the applied descriptors. However, results also showed some differences in the use of descriptors. The use of different descriptors could be associated with a different ‘odour-library’ which often is dependent on geographic origin of the panelists. This would be reasonable, due to

Concept of Riesling wines’ typicality

61

‘odours frequently evoke idiosyncratic, autobiographical memories, which are often learned in childhood’ (Chu and Downes, 2000; cited by Parr et al., 2002). Also semantic memory is assumed to be based on a person’s general knowledge and experience with an odourant (Savic and Berglund, 2000; cited by Parr et al., 2002). Especially the fact, that the German panel was more limited for citrus fruit descriptors and that in another way the French panel did not use elderberry and gooseberry as descriptors, can be discussed in that direction. From a German point of view, elderberries and gooseberries are more abundant in domestic gardens than lemon trees in various variations are. In fact, it could be shown in intercultural studies that odourants which are not particularly familiar to panelists are difficult to identify and that shifts in odour category boundaries might be explained by cultural factors, such as familiarity with specific odours (Chrea et al., 2005). Another study observed ‘richness of language’ as an influential factor to explain differences in the sensory characterisation of jellies between two countries (Blancher et al., 2007). However, it can be considered, that this should not be a major factor between German and French panels. It has to be remarked that these factors were revealed in studies comparing Asian and western cultures (Vietnamese, French, North Americans), but to a smaller extent comparable effects can exist between western cultures (Risvik et al. 1992). Lower mean Ri observed for the Geisenheim panel did reflect less good reproducibility for the descriptors citations in comparison with the French panel. However, the German panels’ mean value was still considered above the exclusion criterion for individual panellists proposed by Campo et al. (2008). It has to be mentioned that both panels have not undergone a specific training and that therefore the use of descriptors was surprisingly stable. Even if single judges’ repeatability was shown to be relatively poor, the global Ri’s for both panels were shown to be 0.59 and 0.68, corresponding to 59% and 68% of descriptor family repeatability for the named descriptors for the German and the French panel respectively. This suggests that, when working with highly experienced wine tasters, a common sense of vocabulary is available, especially within a panel, which is used to work together for a longer period of time, which was the case for both panels in this study. This would be a practical implication of the assumption that ‘wine professionals working with similar products [...] over the course of time, develop conceptual alignment and similar linguistic structures concerning the specific product so as to facilitate within-group communication’ (O’Mahoney, 1991; cited by Parr 2007). Consequently, even if individual repeatability is relatively poor, the result of a global group can be highly repeatable. The same observation was made by Sarrazin (2007b) regarding the rating of Sauternes sweet wines typicality. One interpretation could be that human beings are never in exactly the same

Concept of Riesling wines’ typicality

62

physiological and psychological state, which can lead to low repeatability in sensory judgements in different sensory testing sessions.

2.5.5 Essential outcomes regarding panel performance and methodology Both panels were able to judge wines upon their aromatic expression towards Riesling wines’ typicality in repeatable manner. Both panels had a homogeneous vision of how a typical Riesling wine should express aromatically (2.5.1.2.1), although the observed homogeneity was higher for the German panel than for the French one. Results from both methodical approaches showed that both panels could significantly discriminate Riesling wines from other wine styles (e.g. Bordeaux style white wine blends characterised by Sauvignon blanc aroma). Both methodologies were shown to be tools in typicality determination showing similar results, whereas in this study the frequency based methodology (A Not-A test) only showed reliable differentiation inside a category if the panel was highly experienced with the product category to be tested. This was demonstrated by correlations of the results of both applied methodologies for both panels’ data sets (2.5.1.3.3). There was a high agreement in the products’ characterisation on a semantic level, but it was observed, that in some cases different descriptors seemed to be attached to a certain odour. Regarding each panel’s repeatability for descriptor usage, relatively poor repeatability was observed for individual judges, but good repeatability was observed for the whole panels’ performances, showing tendentially better performance for the French panel (2.5.3.2). Both sensory panels were qualified to fulfil the demanded tasks, as both methods to rate typicality in combination with the demanded descriptor task turned out to be suitable for characterisation of the sensory concept of typical dry Riesling wines.

Concept of Riesling wines’ typicality

63

2.5.6 General sensory concept of typical Riesling wines

2.5.6.1 German panels’ and French panels’ concept of typical Riesling wines A very simple way of comparing the sensory concept of Riesling wines’ typicality of two panels, and therefore if the same Riesling wine is appreciated as a typical Riesling wine, is to plot the results of Riesling wine typicality ratings of wines tested by both panels. Both panels were able to hierarchize the wines being presented according to their typicality and both panels discriminated non-Riesling wines, especially Sauvignon blanc style wines, from Riesling wines. Even if rated at a different typicality level, two out of three Riesling wines rated as most typical in one panel were rated among the three most typical Riesling wines in the other panel as well (Fig. 2-5).

Fig. 2-5 Average Riesling wines’ typicality ratings of both panels, for the 18 wines tested by both, German (GH) and French (Bx) panel (Session 1 respectively). * indicating the three wines appreciated as the most typical by both panels; N.a.N. sample was not assessed in both panels; sample names indicated in Table 2-7;

SBN

ZSB

BX

RIA

L3R

INA1

N.a

.N.

RIR

H4

RIR

H5

RIN

ZR

IAU

T1R

IRH

1R

IRH

2R

IRH

11R

IRH

6R

IMO

1R

IAU

T2R

IRH

9R

IRH

3R

IRH

8

typi

calit

y

0,0

0,2

0,4

0,6

0,8

1,0

GHBX

**

*

***

Concept of Riesling wines’ typicality

64

2.5.6.2 Univariate Pearson product-moment correlation - typicality ratings vs. descriptor family frequencies In order to gain insights into Riesling wines’ typicality’ characteristics, descriptor family frequencies were correlated to its relative values. Pearson product-moment correlations were computed between the descriptor family frequencies, the recognition frequency of Riesling wines (A Not-A test) and the directly measured typicality ratings at a significance level of α = 0.05 on both panels’ total data sets (Appendix 5 and Appendix 6) and on those only comprising Riesling wines. Pearson product-moment correlation tables for the data sets only including Riesling wines are shown in Appendix 7 and in Appendix 8 for the German and the French panel respectively. For the German panel, descriptor families’ frequencies vegetal (r=-0.764; p<0.0001), sauvignon (r = -0.755; p < 0.0001), Ribes ssp. / Sambucus ssp. (r = -0.675; p < 0.0001), fruity (r = 0.631; p = 0.0001), white fruit (r = 0.479; p = 0.007), yellow fruit (r = 0.658; p < 0.0001) and minerality (r = 0.595; p = 0.001) showed significant correlation to relative typicality ratings in the total data set (Appendix 5). Correlations observed in an exclusively Riesling wines containing set, were fruity (r = 0.590; p = 0.003), citrus fruit (r = 0.556; p = 0.006), yellow fruit (r = 0.476; p = 0.022), sulfur (r = -0.588; p = 0.003), and off-flavour (r = -0.459; p = 0.027) (Appendix 6). For the French panel, descriptor families’ frequencies vegetal (r = -0.504; p = 0.005), sauvignon (r = -0.612; p = 0.0001), citrus fruit (r = 0.692; p < 0.0001) correlated significantly with typicality ratings in the total data set (Appendix 6), whereas for the Riesling set descriptor families citrus fruit (r = 0.492; p = 0.015), MLF (r = -0.456; p = 0.025), and off-flavour (r = -0.479; p = 0.018) were significant (Appendix 8). For German panelists Riesling wines’ typicality was shown to be associated with fruity, white and yellow fruit odours. Including non-Riesling wines, which show distinctive varietal aroma such as Sauvignon blanc wines do, Riesling wines typicality is also associated to minerality. The French panel mostly detected citrus fruit descriptors in Riesling wines’ with high typicality ratings. Comparing the obtained results, German and French panelists agreed that the abundance of citrus fruit odours increased with Riesling wines’ typicality, if tested wines were compared to Riesling wines only. Additionally, the German panelists did smell more fruity and yellow fruit odours, like peach or apricot, in wines being rated as more typical. Both panels clearly associated vegetal and Sauvignon blanc related descriptors to wines showing poor Riesling wines’ typicality.

Concept of Riesling wines’ typicality

65

2.5.6.3 Multivariate methodology - Generalised Procrustes Analysis (GPA) As a multivariate tool to compare sensory data of different panels, concerning their sensory concept of a product, Generalised Procrustes Analysis (GPA) can be used and interpreted regarding three statistical characteristics, the so called ‘transformations’: scaling, translation and rotation (Table 2-13). Despite being optimised for quantitative sensory data (from Quantitative Descriptive Analysis or Free Choice Profiling), it can also be applied for frequenced data as obtained in this study, due to lacking accessible alternative. The resulting visualisation might not be optimal for this kind of data but it can provide a good overview of the two sensory spaces anyway. To compare the sensory concepts described by both panels, GPA was applied on the two data sets of mean sensory values of the intersecting samples (18 out of 30 wines) evaluated by both panels. The first two dimensions of the consensus model explained 71.5% (56.9% and 14.5% respectively) of the total variance. All three transformations performed in the GPA showed a significant effect. Significance for the scaling transformation (p < 0.001) uncovered the panels’ usage of different ranges of the measurement scale (Risvik et al., 1992), which in this case means a differing citation frequency range for single descriptor families by the whole panel. For example it was observed that in the German panel’s data for the family sauvignon the range was 0 – 22 and in the French panel’s it was 0 – 10. In an opposite way, the descriptor family citrus fruit ranged from 0 – 4 for the German panel and from 0 - 8 for the French panel, the descriptor family floral ranged from 0 – 5 for the German data and from 0 - 10 for the French data respectively. Table 2-13 Transformations conducted in General Procrustes Analysis (GPA)

source DF sum of squares mean squares F Pr > F

residuals after scaling 149 224.00 1.50 scaling 1 46.08 46.08 30.649 < 0.001residuals after rotation 150 270.08 1.80 rotation 190 422.15 2.22 1.478 0.006 residuals after translation 340 692.22 2.04 translation 20 407.45 20.37 13.551 < 0.001corrected total 360 1099.68 3.05 DF: degrees of freedom; F: f-value; Pr > F: indicating significance at α<0.05 for difference between ‘before’ and after ‘transformation’

Concept of Riesling wines’ typicality

66

The second significant effect was observed for the translation transformation (p < 0.0001) explaining that both panels used different levels of the measurement scale (Risvik, 1992), which indicates in this case differing citation frequency levels of single descriptor families for the whole panel. This effect can be demonstrated by the descriptor family fruity which ranged from 1 – 10 for the German panel and from 4 – 16 for the French panel. For wines RIAUT1 and RIAUT2 the descriptor family frequencies were rated at 7 and 10 by the German panel whereas the French panel rated them at 15 and 16. For the sample SBNZ the descriptor family frequency for this descriptor was 1 for the German panel and 4 for the French panel respectively. Finally the rotation transformation was significant (p = 0.006), which means that the panels were using different expressions and descriptors to describe the same samples. It was expected, that due to the sorting of descriptors, some variance in the semantic descriptor choice was already removed. Even though, this value uncovered a difference in the usage of descriptors for certain odours between the two panels.

Fig. 2-6 General Procrustes Analysis: Correlation circle of the descriptor family frequencies and the relative typicality ratings for the German and the French panel (German (GH) = blue and French (BX) = green) obtained in sensory analysis

vegetalsauvignon

spicy

herbs

Ribes / Sambucus

fruity

citrus fruit

white fruit

yellow fruit

tropical fruit

dried / candied fruit

floral

aging notes

mineralityfermentative aroma

sulfurMLF

off‐flavourothers

typicality

vegetalsauvignon

spicy

herbs

fruity

citrus fruit white fruit

yellow fruit

tropical fruitdried / candied fruit

floral

aging notes

minerality

fermentative aroma

sulfur

MLFoff‐flavour

others

typicality

‐1

‐0,75

‐0,5

‐0,25

0

0,25

0,5

0,75

1

‐1 ‐0,75 ‐0,5 ‐0,25 0 0,25 0,5 0,75 1

F2 (1

4,54

 %)

F1 (56,93 %)

Dimensions (axes F1 and F2: 71,47 %)

GH BX

Concept of Riesling wines’ typicality

67

From the bi-plot of the visualisation for the first two dimensions of GPA and its correlation table, it becomes evident that there is a good agreement for the descriptor families vegetal, sauvignon, yellow fruit, floral, aging notes and for relative typicality ratings. However, disagreement was observed for the descriptor families herbs, Ribes ssp. / Sambucus ssp., citrus fruit, tropical fruit, minerality and off-flavour (Fig. 2-6 / Table 2-14). The most abundant differences in the descriptor family’s correlation (Fig 2-5) to the sensory concept like Ribes ssp. / Sambucus ssp., citrus fruit, and minerality can be traced to the raw descriptor data (2.5.3.1 and Appendix 1).

Fig. 2-7 General Procrustes Analysis: Positioning of wines (samples plot) regarding the descriptor family frequencies and the relative typicality ratings for the German and the French panel (German (GH) = blue and French (BX) = green) obtained in sensory analysis as well as a consensus configuration (Consensus = red)) The appreciation of Riesling wines’ global typicality of the German panel complied highly with the French panels’ global typicality ratings. Though, both seemed to relate other descriptor families with global Riesling wines’ typicality, even if correlations were shown to be moderate. Both related typicality ratings to yellow fruit on the first two dimensions of the GPA

SBBXSBNZ

RIAL3

RINA1

RIRH4RIRH5

RIAL7

RIRH1

RIRH8

RINZ

RIRH3

RIRH9

RIRH2

RIAUT1

RIAUT2

RIRH11

RIRH6RIMO1

SBBX

SBNZ

RIAL3

RINA1

RIRH4RIRH5

RIAL7

RIRH1

RIRH8

RINZ

RIRH3

RIRH9

RIRH2

RIAUT1

RIAUT2

RIRH11

RIRH6

RIMO1

SBBX SBNZ

RIAL3

RINA1

RIRH4RIRH5

RIAL7

RIRH1

RIRH8

RINZ

RIRH3

RIRH9

RIRH2

RIAUT1

RIAUT2

RIRH11

RIRH6

RIMO1

‐7

‐2

3

8

13

‐10 ‐5 0 5 10 15 20 25

F2 (1

4,54

 %)

F1 (56,93 %)

Objects (axes F1 and F2: 71,47 %)

GH BX Consensus

Concept of Riesling wines’ typicality

68

projection, whereas German typicality ratings co-correlated also with minerality and the French typicality ratings did additionally co-correlate with the citrus fruit descriptor family. In the samples plot (Fig. 2-7) it got evident that both panels discriminated the non-Riesling wines SBBX and SBNZ from the Riesling wines. Generally, Riesling wines accumulated, but the distincitive aromatic profiles of single wines were characterised by both panels in a similar way as it was observed for the the wines RINZ, RIAL7, RIRH6, or RIMO1, by their positions in samples plot. These results are not exactly the same as obtained from univariate comparison (2.5.6.2), which can be considered to be caused by the fact that both sets of data is only partially the same, due to GPA being based only on 18 samples out of the 30, which were used for the univariate procedure. Table 2-14 Correlation factors r for each descriptor family and each panel for the first two dimensions (F1 and F2) in the General Procrustes Analysis for 18 intersecting wines from sensory analysis (Correlation factor r indicates correlation with the Factor (F), if for both panels the correlation factors are similar there is good compliance regarding the concerning descriptor family; values are interpretable as for classical Pearson product-moment correlations)

descriptor family correlation factor  r 

German panel  French panel F1 F2 F1  F2

vegetal  0,942 0,015 0,904  ‐0,049sauvignon  0,978 ‐0,012 0,918  ‐0,039spicy  0,187 ‐0,200 0,659  ‐0,138herbs  0,173 0,201 0,829  ‐0,056Ribes ssp. / Sambucus ssp.  0,921 ‐0,001 0,000  0,000fruity  ‐0,321 ‐0,574 ‐0,704  ‐0,262citrus fruit  ‐0,007 ‐0,437 ‐0,646  ‐0,200white fruit  ‐0,345 ‐0,510 ‐0,019  ‐0,143yellow fruit  ‐0,392 ‐0,449 ‐0,366  ‐0,283tropical fruit  0,415 0,548 ‐0,239  0,160dried / candied fruit  0,280 ‐0,158 0,073  0,090floral  ‐0,314 ‐0,417 ‐0,344  ‐0,445aging notes  ‐0,300 0,782 ‐0,243  0,781minerality  ‐0,414 ‐0,464 ‐0,029  0,689fermentative aroma  ‐0,306 ‐0,451 0,213  ‐0,278sulfur  ‐0,100 0,284 0,022  0,400MLF  ‐0,085 0,252 ‐0,170  ‐0,171off‐flavour  ‐0,086 0,559 0,203  ‐0,060others  ‐0,307 0,492 ‐0,135  0,450typicality  ‐0,857 ‐0,278 ‐0,817  ‐0,046Descriptor families showing similar correlation factors for the first to factors of GPA are indicated in bold; Descriptor families showing differing correlation factors for the first two factors of GPA are indicated in bold, italic and underlined for the differing factor; MLF: malolactic fermentation 

Concept of Riesling wines’ typicality

69

2.5.7 Essential outcomes regarding universality of Riesling wines’ typicality Both panels agreed to a high degree on Riesling wines’ typicality appreciation. This got evident through both, univariate and multivariate data comparison (2.5.6.1; 2.5.6.2; 2.5.6.3). In addition to that, both panels’ sensory data revealed a consistent sensory concept for typical dry Riesling wines, which is mainly characterised by citrus fruit, fruity and yellow fruit aromatic expression. Typical sauvignon like or vegetal odours are inversely associated with Riesling wines’ typicality (2.5.6.2; 2.5.6.3). However, some inter panel differences in descriptor - odour fitting got obvious, especially in the Generalised Procrustes Analysis.

Concept of Riesling wines’ typicality

70

2.5.8 Chemical base of Riesling wines’ typicality

2.5.8.1 Univariate methodology

2.5.8.1.1 Pearson product-moment correlation - typicality ratings vs. aroma

compound concentrations

Pearson product - moment correlations were also computed between relative typicality ratings and aroma compounds concentrations at a significance level of α = 0.05 for both panels’ data sets with and without removing non-Riesling wines from data sets (Table 2-15 and Table 2-16, respectively). For the German panel’s total data set significant correlations with dimethyl sulfide (DMS) (r = -0.440; p = 0.015), 2-methylpropan-1-ol (r = -0.373; p = 0.042), 3-methylbutanol (r = -0.480; p = 0.007), 2-methylbutanol (r = -0.378; p = 0.039), trans-linalool oxide (r = 0.596; p = 0.001), cis-linalool oxide (r = 0.450; p = 0.013), α-terpineol (r = 0.411; p = 0.024), nerol oxide (r = 0.553; p=0.002) and nerol (r = -0.369; p = 0.045) were observed. Table 2-15 Significant Pearson product - moment correlations (α = 0.05) between relative typicality ratings and aroma compounds’ concentrations for both panels including all wines (Riesling and non-Riesling)

Variable German – typicality  French ‐ typicality 

r  p  correlation  r  p  correlation Alcohols             

3‐methyl‐1‐butanol  ‐0.480  0.007  slightly negative  nc  ns  ‐ 

2‐methyl‐1‐butanol  0.378  0.039  slightly positive  nc  ns  ‐ 

Monoterpenols             

α‐terpineol  0.411  0.024  slightly positive  nc  ns  ‐ 

nerol  ‐0.369  0.045  slightly negative  nc  ns  ‐ 

Monoterpenol oxides             

cis‐linalool oxide  0.450  0.013  slightly positive  0.421  0.023  slightly positive

trans‐linalol oxide  0.596  0.001  positive  nc  ns  ‐ 

nerol oxide  0.553  0.002  positive  0.396  0.033  slightly positive

Sulfur compounds             

dimethyl sulfide (DMS)   ‐0.440  0.015  slightly negative  nc  ns  ‐ 

methanethiol  nc  ns  ‐  0.394  0.035  slightly positive

ethyl 2‐sulfanylacetate  nc  ns  ‐  ‐0.559  0.002  negative 

4‐methyl‐4‐sulfanylpentan‐2‐one  nc  ns  ‐  0.398  0.032  slightly positive

ethyl 2‐sulfanylpropionate  nc  ns  ‐  ‐0.504  0.005   

ethyl 3‐sulfanylpropionate  nc  ns  ‐  ‐0.617  <0.001  negative 

nc: no correlation; ns: not significant; variables showing significant correlation for both panels indicated in bold   

Concept of Riesling wines’ typicality

71

For the French panel’s total data set significant correlation were observed with methanethiol (r = 0.394; p = 0.035), hexanoic acid (r = 0.370; p = 0.048), cis-linalool oxide (r = 0.421; p = 0.023), nerol oxide (r = 0.396; p = 0.033), ethyl 2-sulfanylacetate (r = -0.559; p = 0.002), 4-methyl-4-sulfanylpentan-2-one (r = -0.398; p = 0.032), ethyl 2-sulfanylpropionate (r = -0.504; p = 0.005) and ethyl 3-sulfanylpropionate (r = -0.627; p < 0.001). Table 2-16 Significant Pearson product-moment correlations (α = 0.05) between relative typicality ratings and aroma compounds’ concentrations for both panels including only Riesling wines

Variable German – typicality  French ‐ typicality 

r  p  Correlation  r  p  Correlation Alcohols             

3‐methylbutanol  ‐0.434  0.038  slightly negative  nc  ns  ‐ 2‐methylbutanol Ethylester 

‐0.486  0.019  slightly negative  nc  ns  ‐ 

ethyl 2‐methylpropanoate  ‐0.524  0.010  negative  nc  ns  ‐ 

ethyl butanoate  ‐0.502  0.015  negative  nc  ns  ‐ 

diethyl succinate  ‐0.699  <0.001  negative  nc  ns  ‐ 

miscellaneous             

2‐phenylethanol  ‐0.507  0.010  negative  nc  ns  ‐ 

Monoterpenols             

nerol  ‐0.426  0.041  slightly negative  nc  ns  ‐ 

geraniol  0.530  0.009  positive  nc  ns  ‐ 

Norisoprenoids             

β‐ionone  0.415  0.049  slightly positive  nc  ns  ‐ 

Sulfur compounds             

3‐sulfanylhexanol  0.498  0.015  slightly positive  0.429  0.036  slightly positive

benzenemethanethiol  nc  ns  ‐  ‐0.711  <0,001  strongly negative

methanethiol  nc  ns  ‐  0.417  0.043  slightly positive

carbon disulfide  nc  ns  ‐  ‐0.554  0.005  negative 

ethyl 2‐sulfanylacetate  nc  ns  ‐  ‐0.600  0.002  negative 

ethyl 2‐sulfanylpropionate  nc  ns  ‐  ‐0.579  0.003  negative 

ethyl 3‐sulfanylpropionate  nc  ns  ‐  ‐0.600  0.002  negative 

nc: no correlation; ns: not significant; variables showing significant correlation for both panels indicated in bold;   

When removing all non-Riesling wines from the data set, for the German panel, 3-methyl-1-butanol ( r = -0.434; p = 0.038), 2-methyl-1-butanol (r = -0.486; p = 0.019), ethyl 2-methylpropanoate (r = -0.524; p = 0.010), ethyl butanoate (r = -0.502; p=0.015), 2-phenylethanol (r = -0.507; p = 0.014), diethyl succinate (r = -0.699; p = 0.0001), nerol (r = -0.426; p = 0.041), geraniol (r = 0.530; p = 0.009), β-ionone (r = 0.415; p = 0.049) and 3-sulfanylhexan-1-ol (r = 0.498; p = 0.015) correlated significant with typicality ratings. When removing all non-Riesling wines, data for the French panel correlations to relative typicality ratings were observed for methanethiol (r = 0.417; p = 0.043), carbon disulfide (r = -0.554; p = 0.005), ethyl acetate (r = -0.528; p = 0.008), diethyl succinate (r = -0.441;

Concept of Riesling wines’ typicality

72

p = 0.031), ethyl 2-sulfanylpropionate (r = -0.579; p = 0.003), ethyl 2-sulfanylacetate ( r = -0.600; p = 0.002), ethyl 3-sulfanylpropionate (r = -0.557; p < 0.005), benzenemethanethiol (r = -0.711; p < 0.0001) and 3-sulfanylhexan-1-ol (r = 0.429; p < 0.036).

2.5.8.1.2 Pearson product-moment correlation - Riesling wines’ typicality

correlated aroma compounds vs. total frequency of descriptor families

A common correlation was identified with the total frequency of a descriptor family for both panels’ data, neither for concentrations of terpenol oxides cis-linalool oxide and nerol oxide (Table 2-17), nor for those of the polyfunctional thiol 3-sulfanylhexan-1-ol (Table 2-18). Table 2-17 Pearson product-moment correlations (α = 0.05) between concentrations of cis - linalool oxide, nerol oxide and the total descriptor family frequencies for both panels including all wines (Riesling and non-Riesling)

  German  French descriptor group  cis ‐ linalool oxide  cis ‐ linalool oxide 

  r  p  correlation  r  p  correlation 

fruity 0.503 0.005 slightly positive nc ns - citrus fruit 0.445 0.012 slightly positive nc ns - neroloxide  neroloxide  r  p  correlation  r  p  correlation 

aging notes 0.507 0.004 slightly positive nc ns - nc: no correlation; ns: not significant at a significance level α = 0.05; variables showing significant correlation for both panels are indicated in bold; 

Table 2-18 Pearson product-moment correlations (α = 0.05) between 3-sulfanylhexan-1-ol concentration and the total descriptor family frequencies for both panels on Riesling wines only

German  French 

descriptor group 3‐sulfanylhexan‐1‐ol  3‐sulfanylhexan‐1‐ol 

r  p  correlation  r  p  correlation 

yellow fruit 0.488 0.031 slightly positive nc ns - nc: no correlation; ns: not significant at a significance level α = 0.05; variables showing significantt correlation for both panels are indicated in bold 

Concept of Riesling wines’ typicality

73

2.5.8.1.3 Pearson product-moment correlation - typicality correlated total

descriptor family frequencies vs. aroma compound concentrations

Fruity aroma descriptors were slightly correlated to 3-methylbutyl acetate for both panels. Neither total frequency for citrus fruit descriptors nor for yellow fruit descriptors showed common correlations in both panels (Table 2-19). Table 2-19 Pearson product-moment correlations (α = 0.05) between Riesling wines’ typicality related descriptor families and aroma compound concentrations for both panels in the Riesling wines’ data set

  German  French variable  fruity  fruity 

  r  p  correlation  r  p  correlation acetate ester        

3-methylbutyl acetate 0.486  0.019  slightly positive  0.449  0.028  slightly positive 

variable citrus fruit  citrus fruit 

r  p  correlation  r  p  correlation ethyl ester            

ethyl butanoate nc  ns  ‐  0.599  0.002  positive 

ethyl hexanoate nc  ns  ‐  0.470  0.021  slightly positive 

Monoterpenols             

geraniol 0.526  0.010  slightly positive  nc  ns  ‐ 

variable yellow fruit  yellow fruit 

r  p  correlation  r  p  correlation acetate ester            

3-methylbutyl acetate 0.502  0.015  slightly positive  nc  ns  ‐ 

2-methylbutyl acetate 0.574  0.004  positive  nc  ns  ‐ 

Sulfur compounds            

3-sulfanylhexan-1-ol 0.488  0.018  slightly positive  nc  ns  ‐ 

nc: no correlation; ns: not significant at a significance level α = 0.05; variables showing significant correlation for both panels are indicated in bold;  

Concept of Riesling wines’ typicality

74

2.5.8.2 Multivariate methodology A combination of Partial Least Square Regression (PLS) and Canonical Correspondence Analysis (CCA) regarding Riesling wines’ typicality ratings was applied as a multivariate approach, to gain insights into implied descriptor families as well as into possible contributions of aroma compounds. Typicality ratings, as global judgements, were explained as product characteristics by the variables ‘aromatic compound concentration’ using PLS (Tenenhaus et al., 2005). The most contributing variables in the received model, from both panels’ data sets, have then been introduced into the CCA as metric variables.

2.5.8.2.1 Partial Least Square Regression (PLS)

Typicality vs. aroma compounds – total Partial Least Square regression was applied on the German and French panels’ data sets to extract volatile compounds implicated in the perception of Riesling wines’ typicality. German data set The resulting PLS regression of the relative typicality ratings (Y’s) on the volatile compounds concentrations (X’s) showed good quality on one component (Q2 (cum) = 0.521; R2Y (cum) = 0.690, R2X (cum) = 0.141). The most important variables in the PLS model were chosen by the Variable Importance in the Projection table by choosing variables with minimum values of 0.8 on the first component (Table 2-20). French data set For the French panel PLS regression of relative typicality ratings (Y’s) on the volatile compounds concentrations (X’s) also showed good model quality on three components (Q2

(cum) = 0.738; R2Y (cum) = 0.972, R2X (cum) = 0.375). Like for the German panel, the most important variables in the PLS model were chosen by the Variable Importance in the Projection table by choosing variables with higher values than 0.8 on the first component (Table 2-20). Common important variables To choose the aroma compounds most likely being involved in typicality related sensory properties, intersecting aroma compounds of both panels’ regression models were chosen. The resulting 15 aroma compounds, presented in Table 2-20 were used for further multivariate analysis.

Concept of Riesling wines’ typicality

75

Table 2-20 Common contributing variables for each panel, the German (GH) and the French (Bx) in Partial Least Square regression models for the prediction of relative typicality ratings with Variable Importance in the Projection (VIP) values; The VIP value is indicating the importance of the variable for the PLS model to fit, values >0.8 are considered to have a major impact (personal Communication of Addinsoft, Andernach, Germany)

variable German  French VIP  VIP 

trans‐linalool oxide  2.191  1.435 

linalool  2.186  0.849 

nerol oxide  2.109  0.960 

cis‐linalool oxide  2.079  1.600 

3‐methyl‐1‐butanol  1.611  0.982 

DMS  1.556  1.006 

2‐methyl‐1‐butanol  1.237  1.002 

TDN  1.073  0.800 

diethyl succinate  0.958  1.519 

β‐ionone  0.951  1.091 

4‐methyl‐4‐sulfanylpentan‐2‐one  0.937  1.568 

3‐sulfanylhexan‐1‐ol  0.921  0.909 

vitispirane  0.897  0.934 

ethyl butanoate  0.864  1.316 

benzenemethanethiol  0.802  1.449 DMS: dimethyl sulfide TDN: 1,1,6‐trimethyl‐1,2‐dihydronaphthalene 

2.5.8.2.2 Canonical Correspondence Analysis (CCA) on common variables

from both panels

Canonnical Correspondence Analysis (CCA) allows combining metric and non-metric data in a multivariate method. It is widely used in ecological research (ter Braak and Verdonschot, 1995). An interesting feature of interpretation for sensory results could be the analogism to that of an ‘ecological niche’ for certain species in ecology (ter Braak and Verdonschot, 1995). Due to the complex nature of odour, it is known that sensory properties do not change linearly with changes in aroma compounds’ concentrations and only appear if certain aroma compounds are present in certain concentration ranges and combinations – a ‘sensorial niche’ in this sense. Therefore CCA is considered to be a useful tool to relate several factors and their importance for an odours perception. Furthermore it can be regarded as a logical extension of Correspondence Analysis. Similar approaches were presented for the correlation of sensory and analytical chemistry data by plotting additional variables into Correspondence Analysis plots (Torres and van de Velden, 2007). This also demonstrates

Concept of Riesling wines’ typicality

76

the need of alternatives to Principle Component Analysis (PCA) and Corespondence Analysis (CA) for the combinated presentation of metric and non-metric data.

2.5.8.2.2.1 General properties of the CCA

The CCA-map (Fig. 2-8) was calculated on the German data set comprising only Riesling wines. This data set was chosen, due to showing more significant variation in global typicality ratings within Riesling wines (2.5.1.3.1). The CCA explained 83.11% of the constrained inertia (ci). The first five factors corresponded to 59.12% of the total inertia (ti), which was mostly explained by the first three factors (67.16% and 47.77% cumulative constrained and total inertia respectively). The eigenvalues of F1 (horizontally) and F2 (vertically) are 0.23 and 0.13, respectively; the eigenvalue of F3 (not displayed in Fig. 2-8) is 0.13.

2.5.8.2.2.2 Factor composition

Whereas the first factor (34.08% constrained inertia (ci); 24.24% total inertia (ti)) is dominated by wines showing high values for the off-flavour descriptor family, which can be seen by its contributions (con) (con 0.79) and is slightly influenced by wines showing higher descriptor family frequencies for floral (con 0.04), yellow fruit (con 0.03) and fruity (con 0.03) descriptors, the second factor (19.37% ci; 13.78% ti) is determined by wines showing high ratings for floral (con 0.59) descriptors and those of the families sulfur (con 0.15), aging notes (con 0.09) and vegetal odours (con 0.06). The third factor (13.72% ci; 9.76% ti) is determined by wines showing aging notes (con 0.45), tropical fruit (con 0.10), citrus fruit (con 0.07), white fruit (con 0.06), off-flavour (con 0.06) and minerality odours (con 0.05).

2.5.8.2.2.3 Explanation of sensory descriptor families by factors

The sensory properties’ (descriptor family appearance) explanation by the factors can be observed regarding the cos2 values for each descriptor family and each factor. Therefore, the first factor explains the cited descriptor families off-flavour (cos2 0.942), fruity (cos2 0.510), yellow fruit (cos2 0.446), dried / candied fruit (cos2 0.287) or fermentative aroma (cos2 0.231). The descriptor families floral (cos2 0.797), sulfur (cos2 0.475), and vegetal (cos2 0.408) are explained by the second factor, whereas the third factor explains the descriptor families tropical fruit (cos2 0.686), aging notes (cos2 0.646), citrus fruit (cos2 0.305) and minerality (cos2 0.290). The A Not-A test derived recognition frequency RIESLING gets explained mostly by the second factor (cos2 0.442).

Concept of Riesling wines’ typicality

77

Fig. 2-8 Map of Canonical Correspondence Analysis (CCA) using PLS chosen aroma compounds’ concentrations (Variables (red)), wines (Sites (green)) and descriptor family frequencies (Objects (black)) on the German panels’ data (Riesling wines); not shown wines RIAl7 (-1.456; -0.208); RIAl8 (-1.179; -0.055) and descriptor family frequency off-flavour (-2.102; -0.237); BMT: benzenemethanethiol; DMS: dimethyl sulfide; 3SH: 3-sulfanylhexan-1-ol; TDN: 1,1,6-trimethyl-1,2-dihydronaphthalene; MLF: malo-lactic fermentation; wine samples are indicated in Table 2-7

RIMO2

RIAL3

RINA1

RIRH12

RIMO3

RIRH14

RIRH13

RIRH4

RIPF

RIRH5

RINZ

RIAUT1

RIRH1

RIRH2 RIRH11

RIRH6RIMO1

RIAUT2

RIRH9

RIRH3

RIRH8

vegetal

sauvignon

spicy

herbs

fruity

citrus fruit

white fruit

yellow fruit

tropical fruit

dried / candied fruit

aging notes

minerality fermentative aroma

sulphur

MLF

others RIESLING

trans‐linalool oxide

nerol oxide

3‐methyl‐1‐butanol cis‐linalool oxide

DMS

α‐terpineol

linalool

TDN

diethyl succinate

ethyl butanoate

3SH

vitispirane BMT

typicality

‐1,1

‐0,6

‐0,1

0,4

‐1,1 ‐0,6 ‐0,1 0,4 0,9

F2 (1

9,37

 %)

F1 (34,08 %)

CCA Map / Symmetric(axes F1 and F2: 53,44 %)

Sites Objects Variables

Concept of Riesling wines’ typicality

78

2.5.8.2.2.4 Interpretation analogous to ecological data – the sensorial niche

Vectors for the metric variables aroma compounds concentrations were graphically interpreted according to ter Braak and Verdonschot (1995) for ecological data. For Canonical Correspondance Analysis (CCA) the wines were defined as sites in an ecological point of view. In these sites (in this study represented by the wines) several descriptors were detected by several panel members, leading, when taking descriptors as species in ecological analogism, to a nominal data pattern for each wine sample, which is comparable to environmental data. In such a model, the metric variables aroma compound concentrations are seen as metric environmental variables. Most variance for each metric variable is observed along the vector with the mean represented in the origin of the plot. Vectors are forming bi-plots either with the wines or the descriptor families. Information on favouring conditions for the appearance of a sensory descriptor can be obtained by interpretation of the data by means of a CCA analogously to the ecological niche as a sensorial niche.

2.5.8.2.2.5 Interpretation of wine samples

Regarding the wines, the CCA map (Fig. 2-8) can be interpreted reflecting the wines on the typicality vector. It is observed that wines like RIRH8, RIRH6, RIMO1, RIRH13, RIRH9, RIRH11, RIRH2, RIRH3, RIAUT1, RIPF, RINZ, RIAL3, RIRH4, RIRH14, RIRH1, and RIRH5 showed higher typicality than RIRH12, RINA1, RIAUT2, RIAL8, RIAL7 or than RIMO3, RIMO2. In the CCA map, RIRH8, RIRH6 and RIMO1 were positioned with increasing 3-sulfanylhexan-1-ol or cis-linalool oxide concentrations, whereas RIRH13, RIRH9 with increasing 3-sulfanylhexan-1-ol concentrations only. The wines RIRH11, RIRH2, RIRH3 were grouped with ethyl butanoate and trans-linalool oxide concentrations. RIRH4 was grouped closer to the vector of nerol oxide concentrations. RIAUT1 grouped with 3-sulfanylhexan-1-ol and cis-linalool oxide. Despite approximately equal typicality levels for the wines RIPF, RIAUT2, RIRH1, RIRH5, RINZ, RIAL3 and RIRH14, they were all positioned differently. On the one hand for RINZ, RIAL3, RIRH5 and RIRH14, this could be related to increasing nerol oxide, TDN and vitispirane concentrations, on the other hand it could be related to increasing linalool and cis-linalool oxide concentrations for RIAUT2 and RIPF, whereas RIRH1 seemed to be impacted by benzenemethanthiol and dimethyl sulfide concentrations. The wines RIAL7, RIAL8, RIRH12 were grouped with diethyl succinate and 3-methyl-1-butanol, whereas RINA1 was more related to α-terpineol. RIMO3 and RIMO2

Concept of Riesling wines’ typicality

79

were positioned with high linalool concentrations showing both lower typicality values than the mean value.

2.5.8.2.2.6 Interpretation for descriptor families

Regarding the descriptor families, it can be extracted clearly from the CCA map (Fig. 2-8), that increasing typicality corresponds to increasing general fruity descriptor families such as citrus fruit, white fruit, and dried / candied fruit. Descriptor families like aging notes and tropical fruits were more likely to be present in wines showing higher typicality than the mean value of the data set, as fermentative aroma and yellow fruit did as well. The descriptor families herbs, others, vegetal, sulfur, and floral inversely were shown to be associated with lower typicality ratings under certain conditions. The interpretation of correlations between the descriptor families and single aroma compounds’ concentrations should be done attentionously. It can be considered, that certain concentration ranges of different, maybe multiple aroma compounds are beneficial to the perception of odours and therefore to the naming of a descriptor of a descriptor family and could create a sensorial niche. Following this approach, the CCA can provide some helpful information by including nominal data in contrast to a PCA, which itself is strictly limited to decoupling of linear relations of metric data. Vegetal descriptors have been cited with higher frequency in wines showing higher α-terpineol concentrations, whereas floral descriptors in wines were showing high linalool concentrations. Another example is the odour family of aging notes, which was associated with higher TDN and nerol oxide concentrations in the wines. The origin of the vectors of a CCA map (Fig. 2-8) is presenting the mean value of each metric variable and therefore certain information is also obtained on the quantity of a variable. So, if the descriptor families are correlated to higher aroma compounds’ concentration, the corresponding odours are more likely to appear at a concentration higher than the mean value of a metric variable for that data set, depending on its position in the CCA map.

2.5.8.2.2.7 Co-correlations of metric variables

Regarding the vectors for aroma compounds concentrations in relation to the typicality vector, it can be deduced that typicality is impacted by 3-sulfanylhexan-1-ol and ethyl butanoate, dimethylsulfide and trans-linalool oxide concentrations and is disfavoured by diethyl succinate, 3-methyl-1-butanol, α-terpineol and high linalool concentrations, whereas nerol oxide, benzenemethanethiol, TDN, vitispirane and cis-linalool oxide were positioned indifferently towards typicality.

Concept of Riesling wines’ typicality

80

2.5.9 Discussion on aroma compounds contribution to Riesling wines typicality

2.5.9.1 Univariate methodology When correlating typicality ratings and aroma compounds’ concentrations using the Pearson product-moment method, the concentrations of the monoterpenol oxides cis-linalool oxide and nerol oxide showed significant correlations in the German data set, containing Riesling wines and non-Riesling wines. In the data sets, from which non-Riesling wines were eliminated, a weak, but significant correlation was observed for 3-sulfanylhexan-1-ol, independently for both panels. The descriptor family of citrus fruit was shown to correlate with Riesling wines’ typicality for both panels (2.5.8.1). For the German panel, Pearson product-moment correlation showed a significant correlation between geraniol concentrations and the citrus fruit descriptor family frequency. For the French panel, ethyl butanoate and ethyl hexanoate concentrations did correlate with that descriptor family, but no common correlation was observed for both panels.

2.5.9.2 Multivariate methodology Like for the univariate approach of comparing results of two different panels, multivariate tools for pointing out aroma compounds being related to typicality ratings were applied on both panels’ total data sets, in order to select common aroma compounds. Some of these 15 selected aroma compounds, which were considered to impact Riesling wines’ typicality, were also observed to correlate within the univariate procedure. Among the other aroma compounds selected by using PLS regression, higher alcohols (2-methyl-1-butanol and 3-methyl-1-butanol), ethylic esters (ethyl butanoate, diethyl succinate), terpenols and terpenol oxides (linalool, α-terpineol and trans-linalool oxide), sulfur compounds (dimethyl sulfide, 4-methyl-4-sulfanylpentan-2-one, and benzenemethanethiol), and norisoprenoids (vitispiranes and TDN) were considered to be important for modelling the Riesling wines’ typicality values, upon their VIP values (Table 2-20). A major difficulty of aroma compounds concentrations correlation to sensory characteristic is the non-linear nature of flavour perception (Meilgaard et al., 2007) and the complexity of aroma compounds involved in a certain odour through synergistic effects (Campo et al., 2005). Therefore it was considered to be useful to apply a novel methodology to explore the nature of odour. A possible implication of the aroma compounds, which were selected upon showing high variation in aromatic expression of typical Riesling wines, was illustrated by applying

Concept of Riesling wines’ typicality

81

Canonical Correspondence Analysis (CCA) on the data set of the German panel containing only Riesling wines, with the aroma compounds selected by a comparative PLS regression procedure. It can be considered that such an approach could help to overcome the concerns regarding strong linear correlation of single aroma compounds’ concentration and the odour qualities’ intensity, which is evoked by some authors (Campo et al., 2005; Meilgaard et al., 2007; Benkwitz et al., 2012). Being interpreted analogously to the ecological niche approach used in ecological science (ter Braak and Verdonschot, 1995) it can be considered to be a sensorial niche. By all means, this approach resulted in similar results compared to univariate data analysis, especially concerning the contribution of 3-sulfanylhexan-1-ol to Riesling wines’ typicality. Neverless, it could not answer the question of the contribution of TDN and linalool concentrations towards perceived Riesling wines’ typicality, but allowed to picture the context. The descriptor family aging notes, which was strongly associated with increasing concentrations of TDN, showed no tendency to be detected at a higher typicality level, than the mean typicality value of the data set, whereas the descriptor family floral, which was strongly associated to increasing linalool concentrations, showed a tendency to be detected at lower typicality level than the mean typicality value. Both aroma compounds’ variations were detected in a concentration range above their sensory thresholds (Darriet et al., 2012; Sacks et al., 2012), but no direct link was observed to typicality.

Concept of Riesling wines’ typicality

82

2.5.9.3 Observations for selected aroma compounds’ concentrations regarding typicality Grouping wines regarding their typicality ratings into four groups of atypical (typicality score 0.0 - 0.2), less typical (0.2 – 0.4), typical (0.4 – 0.6) and very typical Riesling wines (>0.6) (Table 2-21) for the German panels’ total data set, it was observed that the mean concentration of 3-sulfanylhexan-1-ol was highest in the group of atypical wines, which comprised Sauvignon blanc style wines samples. But the concentration of 3-sulfanylhexan-1-ol increased with the level of typicality from the groups of less typical and typical Riesling wines to the group of very typical Riesling wines, indicating an impact of this compound for typical Riesling odour. Table 2-21 Mean concentrations and OAVs for typicality related aroma compounds for the German panels’ total data set after grouping into different levels of typicality (with relative typicality rating ranges: atypical 0 – 0.2; less typical 0.2 – 0.4; typical 0.4 – 0.6; very typical >0.6)

 mean  min  max  SD 

CV [%] 

SEM  n OAV 

  mean  SD atypical           typicality  0.11  0.09  0.15 0.03 25 0.01 6  ‐  ‐3SH [ng/L]  1160  697  1898 425 37 173 6  19.34  7.08linalool [µg/L]  1 nq  6 2 152 1 6  0.07  0.11TDN [µg/L]  0.5  0.0  1.1 0.4 83 0.2 6  0.26  0.21DMS [µg/L]  27.7  8.9  65.9 19.4 70 7.9 6  0.92  0.65less typical     typicality  0.39  0.21  0.49 0.09 23 0.03 7  ‐  ‐3SH [ng/L]  457  218  779 190 42 72 7  7.89  3.51linalool [µg/L]  41  0  147 46 113 17 7  2.57  2.83TDN [µg/L]  3.8  0.0  9.4 3.0 79 1.1 7  1.78  1.11DMS [µg/L]  11.4  7.2  19.2 4.5 39 1.7 7  0.39  0.15typical                   typicality  0.54  0.53  0.56 0.01 2 0.01 5  ‐  ‐3SH [ng/L]  643  247  1189 332 52 148 5  9.42  4.94linalool [µg/L]  36  15  68 20 55 9 5  1.63  0.96TDN [µg/L]  2.4  1.6  3.3 0.7 31 0.3 5  1.51  1.26DMS [µg/L]  11.3  8.5  17.6 3.4 30 1.5 5  0.38  0.12very typical     typicality  0.69  0.61  0.80 0.06 9 0.02 12  ‐  ‐3SH [ng/L]  739  253  1558 379 51 110 12  12.31  6.32linalool [µg/L]  33  8  60 14 43 4 12  1.65  0.71TDN [µg/L]  4.7  0.9  22.1 5.5 117 1.6 12  2.33  2.73DMS [µg/L]  13.2  5.3  22.5 5.0 38 1.5 12  0.44  0.173SH : 3‐sulfanylhexanol; TDN: 1,1,6‐trimethyl‐1,2‐dihydronaphthalene; DMS: dimethyl sulfide; nq: not quantifiable; SD : standard deviation; CV : coefficient of variation; SEM: standard error on means; n: sample number; OAV: odour activity value; OAV > 1 indicated in bold; OAVs based on thresholds: 3SH: 60 ng/La, linalool: 20 µg/La, TDN: 2 µg/Lb, DMS: 30 µg/Lc; aDarriet et al., 2012; bSacks et al., 2012; c Anocibar‐Beloqui, 1995  

Concept of Riesling wines’ typicality

83

For linalool, a very low mean concentration was observed within the group of atypical wines. The group of less typical wines showed a broad range of variation in linalool concentrations (n.q. (<1 µg/L) to 147 µg/L), whereas the mean concentrations increased with the level of typicality, when the outlier sample containing 147 µg/L linalool was removed. Interestingly, it was observed, that the inner group variation of the linalool concentration diminished with increasing typicality (Fig. 2-9). However, no significant relation was observed between perceived typicality and linalool concentrations.

Fig. 2-9 Boxplots of linalool concentrations for German panels’ total data set after grouping into different levels of typicality atypical (0.0-0.2), less typical (0.2-0.4), typical (0.4-0.6), very typical (>0.6); box plots show median, Q0.25 and Q0.75; Whiskers showing minimum and maximum respectively For TDN the mean concentration was lower in the group of atypical wines than in the other groups, whereas no clear pattern in relation to the level of typicality was observed. Taking the newly published detection threshold of 2 µg/L (Sacks et al., 2012) into account, which was confirmed during this project at 2.3 µg/L in model wine solution, and following the odour activity value (OAV) concept, the mean of OAV varied and showed lower OAV mean values for less typical wines (1.78) and typical wines (1.51) in comparison to very typical wines (2.33). Another interesting observation was made for the compound dimethyl sulfide, which slightly increased with the level of typicality and showed a mean concentration of 13 µg/L in the group of very typical wines. Dimethyl sulfide was reported to create synergistic effects for fruity odours at sub-threshold concentrations of 10 µg/L in mixtures, which contained fruity esters, C13-norisoprenoids, and linalool (Escudero et al., 2007; Pineau et al., 2007; Loscos et al., 2007). When computing ANOVA on the single groups for the single variables, no statistical significant difference was observed.

0

20

40

60

80

100

120

140

160

atypical less typical typical very typical

c (lina

lool) [µg

/L]

Concept of Riesling wines’ typicality

84

2.5.10 Essential outcomes regarding aroma compounds and Riesling wines’ typicality The data analysis using different tools, from simple linear correlation to multivariate approaches (PLS), led to the conclusion that several groups of aroma compounds impact Riesling wines’ aroma and that especially 3-sulfanylhexan-1-ol impacts the perception of dry Riesling wines’ typicality. In addition to that, cis-linalool oxide and nerol oxide were shown to correlate to Riesling wines’ typicality related descriptors. The multivariate methodology of applying PLS and subsequently a CCA, in order to extract important aroma compounds and link them to parametric data, as obtained from frequency based descriptive tasks in sensory science, gave interpretable and useful results. This approach was shown to be helpful for the interpretation of complex sensory data and could provide an useful tool in the characterisation of sensory spaces in the future. However, the role of the monoterpenol linalool and the C13-norisoprenoid TDN still remains unclear. Their concentrations showed variation above the sensory odour thresholds and were highly correlated to the descriptor families floral and aging notes.

Concept of Riesling wines’ typicality

85

2.5.11 Consideration of an alternative approach: taking into account dissimilarities

2.5.11.1 Differentiation into factors of similarity and factors of dissimilarity of a category

2.5.11.1.1 Principle

Following the distinction concerning the concept of typicality [french typicité] made by Trognon (see also Paragraph 1.6 and Trognon, 2005), each property of an object can be classified as either a factor of similarity or a factor of dissimilarity. In such a model, a super-category is defined as a group of objects with common properties. Therefore, these properties can be seen as super-categorical factors of similarity. Differing properties of these objects building a super-category can be seen as discriminating properties leading to sub-categories and therefore, they are considered to be super-category factors of dissimilarities. By definition, the sub-category forming group of objects has common properties again. These sub-categorical factors of similarity forming properties can comprise super-categorical factors of similarity as well as super-categorical factors of dissimilarity, which then do not differ in the new group formed by objects of the new sub-category. Differing properties of objects in this new formed sub-category again can be defined as sub-categorical factors of dissimilarity. These sub-categorical factors of dissimilarity, theirselves can comprise super-category factors of dissimilarity and super-categorical factors of similarity, which then differ in the group of objects forming the sub-category. The uncertainty of such an approach would be the question upon which extent a difference of a factor of similarity may differ within a category, until it becomes a factor of dissimilarity of a sub-category. Talking about categories, it also has to be considered, that categories can be formed differently. The question ‘What is a Riesling wine?’ can either be responded by perceived properties, such as sensory attributes in the answer ‘A Riesling wine has to taste/smell like ....’, or by substantial, natural attributes in the answer ‘A Riesling wine has to be made from grapes of the variety Vitis vinifera L. cv. Riesling’ representing either a perceived or a natural category. Perceived categories are build up on similarities with a prototype or idéo-type, an exemplary model, by sharing a maximum of perceived properties with it. Therefore, a measurement of perceived categorical typicality would correlate with the measurement of similarities with a prototype, whereas it would not correlate with the dissimilarities. For that reason it might not be an adequate tool to carve out the not similar properties of a product (the dissimilarities), which ‘make the difference’ between two samples of similar typicality.

Concept of Riesling wines’ typicality

86

One approach to achieve distinction among properties could be to differentiate them in between samples of a natural category into factors of similarity and factors of dissimilarity. By using simple statistics, significant differences of perceived properties on natural categories can be elucidated. The following combinatorial design should lead to differentiation of a natural categories’ property either into a factor of similarity or a factor of dissimilarity – in the case of this study, the sensory attributes of the natural category Riesling wine. The total data set for the total frequencies of the descriptor families of the objects wines, forming the super-category dry white wines, were analysed for significant differences (chi-square (χ2) proportions; α=0.05). Non-significant differences for total frequencies of a single descriptor family were considered to indicate a factor of similarity in the super-category dry white wine, which are underlying a gradient of representativeness in the category according to Rosch and Mervis (1975). Significant differences for descriptor families’ total frequencies in the data set were considered to be identified as a factor of dissimilarity of the super-category dry white wine. Subsequently, the sub-category Riesling wine was formed by removing non-Riesling wines from the data set and re-analysing for significant differences in the frequencies of single descriptor families. Descriptor families showing no significant differences in their frequencies in the Riesling wines’ data set, but have been considered to be factors of dry white wines’ dissimilarity, can be considered as factors of Riesling wines’ similarity, factors of other variety wines’ similarity or as factors of other variety wines’ dissimilarity. Affiliation to these groups of properties then was determined by either showing significant differences in the sub-category other variety wines, which would result in a factor of other variety wines’ dissimilarity or not. Being not identified as a factor of other variety wines’ dissimilarity, it leads subsequently to the classification to be either a factor of Riesling wines’ similarity or a factor of other varieties wines’ similarity, depending on higher absolute values for the frequencies in the data set. Furthermore, factors of dry white wines’ similarity theoretically could act as factors of sub categorical similarity (for Riesling or other variety wines) at the same time. Those ones considered to be factors of dry white wines’ dissimilarities showing significant differences for frequencies in the Riesling wines’ data, were considered to be factors of Riesling wines’ dissimilarity. Finally, this approach leads to patterns of significant results for each descriptor group, which then allows classifying each descriptor family as a factor of similarity or factor of dissimilarity of a certain group of objects.

Concept of Riesling wines’ typicality

87

Table 2-22 Differentiation of descriptor families into factors of similarity or factors of dissimilarity of the natural categories Riesling wines and non-Riesling wines for the German and the French sensory panel

Riesling wines’  factor of dissimilarity 

 Riesling wines’  

factor of similarity  

non‐Riesling wines’ factor of dissimilarity 

 non‐Riesling wines’ factor of similarity 

German  French German French German  French German French

off‐flavour  off‐flavour  

RIESLING  RIESLING  

others  ‐  

‐  others 

MLF  ‐  

‐  others  

‐  floral  

‐  sulfur 

sulfur  ‐  

‐  MLF   Ribes ssp./ 

Sambucus ssp.  ‐  

‐ fermentative 

aroma 

‐  minerality  

‐  sulfur  

sauvignon  sauvignon  

minerality  ‐ 

aging notes  aging notes   fermentative 

aroma fermentative 

aroma  

vegetal  vegetal  

‐  ‐ 

floral  floral  

minerality  ‐  

‐  ‐   dried/candied 

fruit ‐ 

‐ dried/candied 

fruit  

tropical fruit  tropical fruit  

‐  ‐  

‐  tropical fruit 

vegetal  ‐  

yellow fruit  yellow fruit  

‐  ‐  

‐  yellow fruit 

‐  ‐  

white fruit  white fruit  

‐  ‐  

‐  white fruit 

‐  ‐  

citrus fruit  citrus fruit  

‐  ‐  

citrus fruit  citrus fruit 

‐  ‐  

fruity  fruity  

‐  ‐  

  fruity 

‐  ‐  

herbs  ‐  

‐  ‐  

herbs  herbs 

‐  ‐  

‐  spicy  

‐  ‐  

spicy  spicy 

the same differentiation for descriptor family on both panels’ data sets are indicated in bold italic

Concept of Riesling wines’ typicality

88

2.5.11.1.2 Chi-Square (χ2) - descriptor family frequencies

The German panels’ total data set showed significant differences for 13 descriptor families regarding their total citation frequencies, whereas six descriptor families showed no differences. In the French equivalent, significant different total citation frequencies were observed for nine descriptor families, whereas for ten descriptor families no differences were detected. The groups showing significant differences represent the factors of dry white wines’ dissimilarity, whereas those not showing significant differences act as factors of dry white wines’ similarity. Subsequently, further χ2-statistics were carried out on selected data sets representing the natural categories Riesling wines and non-Riesling wines resulting in classifying the descriptor families as presented inTable 2-22.

2.5.11.1.3 Pearson product-moment correlation – total frequency of descriptor

families vs. aroma compounds’ concentrations

To obtain a more profound insight into factors of dissimilarity related aroma compounds, a Pearson correlation was computed (α = 0.05) on the sets of Riesling wines, regarding commonly by the German and the French panels’ data classifying factors of Riesling wines’ dissimilarity – the descriptor families aging notes and floral (Table 2-23). C13-norisoprenoids concentrations of vitispirane isomers and TDN did significantly correlate with the descriptor family aging notes as monoterpenol’s linalool concentration did with the descriptor family floral in both panels’ data sets.

2.5.12 Discussion on the Riesling wines’ typicality related descriptor families Comparison of categorical typicality – perceived category vs. natural category Both methods, direct correlation of typicality ratings and discriminant statistics on full and partial data sets in order to determine typicality related descriptor families, led to comparable results, as both approaches revealed the descriptor family of citrus fruit as a factor of similarity. Direct measurement of one categories’ typicality can be seen as direct measurement of matching properties towards a prototype, and therefore a property’s measurement showing high correlation to typicality, is seen as an important factor of similarity by the assessor. This dimension of implicit measurement does normally not impact an approach of distinguishing measured properties of a natural category in either factors of similarity or factors of dissimilarity by statistical characteristics. Whereas in the Pearson

Concept of Riesling wines’ typicality

89

product-moment correlation approach with directly measured typicality, only the descriptor family citrus fruit was shown to be correlated with Riesling wines’ typicality for both panels. Table 2-23 Significant Pearson product – moment correlations (α<0.05) between factors of Riesling wines’ dissimilarity related descriptor family frequencies and aroma compounds for both panels on Riesling wines

  German  French variable  ageing notes  ageing notes 

  r  p  correlation  r  p  correlation monoterpenoloxides        

trans-linalool oxide 0.558  0.006  positive  nc  ns  ‐ 

cis-linalool oxide 0.452  0.030  slightly positive  nc  ns  ‐ 

nerol oxide 0.577  0.004  positive  nc  ns  ‐ 

norisoprenoids            

vitispirane 0.648  0.001  positive  0.698  <0.001   positive 

TDN 0.768  <0.001  strongly positive 0.835  <0.001  strongly positive 

β-damascenone nc  ns  ‐  0.536  0.007  positive 

Sulfur compounds            

dimethyl sulfide (DMS)  0.450  0.031  slightly positive  nc  ns  ‐ 

variable floral  floral 

r  p  correlation  r  p  correlation monoterpenols            

linalool 0.826  <0.001  strongly positive 0.523  0.009  positive 

norisoprenoids             

α-ionone 0.479  0.021  slightly positive  nc  ns  ‐ 

β-ionone  nc  ns  ‐  0.477  0.018  slightly positive 

nc: no correlation; ns: not significant at a significance level α<0.05; Variables showing significant correlation for both panels are indicated in bold;  

By means of χ2-proportion significance testing in the alternative approach, this descriptor family citrus fruit was identified among others (e.g. RIESLING recognition frequency, fruity, yellow fruit, white fruit, tropical fruit, fermentative aroma) to be a factor of Riesling wines’ similarity. This was consistent for both panels’ data sets. The descriptor family off-flavour was shown to be a descriptor family relevant as a factor of Riesling wines’ dissimilarity, due to some Riesling wines showing odours sorted to this group (e.g. reductive, lactic), these odours were not present in wines from other varieties, but they were not considered to be linked to typicality. The recognition frequency RIESLING and the descriptor families fruity and yellow fruit, which were classified as being a factor of similarity by χ2-proportion statistics, showed significant correlation to direct measured typicality for the German panel also, but only did to a lesser extent for the French one. In conclusion then, both methods could be used for choosing typicality related descriptor families in the sense of being factors of similarity, though it has to be considered that the

Concept of Riesling wines’ typicality

90

information behind this is based on different categorical boundaries. Whereas both methods indicate the shared properties of a category, the method based on the natural category does not indicate any importance towards a single property, whereas the correlation with a direct measured perceived typicality implicitly acts as a rating of importance for a certain factor. Therefore the presence of the descriptor family citrus fruit, and to a lower extent fruity and yellow fruit in Riesling wines, can be considered to be more important for Riesling wines’ typicality than for example white fruit and tropical fruit are, due to being directly correlated to the measured perceived typicality. The descriptor families aging notes and floral were considered to be a factor of Riesling wines’ dissimilarity by both panels. However, their impact on perceived typicality still is not well understood.

2.5.13 Essential outcomes regarding an alternative approach to typicality Classification of descriptor families upon their differences in natural categories is shown to be a complementary tool to direct typicality rating and its correlation to descriptive sensory data. It does not give inverse conclusions on similarity related properties, whereas it gives additional information on the nature of a category’s variation, by pointing out factors of dissimilarity of a category. In this case, the approach allows discussing the consequences of variation in the concentrations of the aroma compounds linalool and TDN in the context of aromatic typicality of Riesling wines. The role of these aroma compounds, which can contribute to Riesling wines’ aromatic expression, can not be explained in the context of typicality. The factors of dissimilarity in a categorical sensory space are likely to be important for the position of a single object within a group of objects of a category, and therefore influencing the feature of ‘being unique’, which is an essential feature in the French notion of typicité (Salette, 1997; Dubourdieu, 2012). How exactly this impacts the measured perceived typicality is not well understood and needs some further attention.

Concept of Riesling wines’ typicality

91

2.6 Conclusion The concept of aromatic typicality was shown to be a very useful and reliable tool to characterise typical Riesling wines’ sensory profile by means of different sensory methods, carried out by two different wine experts’ panels (a German and a French one). Both panels were able to distinguish Riesling wines upon their typicality from other wines in a repeatable manner and the concept of Riesling wines’ typicality was shown to be similar for both groups of judges. Furthermore, the presented frequented methodology, the A Not-A test, showed to be a novel complementary tool for the measurement of a products’ typicality. For the descriptive task, both panels showed good repeatability for citation of descriptors from single descriptor families, whereas some differences were observed on semantic level between both panels. A Riesling wine’s typicality was shown to be significantly associated to citrus fruit descriptors for both panels, while for the German panel, it was also associated to fruity and yellow fruit descriptors. The concentrations of the polyfunctional varietal thiol 3-sulfanylhexan-1-ol were shown to increase with increasing ratings for Riesling wines’ typicality in both panels. Multivariate data analysis using a combination of PLS and CCA, led to interesting insights into the sensory and chemical space of Riesling wines typicality. It verified the results from univariate data analysis and provided an interesting view on the complexity of the sensory space of Riesling wines and implicated aroma compounds. The quite theoretical approach of dividing the wines’ sensory properties into factors of similarities and factors of dissimilarities led to the consideration that the highly variable concentrations of the aroma compounds linalool and TDN could be implicated in Riesling wines’ categorical typicality as a feature of being different.

92

Chapter 3

3 Viticultural studies

Viticultural studies

93

3.1 Introduction Typicality, in the sense of French typicité and terroir is highly correlated by its definition (see also Paragraph 1.6). Terroir is likely to be changed with climate change within its meteorological dimension and therefore this scenario is likely to affect viticulture (see also Paragraph 1.3 and 1.4). Influences through changes in these conditions were observed for various parameters in vines’ phenology and resulting fruit quality, whereas no direct impact on typicality influencing properties like aromatic expression and implicated chemical composition was studied. Most aroma compounds imparting in aromatic typicality of Riesling wines are considered to belong to the chemical families of monoterpenols, C13-norisoprenoids and thiols (Chapter 2). The aim of this part of the study is, to observe the influence of certain viticultural conditions, which are likely to change upon climatologic variation, e.g. drought, irrigation regime and canopy manipulation. Leaf removal could get nescessary in order to improve microclimate in the fruit zone, and therefore impacts of these viticultural variables on these families of aroma compounds were studied. Furthermore precursors’ generation in vines and grape berries and their transformation into volatiles in the winemaking process should be studied as well. Finally the finished wines were evaluated by sensory and instrumental analysis to link viticultural conditions to aromatic perception and to concentrations of impacting marker molecules for the aromatic typicality of Riesling wines. The experimental design was focused on establishing differences in vines’ water status in a vineyard showing low soil water holding capacity by means of drip irrigation. In addition to this, different leaf removal was conducted to diversify microclimatic conditions, which could become necessary in consequence to application of irrigation in viticulture.

Viticultural studies

94

3.2 Material and methods

3.2.1 Vineyard The experimental vineyard was located in the Rheingau Valley on the banks of the Rhine River in Germany and is owned and managed by the winery ‘Hessische Staatsweingüter Kloster Eberbach’. It is characterised as a steep slope vineyard (Table 3-1). The experiments have been conducted during the vintages 2008, 2009 and 2010.

Table 3-1 Vineyard characteristics according to Hessisches Landesamt für Umwelt und Geologie (http://weinbaustandort.hessen.de/viewer.html (29.05.2012)) for the chosen experimental plot: Rüdesheim (Germany) – Burgweg – Berg Schlossberg (Eiseninger) geographical characteristics geo-coordinates 49°58’29.05’’N; 7°53’00.44’’ altitude 120 m NHN surface area 0.71 ha planting characteristics

year of planting 1996 rootstock Börner variety Vitis vinifera L. cv. Riesling – estate own clone selection number of rows 51 row spacing 2.50 m vine spacing 1.10 m irrigation system drip irrigation soil cover permanent – whole surface permanent sward geological characteristics

soil hortic anthrosol on quartzite debris over quartzite weathered slate

carbonate carbonate free to poor carbonate content water holding capacity poor (nFK : 66.32 mm to 2 m max) slope 60 – 90% slope / south exposed drought stress risk class IV – >50% years below water stress threshold climatic characteristics

total direct sunlight radiation between April and October 177 – 181 [KJoule/cm2]

mean daylight temperature during Riesling ripening period 14.8 – 14.9 [°C]

Table 3-2 Viticultural treatments in the experimental vineyard

treatment description experimental

units n (vines per unit)

vintages

C control (non-irrigated, not defoliated) 3 (>20) 2008, 2009, 2010

I irrigated 3 (>20) 2008, 2009, 2010

IdB irrigated, defoliated at flowering and redefoliated at veraison 3 (>20) 2009, 2010

IdV irrigated, defoliated at veraison 3 (>20) 2008, 2009

Viticultural studies

95

3.2.2 Viticultural treatments The experimental setup was realised by the staff of the Department of Viticulture of the Geisenheim Research Centre. In total four different treatments, containing each three experimental replicates, were conducted during three vintages. Generally irrigation was applied during growing season between May and October when pre-dawn leaf water potential (Ψ(PD)) dropped below the irrigation threshold of Ψ(PD) < 0.3 mPa (Gruber and Schultz, 2005). Defoliation was carried out on the eastern side of the canopy by manually removing two primary leafs and secondary leafs in the fruit zone.

Fig. 3-1 Experimental vineyard located in Rüdesheim (Germany) – Berg Schlossberg. Viticultural treatments were carried out in the same design for the years 2008, 2009 and 2010; 1,2,3 indicate experimental units in single treatments

3.2.3 Viticultural measurements

3.2.3.1 Grape vines’ water status Grape vines’ water status was determined using three different methods:

i) pre-dawn leaf water potential Ψ(PD) measurements ii) soil water measurement and application of the Fractionated Transpirable Soil

Water (FTSW) model iii) δ 12C/13C isotope ratio measurement in grape juice

3.2.3.1.1 Pre-dawn leaf water potential Ψ(PD)

Pre-dawn leaf water potential (Ψ(PD)) was measured prior to sunrise with a portable pressure chamber (Soil moisture corporation, Santa Barbara, CA, California) according to Scholander et al. (1965). Leaves used for measurements (n=8) were covered with plastic bags shortly before excision.

Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

quantity of vines 46 45 45 44 44 44 44 43 41 42 42 43 42 41 41 41 41 40 42 43 40 40 40 40 39 39 39 38 38 38 38 37 36 36 36 36 35 34 34 33 33 33 33 32 31 31 31 31 31 30 30

C: non-irrigated controlI: irrigated treatmentIdB: irrigated treatment, defoliated at floweringIdV: irrigated treatment, defoliated at veraison

CIdB

IdV

IdV

IdB

I IC

N

2

1

2

1

2

1

2

1

3 333

Viticultural studies

96

3.2.3.1.2 Fraction of transpirable soil water (FTSW) derived pre-dawn leaf

water potential Ψ(PD)

The soil water content was measured using capacitance sensors (Diviner 2000, Sentek, Adelaide, SA, Australia) according to Gruber and Schultz (2005) and Ψ(PD) was calculated according to Hofmann and Schultz (2008).

3.2.3.1.3 Carbon isotope ratio – δ13C

The samples were prepared according to van Leeuwen et al (2010) and analysed at the James Hutton Institute, Invergowrie, DD2 5DA, United Kingdom.

3.2.4 Temperature - calculation of growing degree days (GDD) The calculation of GDD days was done according to Wiliams et al. (1985) to a base temperature of 10 °C and a maximum temperature of 30 °C.

3.2.5 Canopy micro-climate - ‘Point-quadrat’ measures For the description of canopy’s fruit zone micro climate prior and after defoliation at veraison and at the end of the growing season ‘point-quadrat’ measurements were conducted using the transects technique according to Smart and Robinson (1996).

3.2.6 Samples

3.2.6.1 Berry samples for maturity measures For maturity measures in each experimental unit, 100 berries were sampled randomly (50 berries from west exposed and 50 berries from east exposed cluster) on several dates during the ripening period to determine berry weight. Then the berries got pressed in a small scale press under equal conditions to obtain grape juice for analysis.

3.2.6.2 Berry samples for berry extracts For each experimental unit 200 berries were randomly sampled from bunches throughout the fruit zone (100 berries from east exposed and 100 berries from west exposed cluster) on different dates during the ripening period. The berries were cut at the pedicel without damaging the berry skin, transferred into polyethylene plastic containers with a capacity for approximately 50 berries and covered with carbon dioxide before getting sealed. The sealed containers were immediately dumped into liquid nitrogen in order to shock freeze the berries without direct contact to liquid nitrogen. These samples then were stored at -18 °C prior to berry extract preparation.

Viticultural studies

97

3.2.6.3 Grape must Grape must was sampled after harvested grapes were pressed and racked under cellar conditions prior fermentation and kept frozen at -18 °C until being analysed.

3.2.7 Berry extract preparation To avoid oxidative processes berries were defrosted under a protective gas (nitrogen) and K2S2O5 was added, resulting in a concentration of approximately 100 mg of total SO2 per L berry extract. The berries were minced in a juice mixer for 10 s at the highest speed and then centrifuged (10,000 rpm, 18100 x g, 15 min, 4 °C). The supernatant was aliquoted and stored at -18 °C prior to analysis.

3.2.8 Chemical analysis

3.2.8.1 Total soluble solids Total soluble solids (TSS) was determined in obtained grape juice using a handheld refractometer (MHRW – 190 ATC; Müller Optronic; Erfurt; Germany) graded in °Oe and transformed into °Brix according to Jakob (2012).

3.2.8.2 Reducing sugars The reducing sugars were calculated in g/L from °Oe values according to Troost (1980).

3.2.8.3 pH The pH was determined potentiometrically with a combined electrode using a pH-meter (inoLab® 7110 Serie, WTW GmbH, Weilheim, Germany).

3.2.8.4 Total acidity The total acidity (TA) was determined by titration of 50 mL grape juice with 0.1 M NaOH to pH 8.2 and calculated as tartaric acid equivalents.

3.2.8.5 Primary bound α-amino nitrogen as o-phthaldialdehyde - N-actetyl-L-cysteine derivates (N-OPA) The N-OPA concentrations were measured according to Dukes and Butzke (1998) using a spectrophotometer and calculated as iso-leucine equivalents.

Viticultural studies

98

3.2.8.5.1 Buffer solution

3.837 g NaOH and 8.468 g of ortho-boric acid were dissolved in H2O and adjusted to one liter in a 1 L volumetric flask.

3.2.8.5.2 N-actetyl-L-cysteine (NAC) solution

For 50 samples 122.5 mg NAC were dissolved in 5 mL H2O.

3.2.8.5.3 o-phthaldialdehyde (OPA) solution

For 50 samples 100.8 mg OPA were dissolved in 5 mL ethanol.

3.2.8.5.4 Sample preparation and analysis

The grape juice samples were prepared as indicated in Table 3-3 in disposable UV/VIS cuvettes. Absorbance was measured at a wavelength of λ 340 nm and an optical pathlength of 1 cm using a spectrophotometer (UV/VIS Spectrometer Lambda 2, PerkinElmer, Rodgau, Germany). Table 3-3 Preparation scheme for N-OPA analysis

solution volume [mL]

blind sample

buffer solution (3.2.8.5.1) 2.9 2.9

NAC solution (3.2.8.5.2) 0.1 0.1

H2O 0.05 -

sample - 0.05

1) 2 min after mixing, measurement of absorbance 1 (A1) 2) immediately after measurement, addition of OPA-Solution

OPA-solution (3.2.7.8.3) 0.1 0.1

mixing and measurement of absorbance 2 (A2) after 15 min

3.2.8.6 Analysis by means of Fourier transform infrared spectroscopy (FTIR) A range of parameters in berry extracts, grape juice (must) and wines were measured by means of a FTIR-inferometer using an OenoFossTM (Foss, Hilleroed, Denmark). Parameter included, were total soluble solids as BRIX [°Brix], reducing sugars [g/L], pH, total acidity [g tartaric acid/L] for berry extracts and grape juice (must) and for the wines ethanol [%vol], reducing sugars [g/L] and total acidity [g tartaric acid/L]. Samples were analysed at the

Viticultural studies

99

Department of Wine Chemistry and Beverage Research at the Geisenheim Research Center.

3.2.8.7 Total nitrogen concentration after KJELDAHL - decomposition detected as ammonia Chemical decomposition of organic bound nitrogen is based on the standard procedure of KJELDAHL (Schaller, 2000). 5 mL of berry extracts or grape juice (must) were used as sample size. To the sample 5 mL of the incineration mixture, comprising 0.48 g Se (black grounded), 14.0 g Li2SO4 (suprapure), 420 mL H2SO4 (conc.), and 330 mL H2O2 (p.a.) were added and this mixture was refluxed until being cleared (180 °C; approximately 3 hours). After cooling down the reaction mixture was then transferred quantitatively through a filter into a 50 mL volumetric flask and was adjusted with H2O to 50 mL. This incineration solution was analysed by means of flow injection analysis using a Foss Tecator FIAstar 5000 Analyser (FOSS, Hilleroed, Denmark) according to Schaller (2000). Measurements were carried out at the Department of Soil Sciences and Plant Nutrition at the Geisenheim Research Center.

3.2.8.8 Free ammonia in grape juice and grape must Free ammonia [NH4]+ was measured through flow injection analysis using a Foss Tecator FIAstar 5000 Analyser (FOSS) according to Schaller (2000). The injection volume for grape juice was 40 µL, H2O acted as carrier solvent, 3 M NaOH acted as pH modifier. The indicator solution was prepared according to the suppliers’ manual (FOSS). Wavelengths of λ 590 nm and 720 nm were selected. The standard curves were generated using ammonium chloride solutions. Measurements were carried out at the Department of Soil Sciences and Plant Nutrition at the Geisenheim Research Center.

3.2.8.9 Reduced glutathione The reduced Glutathion was analysed according to Lavigne et al. (2007) as a monobromobimane (MBB) thiol conjugate. Modifications were applied concerning the stabilisation of the MBB-thiol conjugate and the principle of separation and detection (Pons, 2009; unpublished). For stabilisation purposes 40 µL of 70% of trifluoroacetic acid (TFA) was added to the reaction mixture after the derivatisation procedure and left at least for one hour for equilibration. Then the derivatised samples were analysed by means of a C18-RP-HPLC-FLD system (Dionex – Fisher Scientific, Illkirch-Cedex, France) using chromatographic conditions presented in Table 3-4.

Viticultural studies

100

Table 3-4 Chromatographic conditions for analysis of reduced glutathione

Auto sampler Injection Volume 5 µL Flow 0.8 mL/min Gradient Isocratic ; 100% Solvent A Detector Dionex FLD – 3400 RS Excitation λ Em 1: 390 nm; Em 2: 260 nm Emission λ Em 1: 482 nm; Em 2: 400 nm Sensitivity Em 1: 2; Em 2: 3 Flow Cell Temperature 40 °C Column Supelco Inc; LiChrospher SI-60; 5 µm; 250 mm x 4.6 mm Temperature 25 °C Solvent A methanol / H2O (10/90; v/v); (2.5% glacial acetic acid)

3.2.8.10 Phenolic compounds The samples were membrane filtered (0.45 µm) and were analysed by means of a C18-RP-HPLC-PDA system (Dionex - Thermo Scientific, Fisher Scientific GmbH, Schwerte, Germany) under chromatographic conditions shown in Table 3-5 and Table 3-6. The wavelengths for the identification of catechins and hydroxybenzoic acid compounds were λ1: 280 nm, for hydroxycinnamic acids λ2: 320 nm and for flavonol glycosides λ3: 360 nm. Concentrations of caftaric acid and glutathionylcaftaric acid were calculated as caffeic acid equivalents. The analysis was carried out at the Department of Wine Chemistry and Beverage Research at the Geisenheim Research Center.

Table 3-5 Chromatographic conditions for analysis of phenolic compounds

Injection Volume 2 µL

Pump Dionex LPG 3000 HPLC Pump Flow 0,19 mL/min Gradient see gradient Table 3-6 Detector Thermo Scientific Finnigan Surveyor PDA Plus Wavelength λ λ1 : 280 nm; λ2 : 320 nm; λ3 : 360 nm Column NEOS Comp. Ltd.; Fluofix 120E; 5 µm; 2x125 mm

(NEOS Company Ltd., Kobe, Japan) Pre-Column NEOS Comp. Ltd.; Fluofix 120E, 5 µm; 10x4 mm Temperature 20 °C

Solvent A H2O / 85% o-phosphoric acid (995/5; v/v) Solvent B Acetonitril / H2O / 85% o-phosphoric acid (500/495/5; v/v/v)

Table 3-6 Solvent gradient programme phenolic compound analysis

runtime t [min] solvent A [%] solvent B [%] 0 100 0

2.5 100 0 22.5 75 25 32.5 25 75 35 0 100

35.1 100 0 45 100 0

Viticultural studies

101

3.2.8.11 Cysteinylated and glutathionylated precurser of 3-sulfanylhexan-1-ol The samples were purified by means of reversed phase solid phase extraction and analysed by means of C18-RP-UHPLC-HRMS system (Thermo Scientific, Illkirch-Cedex, France) using chromatographic conditions shown in Table 3-7 and Table 3-8. Solid phase extraction SPE Cartridges (LC-18 500 mg 6 mL, Supelco France, Saint Germain-en-Laye Cedex, France) were preconditioned by rinsing with 5 mL methanol, 4 mL H2O / methanol (50/50; v/v) and twice with 2 mL of H2O. 1 mL of grape juice, 1 mL of H2O and 25 µL of the internal standard solution containing the deuterated glutathionylated precursor of 3-sulfanylhexan-1-ol (3-S-(hexan-1-ol)-glutathione-d3) was added, resulting in a concentration of 200 µg/L. This solution was transferred to the reservoir of the SPE cartridge. After the adsorbing step, the SPE packing was washed twice using 1 mL of H2O and then got dried for two minutes by vacuum augmentation (180 MPa). Precursor molecules got eluted with 3 mL of H2O / methanol (70/30; v/v) and were collected into glass tubes. The eluate was evaporated to dryness by means of a RapidVap Vertex Dry Evaporators (Labconco, Kansas City, MO, USA) (speed 60 - 75%; temperature 50 °C; pressure 4 - 10 mbar). The residues got dissolved again by adding succesively 400 µL and 300 µL H2O (0.1% formic acid), resulting in a volume of 700 µL. The solutions were filtered using a membrane filter (0.45 µm) prior injection to HPLC. Table 3-7 Chromatographic conditions for the analysis of cysteinylated and glutathionylated precursor of 3-sulfanylhexan-1-ol

Auto sampler Thermo Scientific Autosampler Injection Volume 5 µL Pump Thermo Scientific Accela Pump Flow 0.6 mL/min Gradient see gradient Table 3.8 Detector Exactive Plus Orbitrap LC-MS Settings MS R: 25000; positive mode Heater temperature 302 °C Capillary voltage 35.5 V Tubelens Voltage 85 V Skimmer voltage 18 V Column Agilent Technologies; Zorbax Eclipse AAA; 5 µm; 4.6 x 150 mm Pre-Column Phenomenex; Seurity Guard; AJO-4492 Temperature 20 °C Solvent A H2O (0.1% formic acid) Solvent B Acetonitrile (0.1% formic acid)

Viticultural studies

102

Table 3-8 Solvent gradient programme analysis of cysteinylated and glutathionylated precursor of 3-sulfanylhexan-1-ol

runtime t [min] solvent A [%] solvent B [%] 0 97 3

1.2 97 3 3.5 87 13 4 86 14

4.4 80 20 4.7 2 98 5.7 2 98 5.8 97 3 6.8 97 3

3.2.9 Analysis of aroma compounds For analysis of the aroma compounds in the finished wines, the methods from Chapter 2 were applied.

3.2.10 Harvest and wine making Harvest The harvest date was decided by the technical staff of the vineyard’s estate. All replicates of the viticultural treatment were harvested manually, keeping separated each viticulture treatment. The yield was determined before pressing. Due to technological reasons replicates of each viticultural treatment were pressed together under comparable conditions (press loading, pressing programme, whole cluster pressing). K2S2O5 was added to free run juice (50 mg/kg pressed grapes). The pressed juice was settled for two days and the supernatant was racked prior to fermentation. Vinification Racked grape musts were measured for reducing sugars, and the musts from the viticultural treatments showing less sugar concentration were adjusted by adding saccharose, to obtain wines of the same ethanol concentrations after fermentation. The vinification was realised in three fold replicated micro scale fermentations for each viticultural treatment and each used commercial yeast strain. 500 g of the grape juice was aliquoted into 750 mL green glass bottles by means of a silicon tube and weighted on a balance to obtain net weight. After rehydrating the yeast in water, the grape juice was inoculated with Saccharomyces cerevisiae yeast strains, either Oenoferm FreddoTM or Laffort X5TM, according to suppliers’ recommendation. These inoculated bottles were closed with airlocks and incubated at 17 °C. The fermentation induced loss of carbon dioxide (CO2) was monitored daily by weighing. Two weeks after the fermentation, wines

Viticultural studies

103

were racked into CO2-filled bottles and K2S2O5 was added corresponding to 50 mg SO2/L. The bottles were sealed by using screw caps and then stored at 15 °C. One set of samples was stored at -18 °C to conserve the status of a young wine and to become analysed later during the project. For the vintages 2009 and 2010 the same procedure was carried out with an increased fermentation volume (1200 g grape must in 1.5 L bottles).

3.2.11 Sensory analysis The sensory analysis was carried out on the finished wines 12 months after bottling for the 2009 vintage. The sensory testing took place in an air-conditioned sensory room (20 °C) equipped with single booths. The wines were presented in ISO standardised glasses and cooled to 10 – 14 °C prior tasting. On each indicated date the sensory analysis took place at 11:30 CET. Triangle test Twelve panellists, experienced in wine tasting, were asked to point out the different sample by orthonasal testing. The wines were distributed randomly for each panellist. Quantitative descriptive analysis Twelve experienced panellists were asked to rate the presented wines for the descriptors apple, lemon, grapefruit, melon, peach, apricot, floral, petrol/kerosene, mineral, oxidative and reductive aroma on an unscaled line anchored with 0 and 9 at the left and right end respectively. Standards were provided before the sensory testing session, to adjust panellists’ odour fitting to odour perception. Prepared standard samples were given in Table 3-9.

Table 3-9 Preparation of standard samples for descriptors for QDA

descriptor preparation

apple a piece of fresh cut apple was placed into a neutral wine the evening before sensory session

lemon a piece of fresh cut lemon was placed into a neutral wine the evening before sensory session

grapefruit wine spiked with an equivalent of 1 µg/L 3SH

melon wine spiked with 200 µL of kitchen melon aroma (www.aromazone.com)

peach wine spiked with 200 µL of kitchen peach aroma (www.aromazone.com)

apricot wine spiked with 200 µL of kitchen apricot aroma (www.aromazone.com)

floral wine spiked with an equivalent of 250 µg/L linalool

petrol/kerosene wine spiked with an equivalent of 10 µg/L TDN

mineral a piece of schist was placed into a neutral wine the evening before sensory session (~ 0,5 cm*2 cm*3 cm / 200 mL)

oxidative wine was saturated with oxygen by passing through compressed air two days before sensory analysis

reduced neutral white wine, with reductive off-flavour

Viticultural studies

104

Statistical analysis i) Triangle tests’ statistical significance was achieved at 8 (p = 0.05) and 9

(p = 0.01) correct answers for the panel of 12 judges (8 male; 4 female) (Lawless and Heymann (2010).

ii) QDA - one factorial ANOVA was carried out using Sigma PlotTM (Systat Software GmbH, Erkrath, Germany) with viticultural treatments as factor.

Viticultural studies

105

3.3 Results and Discussion

3.3.1 General climatic conditions

3.3.1.1 Temperature and precipitation From a meteorological point of view the vintages 2008, 2009 and 2010 showed very different conditions during each growing season. This can be shown with the measure of growing degree days (GDD), which is a measure for temperature favouring photosynthetic activity of a plant (Williams et al., 1985) (Fig. 3-2). The vintage 2008 (1724 GDD) showed relatively cool weather during March and April and high temperatures from May to August, contributing highly to total GDD during that vintage. Therefore early phenological stages as inflorescence, flowering and fruit development were advanced. Declining temperatures during September favoured a long ripening period and an early harvest date in October. For the vintage 2009 (1988 GDD) early phenological stages of the growing season were positively influenced by mild temperatures from April to June. Hot weather in July and August and relatively high temperatures in September and October led to an extremely long ripening period from early August to mid-October. With relatively warm temperatures in March and April early phenological stages were promoted in the vintage 2010 (1694 GDD). Due to cold weather in May and early June during flowering there was no good fertilisation to develop fruits. High temperatures during June and July favoured early veraison. The ripening of the berries was slowed down by relatively cool temperatures during August and September. Due to high pest pressure, the grapes were harvested at the beginning of October. Although the total precipitation quantities during the observed growing period were relatively stable, the precipitation pattern during the growing season of the three vintages showed high variability. Showing a rainfall of 458 mm in 2008, 408 mm in 2009 and 508 mm in 2010 during the growing season, all the quantities were higher than a 30-year-long-period-mean of 364 mm for the corresponding period (Fig. 3-3) (Source: Deutscher Wetterdienst – Außenstelle Geisenheim). For 2008 more rainfall occurred in March and July compared to the long-period-mean. Less rain in relation to that indicator was observed in May, September and October. In 2009 a higher rainfall value towards the long term comparison was only measured in June, whereas low precipitation levels were observed in September and October. In the month of May, July and August of the vintage 2010 high total rainfall is reflected in high precipitation amounts. Only April showed extremely low rain quantity in relation to the long term mean value.

Viticultural studies

106

Fig. 3-2 Monthly (bars) and accumulated growing degree days (lines) of the three single seasons 2008, 2009 and 2010

Fig. 3-3 Monthly rainfall pattern (bars) and accumulating rainfall (lines) of the three single seasons 2008, 2009 and 2010

3.3.1.2 Water status in three years For irrigation scheduling, pre-dawn leaf water potential Ψ(PD) was measured, because it reflects the instantaneous plants’ water status (van Leeuwen, 2010). A combination of soil moisture derived pre-dawn leaf water potential Ψ(PD) values (Lebon et al., 2003; Gruber and Schultz, 2005; Hofmann and Schultz, 2008) and an integrated

March

AprilMay

June

July

August

Septem

ber

October

GD

D

0

100

200

300

400

500

accu

mul

ated

GD

D

0

500

1000

1500

20002008 monthly GDD2009 monthly GDD2010 monthy GDDaccumulated GDDaccumulated GDDaccumulated GDD

March

AprilMay

June

July

August

Septem

ber

October

rain

fall

[mm

]

0

20

40

60

80

100

120

140

accu

mul

ated

rain

fall

[mm

]

0

100

200

300

400

500

2008 monthly rainfall2009 monthly rainfall2010 monthly rainfallaccumulated rainfall 2008accumulated rainfall 2009accumulated rainfall 2010

Viticultural studies

107

measure of δ 13C (van Leeuwen et al., 2010) was applied to compare vintages and viticultural treatments and determine their water status level over the ripening periods.

3.3.1.2.1 Pre-dawn leaf water potential Ψ(PD)

For the vintages 2008 and 2009 significant differences in plants’ pre-dawn leaf water potentials could be established during the ripening period by drip irrigation. Due to excessive rainfall from July to October in 2010 no water deficit was observed, therefore irrigation was not applied.

3.3.1.2.2 Fractionated Transpirable Soil Water (FTSW) derived Ψ(PD)

Comparing the non-irrigated control treatments of the concerned vintages, different trends in vines’ water status during the growing periods were observed. In 2008 the pre-dawn leaf water potential (Ψ(PD)) dropped below the irrigation threshold (Ψ(PD-IT) = - 0.3 MPa) early as in the mid of June until late July. Strong rainfall favoured a backfill of the soil water reservoir and therefore water availability. The vines Ψ(PD) dropped below irrigation threshold again at late August and were kept stable for a month before dropping to the growing seasons’ minimum of Ψ(PD) = - 0.8 MPa in late September.

Fig. 3-4 Vintage 2008: pre-dawn leaf water potential Ψ(PD) (n = 8) trend of the non-irrigated control treatment (C) in relation to daily mean temperature and rainfall In 2009 the vines’ Ψ(PD) dropped to - 0.3 MPa (at flowering) from mid May to early June. After water status’ recovery until the mid of June (Ψ(PD) > - 0.2 MPa) (flowering to fruit set) Ψ(PD) declined successively until the end of August to Ψ(PD) = - 0.7 MPa before rising above the

Vintage 2008

julian day [d]

100 150 200 250 300

rain

fall

[mm

]

0

10

20

30

40

50

60

tem

pera

ture

[°C

]

0

5

10

15

20

25

30

Ψ(P

D) [

MPa

]

-1,4

-1,2

-1,0

-0,8

-0,6

-0,4

-0,2

0,0

rainfall [mm] mean temperature [°C]pre-dawn leaf water potential Ψ(PD)

Viticultural studies

108

Fig. 3-5 Vintage 2009: pre-dawn leaf water potential Ψ(PD) (n = 8) trend of non-irrigated control treatment (C) in relation to daily temperature and rainfall

irrigation threshold (Ψ(PD) > - 0.3 MPa) again for a short period due to rainfall. From then Ψ(PD)

dropped to the seasons’ minimum of Ψ(PD) = - 1.1 MPa in late September and recovered towards the end of the growing season to Ψ(PD) = - 0.5 MPa.

Fig. 3-6 Vintage 2010: pre-dawn leaf water potential Ψ(PD) (n = 8) trend of non-irrigated control treatment (C) in relation to daily temperature and rainfall

Vintage 2009

julian day [d]

100 150 200 250 300

rain

fall

[mm

]

0

10

20

30

40

50

60

tem

pera

ture

[°C

]

0

5

10

15

20

25

30

Ψ (P

D) [

MPa

]

-1,4

-1,2

-1,0

-0,8

-0,6

-0,4

-0,2

0,0

rainfall [mm]mean temperature [°C]pre-dawn leaf water potential Ψ(PD)

Vintage 2010

julian day [d]

100 150 200 250

rain

fall

[mm

]

0

10

20

30

40

50

60

tem

pera

ture

[°C

]

0

10

20

30

40Ψ

(PD

) [M

Pa]

-1,4

-1,2

-1,0

-0,8

-0,6

-0,4

-0,2

0,0

rainfall [mm]mean temperature [°C]pre-dawn leaf water potential Ψ(PD)

Viticultural studies

109

In 2010 vines Ψ(PD) dropped to - 0.3 MPa (at inflorescence and flowering) from late May to the mid of July. Due to excessive rainfall from August to September Ψ(PD) was kept above Ψ(PD) = - 0.2 MPa until late September, when it dropped to Ψ(PD) = - 0.3 MPa towards early in October. Different water potential trends in the three vintages led to different vines’ water status classification concerning vines’ phenological stages as listed in Table 3-10. Table 3-10 Vines’ water status classification at different phenological stages for non-irrigated treatments (Control), based on FTSW derived pre-dawn leaf water potential Ψ(PD) values (according to van Leeuwen et al., 2010)

phenological stage

vine water deficit level

vintage 2008 vintage 2009 vintage 2010

inflorescence / flowering no water deficit weak water deficit no water deficit

fruit development weak water deficit no water deficit weak water deficit

veraison no water deficit moderate water deficit no water deficit

berry ripening (beginning / end)

moderate to severe water deficit / moderate

water deficit

moderate to severe water deficit / severe water

deficit

no water deficit / weak water deficit

harvest moderate water deficit moderate water deficit weak water deficit

3.3.1.2.3 δ 13C – isotope Ratio

The carbon isotope ratio 12C/13C (called δ 13C) can function as an integrated indicator for the average vine water status during grape ripening, when measured on grape sugar at ripeness (van Leeuwen et al., 2010), due to the effect of isotope discrimination of 13C during the photo assimilation of CO2 and its reduction as a consequence of plants’ water deficiency (Farquhar et al., 1989). The grape musts showed differences in δ 13C values regarding different irrigation treatments as well as different vintages (Fig.3-7). Classifying the vines’ water deficit during the ripening period regarding δ 13C values for 2008 according to van Leeuwen et al. (2010) weak to moderate water deficit was observed for the non-irrigated control (C) (δ 13C = -23.370) and only weak water deficit was observed for the irrigated treatment (I) (δ 13C = -25.220). For 2009 the non-irrigated control (C) showed moderate to severe water deficit (δ 13C = -22.620), whereas for the irrigated treatments (I, IdB, IdV) weak to moderate water deficit was detected (δ 13C = -23.160, δ 13C = -23.370, δ 13C = -23.240). For 2010 no water deficit was observed and therefore no differences

Viticultural studies

110

regarding the vines’ water status were established (C: δ 13C = -26.340, I: δ 13C = -26.410, IdB: δ 13C = -26.320).

Fig. 3-7 δ 13C values for grape musts from viticultural treatments of the three vintages 2008, 2009, 2010 for the grape variety Riesling. Musts’ δ 13C values indicating different water deficit levels during ripening period for vintages and viticultural treatments; with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison; error bars indicate standard deviation (n=3)

The comparison between the three vintages regarding the vines’ water status during the ripening period and its influence on the physiological or biochemical parameters should therefore only be done between the non-irrigated control treatment (C) 2008 (δ 13C = -23.370) and the irrigated treatment (I) of 2009 (δ 13C = -23.160) and between the irrigated treatment (I) of 2008 (δ 13C = -25.220) and the non-irrigated control treatment (C) of 2010 (δ 13C = -26.340).

δ 13

C *(

-1)

20

21

22

23

24

25

26

27

viticultural treatments

C I IdB IdV

no water deficit weak water deficit moderate water deficit severe water deficit 2008 2009 2010

Viticultural studies

111

3.3.1.3 Canopy micro-climate - effects of defoliation Defoliation effects in the fruit zone of the canopy express through an elevated percentage of sun exposed grapes. For the vintages 2009 and 2010 the percentage of sun exposed grape berry clusters are shown for different viticultural treatments at different dates.

Fig. 3-8 ‘Box Plots’ for the percentage of grape cluster sun exposure (n = 6) in vintage 2009 in Riesling vines (with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison)

Fig. 3-9 Box Plots for percentage of cluster sun exposure (n = 6) in vintage 2010 in Riesling vines (with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison)

20

40

60

80

100

C I IdB IdV

expo

sed cluster[%]

pre‐veraison

20

40

60

80

100

C I IdB IdV

expo

sed cluster[%]

post‐veraison

20

40

60

80

100

C I IdB IdV

expo

sed cluster[%]

harvest

0

20

40

60

80

100

C I IdB

expo

sed cluster [%]

pre‐veraison

0

20

40

60

80

100

C I IdB

expo

sed cluster [%]

post‐veraison

0

20

40

60

80

100

C I IdB

expo

sed cluster [%]

harvest

Viticultural studies

112

3.3.1.4 Infection by Botrytis cinerea at harvest The infection of Botrytis cinerea showed differences in vintages 2009 and 2010, whereas in 2008 botrytis infection did not affect grape quality (< 5% for all viticultural treatments). In 2009 no significant differences were observed between the three irrigated treatments, though the non-irrigated control (C) showed a tendency to lower infection. For 2010 the irrigated treatment (I) showed significant higher infection, than the non-irrigated control (C) and the defoliated (IdB) treatments (Table 3.11). Table 3-11 Infection of the grape cluster by Botrytis cinerea at harvest for the vintages 2008, 2009 and 2010 of the different viticultural treatments. Description of Botrytis cinerea infection: frequency: percentage of infected clusters, strength: percentage of berries in one cluster, total: frequency * strength; n = 200 cluster. Viticultural treatments are C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison

Botrytis cinerea infection C [%] I [%] IdB [%] IdV [%]

2008 frequency 23 (5) 24 (6) nt 22 (8) strength 16 (2) 15 (2) nt 15 (2) total 4 (1) 4 (1) nt 3 (2)

2009 frequency 40 (12) 61 (15) 58 (22) 53 (11) strength 15 (2) 21 (6) 29 (9) 23 (10) total 6 (2) 14 (7) 19 (11) 13 (8)

2010 frequency 92 (4) 98 (1) 90 (1) nt strength 30 (4) 54 (6) 37 (5) nt total 27 (3) 53 (6) 33 (4) nt

nt: no treatment established in that vintage

3.3.2 Development of grapes’ maturity Berry weight at harvest A tendency to lower berry weight was observed for the non-irrigated treatments (C) in all vintages, whereas it showed only a significant difference in the 2010 vintage. For the vintages 2009 and 2010, berries were sampled and their juice was analysed on five and three dates respectively. In 2009 the sugar assimilation showed a significant difference between the irrigated and non-irrigated treatments during the second half of the berry ripening period. Total acidity (TA) showed significant differences only directly before harvest, between the non-irrigated control (C) and the irrigated (IdB) treatment (Fig. 3-10).

Viticultural studies

113

Fig. 3-10 Berry weight [g/berry] of different treatments at harvest of the three vintages 2008, 2009 and 2010 (with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison)

Noticeable in this context, was a decline of the TSS slope for the non-irrigated treatment (C) between the second (01.09.2009; JD 244) and the third (15.09.2009; JD 258) sampling date. Irrigated treatments (I, IdB, IdV) showed an uniform increase in TSS values during this period.

Fig. 3-11 Development of TSS and total acidity (TA), calculated as tartaric acid equivalents, in berries during the ripening period of viticultural treatments in vintage 2009 (with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison)

vintage2008 2009 2010

berr

y w

eigh

t [g/

berr

y]

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

1,8treatment Ctreatment Itreatment IdBtreatment IdV

julian day [d]

220 230 240 250 260 270 280 290

TSS

[°B

RIX

]

11

12

13

14

15

16

17

18

19

TA [g

/L]

6

8

10

12

14

16

18

20

treatement C TSStreatment I TSStreatment IdB TSStreatment IdV TSStreatment C TA treatment I TA treatment IdB TA treatment IdV TA

Viticultural studies

114

In this time of the ripening period a decrease in TA was observed also for the non-irrigated control (C), which was more important than for the irrigated treatments (I, IdB, IdV).

Fig. 3-12 α-Amino bound nitrogen (N-OPA) concentration and berry weight development during the ripening period for the vintage 2009 (with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison)

Regarding the viticultural treatments of 2010, no differences were observed concerning TSS and TA in the berry extracts during the ripening period. In 2009 the non-irrigated control treatment (C) showed higher organically bound α-amino nitrogen concentrations in the berry extracts after veraison than the irrigated treatments (I, IdB, IdV) at the same time. This effect coincidenced with lower berry weight for the non-irrigated control treatment in comparison to irrigated ones (Fig. 3-12).

julian day [d]

230 240 250 260 270 280 290

N-O

PA [m

g/L]

0

50

100

150

200

250

berr

y w

eigh

t [g/

berr

y]

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

treatment C N-OPA treatment I N-OPA treatment Idb N-OPA treatment IdV N-OPA treatment C berry weight treatment I berry weight treatment IdB berry weight treatment IdV berry weight

Viticultural studies

115

3.3.3 Monoterpenol and C13-norisoprenoid concentrations

3.3.3.1 Monoterpenols in grape must In the grape musts, obtained from each viticultural treatment, total terpene alcohols (free and acid hydrolysed terpenols including geraniol, α-terpineol, linalool, cis-linalool oxide, trans-linalool oxide) (Table 3-12) showed significant differences for the non-irrigated treatments, except for the 2010 vintage, when no water deficit was detected. Significant differences were observed for the vintage 2009 regarding the leaf removal treatments. The treatment, which was irrigated and defoliated at flowering (IdB) showed significant higher total terpenol concentration than the irrigated treatment, which was not defoliated at all (I) (303 µg/L and 224 µg/L respectively). On the other hand, the treatment, which was irrigated and defoliated at veraison (IdV), showed no difference in comparison with the irrigated treatment, which was not defoliated (I) (195 µg/L and 224 µg/L respectively). For the vintage 2010 no significant difference was observed regarding the leaf removal treatment. Results are presented in Table 3-12. Table 3-12 Total terpenols* in grape must obtained from the viticulture experiments; with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison

vintage treatment total terpenols* [µg/L]

2008 C 162 (5)a I 143 (8)b

2009

C 511 (29)a I 224 (6)b

IdB 303 (25)c IdV 195 (6)b

2010 C 227 (5)a I 226 (13)a

IdB 195 (30)a *total terpenols: sum of free and acid hydrolysed terpenols including geraniol, α-terpineol, linalool, cis-linalool oxide, trans-linalool oxide; n = 3; standard deviation is indicated in parenthesis; differing letters indicate statistical significance within a vintage (Holm-Sidak method; p<0.05)

3.3.3.2 Monoterpenol composition of wines after 12 months For the vintages 2008 and 2009, α-terpineol and linalool concentrations in wines from the non-irrigated control treatments (C) generally were increased in comparison to the wines obtained from the irrigated treatments, which were not defoliated (I). Linalool concentrations then were significantly higher in wines from the non-irrigated control treatments (C), showing concentrations of 64 µg/L, 51 µg/L and 129 µg/L, 87 µg/L respectively. The concentrations of

Viticultural studies

116

linalool in the wines from the non-irrigated control treatment (C) also were higher than in those wines from the treatments being irrigated and defoliated (IdB and IdV). This effect was less abundant in 2008. Diferences in the results for α-terpineol were not statistical significant. For the vintage 2010 no differences were observed for the concentrations of α-terpineol and linalool between the wines from the non-irrigated control treatments (C) and the irrigated treatments without leaf removal (I). No difference in the water status was observed for that growing season. Interestingly, a difference for geraniol concentrations in the samples was observed only for the vintage 2010. The wines obtained from the viticultural treatment, where leaf removal was carried out at flowering and veraison (IdB) showed higher concentrations than those from the not defoliated treatments (C and I). For all vintages, linalool and α-terpineol concentrations tended to be higher in wines from viticultural treatments being irrigated and defoliated (IdB, IdV), when getting compared to the wines from the treatments being irrigated, but not defoliated (I). The wines from the defoliated treatments (IdB, IdV) showed significant higher linalool concentrations than the corresponding treatment without leaf removal (I) for the vintage 2009 (104 µg/L; 110 µg/L and 87 µg/l respectively) as well as for the vintage 2010 (22 µg/L and 29 µg/L respectively). In wines from the vintage 2008 the differences in concentrations were not statistical significant but the same tendency was observed (51 µg/L and 57 µg/L). For α-terpineol a similar pattern was observed. α-Terpineol concentrations in the wines of the vintages 2008 and 2010 were 37 µg/L and 24 µg/L for the defoliated treatments (IdV in 2008, IdB in 2010) and 19 µg/L and 30 µg/L for the not defoliated irrigated treatments (I). The same tendency was observed for 2009, despite differences were not significant. Globally the concentration level of analysed free monoterpenols was highest in wines from the vintage 2009, whereas the wines from the vintage 2010 showed the lowest free monoterpenol concentration level. Results are shown in Table 3-14.

3.3.3.3 C13-norisoprenoids in grape must For the vintages 2008 and 2009, significant differences between different viticultural treatments were observed for total concentrations of free and acid hydrolysed norisoprenoids (vitispirane isomers and TDN) in grape musts. No differences were observed for the vintage 2010 (Table 3-13) For the 2008 vintage, total vitispirane and total TDN were significantly higher in the grape must of the non-irrigated control treatment (C) (42.5 µg/L and 4.5 µg/L respectively) than in the one of the irrigated treatment (I) (40.0 µg/L and 1.9 µg/L respectively).

Viticultural studies

117

Table 3-13 Total vitispirane* and total TDN** concentrations in grape musts [µg/L] obtained from the viticulture experiments, as sum of from free and acid hydrolysed vitispiranes and TDN; with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison

vintage treatment total vitispirane total TDN

2008 C 42.5 (0.2)a 4.5 (0.6)a I 40.0 (0.3)b 1.9 (0.8)b

2009

C 125.3 (1.7)a 1.1 (0.2)a I 46.5 (0.5)b 9.3 (0.6)b IdB 48.7 (3.3)b 6.5 (2.5)b IdV 42.0 (0.7)c 26.6 (2.4)c

2010 C 67.0 (1.9)a 1.0 (0.4)a I 70.8 (2.6)a 1.0 (0.1)a IdB 64.9 (2.8)a 0.7 (0.3)a

*total vitispirane include free and acid hydrolised vitispirane isomers; **total TDN include free and acid hydrolised TDN; n = 3; standard deviation is indicated in parenthesis; differing letters indicate statistical significance within a vintage (Holm-Sidak method; p<0.05);

In the vintage 2009, the grape must showed high total vitispirane concentration (125.3 µg/L) for the non-irrigated control (C), whereas the total TDN concentration was low as 1.1 µg/L. In grape musts from irrigated treatments (I, IdB and IdV) total vitispirane concentration was lower (46.5 µg/L, 48.7 µg/L and 42.0 µg/L respectively) than in the one of the non-irrigated viticultural treatment (C). The lowest concentration was observed in the grape must obtained from the treatment, which was irrigated and defoliated at veraison (IdV). For the concentrations of TDN an inverse pattern was observed, showing higher concentrations for grape musts from irrigated treatments I, IdB and IdV (9.3 µg/L, 6.5 µg/L and 26.6 µg/L respectively) than for the grape must of the non-irrigated control treatment C (1.1 µg/L). The highest total TDN concentration was analysed in the grape must of the irrigated treatment, which was defoliated at veraison (IdV). Generally, observed concentration levels of C13-norisoprenoids were higher for the vintage 2009, than for the vintages 2010 and 2008, which showed the lowest concentration level.

3.3.3.4 C13-norisoprenoid composition in wines after 12 months After being bottled for twelve months, vitispirane (the sum of isomeric vitispiranes) concentrations did not show any difference in the wines from the different viticultural treatments for the vintage 2008. The concentrations of TDN in these wines were below the limit of quantification of the analytical method (LoQ 0.1 µg/L). The concentrations of vitispirane in the wines from the vintage 2009 showed similar effects and did not differ. Statistical significant differences in the wines’ TDN concentrations were detected regarding some viticultural treatments. TDN concentrations in the wines of the non-irrigated treatment (C) (0.7 µg/L) were generally lower than in the irrigated treatments (I, IdB, IdV) (1.2 µg/L, 1.9 µg/L, 1.6 µg/L respectively).

Viticultural studies

118

Table 3-14 Monoterpene (after 12 months) and C13-norisoprenoid concentrations in Riesling wines (after 12 months and 22 months) from viticulture experiments fermented with yeast A (Saccharomyces cerevisiae); with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison

vintage treatment geraniol1 α-terpineol2 linalool3 cis-linalool oxide4

trans-linalool oxide5

12 months 22 months

vitispirane6 TDN7 vitispirane6 TDN7

2008 C 8 (0)a 33 (4)a,b 64 (7)a nd nd 17 (0.0)a nd 63.9 (15.0)a 7.8 (5.1)a

I 9 (0)b 30 (2)a 51 (2)b nd nd 16.9 (0.1)a nd 82.4 (1.9)a,b 16.0 (0.8)a

IdV 9 (0)b 37 (2)b 57 (0)a,b nd. nd 17.2 (0.4)a nd 99.8 (17.2)b 17.5 (8.2)a

2009

C 12 (4)a 41 (1)a 129 (9)a 1 (2)a nd 26.5 (2.4)a 0.7 (0.3)a 100.2 (7.8)a,b 25.8 (4.5)a

I 9 (1)a 31 (6)a 87 (5)b 2 (0)a nd 26.9 (0.9)a 1.2 (0.1)a,b 97.2 (5.0)a 29.2 (2.9)a

IdB 9 (1)a 35 (8)a 104 (3)c 2 (0)a nd 32.3 (6.5)a 1.9 (0.6)c 125.8 (14.7)b 34.3 (11.4)a

IdV 10 (1)a 37 (9)a 110 (8)c 2 (1)a nd 28.6 (1.4)a 1.6 (0.5)b,c 114.4 (8.0)a,b 35.7 (6.6)a

2010 C 11 (0)a 21 (1)a 24 (2)a 8 (1)a 23 (2)a 23.7 (1.9)a 1.1 (0.3)a na na

I 11 (0)a 19 (1)a 22 (1)a 9 (0)a 23 (0)a 25.3 (0.7)a 1.0 (0.1)a na na

IdB 13 (1)b 24 (1)b 29 (2)b 8 (2)a 26 (1)a 28.5 (1.9)b 1.9 (0.1)b na na Results obtained with yeast A (Oenoferm Freddo); n = 3; standard deviation is indicated in parenthesis; differing letters indicate statistical significance within a vintage (Holm-Sidak method; p<0.05); nd: not detectable; na: not analysed; Limit of detection: 1,2,3,4,5 LoD: < 0.1 µg/L; Limit of quantification: 1,2,3,4,5 LoQ: < 1.0 µg/L; 6,7 LoD: < 0.05 µg/L, 6,7

LoQ: < 0.1 µg/L

Viticultural studies

119

The wines produced from the irrigated and defoliated treatments (IdB, IdV) generally showed significant higher TDN concentrations in comparison to those from the non-irrigated treatment (C). Within the irrigated viticultural treatments, only wines from the treatment defoliated at flowering (IdB) showed significantly higher TDN concentrations in comparison to those from the treatment not being defoliated at all (I). In the vintage 2010 no significant differences of total vitispirane and TDN concentrations were observed in the wines within the non-irrigated control treatment (C), and the treatment, which was supposed to get irrigated (I). In fact this treatment was not irrigated due to excessive rainfall during the ripening period. The wines from the viticultural treatment, in which leaves were removed at flowering (IdB) showed significant higher total vitispirane and TDN concentrations. The concentration levels of vitispirane and TDN were generally different in the wines dependent on the vintage, showing the highest concentration for the vintage 2009 and the lowest concentration for the vintage 2008 (Table 3-14). Different yeast strains of Saccharomyces cerevisiae used for fermentation showed no significant impact on vitispirane and TDN concentrations for all vintages.

3.3.3.5 C13-norisoprenoid composition in wines after 22 months After 22 months of bottle aging the vitispirane concentrations in the wines from the 2008 vintage showed significant lower concentrations in the wines from the non-irrigated field trial (C), than in those from the irrigated treatment, which was defoliated at veraison (IdV) (63.9 µg/L and 99.8 µg/L respectively). No significant difference was observed in the wines from the irrigated treatments with and without leaf removal at veraison. The same was observed for TDN, whereas no statistical significant difference was observed, although wines from the non-irrigated treatment (C) tended to contain lower TDN quantities. For the wines from the vintage 2009, the results were ambiguous, due to high variability in analysed samples. However, there were significant differences for vitispirane concentrations regarding the viticultural treatments. High concentrations were observed in the wines from the irrigated treatment, which were defoliated at flowering (IdB), showing 125.8 µg/L vitispirane isomers. These wines showed significant higher vitispirane concentrations than those of the irrigated treatment, which was not defoliated (I) resulting in 97.2 µg/L. The vitispirane concentrations in the wines obtained from the non-irrigated viticultural control treatment (C) and the treatment being irrigated and defoliated at veraison (IdV) were in between these concentrations (100.2 µg/L and 114.4 µg/L respectively) and were not statistically significant different.

Viticultural studies

120

The concentrations of TDN in these wines showed a tendency to lower concentrations in the wines from non-defoliated modalities (C, I) independently from water status, though this difference was not statistically significant. As for the wines stored for 12 months, the concentration levels of vitispirane and TDN were generally different in the wines depending on the vintage when being stored for 22 months. The highest concentration level was observed in the wines from the vintage 2009 and the lowest concentration level in the wines from the vintage 2008 (Table 3-14). Different yeast strains of Saccharomyces cerevisiae used for fermentation showed no significant impact on vitispirane and TDN concentrations in wines for all vintages.

3.3.4 Discussion on monoterpenols in grape must The influence of the vineyard management like canopy manipulation on monoterpenol and C13-norisoprenoid concentrations in the berries, wines and must was extensively studied, whereas only little data is available with regard to the effect of water stress. Regarding the summed parameter of free and acid hydrolised monoterpenols in grape must, higher concentrations were observed for the treatments grown under water stress (compare 3.3.1.1 and 3.3.1.2). These findings should be discussed in three directions:

i) concentration of berry monoterpenols in smaller volume of grape juice, due to a smaller berry size and therefore altered skin-juice ratio

ii) increase of monoterpenol synthesis or storage, as a response to the vines’ water deficit

iii) indirect influence of water status on the monoterpenol concentration by auto-defoliation and thereof higher grape cluster sun exposure and temperatures

i) concentration of berry monoterpenols in smaller volume of grape juice, due to

a smaller berry size and therefore altered skin-juice ratio Regarding the concentration effect, it is obvious that in the vintages 2008 and 2009 the berry weight, albeit not being statistical significant, was negatively affected by the water deficit, which is in compliance with earlier reported results (Hardie and Considine, 1976; Matthews et al., 1987; Ojeda et al., 2002). Therefore, in the same mass grapes, more berries, and therefore more berry surface - namely berry skin - were present in one batch of pressed grapes. This led to the assumption that there was a concentration effect, due to altered skin-juice ratio (Marais, 1983) and / or altered pressing conditions influencing the monoterpenol extraction capacity (Versini et al., 1981). For vintage 2010, no such effect was observed, despite the smaller berry size in the control treatment (C). This is in accordance with the results of a study on anthocyanins and tannins

Viticultural studies

121

in Cabernet Sauvignon grapes, which concluded that the sources of variation on berry size (like environmental factors) are the more important factor to determine the berry composition than the berry size itself (Roby et al., 2004).

ii) increase of monoterpenol synthesis or storage, as a response to the vines’ water status

In published literature, increased monoterpene concentrations were observed for Muscat grapes, grown under post-veraison water deficit, independent from the berry size. However, in that study, analysis was undertaken on the headspace of intact berries and the increase was not significant for most of the analysed compounds (El-Ansary et al., 2005). For Sauvignon blanc Giorgessi et al. (2007) showed an increase of free and bound monoterpenes in the berry skins with an increasing water deficit. Similar results were observed in trees from Cupressus sempervirens, when an accumulation of glycosidic bound linalool was observed in leaves peaking short time after an induced water stress (Yani et al., 1993). These authors mention that terpene compounds may have an important role in the process of drought resistance of trees (Yani et al., 1993). Other results regarding the terpenoid emission rates of four different Mediterranean plant species (Rosmarinus officinalis, Cistus Albidus, Quercus coccifera and Pinus halepesis) showed lower emission rates as a response to water stress (Ormeno et al., 2007). The same authors observed the change of percentual contribution for single monoterpenes and sesquiterpenes on the total terpenic emission along the duration of water deficit, showing a shift from high sesquiterpene, but also monoterpenol, emission after four days towards monoterpene emission to the end of the eleventh day of the experiment (Ormeno et al., 2007). These findings suggest, that even short term and already mild water deficit can directly influence the terpenes’ metabolisation and possibly the delocalisation in plants and can increase monoterpenol concentrations in grape musts from water deficit driven vineyards.

iii) indirect influence of water status on the monoterpenol concentration by auto-defoliation and thereof higher grape cluster sun exposure and temperatures

Water deficit modulates the vines canopy structure in order to limit transpiration, through vigour reduction in the pre-veraison stage (Hardie and Considine, 1976; Matthews and Anderson, 1989; Keller, 2005). Furthermore, senescence of leaves can be induced by long term water deficit at the stage of post-veraison also, including recycling of organic matter into other plant organs like fruits (Keller, 2005). This effect was observed during the vintage 2009, when the non-irrigated control basal leaves showed decolouration and abscission took place three weeks before the harvest. Therefore, grape cluster in the non-irrigated control treatment (C) at harvest showed grape berry cluster sun exposure rates comparable to those treatments, which had been defoliated (IdB, IdV). As grape berry sun exposure was shown to affect terpene concentration in berries (Reynolds et al., 1995; Belancic et al., 1997; Kozina et

Viticultural studies

122

al., 2008), higher concentrations of monoterpenols in non-irrigated treatments can possibly be explained by this factor. The fact that monoterpenols’ accumulation in grape berries was shown to be not coupled to sugar accumulation and is known to be timed at the end of the ripening period, supports this hypothesis (Hardy, 1970; Günata et al., 1985a; Marais et al., 1987; Park et al., 1991). In addition to that, an increase in berry temperature - possibly through sun exposure - was observed to be associated with higher monoterpenol concentrations (Reynolds et al., 1995; Reynolds et al., 1996). However, with regard to the leaf removal experiments, results for Muscat, Sauvignon blanc and Riesling grapes and wines showed total terpene levels to decrease, if the sun exposure was 100% (Belancic et al., 1997) or ‘highly’ leaf removed (Kozina et al., 2008). Leaf removal at veraison (IdV) in 2009 and leaf removal at flowering (IdB) in 2010 led to equal or even lower total terpene concentrations in the grape musts compared to their irrigated non-defoliated counterpart (I). Grape berry cluster sun exposure rates of these defoliated treatments showed similar levels to those of the non-irrigated control, but showed lower monoterpenol concentrations. Therefore the increase of monoterpenol concentration in grape berry extracts from non-irrigated treatments is not likely to be indirectly influenced by water stress through auto-defoliation and higher grape cluster temperatures.

3.3.5 Discussion monoterpenol composition of wines after 12 months Like for monoterpenol concentration in grape must as well, results of single terpenols in the wines after finishing fermentation showed a statistical significant increase for the non-irrigated treatments (C), which were grown under higher water deficiency, than the irrigated treatments (I). In contrast to the results for the total monoterpenols obtained in musts, monoterpenol concentrations, especially free linalool, were higher in the wines from the defoliated treatments, than in those from the non-defoliated treatments grown at the same water status. These results are in conformity to other published data (Reynolds et al., 1995; Belancic et al., 1997; Kozina et al., 2008). The contrasting tendency for total terpenols in grape must and free terpenols in corresponding wines already was shown in the data of other studies concerning Riesling and different fertilisation quantities. There, the trend for volatile terpenols was in the opposite direction of that of bound terpenols, showing that even under the same total terpenol concentrations, the quantity of free terpenols can be significantly different (Webster et al., 1993). In the vintage 2010, the lowest α-terpineol and linalool concentrations were observed in the wines, but higher linalool oxide concentrations of all vintages were detected. Terpenol oxides were observed to be formed in over-ripe berries (Marais, 1987). An increase of linalool oxides was also observed in presence of Botrytis cinerea (Fedrizzi et al., 2011). Both

Viticultural studies

123

can be considered, due to high soluble solids’ concentrations and also relatively high infection with Botrytis cinerea in that year in comparison to the vintages 2008 and 2009.

3.3.6 Essential outcomes for monoterpenols and their precursor molecules In conclusion both, water status and microclimate manipulations, have an important influence on monoterpenol concentrations and their pattern in grape must and wines. In white wines’ vinification decreased berry size, as an effect of water deficit, is presumably a less important factor, due to short skin contact times of grape must during the vinification process. The influence of water deficit on the biosynthesis and the metabolisation of terpenoide compounds is more likely to account for variations in grape musts and wine monoterpenol concentrations. On the one hand, the effect of leaf removal and sun-exposure is clearly demonstrated in the grape must and in the finished wines, but on the other hand the transformation during vinification remains unclear. Upon the data from this project, it is likely that the ratio of free and bound terpenes in the wines somehow gets modified through viticultural practices.

3.3.7 Discussion on C13-norisoprenoids C13-norisoprenoids vitispirane and TDN free and acid liberated quantities differed upon different viticultural treatments in grape must. The water status seemed to influence vitispiranes’ and TDN’s concentrations, favouring both vitispirane and TDN precursor formation in musts obtained from berries grown under weak to moderate water deficit levels during the warmer vintage 2009 (compared within 2008: C and 2009: I). In grape must obtained from the treatment grown at a moderate to severe water deficit level, the potential of vitispirane increased, whereas at the same time the potential of TDN decreased (compared within 2009: C and 2009: I). Similar observations were made for the finished wines after 12 months and after 22 months of bottle storage for free C13-norisoprenoids. Interestingly no differences for vitispirane and TDN concentrations were observed between the wines from the non-irrigated control (C) and the irrigated treatment (I) within the 2009 vintage regarding the water deficit level after 12 months, and then after 22 months. In grape musts from the viticultural treatments, which showed only weak or no water deficit during the ripening period (2008: I; 2010: C, I) vitispirane concentrations formed from precursors after acid hydrolysis were lower in the warmer vintage 2008, which showed higher Growing Degree Days values (GDD). An inverse effect was observed for TDN precursors, which were increased in years with higher GDD values. In the finished wines from these

Viticultural studies

124

viticultural trials the same trend was shown for the vitispirane concentrations after 12 months of bottle storage, whereas again it was not for TDN concentrations. Under moderate water deficit the potential of acid hydrolysable vitispirane and TDN in grape musts varied with regard to the timing of defoliation (2009: IdB; IdV). This was observed for wines from these viticultural experiments after 12 months and 22 months also, whereas none of these variations was shown to be statistically significant. In order to interpret these viticultural effects, such as canopy manipulations, water deficit and environmental effects on C13-norisoprenoid concentrations in must and wine, the results in grape musts will be considered before discussing C13-norisoprenoid concentrations in finished wines. The discussion should concern the following possibly influencing factors on the potential of C13-norisoprenoid concentration in grapes and volatile C13-norisoprenoid concentrations analysed in aged wine

i) the effect of water status ii) the effect of micro climate manipulation iii) the effect of vintage and temperature

3.3.7.1 Discussion on C13-norisoprenoids in grape must i) the effect of water status

Comparing the results with the few published data in respect to the influence of water deficit on C13-norisoprenoids, the obtained data were contrary to the results obtained for a Cabernet Sauvignon field trial, where water deficit was provoked by partial root zone drying (PRD) (Bindon et al., 2007). In the latter work, the concentrations of acid liberated TDN in berry homogenates increased with water deficit. Water deficit was measured as midday stem water potential (Ψ(S)) and as midday leaf water potential (Ψ(L)) in that study. The water deficit indicated (up to Ψ(S) -1.1 MPa for irrigated control and Ψ(S) 1.4 MPa for PRD) can be classified as moderate to severe stress for both – irrigated control and PRD – treatments according to the classification scheme of van Leeuwen et al. (2010). Generally it is difficult to compare data from different methods of water potential measurements, as they reflect different situations in the vines’ daily water status. Whereas pre-dawn leaf water potential Ψ(PD) is determined in a situation of a stable equilibrium, Ψ(S) and Ψ(L) represent the dynamic status of a transpiring plant (Gruber and Schultz, 2005). However, the measurement of the midday stem water potential Ψ(S) was shown to be an accurate and robust estimation of the vine water status (Choné et al., 2001). In addition to this, it can be assumed that the vines’ reaction on PRD is different to ‘naturally’ occurring water deficit and that different Vitis vinifera varieties show different biochemical reactions at equal water deficit levels.

Viticultural studies

125

ii) the effect of micro climate manipulation on C13-norisoprenoid precursor concentration

It has been shown that C13-norisoprenoid precursor concentration in grape juice increased with sun exposure (Marais et al., 1992a; Marais et al., 1992c; Gerdes et al., 2002; Lee et al., 2007), but this was shown also to be dependent on the timing of the leaf removal (Kwasniewski et al., 2010). Gerdes et al. (2001) showed an increase of free and bound TDN and Riesling acetal in sun exposed Riesling grapes which have been defoliated at veraison and then shaded by cloth. They showed also that naturally shaded berries had higher TDN levels than those covered by cloth after leaf removal (Gerdes et al., 2001). These observations are in accordance with the obtained results from the present study for the 2009 vintage. The irrigated treatment, which was defoliated at veraison (IdV) showed higher acid liberated TDN concentration in grape must than the irrigated treatment without leaf removal (I). Grape juices of grapes from the modality where leaves have been removed at the stage of flowering (IdB) did not show different acid hydrolysable vitispirane and TDN potential level in comparison to the corresponding non-defoliated treatment during 2009 and 2010. In comparable studies for Cabernet Sauvignon and Riesling these observations were made as well (Lee et al., 2007; Kwasniewski et al., 2010).

iii) the effect of vintage and temperature An effect of vintage and vineyard site was reported on C13-norisoprenoid precursor inclusively on carotenoids in grapes and grape musts (Razungles et al., 1987; Marais et al., 1992a; Marais et al., 1991; Lee et al. 2007), but mostly included the factor sunlight as sun-hours or microclimate manipulation of the canopy. Therefore it is difficult to separate these two factors and subsequently this should be considered when interpreting this kind of data. Further discussion was already done under ii).

Viticultural studies

126

3.3.7.2 Discussion C13-norisoprenoid composition in wines after 12 months and 22 months

i) the effect of water status In this study higher TDN concentrations were analysed in wines obtained from the irrigated viticultural treatments, which were defoliated, whereas a strong tendency towards lower TDN concentrations in the wines from the non-irrigated treatments was observed after 12 months and 22 months of bottle storage. With respect to water deficit in the present Riesling study no significant alternation for volatile vitispirane concentration was detected regarding water deficit in wines after 12 months and 22 months of bottle storage. These results were in contrast to a study of Merlot wines, which were grown under different water status. In that study wines from grapes, grown under a deficit irrigation regime, showed higher concentration for vitispirane, but no results were given for TDN concentrations and the age of the wines was not mentioned in the paper (Qian et al., 2009).

ii) the effect of micro climate manipulation Regarding the effect of leaf removal, the wines of the vintages 2009 and 2010 showed a significant increase in vitispirane and TDN concentrations for treatments, which were defoliated at flowering (IdB), in comparison to the wines from the non-defoliated treatments. In 2009 only the treatment defoliated at veraison (IdV) showed no significant higher concentrations of these compounds in the wines compared to those from the non-defoliated treatment (I). Regarding defoliation the results confirmed the effect of leaf removal on the concentration of free vitispirane and TDN of another study on Merlot wines, when vitispirane and TDN concentrations increased with light exposure, but additionally were influenced by leaf layer number (Lee et al., 2007). Contrary results to the results from this study were obtained for Riesling wines from east-coastal Northern America for the consequences of cluster exposure (Kwasniewski et al., 2010). In that study grape vines’ defoliation was carried out at different phenological stages and was allowed to be compensated by secondary leaf growth, which was evident in the cluster exposure rates. In that study only defoliation at pre-veraison showed higher TDN concentrations in finished wines, which then was not observed for the defoliation at berry set and the defoliation at veraison. In our present study, defoliation treatments were carried out at flowering to be then re-defoliated at veraison (IdB) and at veraison only (IdV). Both showed higher free TDN concentrations in the finished wines in comparison to those, from the not defoliated control treatment (I). Interestingly, absolute TDN concentrations in the two months bottle stored wines from the North American study of Kwasniewski and co-workers were much more elevated (7 - 20 µg/L) than the ones in the present study after twelve

Viticultural studies

127

months of bottle storage in all vintages (<0.1 - 1.9 µg/L). This could either be explained by differing storage conditions, which were observed to influence formation and break down of norisoprenoid compounds (Marais et al., 1992), or by general differing aging behaviour depending on the origin of the wine and therefore its composition. It was already observed that the ratio of free TDN and its TDN potential determined in forced aging experiments in wines from east coastal Northern America was increased in comparison to the wines from the Rheingau region in Germany. These findings showed that the concentration of free TDN was more abundant in American wines compared to wines from Germany, whereas the potential of TDN concentration determined by a forced aging experiment was even higher in German Riesling wines (Sponholz and Hühn, 1997). These authors also suggested that a different, structurally only partially identified precursor molecule could be engaged in TDN formation in wine (Versini et al., 1996). The formation via this precursor is suggested to take place in a slower reaction than via the pathway proposed by Waldmann and Winterhalter (1992). Therefore, precursor molecules’ distribution could be crucial for TDN formation kinetics in Riesling wines and thus be responsible for earlier or later increase in TDN concentration during the aging process. Comparing the results of the wines after twelve months and after 22 months of bottle storage, this suggestion gets supported, by equalising TDN concentrations with prolonged aging. In addition to this, the fact that measurements of the potential of volatile TDN in grape musts did not reflect the situation in the wines could also be a supporting element for this hypothesis.

iii) the effect of vintage and temperature The effect of vintage generally is ambivalent. Concentrations of vitispirane and TDN in the wines from treatments of the same water status (2010: C, 2008: I and 2008: C, 2009: I) were linked differently to temperatures during the growing season expressed as GDD. For viticultural treatments showing no or weak water deficit, vitispirane and TDN concentrations were found to be elevated in the wines of the colder vintage 2010, whereas for the treatments showing weak to moderate water deficit higher vitispirane and TDN concentrations were analysed in the wines for the warmer vintage 2009 by means of GDD. In fact, this would suggest that the vines’ water status can alter the effect of temperature during the growing season on free C13-norisprenoid concentrations in wines either due to direct precursor metabolisation or due to secondary factors as substrate delivery for vitispirane and TDN formation, which could be used in the transformation reaction during fermentation and bottle aging.

Viticultural studies

128

3.3.8 Essential outcomes for C13-norisoprenoids and their precursor molecules In conclusion it can be stated that C13-norisoprenoid concentrations in wines are strongly influenced by viticultural factors, especially by water status. It is considered that the aging behaviour of Riesling wines is significantly influenced by viticultural factors, due to alter C13-norisoprenoids concentration pattern and determines the rate of TDN formation.

3.3.9 Thiol concentrations in wines and thiol precursor concentrations in berry extracts and grape must

3.3.9.1 Observations on thiol precursor molecules and other related metabolites in berry extracts Cysteinylated and glutathionylated precursor molecules S-3-(hexan-1-ol)-L-cysteine (cys-S-prc) and S-3-(hexan-1-ol)-glutathione (GSH-S-prc) of the aroma compound 3-sulfanylhexan-1-ol (3SH) were assayed in berry samples at different stages of the ripening period. In addition to these compounds, glutathione (GSH), one of its main reaction partners in grape must caftaric acid (CA), and the resulting product S-glutathionylcaftaric acid – the so called Grape Reaction Product (GRP) (Singleton et al., 1985) – were analysed in berry extracts during the grape maturation and in musts (Table 3-15 and Table 3-16). During the vintage 2009 in the early stage of berry ripening (14 days after veraison) only very little quantities of cys-S-prc were measured for all viticultural treatments (1 - 2 µg/L berry extracts). The concentrations of the GSH-S-prc, ranging from 68 - 134 µg/L were generally higher than those of cys-S-prc. Berry extracts from defoliated vineyard treatments showed significant lower concentrations in comparison to treatments without leaf removal. Regarding the GSH concentrations, berry extracts from non-irrigated vines (C) showed significant higher values than in the samples from the irrigated treatments (I) (33.0 mg/L and 25.5 mg/L respectively). Defoliation manipulation corresponded to an increase of GSH resulting in concentrations of 30.7 mg/L and 28.5 mg/L for the different timing of leaf removal, at flowering and at veraison (IdB and IdV respectively). No GRP was detectable and CA concentrations in the berry extracts ranged from 135.0 mg/L to 147.8 mg/L, without significant differences at this stage of maturity. A tendency to higher CA levels in the berry extracts was observed for the non-irrigated and not defoliated treatment (C) and the irrigated treatment, which was undertaken leaf removal at flowering (IdB). The berries sampled one day before harvest, still did show low concentrations of cys-S-prc of 3 - 4 µg/L in the extracts, whereas the concentration of GSH-S-prc was increased in all viticultural treatments, when being compared to the results from the first sample date.

Viticultural studies

129

Table 3-15 Concentrations of S-3-(hexan-1-ol)-L-cysteine (cys-S-prc) , S-3-(hexan-1-ol)-gluthatione (GSH-S-prc), glutathione, S-glutathionylcaftaric acid and caftaric acid in berry extracts and grape must for the 2009 vintage; with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison

sample date

treatment

cys-S-prc

[µg/L]

GSH-S-prc

[µg/L]

GSH

[mg/L]

GRP

[mg/L]

caftaric acid

[mg/L]

veraison + 14 days

C 2 (2)a 134 (42)a 33.0 (2.6)a nd 145.6 (7.7)a

I 1 (0)a 119 (20)a 25.5 (2.7)b nd 135.8 (7.8)a

IdB 1 (0)a 68 (10)b 30.7 (2.2)a,b nd 147.8 (2.9)a

IdV 1 (0)a 71 (5)b 28.5 (1.3)a,b nd 135.0 (5.5)a

harvest

C 4 (2)a 174 (10)a 25.7 (3.1)a nd 113.2 (8.3)a,b

I 3 (1)a 129 (4)b 36.7 (3.9)b nd 118.2 (2.5)a,b

IdB 4 (1)a 180 (23)a 33.6 (2.6)a,b nd 123.0 (1.3)a

IdV 4 (1)a 109 (0)b 37.8 (4.2)b nd 106.8 (5.8)b

must C 64 (21)a 214 (8)a 1.0 (0.1)a 9.9 (0.2)a 50.8 (1.1)a I 55 (0)a 679 (21)b 1.1 (0.1)a 5.6 (0.6)b 23.2 (1.5)b IdB 123 (10)b 492 (1)c 0.9 (0.2)a 11.9 (0.3)c 68.9 (3.5)c IdV 67 (26)a 603 (8)d 1.0 (0.1)a 15.0 (1.2)d 54.4 (0.9)a

n = 3; standard deviation is indicated in parenthesis; differing letters indicate statisticalll significance (Fisher LSDp<0.05); nd: not detectable

The highest concentrations within the treatments were observed for the non-irrigated treatment (C) and the irrigated one, which was defoliated at flowering (IdB), containing 174 µg/L and 180 µg/L respectively. Lower concentrations were observed in the irrigated treatments with leaf removal at veraison and without leaf removal (IdV; I), namely 109 µg/L and 129 µg/L respectively. This concentration pattern was exactly inversed to that of the GSH concentrations, which were lowest in the berry extracts from the non-irrigated control treatment (C) and tendentially showed lower concentration for the irrigated treatment, which was defoliated at flowering (IdB). GSH concentrations generally showed higher values in the berry extracts for the irrigated treatments (I, IdB, IdV) than for the non-irrigated control treatment (C), regardless of microclimate manipulation. For the vintage 2009, GRP was not detected in any of the berry extracts, although CA concentrations decreased in comparison to the first sampling date. CA concentration was shown to be higher in the berry extracts of the irrigated treatment, which was defoliated at flowering (IdB) (123.0 mg/L) than in those ones from the irrigated treatment, which was defoliated at veraison (IdV) (106.8 mg/L). In 2010, three sampling dates were realised - the first one at veraison + 14 d, the second one at veraison + 22 d and the last one at harvest. As in 2009, there were very small concentrations of cys-S-prc in all berry samples at all sampling dates ranging from not detectable to 2 µg/L in the berry extracts.

Viticultural studies

130

Table 3-16 Concentrations of S-3-(hexan-1-ol)-L-cysteine (cys-S-prc), S-3-(hexan-1-ol)glutathione (GSH-S-prc), glutathione, S-glutathionylcaftaric acid and caftaric acid in berry extracts and grape must for the 2010 vintage; with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison

sample date

treatment

cys-S-prc

[µg/L]

GSH-S-prc

[µg/L]

GSH

[mg/L]

GRP

[mg/L]

caftaric acid

[mg/L]

veraison + 14 days

C 1 (1)a 146 (34)a 33.4 (2.6)a na na I 2 (1)a 140 (13)a 25.0 (8.8)a na na

IdB 2 (2)a 123 (20)a 11.4 (9.3)a na na

veraison + 22 days

C 1 (0)a 120 (1)a 37.6 (5.0)a na naI 1 (1)a 85 (7)b 43.9 (3.7)a na na

IdB 2 (2)a 138 (20)a 42.4 (0.8)a na na C 1 (0)a 111 (26)a 47.5 (4.1)a na na

harvest I 1 (0)a 150 (22)a 47.6 (1.5)a na na IdB 0 (0)a 150 (25)a 51.1 (9.0)a na na

must C 17 (0)a 205 (10)a 0.9 (0.1)a 5.2 (0.2)a 19.2 (0.9)a

I 21 (0)b 232 (3)b 1.3 (0.3)a 5.7 (0.4)a 21.2 (1.2)a

IdB 22 (3)b 267 (5)c 1.0 (0.2)a 6.6 (0.3)b 18.8 (2.1)a

na: not analysed; n = 3; standard deviation is indicated in parenthesis; differing letters indicate statistical significance (Fisher LSD; p<0.05) The concentrations for GSH-S-prc showed no significant differences for the first date at veraison + 14 d and at harvest, whereas significant lower concentrations were detected in the treatment supposed to be irrigated (I), at the second sampling date at veraison + 22 d. No general increase of GSH-S-prc in berry extracts was detected with increasing time of ripening. In contrast to that no statistically significant differences were observed for GSH concentrations in the berry extracts regarding the viticultural treatments at all sampling dates, whereas a general increase was observed with increasing time of the berry ripening.

3.3.9.2 Thiol precursor molecules and related metabolites in grape musts Different concentrations of cysteinylated and glutathionylated S-bound precursor (cys-S-prc and GSH-S-prc respectively) molecules of the odoriferous thiol 3-sulfanylhexan-1-ol (3SH) and assimilable nitrogen (sum of N-OPA and Ammonium concentrations) were detected in grape musts from the different experimental treatments in all vintages regarding to viticulture treatments (results compiled in Table 3-17). For the vintage 2008, a difference was observed for the concentration of GSH-S-prc and assimilable nitrogen (N). Assimilable nitrogen showed higher proportions in the must from the non-irrigated control treatment (C) in comparison to the must from the irrigated treatment without defoliation (I) (GHS-S-prc: 216 µg/L and 193 µg/L; N: 345 mg/L and 271 mg/L

Viticultural studies

131

respectively). At the same time, no significant difference was observed for the concentration of cys-S-prc in these samples (10 µg/L and 4 µg/L). In the vintage 2009, the GSH-S-prc concentrations were significantly different in the musts from all viticulture experimental trials, showing lower values for the non-irrigated control treatment (C) (214 µg/L), whereas the irrigated treatments (I, IdB, IdV) generally showed higher concentrations, independent from leaf removal. The highest concentration of GSH-S-prc within these treatments was observed in grape must of the irrigated not defoliated treatments (I) (679 µg/L). The concentrations decreased upon the timing of leaf removal, showing lower GSH-S-Prc concentration in grape must from the treatment already defoliated at flowering (IdB) in comparison to the must of the treatment defoliated at veraison (IdV) (492 µg/L and 603 µg/L respectively). Grape musts’ concentrations of cys-S-prc only showed a significant difference for the irrigated treatment, which was defoliated at flowering (IdB) (123 µg/L). This concentration was elevated in comparison to the musts from all other treatments (C: 64 µg/L; I: 55 µg/L and IdV: 67 µg/L). The non-irrigated control treatment (C) showed the highest assimilable nitrogen content (226 mg/L) in grape must within all viticultural treatments (I: 193 mg/L; IdB: 167 mg/L; IdV: 206 mg/L), with the lowest value observed for the treatment, which was irrigated and defoliated at flowering (IdB).

Table 3-17 3SH-precursor and assimilable nitrogen (N) concentration (sum of N-OPA and free ammonium concentrations) in grape must (n=3) and free 3SH in corresponding finished wines (n=3); with C: non-irrigated control; I: irrigated; IdB: irrigated, defoliated at flowering and defoliated at veraison; IdV: irrigated and defoliated at veraison

vintage

treatment

cys-S-prc

[µg/L]

GSH-S-prc

[µg/L]

3SH in wines

[ng/L]

assimilable N

[mg/L]

2008 C 10 (1)a 216 (1)a 206 (65)a 345 I 4 (5)a 193 (13)b 313 (55)b 271

IdV na na 275 (45)a,b na

2009

C 64 (21)a 214 (8)a 156 (56)a 226 I 55 (0)a 679 (21)b 328 (69)b 193

IdB 123 (10)b 492 (1)c 412 (168)b 167 IdV 67 (26)a 603 (8)d 327 (68)b 206

2010 C 17 (0)a 205 (10)a 326 (41)a 215 I 21 (0)a,b 232 (3)b 984 (18)b 267

IdB 22 (3)b 267 (5)c 287 (12)a 254 na: not analysed; standard deviation is indicated in parenthesis; differing letters indicate statistical significance (Fischer LSD method; p<0.05)

In the vintage 2010, the concentrations of GSH-S-prc in musts of the - in that year two - non-irrigated and not defoliated treatments C and I were surprisingly statistically significantly different (205 µg/L and 232 µg/L respectively). Anyway, both grape musts showed a significant lower GSH-S-prc concentration in comparison to the one obtained from the

Viticultural studies

132

defoliated treatment (IdB), which was showing a concentration of 267 µg/L. The same was observed for the cys-S-prc, whose concentration was lowest in grape musts from the non-irrigated control treatment, without leaf removal (C) (17 µg/L) in comparison to all others (21 µg/L and 22 µg/L). For the concentrations of assimiliable nitrogen an equivalent pattern was observed (C: 215 mg/L; I: 267 mg/L; IdB: 254 mg/L).

3.3.9.3 Comparison of berry extracts and grape musts composition Regarding the composition of the berry samples at harvest and the corresponding grape musts, significant differences could be observed in both, vintage 2009 and vintage 2010.

i) for GSH, caftaric acid and GRP Compared to the berry extracts, lower GSH concentrations were measured in grape musts, corresponding to a decrease by the factor 30 - 50. In the vintage 2009, the berry extracts from the non-irrigated control treatment (C) showed at harvest date a concentration of 25.7 mg/L GSH. At the same stage, GSH concentrations ranging from 33.6 mg/L to 37.8 mg/L were observed in the berry extracts of the irrigated viticultural treatments (I, IdB, IdV). After pressing, no grape must from the corresponding viticultural trials was shown to exceed GSH concentration of 1 mg/L. The same effect was observed for the vintage 2010, when the concentration of GSH in berry extracts ranged from 47.5 mg/L to 51.1 mg/L and did not exceed 1.3 mg/L in grape musts after pressing. Caftaric acid (CA) concentrations decreased in 2009 from berry samples to grape must by factors of 1.8 to 5.1. The strongest decrease was observed for the irrigated treatment, which was not defoliated (I), leading to lowest CA concentrations in grape must (23.2 mg/L) within all viticultural treatments (C: 50.8 mg/L; IdB: 68.9 mg/L; IdV: 54.4 mg/L). For the non-irrigated control (C) and those treatments being irrigated and defoliated (IdB; IdV) the factors of decrease were 2.2, 1.8 and 2.0 respectively and led to the same concentration pattern like in the berry extracts, but at a lower level. For the vintage 2009, the GRP was not detected in any of the berry samples, whereas it was abundant in obtained grape musts at concentrations between 5.6 mg/L and 15 mg/L. The concentrations differed significantly for all viticultural treatments. The non-irrigated control treatment (C) showed significant higher GRP concentration than the irrigated treatment without defoliation (I) containing 9.9 mg/L and 5.6 mg/L respectively. The irrigated treatments, which were leaf removed (IdV; IdB), theirselves contained significant higher concentrations of the GRP than the not defoliated treatments (C: 11.9 mg/L and I: 15 mg/L respectively).

ii) for GSH-S-prc and cys-S-prc Interestingly, an increase by factors from 16 to 31 was observed for concentrations of cys-S-prc, when comparing berry extracts and the grape musts. For the vintage 2009 and the

Viticultural studies

133

vintage 2010 for none of the grape extracts, a cys-S-prc concentration higher than 4 µg/L and 1 µg/L was detected at harvest respectively. In contrast to that, concentration ranges between 55 µg/L to 123 µg/L and 17 µg/L to 22 µg/L were analysed in the grape musts for the vintages 2009 and 2010 respectively. In the same time GSH-S-prc concentrations were observed to increase only up to factor 5, depending on the viticultural treatment. In the vintage 2009, GSH-S-prc concentrations in the berry extracts from the non-irrigated control treatment (C) were 174 µg/L. In the corresponding grape must, the GSH-S-prc concentration was at 214 µg/L. For irrigated treatments (I; IdV; IdB) GSH-S-prc concentrations were shown to range from 106 µg/L to 180 µg/L in the berry extracts and 492 µg/L to 679 µg/L in the corresponding grape musts. In the vintage 2010 the concentrations for this parameter ranged from 111 µg/L to 150 µg/L in berry extracts and 205 µg/L to 267 µg/L in the corresponding grape musts.

3.3.9.4 Free thiol concentrations in young wines Concentrations of the thiol 3-sulfanylhexan-1-ol (3SH) were analysed in the finished wines from fermented grape musts of the viticulture experiments. In all vintages significant differences were observed in the concentrations of 3SH regarding viticultural treatments (Table 3-17). For 2008, wines showed significant lower 3SH concentration in those from the non-irrgated treatment (C) (206 ng/L) in comparison to those from the irrigated treatment without leaf removal (I) (313 ng/L). In wines from the irrigated and defoliated treatment (IdV) 3SH concentrations were observed to be lower in comparison to those from the irrigated treatment without leaf removal (I) (275 ng/L and 313 ng/L respectively). In the vintage 2009, low 3SH concentrations were observed in wines from the non-irrigated control treatments (C) (156 ng/L). Within the irrigated treatment of the vintage 2009, the one, which was defoliation at veraison (IdV), was not shown to result in lower 3SH concentration in comparison to wines from the non-defoliated treatment (I) (328 ng/L and 327 ng/L respectively). Interestingly wines obtained from the viticultural treatment, which was irrigated and then defoliated at flowering (IdB) showed higher 3SH concentrations (412 ng/L) in comparison to the two other irrigated treatments (I: 327 ng/L; IdV: 328 ng/L), whereas this difference was not statistically significant. In 2010, the non-irrigated treatment (C) produced wines showing significant lower 3SH concentration than the ‘irrigated’ treatment (I), which in fact was not irrigated in that year (326 ng/L and 984 ng/L respectively). Defoliation at flowering (IdB) corresponded to lower 3SH concentrations in the finished wines regarding the non-defoliated irrigated treatments (I) (287 ng/L and 984 ng/L respectively).

Viticultural studies

134

3.3.10 Discussion on observations on thiol precursor molecules and other related metabolites in berry extracts Regarding the biogenesis of cysteinylated and glutathionylated precursor molecules the biochemical process has to be reminded (Fig. 3-13) as well as the effects of environmental factors on several metabolites which are involved in this process. Therefore possible effects of environmental conditions on

i) the concentration of glutathione ii) the precursor GSH-S-prc and cys-S-prc

will be discussed.

Fig. 3-13 Hypothetical S-3-(hexan-1-ol)-glutathione (GSH-S-prc) and S-3-(hexan-1-ol)-cysteine (cys-S-prc) formation pathway in grapevines, induced by abiotic or biotic (B. cinerea) stressors. Free 3-sulfanylhexan-1-ol (3SH) in wine is produced during alcoholic fermentation (reproduced after Thibon et al., 2011)

i) the concentration of glutathione In the berry samples, a general increase in reduced glutathione (GSH) was observed during the berry ripening for the berries grown under weak to moderate or no water deficit in the vintages 2009 and 2010. Only in 2009, the concentration of reduced GSH in the berry extracts from the non-irrigated treatment, which showed moderate to severe water deficit,

(E)-2-hexanal

glutathionyl-3-S-hexan-1-ol

cysteinyl-3-S-(hexan-1-ol

glutathione

grape, mustaroma precursors

glutathione-S-transferase

γ-glutamyltransferaseCarboxypeptidases

stressors(Botrytis cinerea) wine

(volatile thiol)

alcoholicfermentation

Yeast

Yeast

+

+

?

+

+

S. cerevisiae β-lyases

3-sulfanylhexan-1-ol

3-sulfanylhexan-1-ol

Viticultural studies

135

decreased during the berry maturation. No significant effect on reduced GSH concentrations was observed regarding the leaf removal treatments. Generally, an increase of reduced GSH during the berry ripening is in accordance with previous findings, which showed an increase of reduced GSH with the onset of berry ripening and that this increase correlated linearly with soluble solids concentrations, which also increases with berry ripening (Adams et al., 1983). Lower reduced GSH level in berry extracts obtained from vines showing higher water deficit could be explained by the fact of changing reduced GSH to oxidised GSSG ratio as general plants stress response (reviewed in Foyer et al., 2001) or could occur due to increased glutathionylation of cell xenobiotics as hexenal (reviewed in Foyer et al., 2001 and in Noctor et al., 2011). For example, soybean showed decreasing GSH and homoGSH concentrations in leaves under drought conditions, whereas leaves approaching senescence showed altered GSH and homoGSH concentrations (Noctor et al., 2011). Anyway, this effect was not observed in grape vine leaves for water deficit, which did not increase reduced GSH concentration (Kobayashi et al., 2011), whereas this author showed significant effects on GSH through UV-C induced stress on leaves and berries (Kobayashi et al., 2011).

ii) the precursor GSH-S-prc and cys-S-prc In the 2009 vintage, lower concentrations of reduced GSH in berries at harvest correlated with higher concentrations of the GSH-S-prc. Inversely, tendentious higher reduced GSH concentrations coincidenced with lower GSH-S-prc concentrations 14 days after defoliation. This can support the hypothesis of the metabolism of the 3SH precursor molecules as a part of the pathway of a relatively dynamical process of detoxification in the berry, due to berries being exposed to excessive sun light. However, results for the vintage 2010 were less significant and therefore did not confirm the 2009 results, but one has to take into consideration that both vintages showed completely different climatic and viticultural conditions. The concentrations of cysteinylated precursor molecules of 3SH were surprisingly low in the berry extracts, which were obtained by a very reductive sampling procedure. No significant GSH-S-prc accumulation was observed in the berries during the ripening period. Therefore it can be considered, that higher GSH-S-prc in berries at the stage of maturity would be more or less coincidently related to stress conditions. This would suggest that every moment of sampling is just one snapshot in a highly dynamic and complex process.

Viticultural studies

136

3.3.10.1 Discussion on comparison of berry extracts and grape musts composition For the observation of the influence of the passage from berry to must, berries were sampled carefully under reductive conditions to avoid oxidation processes and shock frozen in liquid nitrogen in order to preserve the status of the berry as it was in the vineyard. GRP, which is known to form from caftaric acid under oxidative conditions by conjugation of GSH (Singleton et al., 1985), was measured as a marker of grape extract’s and grape must’s oxidation. For 2009, the berry extracts at harvest showed no GRP and thus were not undergone visible and measurable oxidation process. Only very small quantities of the cys-S-prc were detectable in the berry extracts. After pressing, the grape musts of the corresponding treatments showed highly increased cys-S-prc concentrations and similar pattern to GSH-S-prc in the berry extracts. Moreover, GSH-S-prc in grape must increased proportionally to the GSH concentrations in the berry extracts. Same was observed during the 2010 vintage. GRP formation and caftaric acid’s decrease were not observed to be linked to precursor formation. These results confirm earlier findings on the effect of must oxidation (Roland et al., 2010; Mattivi et al., 2012), on the effect of travelling berries over long distances and harvest procedures regarding concentrations of GSH-S-prc (Capone and Jeffery, 2011; Allen et al., 2011). Capone and co-workers demonstrated in an experiment of grapes’ enzymatic activity inhibition by liquid nitrogen, that the concentration of cys-S-prc then was constant. They concluded that the formation of this metabolite is berry endogen, whereas the main part of the GSH-S-prc was formed through berry crushing (Capone and Jeffery, 2011). In the present study, the hypothesis that cys-S-prc is formed by enzymatic processes of the berry endogenously could be supported. However, it seems that cys-S-prc is endogenously formed in the berry, but accumulates only in post-harvested berries. When comparing the sampling procedures, it can be observed that rending inert the berry samples with liquid nitrogen was not done before the grapes arrived in the laboratory and therefore the biological tissue could possibly still be active after being picked (Capone et al., 2011), whereas in the present study inertness and therefore enzymatic reaction inhibition, was achieved directly after berry sampling in the vineyard. This could lead to the assumption that cys-S-prc is formed as an intermediate metabolite whose concentration is maintained constant in the plant tissue as long as the berry is connected to sap flow and transport systems. Only after being disconnected from the sap flow, the cys-S-prc starts to accumulate due to not being delocalised. The data from 2010 confirmed the 2009 results. They also showed very little cys-S-prc constantly at all sampling dates for all viticulture treatments and independent from GSH-S-prc concentrations, but showed an increase in pressed grape must. Cys-S-prc

Viticultural studies

137

concentrations in grape must seemed to be highly dependent on endogen GSH-S-prc concentrations in berries.

3.3.10.2 Discussion on thiol precursor molecules and related metabolites in grape musts Generally, results from grape musts for cys-S-prc and GSH-S-prc indicated an influence of the water status on GSH-S-prc, whereas no significant difference was observed for the cys-S-prc. In 2008, when the water status of the vine during the growing period was classified as weak to moderate water deficit for the non-irrigated control (C), grape musts GSH-S-prc concentration was significantly higher in comparison to the irrigated treatment (I), for which the water status was classified from weak water deficit to no water deficit at all. In 2009, the vines’ water status for the non-irrigated treatment was determined to show moderate to severe water deficit, whereas the irrigated vines (I; IdB; IdV) showed weak to moderate water deficit. In that vintage, the grape musts of the irrigated treatments showed higher GSH-S-prc concentrations than those obtained from the non-irrigated treatments. This, in fact, is in accordance with results deduced from the aroma potential, measured as enzymatically liberated volatiles, for 4-sulfany-4-methylpentan-2-one and 3-sulfanylhexan-1-ol in Sauvignon blanc grapes. These volatile thiols showed a maximum in grape musts of vines facing mild water stress in the ripening period (Peyrot des Gachons et al., 2000; Choné et al., 2006; Peyrot des Gachons et al., 2005). Another study, carried out on grape vine fruiting cuttings, also showed that the GSH-S-prc concentrations increased as a reaction to water deficit in the varieties Chardonnay, Koshu and Merlot, but in this study the precursor were detected in vine leaf extracts (Kobayashi et al., 2011). In this context, it can be considered that the water status of the non-irrigated vines, especially in the vintage 2009, was not favourable for an increased biosynthesis of GSH-S-prc concentrations in the grape must. In 2009, regarding defoliation treatments on precursor molecules cys-S-prc concentrations, an increase was observed with increased time of defoliation, whereas inversely a decrease was observed for GSH-S-prc concentrations. For 2010, grape must of the treatment being defoliated at flowering showed higher precursor concentration, but a slightly lower concentration of assimilable nitrogen concentration, in comparison to the not defoliated treatments. Higher assimilable nitrogen concentrations were observed in the grape musts obtained from treatments showing lower water status especially for the vintage 2009. As indicated earlier, this could either be caused by smaller berry size or by advanced maturity and senescence (Keller, 2005). In the case of distribution of nitrogen compounds into the berries on a per berry ratio, this would lead to a concentration of nitrogen compounds

Viticultural studies

138

in the berries, due to observed smaller berry size. As senescence and leaf abscission was observed late in the ripening period for the vintage 2009, it seems to be more likely that this effect is caused by redistribution of nitrogen from the leaves into the berries, due to senescence as outlined in Keller (2005). The amount of assimilable nitrogen was not directly correlated to precursor concentrations, which could indicate that different viticultural parameters, such as water status, leaf removal and mould are likely to interact in precursor’s biosynthesis. Furthermore, one has also to consider that precursor concentrations in grape must can undergo changes due to pre-fermentative steps between the vineyard and the fermentation tank. It is reported that harvest technology (Allen et al., 2011), grape transport (Capone et al., 2011), pressing conditions (Allen et al., 2011; Nikolantonaki et al., 2012a; Mattivi et al., 2012) and must oxidation (Roland et al., 2010) can impact precursor concentrations. In all these studies, the changes are mostly observed concerning the concentartions of GSH-S-prc but less for cys-S-prc.

3.3.10.3 Discussion on free thiol concentrations in young wines In young wines, significant differences were observed with respect to viticultural treatments. In the vintages 2008 and 2009 significant lower concentrations of 3SH were observed in wines from non-irrigated treatments in comparison to wines from irrigated treatments. The results were repeatable for two different yeast strains for each vintage. Until now, no information on the effect of the vines’ water deficit on free thiols in wines was published. Publications on water stress concerning thiols and their precursor rely on grapes’ aromatic potential (Peyrot des Gachons et al., 2005) or on direct precursor measurement in grapes (Kobayashi et al., 2010). According to these results, one would conclude that more thiols would have been released, when precursor concentrations were elevated in grape musts (Kobayashi et al., 2010). This point already was criticised and alternatively a correlation between free thiols and the proportion of metabolised cysteinylated precursors was considered by different authors (Tominaga, 1998c; Peyrot des Gachons et al., 2000; Pinu et al., 2012). In order to contemplate the results for free thiols in wines, which were obtained from grape must of different viticultural treatments and conditions, the two main strands

i) formation of free thiols by yeast during fermentation and

ii) stability of free thiols in wine after release will be discussed.

Viticultural studies

139

i) formation of free thiols by yeast during fermentation Grape musts from the viticultural trials were shown to be modified in assimilable nitrogen concentrations by water deficit, likely as a consequence of advanced senescence with abscission. Therefore nitrogen source for yeasts during fermentation is modified through viticultural treatments, especially water deficit. For the transformation of cys-S-prc into 3SH and A3SH, transformation rates of 0.1% - 12% and for GSH-S-prc less than 5% are reported (Dubourdieu et al., 2006; Masneuf-Pomerade et al., 2006; Howell et al., 2004; Subileau et al., 2008; Thibon et al., 2008). Conversion rates in synthetic medium were observed to be lower than for real grape must (Masneuf-Pomerade et al., 2006; Howell et al., 2004). Moreover, an increase of transformation from cys-S-MSP to 4MSP and cys-S-3SH to 3SH by yeast was observed, when nitrogen source was suboptimal for yeast metabolism, influenced by the effect of the so called Nitrogen Catabolic Repression (Thibon et al., 2008). In the vintage 2008, wines from the non-irrigated control treatment (C) showed the lowest 3SH concentrations compared to those from the irrigated treatment (I), but inversely the grape must from the non-irrigated control treatment (C) showed higher assimilable nitrogen concentration than the one from the irrigated treatment (I). The same observation was made in vintage 2009. Microclimate manipulation in this vintage seemed to impact the concentrations of free thiols in wines in the same way. These results support, that the effect of Nitrogen Catabolic Repression possibly influences metabolism of volatile thiols during fermentation, although no relation between precursor molecules and liberated thiols was observed. However, it could also be considered that in those musts, which show high free thiol concentration after fermentation, other biosynthesis pathways including other precursor molecules can impact free thiol concentrations in finished wines (Subileau et al., 2008; Pinu et al., 2012).

ii) stability of free thiols in wine after release Reductive oenological practice is recommended in order to stabilise volatile thiols in wine and to avoid losses through oxidation and trapping by Michael addition reactions on quinones (Ribéreau-Gayon et al., 2006). First, addition of sulfur dioxide to grape must and to wines is one aspect of antioxidative actions applied in oenology. Water deficit was shown to increase phenolic compounds in grapes (Roby et al., 2004). Second, phenolic compounds, like hydroxycinnamic acids and catechols, can be oxidised to quinones mostly by means of the grape’s polyphenoloxidase and heavy metal catalysed Fenton reactions. These oxidation products theirselves can react with free thiols (Michael addition reaction) and therefore decrease their concentration in wines (Nikolantonaki et al., 2012). These phenol oxidations can be limited by the use of carbon dioxide and sulfur dioxide (Nikolantonaki et al., 2012a).

Viticultural studies

140

Therefore water deficit on grape vines could theoretically indirectly lead to scalping of free thiols in corresponding finished wines. Considering the way of wine elaboration, these points should have limited the impact of free thiols stability in wines.

3.3.11 Essential outcomes regarding thiols and their precursor molecules This part of the project permitted to study effects of viticultural conditions, focused on the water status and the microclimate manipulation, on several metabolites being susceptible to impact volatile thiol concentrations in wines. These observations were done in different stages of the growing period including berry samples, grape must, finished wines, and aged wines. Concentrations of reduced glutathione in berries generally were shown to increase after veraison, whereas they were shown to decrease, when the grape vines suffered from severe water deficit. No significant effects have been observed on this parameter related to leaf removal in the fruit zone. Low concentrations of reduced glutathione were correlated to high concentrations of the glutathionylated precursor of 3-sulfanylhexan-1-ol in berries when vines were grown under water stress, compared to berries issued from vines grown under moderate water deficiency. No statistical significant accumulation of the glutathionylated precursor of 3-sulfanylhexan-1-ol was shown, albeit a tendency to do so can be observed. Surprisingly, very low concentrations of the cysteinylated precursor of 3-sulfanylhexan-1-ol were present in grape berries. The results of this study indicate that the biosynthesis and accumulation of cysteinylated precursor of 3-sulfanylhexan-1-ol takes place in the berry after harvest and before pressing. The crucial role of endogenous glutathionylated precursor in the berries as a precursor for the cysteinylated precursor of 3-sulfanylhexan-1-ol is considered, due to matching concentration pattern of these two metabolites in berry extracts and in grape must. During this process the formation of glutathionylated precursor of 3-sulfanylhexan-1-ol seemed to be linked to the concentrations of reduced glutathione in grape berries. The formation of the Grape Reaction Product (GRP; 2-S-glutathionyl caftaric acid) and the decrease in caftaric acid concentration did not show any relation neither to cys-S-prc nor to GSH-S-prc metabolism. In grape musts glutathionylated precursor concentrations were higher in grape musts obtained from vines grown under weak to moderate water deficit, whereas severe water deficit and no water deficit at all was not favourable. Defoliation seemed to decrease glutathionylated precursor concentrations, whereas it increased cysteinylated concentrations in grape musts.

Viticultural studies

141

Moreover, no link was observed between assimilable nitrogen concentrations and concentrations of cysteinylated or glutathionylated precursor molecules, whereas assimilable nitrogen seemed to be linked to free thiols concentrations in wines. In total this part of the study showed that viticultural conditions like water deficit and microclimate manipulation directly and indirectly impact free thiol concentrations in wines. Weak to moderate water deficit during the ripening period were considered to favour 3-sulfanylhexan-1-ol formation in wines, whereas effects of microclimate manipulation did impact indifferently.

3.3.12 Sensory aspects In triangle tests the wines from the non-irrigated treatment (C) were significant different from the irrigated treatment (I) and the one which was irrigated and defoliated at veraison (IdV) at a significance level of α = 0.05 (p < 0.05).

Fig. 3-14 Spider web of results of QDA for finished wines of the viticulture experiment from the vintage 2009 after 12 months of bottle storage; with C: non-irrigated control; I: irrigated; IdV: irrigated and defoliated at veraison; no statistical significant differences in ANOVA for single descriptors Then Quantitative Descriptive Analysis (QDA) was carried out on the wines after 12 months of storage in order to clarify the perceived differences in the triangle test. In QDA (Fig. 3-14), it was proposed to rank the level of intensity of various descriptors used for describing Riesling wines. No significant different rating for descriptors was observed, whereas the

0

1

2

3

4

5apple

lemon

grapefruit

melon

peach

abricotfloral

petrol / kerosene

minerality

oxidative

reduced

non-irrigated control

irrigated treatment

irrigated treatment, defoliated at veraison

Viticultural studies

142

treatments C and I were less pronounced in petrol / kerosene attribute than IdV. Grapefruit and lemon aroma were tendencially strongest in C, whereas IdV had tendencially strongest minerality character and was less pronounced in floral and fruity attributes as abricot, peach and melon.

3.3.13 Discussion on sensory aspects Varietal character of wines and therefore viticultural influences on wines’ aromatic expression is more likely to be observed in wines, which have been bottle stored for a certain period of time than in wines directly tested after finished fermentation. It was observed that acetate esters, which are mainly responsible for the so called ‘fermentative aroma’, are formed during alcoholic fermentation and degrade during the storage by hydrolysis, due to equilibrating the chemical equilibrium in wines (Rapp et al., 1984). Then, when fermentative aroma deminuishes and aging aroma has not developed strongly at the same time, varietal character of wines is considered to be perceived more intense than in very young wines, showing intense fermentative aroma, or than in very old wines, showing intense aging aroma. In twelve months bottle stored wines significant differences were detected orthonasally between wines from the viticulture treatments C and I as well as between wines from C and IdV in their global aromatic expression. Compared to analytical data this coincidenced to significant differences for linalool (129 µg/L, 87 µg/L and 110 µg/L for C, I and IdV respectively), TDN (0.7 µg/L, 1.2 µg/L and 1.6 µg/L) and 3SH (156 µg/L, 328 µg/L and 327 µg/L). Despite differences for single descriptors have not been significant, it was observed that the defoliated treatment (IdV) showed a tendency towards more pronounced petrol or kerosene like odour, which coincidenced with higher TDN concentrations close to the recently determined odour detection threshold of 2 µg/L (Sacks et al., 2012), which was confirmed during this study at 2.7 µg/L. Moreover, despite lower concentration of 3SH, wines from the non-irrigated control treatment showed a tendency to more pronounced grapefruit and lemon odour. In conclusion, the present study shows the impact of viticultural conditions and practices on the final olfactory properties of the product wine. Through influence of these factors on final concentrations of typicality related aromatic compounds in Riesling (compare chapter 1), the overall aromatic perception changes and interacts with typicality appreciation of Riesling wines.

Viticultural studies

143

3.4 Conclusion on viticultural influence on Riesling wines’ aromatic aromatic expression

Following to a number of works, this study helps to demonstrate the impact of changes in viticultural conditions on modifications on the aromatic expression of finished wines, which could impact varietal typicality, by both chemical and sensory analysis. Particularly, through the establishment of a sampling procedure which allowed to monitor changes of aroma compounds and precursor molecules concentrations at several levels of production (during grape maturation, during the pre-fermentative operations, after wine elaboration and aging) this research led to valuable insights. Regarding terpenols and C13-norisoprenoids, an effect of grape vines water status on their concentrations in the finished wines was clearly demonstrated. Especially in the case of the C13-norisoprenoid 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN), the influence of viticultural practices is considered to play an important role in the formation kinetic of this compound during bottle aging, highlighting also the interest of integrating bottle aging of experimental wines into these kind of studies. This got more obvious as the value of an aromatic potential regarding these aroma compound classes, obtained from grape must, did not picture the situation in the finished wines sufficiently. Regarding the powerful aroma compound 3-sulfanylhexan-1-ol, it was demonstrated that the production step between harvest and pressing is likely to influence concentrations of the cysteinylated precursor molecules of 3-sulfanylhexan-1-ol, and that its concentration in grape must is somehow linked to its glutathionylated precursor in the grape berries. However, in the finished wines concentrations of 3-sulfanylhexan-1-ol were shown to be related to viticultural conditions, whereas the mechanism of the influence still remains unclear. A possible indication of yeast assimilable nitrogen concentrations in grape musts on 3-sulfanylhexan-1-ol concentrations in wines is considered.

144

Chapter 4

4 Aroma compound identification

Aroma compound identification

145

4.1 Introduction In chapter 1, it was shown by a targeted study that several aroma compounds are likely to influence Riesling wines aromatic typicality. In chapter 2 some of their concentrations in wine were shown to be influenced by environmental factors in the vineyard. Generally it should be considered, that other aroma compounds than those included in the targeted approach, are involved in the typical aroma of Riesling wines. In order to progress on this subject, various analytical techniques such as wine extract fractionation associated with sensory analysis, hyphenated gas chromatographic technique coupled to both, mass selective (GC-MS) and olfactory detection (GC-O) can be used. In the past, GC-O approaches were used to characterise odour active compounds in Riesling wines from North America and Croatia. As results, mainly odoriferous compounds as ethyl and acetate esters, terpenols and higher alcohols were evidenced (Chisholm et al., 1994; Komes et al., 2006). These compounds were included in the targeted approach, which was applied in chapter 1. But, it should be noted that while Chisholm and co-workers used a comparative analysis between Riesling and French-American hybrid wines, they did not develop an approach based on the concept of typicality. In this study, it was demonstrated that under these conditions wines from the variety Riesling and Sauvignon blanc style wines differed extremely in their perceived typicality. So, it is considered that in order to decrypt typical Riesling wines’ aroma compounds a comparative sensory approach between representative wines from those two varieties (Riesling and Sauvignon blanc) through all the steps of wine extraction, fractionation and analysis (particularly GC-O and GC-MS) can be used as a novel strategy. For example, this approach was successfully applied to elaborate aroma compounds abundant in sweet dessert wines from the appellation of Sauternes in comparison to dry white wines from the Bordelais (Sarrazin, 2007; Sarrazin et al., 2008). It is considered that this comparative strategy, implicating sensory analysis at all stages, would facilitate the identification of new marker compounds implicated in Riesling wines’ typical aroma.

Aroma compound identification

146

4.2 Material and methods

4.2.1 Global aroma extracts 200 mL (500 mL; or 750 mL) of the wines (see Table 4-1) were extracted successively with 30 mL, 30 mL and 20 mL (50 mL, 50 mL and 30 mL; or 70 mL, 70 mL and 50 mL respectively) of dichloromethane using magnetic stirring (at 500 rpm) in flasks and decantation in a separation funnel. The combined organic phases were dried using anhydrous sodium sulfate and then got gently concentrated to a volume of approximately 20 mL using a rotary evaporator (waterbath 20 °C; 500 kPa). Subsequently, the aroma extract was concentrated under a gentle nitrogen streem (~ 100 mL/min) to 0.2 mL (0.5 mL; 1 mL) and stored at -18 °C for further use. Table 4-1 Wines used for obtaining global aroma extracts

no. wines variety origin winery vintage

1 Von-Lade QbA trocken Riesling D-Rheingau Weingut der

Forschungsanstalt Geisenheimm

2009

2 Von-Lade QbA trocken Riesling D-Rheingau Weingut der

Forschungsanstalt Geisenheimm

2010

3 Berg Rottland Riesling D-Rheingau Hessische

Staatsweingüter – Kloster Eberbach

2009

4 Berg Rottland Riesling D-Rheingau Hessische

Staatsweingüter – Kloster Eberbach

2010

5 Château Moutin Sauvignon blanc (90 %), Semillon

(10 %)

F-Bordeaux (Graves) Vignobles Darriet 2008

6 Domaine de la Maroutine, Bordeaux blanc sec Sauvignon blanc F-Bordeaux

(Loupiac) Vignobles Darriet 2010

7 Château Bonnet blanc

Sauvignon blanc (55 %),

Semillon (35 %), Muscadelle (10 %)

F-Bordeaux (Entre Deux

Mers) Château Bonnet 2009

8 Château Bonnet blanc Sauvignon blanc

(65 %), Semillon (35 %)

F-Bordeaux (Entre Deux

Mers) Château Bonnet 2010

4.2.2 Thiol selective aroma extracts Thiol selective aroma extracts were prepared according to the method previously described (Tominaga et al., 1998b).

Aroma compound identification

147

4.2.3 Fractioning global aroma extracts by means of high pressure liquid chromatography (HPLC) according to Ferreira et al. (1999) and Pineau et al. (2009) 200 µL (250 µL) of the global aroma extract (4.2.1) were injected on a C18-semi preparative column and separated along a H2O - ethanolic gradient. After passing the column, the effluent was collected changing the recipient (capped vials preloaded with nitrogen as inert gas to avoid oxidation) every minute. In total 60 fractions were collected and kept at -18 °C in the dark until further usage (conditions shown in Table 4-2 and Table 4-3) Table 4-2 Chromatographic conditions for the fractionation of global aroma extracts

Auto sampler Dionex Ultimate 3000 Autosampler Injection Volume 200 µL (250 µL)

Pump Dionex Ultimate 3000 Pump Flow 1 mL/min Gradient see gradient table 4-3

Fraction collector Dionex Ultimate 3000 Autosampler

Column Waters Nova-Pak® HR C18, 6 µm 60 Ǻ, 7.8 x 300 mm Pre-Column Varian MetaGuard 4.6 mm, Polaris C18-Ether 5 µm Temperature 20 °C

Solvent A H2O Solvent B ethanol (distilled freshly)

Table 4-3 Solvent gradient programme for the fractionation of global aroma extracts

runtime t [min] solvent A [%] solvent B [%] 0 100 0 2 100 0 19 76 34 85 0 100 95 0 100 96 100 0 106 100 0

4.2.4 Sensory analysis of aroma extract fractions The fractions obtained were poured into wine tasting glasses (ISO, 1977), and evaluated in batches of ten glasses by a sensory panel consisting of 3 trained panelists for differences between extracts from Riesling wines and Sauvignon blanc style wines. The fractions obtained in 4.2.3 were diluted using H2O resulting in 8.7 mL of 12% (v/v) hydro-alcoholic solutions. These solutions were poured into wine tasting glasses (ISO, 1977), and

Aroma compound identification

148

evaluated in batches of 10 glasses by a sensory panel consisting of 3 trained panelists for differences between extracts from Riesling wines and Sauvignon blanc style wines.

4.2.5 Triangle tests of aroma extract fractions and reconstituted aroma extracts

i) Aroma extract fractions of one typical Riesling wine and one typical Sauvignon blanc style wine (4.2.3) were diluted using H2O resulting in 15 mL of 12 % (v/v) hydro-alcoholic solutions.

ii) Aroma extract reconstitution solutions (one partially reconstituted (RI-F54/F55), one totally reconstituted (RI) and one totally reconstituted and supplemented aroma extract (RI+F54/F55)) of one typical Riesling wine were prepared from the collected aroma extract fractions (4.2.3) in 15% hydro-alcoholic solution, representing an aroma extract equivalent of 180 mL of wine.

Table 4-4 Preparation of aromatic reconstitution solutions

reconstituted solutions fractions % EtOH (v/v) RI-F54/F55 F1 to F53 + F56 to F60 15

RI F1 to F60 15 RI+F54/F55 F1 to F60 + F54 + F55 15

These solutions were presented in wine tasting glasses (ISO, 1977) in form of triangle test. These triangle tests were conducted in order to detect differences in the global aromatic expression by 22 and 16 experienced judges for i) and ii) respectively. Statistical significance was evaluated according to Lawless and Heymann (2010).

4.2.6 Quantitative Descriptive Analysis (QDA) of aroma extract fractions and reconstituted aroma extracts In the descriptive task, the solution prepared in 4.2.5 were presented in wine tasting glasses (ISO, 1977) and the panels were asked to rate the solutions for the descriptors fruity and floral on a non-segmented scale of 10 cm length anchored with ‘not’ and ‘very’ at the left and right end respectively. One factorial ANOVA was carried out using Sigma PlotTM (Systat Software GmbH, Erkrath, Germany) on the results.

4.2.7 Liquid-liquid micro extraction of aroma extract fractions The fractions (4.2.3) were diluted using H2O resulting in 8.7 mL of 12 vol.% alcoholic solutions. These solutions were re-extracted successively three times with 1 mL of dichloromethane. The organic phases were unified and dried over anhydrous sodium sulfate

Aroma compound identification

149

and then got concentrated under a gentle nitrogen stream to 200 µL. These extracts were stored at -18 °C until further usage.

4.2.8 Gas chromatography coupled to olfactometric detection (GC-O) For gas chromatography olfactory analysis 2 µL of the organic aroma extract or aroma extract fraction micro-extract were injected and were sniffed by 3 trained panelists using chromatographic conditions according to Table 4.5 on two capillary columns of different polarity. The detection was performed at the exit of the capillary column by trained panelists. The sniffing port ODO 1 allows olfactory sniffing of a cooled and humidified gaseous effluent thanks to both, a Venturi effect and the bubbling of nitrogen in distilled water. Capillary column comes just at the upper limit of the ODO 1 installation and there is no necessity of make up gas. The flow of nitrogen used for the Venturi effect is close to 600 mL/min and for humidification, the nitrogen flow bubbling through water is close to 90 mL/min. The flows are controlled by needle valves. This strenuous exercise can be performed for more than forty minutes (Darriet et al., 1991). Each panelist sniffed for 30 Minutes. They were asked to indicate a detected stimulus and name a descriptor. Stimuli detected by 2 of 3 panelists were considered as odouriferous zones (OZ). Table 4-5 Chromatographic conditions for 1D-GC-O

Autosampler Injector Mode Purge Flow Temperature Liner Gas Chromatograph Capillary Column Carrier Gas Oven Program Detector

manual Injection split - splitless Injector (Agilent technologies, USA) splitless (1 min); constant temperature 50 mL/min

230 °C SGE Flow Gooseneck GC 5890 Series II (Hewlett Packard, USA) SGE – BP20 (SGE Europe Ltd, UK) (50m x 250 µm ID x 0.22 µm FT) Agilent-DB-5 (Agilent technologies, USA) (50m x 250 µm ID x 0.25 µm FT) Hydrogen; const. pressure 250 kPa; 3 bar forepressure 45 °C (1 min) – 230 °C at 4 °C/min; Hold 15 min 45 °C (1 min) – 280 °C at 4 °C/min; Hold 15 min olfactometric port ODO-1 – SGE – Ringbow, Australia

Aroma compound identification

150

4.2.9 Two-dimensional heart-cut gas chromatography coupled to olfactometric detection and high-resolution mass-spectrometry (GC-GC-O-HRMS) The chromatographic setup and conditions are presented in Table 4-6: Table 4-6 Instrumental settings and chromatographic conditions for GC-GC-O-HRMS

Injection mode Injector Mode Purge Flow Temperature Liner Gas Chromatograph 1 Capillary Column Carrier Gas Oven Program MCS 2 Counter flow MCS Program CTS 1 CTS Program Transferline Gas Chromatograph 2 Capillary Column Carrier Gas Oven Program Detector ODP 2 Transferline HRMS Mode Ionisation voltage ion chamber

Manual injection Split – splitless Injector (Agilent technologies, USA) Splitless (1 min); constant temperature 50 mL/min 230 °C SGE Flow Gooseneck GC 7890 A (Agilent technologies, USA) SGE – BP20 (SGE Europe Ltd, UK) (30m x 250 µm ID x 0.22 µm FT) Helium (Linde, France) Initial pressure 213 kPa (1 min) – 271.5 kPa at 1.57 kPa/min; Hold 45 min 45 °C (1 min) – 190 °C at 4 °C/min 190 °C – 240 °C at 10 °C/min; Hold 20 min 240 °C – 45 °C at 5 °C/min Multi Column Switching System (Gerstel, Mülheim a.d.R., Germany) 10 mL/min 0 – 32 min:Vent 1 ON at 107 kPa 32 – 37 min: Vent 1 OFF 37 – 94.67 min: Vent 1 ON at 107 kPa, Ramp: 0.6 kPa/min 94.67 min – 114.67 min: Vent ON at 144 kPa Cryo Trap System (Gerstel, Mülheim a.d.R., Germany) enabled constant at 210 °C GC 7890 A (Agilent technologies) Supelco SPB-1 (Supelco – Analytical, USA) (30m x 250 µm ID x 1.0 µm FT) Helium; Initial pressure 167 kPa (1 min) – 219 kPa at 1.79 kPa/min; Hold 30 min 45 °C (1 min) – 280 °C at 4 °C/min; Hold 15 min Split: 50% ODP; 50% HRMS Olfactory Detector Port (Gerstel, Mülheim a.d.R., Germany) (Split 50:50) 290 °C JEOL AccuTOFGC JMS-T 100 GC (JEOL (Europe) SAS, France) EI positive 70 eV 303 200 °C

4.2.10 Calculation of lineary retention indices (LRI) Lineary Retention Indices (LRI) were calculated according to the method proposed by Kovats (1958). The LRI is a characteristic, which allows comparing volatile compound seperation behaviour on a stationary phase, towards a series of n-alcanes. This function as a characteristic, which allows comparing LRI’s of single compounds independendly from instrumental setups. To calculate LRIs a solution of n-Alcanes is injected at the same

Aroma compound identification

151

condition as the aroma extract. Using the retention times of the single n-alcanes a compound LRI is calculated as:

100 100

with: Rt Retention time n number of C-atoms in the n-Alcane, which elutates before the analyte

4.2.11 Quantification of trans-ethyl cinnamate by means of HS-SPME-GC-MS The preparation of the sample and the chromatographic conditions were applied as described previously by Antalick et al. (2011). Quantification was done by calculating the response factor to the isotopically labelled trans-ethyl cinnamate-d5. Analysis was carried out in the Laboratory of Professor Dr. Gilles de Revel at the ISVV in Bordeaux.

4.2.12 Quantification of trans-ethyl cinnamate by means of SBSE-GC-MS Sample preparation and chromatographic conditions were according to 2.3.3. The identification and quantification was carried out using mass fragment m/z 176 as quantifier ion and m/z 131 and m/z 147 as qualifier ions. Calibration parameters are listed in Appendix 16. Analysis was partially carried out by Mrs Johanna Elise Hoppe during her Bachelor thesis at the Hochschule RheinMain in the Department of Microbiology and Biochemistry of the Geisenheim Research Center.

4.2.13 Determination of the odour - detection threshold of trans-ethyl cinnamate In order to determine the detection threshold, a wine like model solution (10 vol.% ethanol in H2O; 3 g/L tartaric acid; adjusted to pH 3 with NaOH solution) was prepared. For determination of the detection threshold in wine, a Riesling wine containing initially a low concentration of trans-ethyl cinnamate (0.2 µg/L) from the winery at the Geisenheim Research Center was used (Riesling Classic 2010). For the determination of the detection threshold in the wine like model solution and in the Riesling wine, a series of rising concentrations was orthonasally tested in duplicate by a trained sensory panel (n 26) in a paired comparative test against the not supplemented matrix.

Aroma compound identification

152

Concentration levels in wine like model solution (MS) and in Riesling wine (RW) were 0 µg/L (MS) 0.2 µg/L (RW), 0.5 µg/L, 1.0 µg/L, 2.5 µg/L, 5.0 µg/L, 7.5 µg/L and 10 µg/L. The odour detection threshold was calculated as the concentration, when 50% of panelists detected a difference between the sample and the sample’s matrix above chance level. According to Lawless and Heymann (2010) the calculated percentage of detected differences in the applied design was should be preq 75%. Experiments were conducted by Mrs Johanna Elise Hoppe during her Bachelor thesis at the Hochschule RheinMain in the Department of Enology and Wine Technology and the Department of Microbiology and Biochemistry of the Geisenheim Research Center.

4.2.14 Effect of trans-ethyl cinnamate concentrations on Riesling wines’ typicality using the A Not-A method Different wines were spiked with trans-ethyl cinnamate resulting in wines with sub-odour detection threshold concentration (no addition 0.2 µg/L - 0.3 µg/L), the concentrations around the odour detection threshold (1.5 µg/L) and the super-odour detection threshold concentration (3.0 µg/L). These wines were tested orthonasally for Riesling wines’ aromatic typicality by deciding if the presented sample is a good example of the category Riesling or not. Experiments were conducted by Mrs Johanna Elise Hoppe during her Bachelor thesis at the Hochschule RheinMain in the Department of Enology and Wine Technology and the Department of Microbiology and Biochemistry of the Geisenheim Research Center.

Aroma compound identification

153

4.3 Results and discussion

4.3.1 Comparison of the odoriferous zones in aroma extracts of typical and not typical Riesling wines by means of GC-O Following the hypothesis, that aroma extracts of representative Riesling wines show specific odoriferous zones in GC-O aromagrams in comparison to non-representative Riesling wines, comparative GC-O analysis was carried out on global organic aroma extracts. Table 4-7 Odoriferous Zones (OZ) for aroma extracts from Riesling wines of different typicality on a polar and non-polar stationary phase (SGE BP-20, Agilent DB-5)

LRI compound

odour descriptor OZ BP-20 DB-5 typical Riesling not typical Riesling

1 1131 rubber rubber 2 1148 828 2-methylpyrazine popcorn 3 1219 917 4-methoxy-2-methyl-2-sulfanylbutane catty catty 4 1223 roasted / broth catty / broth5 1232 906 3-sulfanylpentan-2-one vegetal broth – cooked onion vegetal broth6 1246 floral floral 7 1261 broth 8 1294 broth / yeasty 9 1321 chicken broth chicken broth10 1331 931 2-methyl-3-sulfanylfurane roasted meat roasted meat11 1345 chicken broth chicken broth12 1350 spicy / cleaning agent spicy / cleaning agent13 1384 965 4-methyl-4-sulfanylpentan-2-one catty / box tree catty / box tree / floral14 1407 roasted reduced / sulfur off-flavour15 1423 1072 cis-linalool oxide fruity fruity / floral16 1441 928 2-furfurylthiol roasted / spicy vegetal / cooked cauliflower17 1452 clinical / vinegar clinical / dentist / vegemite18 1462 909 vegetal vegetal / vegemite19 1492 sweat / citrus 20 1502 roasted roasted / reduced / sulfur off-flavour21 1517 1128 linalool fruity fruity 22 1527 1190 hexyl butanoate green apple (granny smith) green 23 1535 potato chips / spicy / red capsicum24 1543 green capsicum25 1559 1193 iso-butylmethoxypyrazine green capsicum green capsicum / chicken broth26 1572 broth / clinical / dentist clinical / dentist27 1586 broth 28 1622 green tea green tea 29 1626 burnt strong broth fond / burnt30 1676 ripe fruit ripe fruit 31 1681 broth 32 1695 1220 α-terpineol spicy spicy 33 1703 sweat strong spiciness34 1730 1381 3-hexenylhexanoate fruity 35 1733 sweat / vegetal sweat 36 1747 sweat / vegetal / green tea / iodine 37 1759 1244 3-sulfanylhexanyl acetate grapefruit grapefruit 38 1797 popcorn popcorn 39 1824 sea shells / ocean crustacean / ocean40 1858 1169 3-sulfanylhexan-1-ol grapefruit grapefruit 41 1881 terpenic / citrus like citrus like 42 1894 ripe fruit 43 1918 spicy / fruity citrus like / orange44 1927 1126 2-phenylethanol rose 45 1943 cabbage cabbage 46 1956 plastic plastic 47 1983 cooked fruits 48 1992 waxy 49 2058 decent citrus odour50 2121 bread crust bread crust51 2127 cooked fruits 52 2188 bread crust burnt / panfried53 2247 1613 δ-undecalactone peach / abricot dried fruit / abricot / peach54 2327 malt malt 55 2415 floral 56 2434 burnt

compounds tentatively identified by LRI on BP-20, LRI on DB-5 and odour quality in GC-O

Aroma compound identification

154

Using a polar polyethylene glycol stationary phase (SGE BP-20), a total of 56 odoriferous zones (OZ) were detected in the global Riesling wines’ aroma extracts, when sniffed by one panelist. Typical Riesling wines’ aroma extracts showed a total of 51 OZ, whereas not typical Riesling wines’ aroma extracts showed only 43 OZ. In total 40 OZ were abundant at the same LRI for both categories. Many differences were detected at high GC oven temperatures (>200 °C; LRI > 1900 (BP-20)). The odoriferous zones obtained by GC-O on a polar (SGE BP-20) and a non-polar stationary phase (Agilent DB-5) are given in Table 4-7. The reproducibility in the temperature range of > 200 °C (LRI > 1900 (SGE BP-20)) was generally poor, between the 3 panelists as well as it was for the global extracts, and therefore differences were difficult to verify. One major problem seems to be the complexity of the aroma extracts’ aromagram in the GC-O and the high number of OZ which lead to fatigue especially at the end of the 30 minutes sniffing time. From these preliminary GC-O results, it cannot be deduced that there are specific odoriferous zones in the aromagrams of the wines obtained by GC-O of global organic aroma extracts of typical Riesling wines in comparison to those of non-typical Riesling wines (Table 4-7).

4.3.2 Comparison of the odoriferous zones in thiol selective aroma extracts (TSAE) of typical Riesling and typical Sauvignon blanc style wines by means of GC-O As the complexity of global organic aroma extracts was causing problems in the reliability of detecting odours at higher oven temperatures during GC-O, preparing selective extracts by enrichment of molecules showing a certain chemical property or functional group (e.g. thiol) can be a way to eliminate complexity in the aromagram during GC-O. Sauvignon blanc style wines were shown to express poor Riesling wines typicality (chapter 1) and the thiols 4MSP, A3SH and 3SH are well understood to be aromatic marker molecules for Sauvignon blanc style wines. Therefore it was expected that thiol selective aroma extracts of representative Riesling wines would show different odoriferous zones during GC-O analysis, than representative Sauvignon blanc style wines. In total 57 OZ were detected by means of GC-O on polar polyethylene glycol stationary phase (BP20) in thiol selective aroma extracts (TSAE) of typical Riesling and typical Sauvignon blanc style wines. Typical Riesling wines TSAE showed 38 OZ, whereas those of typical Sauvignon blanc style wines showed 53 OZ on polar polyethylene glycol stationary phase (BP20) (Table 4-8).

Aroma compound identification

155

Table 4-8 Odoriferous Zones (OZ) for thiol selective aroma extracts (TSAE) from typical Riesling wines and from typical Sauvignon blanc style wines on a polar and non-polar stationary phase (SGE BP-20, Agilent DB-5)

LRI compound

odour descriptor OZ BP-20 DB-5 typical Riesling typical Sauvignon blanc

1 1168 888 ethyl pentanoate fresh / fruity sulfurous 2 1196 reduced 3 1206 863 1-hepten-3-one yeast / metallic / beer beer / beer in green bottles4 1222 911 4-methoxy-2-methyl-2-sulfanylbutane bread crust / yeast / broth bread crust5 1292 891 ethyl-2-sulfanylpropionate broom / leach / onion boxtree / sweat / broom6 1318 970 4-methyl-4-sulfanylbutan-2-one orange / floral / spicy / herbs fruity 7 1342 861 citrus onion 8 1352 861 2-methyl-3-sulfanylfurane metallic / onion yeast / metallic / brewery9 1369 988 4-methyl-4-sulfanylpentan-2-one broom box tree / broom

10 1410 bread crust / meaty / yeast meaty / broth / reduced11 1413 909 2-furfuryl thiol / 2-furanmethanthiol chicken broth / coffee roasted chicken12 1431 broom / floral floral / hey 13 1435 cabbage 14 1453 popcorn cats pee15 1465 brewery / beer cats pee 16 1472 1067 4-methyl-4-sulfanylpentan-2-ol spicy / box tree box tree / grapefruit17 1498 944 methional brewery / beer beer / yeast / bread crust / leach18 1509 998 2-methyltetrahydrothiophen-3-one bread crust broth 19 1522 metallic 20 1526 bread crust21 1536 1193 2,3-diethyl-5-methylpyrazine meaty / roasted chicken22 1554 grilled onions / coffee / breadcrust bread crust23 1565 mineral / flinty / brewery24 1580 fruity 25 1590 yellow fruit melon / papaya / spicy26 1623 mineral / slate metallic 27 1651 spicy / broth28 1672 empyromatic / flinty metallic / mineral / vegetal29 1725 empyromatic / flinty 30 1748 roasted coffee31 1769 floral / violet honey / acacia32 1779 1244 3-sulfanylhexanyl acetate broom leach 33 1803 lactic / cheese / sweat sweat / cheese34 1823 sweat / animal / cheese roasted onion / cheese35 1845 1169 3-sulfanylhexan-1-ol anis / broom / grapefruit anis / broom36 1864 spicy / broom / celery / leach anis / animal / onion37 1873 roasted onion animal / bread crust38 1891 yellow fruit 39 1902 fruity / melon fruity 40 1930 earthy / thiol passion fruit41 1941 leach / onion42 1947 potato chips43 1956 citrus cest orange extract44 1972 candied fruit / grapefruit45 1984 grapefruit fruity 46 1998 grapefruit grapefruit 47 2024 fruity / mango / passion fruit48 2058 flinty 49 2112 honey / quince yelly50 2123 grapefruit 51 2154 herbal / onion celery / maggi spicer52 2172 vegetable soup53 2224 spicy 54 2238 fruit jam 55 2267 candied fruit cooked carrots56 2285 fresh fruit / candied fruit57 2313 fruity

compounds tentatively identified by LRI on BP-20, LRI on DB-5 and odour quality in GC-O

On a non-polar (5%-Phenyl)-methylpolysiloxane stationary phase (DB-5-MS) only a total of 25 OZ were detected in GC-O analysis. In typical Riesling wines TSAE 22 OZ were detected, whereas 21 OZ were obtained in typical Sauvignon blanc style wines’ TSAE. TSAE of typical Sauvignon blanc style wines showed a higher number of OZ than TSAE from typical Riesling wines. This observation was more abundant using the polar stationary phase (SGE BP-20) than using the non-polar stationary phase (Agilent DB-5-MS). With the polar configuration, especially in the regions from LRI’s 1522 to 1580, LRI’s 1930 – 1984, and

Aroma compound identification

156

LRI’s 2024 – 2238, more OZ were detected for Sauvignon blanc style wines’ TSAE, whereas the three marker molecules for Sauvignon blanc style wines were also detected in the Riesling wines’ TSAE (Table 4-8). Generally, the hypothesis which suggests different OZ in TSAE of typical Riesling wines in comparison to typical Sauvignon blanc / Semillon blended wines could be confirmed. Albeight there was no OZ reminding specificly typical Riesling wines’ aromatic expression and which then was not abundant in Sauvignon blanc style wines’ TSAE.

4.3.3 Comparison of the aroma extract fractions (AEF) of typical Riesling wines and typical Sauvignon blanc style wines Following an alternative approach in order to compare varietal wines’ aroma, global organic aroma extracts were fractioned using a semi preparative HPLC configuration (RP-C18, H2O - ethanol gradient), resulting in an order of elution from polar compounds in the earlier fractions and less polar compounds in the late fractions. Developed by Ferreira et al. (1999), and then adapted by Pineau et al. (2009), this approach has been used with success in the work group in some following published works, in order to evidence odoriferous compounds contributing to qualitative aspects of wine flavour or off-odours (Pons, 2006; Pineau et al., 2009; Sarrazin et al., 2010; Nikolantonaki and Darriet, 2011). After fractionation, the obtained fractions of the wines’ aroma extracts (AEF) were evaluated for differences between fractions obtained from Riesling wine global aroma extracts and those obtained from Sauvignon blanc style wines global aroma extracts by means of orthonasal assessment by a panel of 3 trained panelists with the objective to describe the single fractions’ odours. The same fractions of different wine styles were supposed to express different odours. The orthonasal descriptors of the AEF from typical Riesling wines were compared to those of the corresponding AEF of typical Sauvignon blanc style wines. The fractions showing different odours and being recognised as typical, either for Riesling or Sauvignon blanc style wines, were chosen to be candidates for comparative GC-O analysis. The 3 panelists agreed in this preliminary screening, that from 41 tested fractions 6 fractions clearly showed differences (Fractions 27; 30; 40; 41; 54; 55), 4 fractions showed similar odoriferous expression (Fractions 29; 36; 42; 43) and for the other 31 fractions no consensus was achieved between the panelists. Especially fractions 54 and 55 showed a pleasant fruity expression in Riesling wines’ extracts, which was clearly not abundant in Sauvignon blanc style wines’ extracts (Table 4-9). Riesling aroma extracts presented more caramel related odours than Sauvignon blanc style wines’ aroma extracts, especially in fractions 27 and 30. Fractions 40 and 41 expressed aromatic nuances of torrified coffee and phenolic odour in the Riesling wines’

Aroma compound identification

157

extracts, which were not detectable in the corresponding fractions of Sauvignon blanc style wines aroma extracts. Table 4-9 Odour descriptors for aroma extract fractions of typical Riesling and typical Sauvignon blanc wines

Fraction Odour descriptor

typical Riesling typical Sauvignon blanc

19 fruity - soapy / fruity / floral - terpenic lemon – grapefruit / fruity – passionfruit – citrus fruit / lemon -mineral

20 soapy / - / citrus fruit - / - / - 21 - / caramel – butter / cheese - champignon floral / - / butyric acid 22 - / fatty acids / cheese - champignon - / - / - 23 plastic / methional / dust - / - / - 24 - / fatty acids / citrus fruit - / - / - 25 - / biscuit – meaty / ethylacetate plastic / methional / herbal 26 ethyl acetate / ethyl acetate / ethyl acetate - / vegetal / cheese 27 candy floss / caramel – biscuit / - - / - / - 28 - / biscuit – caramel – butter / citrus fruit tomato leaf / caramel – butter / heavy 29 fruity / fruity – melon / red fruits – strawberry – cherry fruity / melon / ethyl acetate 30 caramel / fruity – peach / sweaty - / - / champignon 31 caramel / - / - - / - / - 32 - / fatty acids / fruity - / - / fruity 33 iso-amyl alcohol / iso-amyl alcohol / iso-amyl alcohol lemon – mint / lactic – yoghurt – lemon / lemon34 iso-amyl alcohol / cheese – sweaty feet / rose iso-amyl alcohol / iso-amyl alcohol / sweaty - candy floss-

iso-amyl alcohol 35 rose / floral – lemon /rose - / - / - 36 lemon / fruity / greenish lemon / grapefruit / spicy 37 - / - / spicy - / - / - 38 fruity / heavy / - - / meaty – chicken brooth / - 39 vanille / - / fruity - / phenol / citrus fruit 40 herbal / torrified coffee / caramel spicy / - / - 41 mint / fruity – phenol / - butter / - / - 42 mint / fruity – passion fruit / fruity fruity / grapefruit / - 43 fruity – mint / fruity – melon – peach / buttery fruity - pear / fruity – melon – peach / caramel 44 herbal / lemon / - fruity / - / - 45 floral - sugar / anis – mint / plastic - / anis – mint / - 46 herbal / lime / - - / lime / - 47 floral / grapefruit / greasy - / lemon / - 48 fruity – linalool – geraniol / lime anis / citrus fruit lemon / anis / phenol 49 fresh fruit – lemon / fennel / fruity - apple plastic – spicy / fruity / fruity – spicy – mint 50 - / lemon / iso-amylacetate - / fruity – floral / iso-amylacetate 51 - / - / fruity herbal / - / - 52 fruity / floral / - fruity / herbal – lemon / herbal - thiol 53 fruity – floral / fruity – floral / - floral / floral / phenol 54 fruity / lemon – fruity / - - / floral / - 55 fruity / fruity / lemon - / floral / - 56 candied / candied fruit / terpenic acetoin / fruity / lemon - linalool 57 ripe fruit / fruity / terpenic fresh fruit / fruity / terpenic 58 herbal / - / - fruity / - / terpenic 59 - / - / - - / - / - 60 - / - / - - / - / -

fractions indcated in bold show differences; Fractions indicated in italics and bold show similar odours

Although wine styles from the varieties Riesling and Sauvignon blanc style wines showed highly different aromatic expression, the number of AEF showing no clear differences was surprisingly high. Anyway, these results led to the confirmation of the hypothesis that some AEF of typical Riesling wines showed different aromatic expression in comparison to the corresponding AEF obtained from typical Sauvignon blanc style wines. Especially fractions 54 and 55 in Riesling wines extracts showed odours reminiscent to Riesling wines. Prior to selection for GC-O analysis, these fractions were tested by a bigger sensory panel in order to verify these preliminary results.

Aroma compound identification

158

4.3.4 Comparison of the aroma extract fractions (AEF) of typical Riesling wines and typical Sauvignon blanc style wines by means of triangle tests and Quantitative Descriptive Analysis (QDA) – Fractions 54 and 55 Hydroalcoholic solutions of AEF 54 and 55 of a typical Riesling and a typical Sauvignon blanc style wine, representing an equivalent of 180 mL of wine, were tested orthonasally in triangle tests by 22 judges. These fractions were judged to be different at a significance level of 1%, verifying the results from 4.3.3. The hydroalcoholic solutions containing AEF 54 and AEF 55 showed tendencially stronger fruity and floral odours for Riesling wines aroma extract fractions (Fig. 4-1), although results were not statistically significant (α = 0.05).

Fig. 4-1 Boxplots showing QDA results for descriptors ‘fruity’ and ‘floral’ for 12.7%vol model solutions of fractions 54 and 55 from one typical Riesling (RI) and one typical Sauvignon blanc style wine (SB) (wines 1 and 8; Table 4-1); judgements done on non-segmented scale, normalised to 100 in the figure

4.3.5 Comparison of the aroma fraction reconstitution for typical Riesling wines and typical Sauvignon blanc style wines by means of triangle tests and Quantitative Descriptive Analysis – Fractions 54 and 55 Hydroalcoholic solutions of total reconstituted aroma extracts (RI), partial aroma extract reconstitution (RI-F54/F55) and total aroma extracts reconstitution supplemented with the fractions 54 and 55 (RI-F54/F55) of a typical Riesling wine, representing an equivalent of 180 mL of wine, were tested orthonasally in triangular tests by 16 judges. Aromatic reconstitution solutions RI and RI+F54/F55 were rated to be significantly different from aromatic reconstitution solution RI-F54/F55 at a significance level of (α = 0.05).

0

10

20

30

40

50

60

70

80

90

100

RI SB

fruity

0

10

20

30

40

50

60

70

80

90

100

RI SB

floral

Aroma compound identification

159

In these aromatic reconstitution experiments, hydroalcoholic solutions without AEF 54 / 55 showed a tendency to lower ratings for the descriptors ‘fruity’ and ‘floral’ than total reconstituted aroma extract solutions. Doubling these fractions led to higher ratings for ‘fruity’ also, but inversely led to lower ratings for the descriptor ‘floral’ (Fig. 4-2). According to these results it can be concluded that the AEF 54 and AEF 55 likely impact Riesling wines’ aromatic expression. Therefore, the aroma compounds in these AEF should be considered to have an influence on Riesling wines’ aromatic typicality.

Fig. 4-2 Boxplots showing QDA results for the descriptors ‘fruity’ and ‘floral’ for reconstitution experiments with and without fractions 54 and 55 from typical Riesling (R) (wine 1; Table 4-1); judgements done on non-segmented scale, normalised to 100 in the figure

4.3.6 Comparison of the odoriferous zones (OZ) in aroma extract fractions (AEF) of typical Riesling wines and typical Sauvignon blanc style wines using GC-O – Fraction 55 Supposing that fractions of aroma extracts showing different sensory properties, exhibit also different odoriferous zones in GC-O analysis, fraction 55 obtained from Riesling wines’ aroma extracts and from Sauvignon blanc style wines were used for GC-O analysis. For the fraction 55 of typical Riesling and typical Sauvignon blanc style wines in total 27 OZ were detected by means of GC-O on polar polyethylene glycol stationary phase (BP20). Typical Riesling wines AEF 55 showed 20 OZ, whereas those of typical Sauvignon blanc style wines showed 19 OZ on polar polyethylene glycol stationary phase (SGE BP20) (Table 4-10). Whereas most powerful OZ were detected in AEF 55 of both wine styles (e.g. OZ 3, 4, 6, 10, 12, 18 and 23) only one powerful OZ was detected in AEF 55 obtained from Riesling wines (OZ 27) and exhibited a strong odour described as ‘floral’ and ‘fruity’.

0

10

20

30

40

50

60

70

80

90

100

Ri‐F54/F55 RI RI+F54/F55

fruity

0

10

20

30

40

50

60

70

80

90

100

Ri‐F54/F55 RI RI+F54/F55

floral

Aroma compound identification

160

Table 4-10 Odoriferous Zones (OZ) for aroma extracts from AEF 55 obtained from typical Riesling and typical Sauignon blanc style wines on a polar stationary phase (SGE BP-20)

OZ LRI odour descriptor

BP-20 typical Riesling typical Sauvignon blanc 1 855  solvent solvent 2 882  ethyl acetate ‐ 3 890  ethyl acetate ethyl acetate 4 894  ethyl acetate, fruity ethyl acetate, fruity 5 956  ‐  reduced, metallic 6 1029  metallic, beer metallic, beer 7 1132  ‐  onion 8 1153  onion  dusty, metallic 9 1222  ‐  burnt, caramel 10 1234  garlic  reduced, sulfurous 11 1245  sweet, burnt sugar ‐ 12 1258  fruity, orange fruity, box tree 13 1283  green, old cut grass ‐ 14 1304  ‐  leach 15 1319  beer, metallic beer, metallic, yeast 16 1385  ‐  melon, fresh peach 17 1413  herbal, dusty, hey burnt 18 1552  reduced leach, onion 19 1564  garlic  ‐ 20 1665  ‐‐  flint stone 21 1741  flint stone ‐ 22 1816  cheese broom, fruity 23 1866  fruity  fruity 24 1910  fried potato ‐ 25 2007  floral, rose floral 26 2020  ‐  pine tree 27 2048  floral‐fruity ‐ 

OZ in bold indicating similar odours by 3 panelists; OZ in bold and italic indicating differing odours by 3 panelists 

Therefore, the hypothesis, which suggested different OZ in differing aroma extract fractions of typical Riesling wines in comparison to typical Sauvignon blanc style wines, could be confirmed for the AEF 55.

4.3.7 Identification of trans-ethyl cinnamate

4.3.7.1 One-dimensional GC-O/HR-TOF-MS In order to identify the aromatic compound responsible for OZ 27, one dimensional GC-MS experiments were carried out on AEF 55 using a polar polyethylene glycol stationary phase (Agilent DB-Wax).

Fig. 4-3 TIC of monodimensional GC-O/MS (DB-Wax; 50m*0,22 mm*0,25 µm) of AEF 55 (with RT 41.46 corresponding to LRI 2064)

�TIC[1]; / EI+ / snif f 2

0

1000

2000

3000

Intensity (3304000 )x 103

42.51

39.62

41.4640.82 42.8042.6741.9040.44 42.2141.3540.74 41.6139.53 41.8140.54 41.1839.38 40.34 42.0039.96

OZ 27

Aroma compound identification

161

As on the equivalent column used in paragraph 4.3.6 (SGE BP-20), the selected OZ 27 was identified at similar LRI 2064 using a Agilent DB-Wax capillary column. In the total ion chromatogram (TIC) of the mass spectrometric detection a small signal could be attached to the OZ. The correspondant full scan mass spectrum (m/z 30 – m/z 250) was evaluated showing most abundant mass fragments at m/z 181, 176, 151, 147, 146, 121, 105, 104, 91, 73, 59, 45 and 43. In addition to this, the mass spectrum was quite noisy, which did not favour a reliable mass spectra comparison to a mass spectra data base (NIST® data base) and/or a calculation of molecular formula.

4.3.7.2 Two-dimensional heart-cut GC-GC-O/HR-TOF-MS As the next step in the identification of responsible odoriferous compounds in OZ 27 two-dimensional heart-cut GC-GC-O/HR-TOF-MS was applied using a polar polyethylene glycol stationary phase (Agilent DB-WAX) as a first separation step and a non-polar dimethyl siloxane stationary phase (Supelco SPB-1) in the second separation step after heart cutting. Both separations were monitored by means of an olfactometric port.

Fig. 4-4 Mass Spectra of trans-ethyl cinnamate a) Peak at LRI 2155 / LRI 1500 from AEF 55 b) NIST® data base The characteristic fruity-floral odour of OZ 27 was identified at LRI 2155 in the first separation and transferred to the second column by means of the heart-cut technique with a window of 3.5 min. The same odour was detected at LRI 1500 again in the second separation. The corresponding TIC showed a small, but good seperated signal corresponding to the LRI 1500 and the OZ in the second separation. The full scan mass spectrum showed the most abundant mass fragments at m/z 176, 148, 147, 131, 103, 91 and 77. Comparison to the

m/z m/z50 100 150 200

m/z

0

1000

Area (1396)

131.0463

176.0797

103.047077.0337

70.0738 147.0861104.0538

a) b)

trans‐ethyl cinnamate

Aroma compound identification

162

NIST® data base resulted with a 98.2% propability that the compound sought after is trans-ethyl cinnamate. This then was verified by spike experiments.

4.3.8 Impact of trans-ethyl cinnamate on Riesling wines’ aromatic expression

4.3.8.1 Quantification of trans-ethyl cinnamate in Riesling wines and Sauvignon blanc style wines In order to verify that the concentrations of trans-ethyl cinnamate were higher in Riesling wines, wines from different grape varieties (16 Riesling wines and 13 wines made from Sauvignon blanc style wines) were analysed for trans-ethyl cinnamate by means of stable isotope dilution analysis according to Antalick et al., 2011.

Fig. 4-5 Box-Plots showing concentration ranges of trans-ethyl cinnamate in Riesling (n 16) and Sauvignon blanc / Sémillon blended wines (n 13). Different letters indicate statistical difference in median values (Mann-Whitney Rank Sum Test, Confidence Intervals 95%, p < 0.001) The 16 Riesling wines analysed showed a concentration range from 0.6 µg/L to 4.5 µg/L resulting in a mean of 1.8 µg/L and a standard deviation of 1.3 µg/L. On the other hand, trans-ethyl cinnamate concentrations ranged from 0.5 µg/L to 1.1 µg/L for those 13 Sauvignon blanc style wines analysed resulting in a mean concentration of 0.7 µg/L with a standard deviation of 0.1 µg/L. These concentrations of trans-ethyl cinnamate were significantly different (p<0.001) according to ‘Mann-Whitney Rank Sum Test’ at a confidence interval of 95 % (Fig. 4-5), and showed that this compound is more likely to impact Riesling wines’ aroma, due to higher concentrations.

0

1

2

3

4

5

Riesling Sauvignon Blanc

tran

s–ethylcinn

amate[µg/L]

a b

Aroma compound identification

163

4.3.8.2 Determination of the odour detection threshold of trans-ethyl cinnamate The odour detection threshold was determined twice in both, wine-like model solution and neutral Riesling wine. In a wine-like model solution, the detection threshold was at 0.9 µg/L and 1.2 µg/L respectively (Fig. 4-6) whereas in a neutral Riesling wine, the odour detection threshold was determined at 1.6 µg/L and 2.4 µg/L respectively (Fig. 4-7).

Fig. 4-6 Determination of the odour detection threshold of trans - ethyl cinnamate in model wine solutions

Fig. 4-7 Determination of the odour detection threshold of trans - ethyl cinnamate in a Riesling wine

4.3.8.3 Correlation of trans-ethyl cinnamate concentrations with Riesling wines’ typicality In addition to the trans-ethyl cinnamate concentrations’ analysis in Sauvignon blanc style wines and Riesling wines, trans-ethyl cinnamate concentrations were also determined in

y = 0,0073e0,0639x

R² = 0,9416

y = 0,0072e0,0681x

R² = 0,9633

0

2

4

6

8

10

12

0 10 20 30 40 50 60 70 80 90 100

tran

s‐ethylcinna

mate[µg/L]

proportion of assessors detecting a difference [%]Recognition Threshold ‐model solution 1 Recognition threshold ‐model solution 2

50% above chance

c (DT) 0.9 µg/Lc (DT) 1.2 µg/L

y = 0,0127e0,0704x

R² = 0,9542

y = 0,0126e0,0642x

R² = 0,9809

0

2

4

6

8

10

12

14

16

0 10 20 30 40 50 60 70 80 90 100

tran

s‐ethylcinna

mate[µg/L]

proportion of assessors detecting a difference [%]

Recognition Threshold ‐ Riesling wine 1 Recognition threshold ‐ Riesling wine 2

50% above chance

c (DT) 2.4 µg/L

c (DT) 1.6 µg/L

Aroma compound identification

164

wines from Chapter 1 and the results were plotted against their typicality ratings assessed by the German and the French panel (Fig. 4-8 and Fig. 4-9 respectively). This data confirmed that trans-ethyl cinnamate concentrations were lower in Sauvignon blanc style wines (up to 0.8 µg/L) while they were usually higher in wines from Riesling (up to 6.4 µg/L) and for Chardonnay (up to 2.7 µg/L) (Fig. 4-8 and Fig. 4-9).

Fig. 4-8 Data set from the German panel showing trans-ethyl cinnamate concentrations in relation to their perceived Riesling wines’ typicality (Chapter 1)

Fig. 4-9 Data set from the French panel showing trans – ethyl cinnamate concentrations in relation to their perceived Riesling wines’ typicality (Chapter 1) For both panels, data show that there is no direct linear correlation between trans-ethyl cinnamate concentrations and Riesling wines’ typicality ratings. With the German panel, Riesling wines were found to be quite typical (typicality ratings from 0.5 to 0.7), when concentration ranges passed 1 µg/L, but also very typical Riesling wines (typicality ratings > 0.7) showed lower concentrations than that.

0

1

2

3

4

5

6

7

0 0,2 0,4 0,6 0,8 1

tran

s‐ethylcinna

mate[µg/L]

typicality

WinesChardonnay

Sauvignon blanc

0

1

2

3

4

5

0 0,2 0,4 0,6 0,8 1

tran

s‐ethylcinna

mate[µg/L]

typicality

Wines

Gewürz‐traminer

Pinot gris

Sauvignon blanc

Aroma compound identification

165

In contrast to the German panels’ data, in the French data set, the most typical Riesling wines (typicality ratings from 0.5 to 0.7) showed relatively high concentrations of trans-ethyl cinnamate (2.1 µg/L to 4.6 µg/L), whereas generally no correlation could be established between Riesling wines’ typicality and trans-ethyl cinnamate concentrations.

4.3.8.4 Effect of trans-ethyl cinnamate additions on Riesling wines’ typicality In order to evaluate shifts in perceived Riesling wines’ typicality, depending on different trans-ethyl cinnamate concentrations, four wines, naturally showing sub-threshold concentrations of this compound, were spiked with trans-ethyl cinnamate resulting in different concentration levels. These wines were orthonasally evaluated by a sensory panel of wine experts according to the A Not-A method as described previously (chapter 1). Results showed that additions of trans-ethyl cinnamate can impact Riesling wines’ typicality appreciation measured as proportion of the judges who indicate wines as a Riesling wine, but the effects were not shown to be significant in chi-square (χ2) statistics (α = 0.05).

Fig. 4-10 Proportions of Judges indicating wines as being a Riesling wine for four different wines being supplemented with trans-ethyl cinnamate resulting in sub-odour detection threshold (c (sub-threshold) 0.2 µg/L), odour detection threshold (c (threshold) 1.5 µg/L) and super-odour detection threshold concentration (c (super-threshold) 3.0 µg/L); error bars indicating standard errors; Ri indicating Riesling wine 1 to 3; WB indicating Pinot blanc (Weissburgunder) wine For 2 out of 3 Riesling wines (Ri1, Ri3), super-odour detection threshold concentrations of trans-ethyl cinnamate increased Riesling wines’ typicality in comparison to not supplemented wines, whereas for one Riesling wine (Ri2), initially much less typical than Ri1 and Ri3, typicality decreased. For a wine from the variety Pinot blanc (Weissburgunder), the sample

0

10

20

30

40

50

60

70

80

90

100

Ri1 Ri2 Ri3 WB

prop

ortion

sof

judges

indicating

wines

asRieslin

g [%

]

c (sub‐threshold)

c (threshold)

c (super‐threshold)

Aroma compound identification

166

containing super-odour detection threshold concentration (3 µg/L) of trans-ethyl cinnamate was more frequently attributed to be a Riesling wine (Fig 4-10). The effect that for three wines, particularly one, Riesling wines’ typicality increased with increasing trans-ethyl cinnamate concentration, whereas it decreased for another wine with increasing concentration, led to the assumption that the effect of trans-ethyl cinnamate on Riesling wines’ aromatic typicality depends on the wine’s matrix (Hoppe 2012). This could be explained by synergistic or antagonistic effects with other volatile compounds as described earlier in published works. Previous studies revealed interaction effects with other compounds. Pineau et al. (2007) showed that the presence of β-damascenone at subthreshold levels decreased the detection threshold of trans-ethyl cinnamate and Escudero et al. (2007) demonstrated that the addition of 1.22 µg/L trans-ethyl cinnamate among other compounds showed an effect towards stronger ripe fruit, honey and sweet descriptors in neutral wine. Also, Loscos et al. (2007) showed that the addition of a mixture of trans-ethyl cinnamate and ethyl dihydrocinnamate to a synthetic wine showed no significant difference towards a control, but did, when being added in combination with other goups of aroma compounds like terpenes, lactones and/or vanillins, which showed no significant difference themselves before.

4.3.9 Discussion on the role of trans-ethyl cinnamate in wine aroma Trans-ethyl cinnamate is an odoriferous ester formed during alcoholic fermentation from cinnamic acid. This latter compound is originated in the grape berry and is formed via the phenylpropanoid pathway by the action of PAL enzyme (Phenylalanine Ammonia Lyase) on phenylalanine (Versini et al., 1983). The contribution of this compound to wine aroma was considered at the end of the 1970’s, due to the observation being made that concentrations of this ester increased in red wine which were fermented by carbonic maceration, in comparison with red wines obtained by a classical crushing protocol (Versini et al., 1983; Flanzy, 1987; Bitteur et al., 1992). During the 1990’s, Etievant and co-workers in Dijon (Moio et Etievant, 1995; Aubry et al., 1997) reconsidered the organoleptic impact of trans-ethyl cinnamate and three other esters (ethyl dihydrocinnamate, methyl anthranilate and ethyl anthranilate) in Pinot noir wines through the development of a Stable Isotope Dilution Assay. With regards to trans-ethyl cinnamate concentrations ranged from 500 ng/L to 1.6 µg/L (average of 800 ng/L) a content lower than the olfactory perception threshold of this compound. Thus, the evidence for concentrations of this compound in Riesling wine, at levels it can contribute to these wines’ aroma, directly or more presumably through perceptual interaction phenomena is quite original, surprising and illustrative of the interest to correlate sensory analysis approaches to analytical chemistry.

Aroma compound identification

167

4.3.10 Effect of viticultural factors on trans-ethyl cinnamate concentrations in finished wines The influence of climatic and viticultural factors, like water deficiency and sun exposure on the concentrations of aroma compounds and aromatic expression already was demonstrated in the second chapter. In order to discuss the influence of these environmental factors on trans-ethyl cinnamate concentrations in the finished wines of different vintages, the wines were analysed using SBSE-GC-MS after 10 and 22 months of storage respectively.

Fig. 4-11 Concentrations of trans-ethyl cinnamate in finished wines depending on the vintage, water status of grape vines and grape berries’ sun exposure. (C Control; not irrigated and not defoliated experimental plot, I irrigated experimental plot, IdB experimental plot of vines, which is irrigated and defoliated at flowering and at veraison, IdV experimental plot of vines, which is irrigated and defoliated at veraison) Despite being relatively small, significant differences – except for 2008 vintage – in trans-ethyl cinnamate concentrations were detected in wines after 12 months of storage. Wines obtained from the viticultural experimental plots being not irrigated (C) showed lower concentrations in comparison to their irrigated pendants (I) (Fig. 4-11). Furthermore, wines from plots being defoliated showed significant lower concentrations in comparison to those from not defoliated experimental plots for the vintages 2009 and 2010, resulting in lowest trans-ethyl cinnamate concentrations in wines from vineyard plots being defoliated at flowering (Table 4-11; Fig. 4-11). Other studies showed alternating concentrations of trans-ethyl cinnamate in wines obtained from grapes harvested at different maturity. Concentrations decreased with higher maturity in wines from Pinot Noir (Fang and Qian, 2006) as they did also in Canadian ice wines from Riesling grapes (Bowen, 2010). It can be considered that impacted by the level of maturity, the content of trans-ethyl cinnamate in wine should be impacted by water deficiency and

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

2008 2009 2010

tran

s‐ethylcinna

mate[µg/L]

C I IdB IdV

grape cgrape c Table 4regardin

During months1.6 to 2the irrigwere ob

Fig. 4-1experimError ba The incaroma c

cluster sun composition

4-11 Conceng the facto

vintage

2008

2009

2010

C: not irrigatveraison; IdV(n = 6); diffp < 0.01); sig

bottle stora, concentra, in compar

gated viticubserved (Ta

12 Developmental plots ars represen

crease of trcompounds

exposure and aroma

entrations ors water de

treatment

C I

IdV C I

IdB IdV C I

IdB ted control; I: irV: irrigated treatferent letters ingnificance: indic

age, trans-eations of trarison to conlture treatm

able 4-11; F

pment of tra(chapter 2)

nt standard

trans-ethyl cs during the

also, as tha potential.

of trans -eficiency an

12 motrans –cinna

[µg 0.2 ( 0.1 ( 0.1 ( 0.2 ( 0.7 ( 0.4 ( 0.6 ( 0.4 ( 0.6 ( 0.5 (

rrigated treatmetment, defoliate

ndicating significating significan

ethyl cinnamans-ethyl cinncentrationsment (I) froig. 4-12).

rans – ethyl) during bot deviation (

cinnamate ie storage i

hese param

ethyl cinnad sun expo

onths – ethyl amate g/L] 0.0)a

0.0)a

0.0)a

0.0)a

0.0)b

0.0)c

0.1)d

0.0)a

0.0)b

0.0)c

ent; IdB: irrigateed at veraison; scant differencet difference in t-

mate concennamate in s analysed am the 200

cinnamate ttle storage (n=6)

is in accorn alcoholic

A

meters gene

amate in wosure (chapt

22 montrans –cinnam

[µg/L0.4 (00.2 (00.2 (00.4 (01.2 (00.8 (01.0 (0

---

ed treatment, dstandard deviat

es in ANOVA -tests with * p <

ntrations inthe wines r

after 12 mo9 vintage c

concentratfor a) Vinta

rdance withbeverages

Aroma com

erally have

wines from ter 2) during

nths

sigethyl mate L] .1)a

.0)b

.0)b

.0)a

.1)b

.1)c

.0)d

efoliated at flowtion indicated in

(post-hoc: Ho0.01, ** p < 0.0

creased sigrised signifinths. For wconcentratio

tions in winage 2008 an

h previous s. A genera

pound iden

a great im

viticulture g bottle stor

gnificance

* * ** ** ** ** ** - - -

wering and at n parentheses lm-Sidak-test;

001

gnificantly. cantly by th

wines, obtainons about

nes from vnd b) Vintag

studies conal increase

tification

168

mpact on

assays rage

After 22 he factor ned from 1.2 µg/L

iticulture ge 2009.

ncerning of ethyl

Aroma compound identification

169

esters in Riesling wines was shown by Rapp et al. (1985), whereas Gijs et al. (2002) showed an increase of trans-ethyl cinnamates’ flavour dilution value in the Aroma Extract Dilution Analysis (AEDA) in aged beer in comparison to fresh beer. Loscos et al. (2007) also showed an increase for trans-ethyl cinnamate during an accelerated ageing experiment in wines, which were obtained by fermentation of model solutions supplemented with flavour extracts obtained from different grape varieties.

4.4 Conclusion The applied comparative approach of a sensory guided selection of organic aroma extract fractions and subsequent comparative gas chromatography, coupled to olfactometry, led to the identification of the compound trans-ethyl cinnamate. This compound is a well known compound originating from the esterification of trans-cinnamic acid from grapes. Interestingly, the quantification showed significant higher concentrations of this compound in Riesling wines in comparison to Sauvignon blanc style wines. The contribution of trans-ethyl cinnamate to Riesling wines’ aromatic typicality was considered, taking into account the self-determined odour detection threshold, in model wine solution and in Riesling wine. Direct correlation of trans-ethyl cinnamate concentrations and perceived aromatic Riesling wines’ typicality could not be demonstrated, but adding trans-ethyl cinnamate did show a tendency to impact perceived Riesling wines’ aromatic typicality, which seemed to be matrix dependent. Moreover, environmental factors like water deficit and grape cluster sun exposure during berry ripening influenced the concentrations of trans-ethyl cinnamate in wines, which then increased during bottle aging. Therefore one can conclude that under certain conditions trans-ethyl cinnamate can influence perceived Riesling wines’ typicality. Even if concentrations are below the odour detection threshold in wine at 1.6 µg/L after fermentation, the compound can become important after a period of bottle aging.

General conclusion and perspectives

170

General conclusion and perspectives Aromatic perception is a major factor for the quality of wines. A wine embodies a whole bunch of intrinsic factors through its aromatic expression, but in the first place, a good wine always represents the knowledge and skills of grape vine growers and winemakers to cope with climatic conditions and to carve out varietal and wine style characteristics by their actions. Ideally, a wine shows to a high degree a characteristic flavour and taste for its category, which should be accompanied by a certain complexity as an element of originality. The concept of this ideal is the base of aromatic typicality [french typicité]. The aim of this thesis was to find answers to the initial question, whether probable changes in climate would impact the contemporary aromatic typicality of wines in general and specifically for those made from Vitis vinifera L. cv. Riesling. This study provided deeper insights on aromatic expression of typical Riesling wines from a sensory, chemical and viticultural point of view. 1) Using the concept of typicality, not only a sensory space of typical Riesling wines was demonstrated, moreover, thanks to adopting novel methodical elements, the sensory space was characterised in its descriptive dimensions. During this sensory part of the project, fundamental aspects of typicality, as what it is defined in oenology, were highlighted and have been demonstrated. Especially the possibility to work in two locations, having two panels with great expertise on different wine styles, led to valuable insights into intercultural consensus on varietal aromatic spaces of wines. This got clearly evident by including a confrontation of Riesling wines to Bordeaux white wines, which are aromatically dominated by the varietal character of Sauvignon blanc. It got obvious that a panel of wine experts, whether experienced with the wine style or not, works highly reproducible on the task of product categorisation and description, when being considered as one unit of measurement. Not surprisingly, it was clearly demonstrated, that Riesling wines’ typicality was associated with citrus fruit odours, whereas vegetal odours, including box tree odour, were more associated with Sauvignon blanc style wines. Regarding the odour naming, some differences were detected between both panels, which leads to the consideration that cultural differences in odour perception of wine experts can already occur within a European context. The application of the frequency based recognition task in form of the A Not-A test proved to be a suitable tool for typicality judging when working with a panel, which is highly trained on a product. Upon results from this sensory study and quantitative analysis of aroma compounds, marker molecules for aromatic typicality of Riesling wines were pointed out by different statistical tools. In this part, a novel multivariate approach was used to link frequented descriptive data to metric quantitative data and gave valuable information on relations between aroma compounds concentrations and odour perception. It was demonstrated for the results of both

General conclusion and perspectives

171

panels by means of Pearson product - moment correlation, that the concentration of 3-sulfanylhexan-1-ol, which is also an important marker molecule in Sauvignon blanc style wines’ aroma, correlates positively to Riesling wines’ typicality. Surprisingly, it did not correlate with the descriptor family of citrus fruit. The utilisation of a novel combination of two multivariate methodologies, including Partial Least Square Regression (PLS) with subsequent application of a Canonical Correspondence Analysis (CCA), verified these observations concerning 3-sulfanylhexan-1-ol and Riesling wines’ typicality appreciation and the lacking correlation to citrus fruit descriptors. Further it provided interesting observations concerning other aroma compounds such as linalool and 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN). Despite both aroma compounds were shown to be important variables for modelling typicality ratings by means of PLS, their concentrations were not directly correlated to the wines’ typicality ratings. From the CCA map it could be deduced that for these compounds the presence of concentrations below the mean concentrations of the data set seem to favour higher typicality ratings. The role of linalool and TDN and the strongly associated odours of the descriptor families ‘floral’ and ‘ageing notes’, was elucidated by an alternative, complementary conceptional approach of dividing typicality into factors of similarity and factors of dissimilarity. This rather theoretical perspective, adapted from fundamental considerations of an economical model, led to the conclusion that these odours and these aroma compounds could also act as important factors of perceived typicality and that this approach as well as the used multivariate aproach in this study can function as a helpful tool in typicality research in the future. 2) The viticultural part of this project permitted to study the impact of environmental and viticultural factors on Riesling wines’ typicality influencing aroma compounds and their precursor molecules at three stages of wine production (berries during the ripening period, the grape must, and the finished wines) over three years with different climatic conditions. Thanks to the installed drip irrigation system in the experimental vineyard in the Rheingau Valley in Germany, it was possible to study the impact of different water status of the vines and micro climate variation decoupled from a number of interacting factors, which often complicate studies, such as geological and geographical variation between vineyards, clonal diversity, regional climatic variations, and variations in canopy and phytosanitary actions. Moreover the integrated approach, ‘from berry set to aged wine’, allowed numerous insights into the relations between viticultural conditions during the growing season, especially the ripening season of the berries, and the finished and bottle aged wines. It was shown that sensory properties of wines differed in their holistic perception, depending on grape vines’ water status and micro climate manipulations during the growing period. Furthermore, monoterpenol concentrations, especially for linalool, were shown to increase in

General conclusion and perspectives

172

the finished wines as a consequence of the grape vines’ water deficit and, to lesser importance, of canopy micro climate manipulation. C13-norisoprenoid concentrations, namely vitispirane isomers and TDN, in the finished wines were strongly influenced by the grape vines’ water status and the canopy micro-climate manipulations, as fruit zone leaf removal at flowering or at veraison. Taking bottle aged wines into account, it is hypothesised from the results of this work that viticultural factors, such as grape vines’ water status, influence the metabolism of C13-norisoprenoids in grape berries. As a result this can lead to differences in aging behaviour of Riesling wines regarding this molecule class. Consequently, the moment of appearance of the ‘kerosene’ odour during bottle aging of Riesling wines is considered to be highly influenced by grape vines’ water status. The concentrations of the thiol 3-sulfanylhexan-1-ol (3SH) in wines was demonstrated to be highly influenced by the grape vines’ water status as it was by canopy micro climate manipulations. It was demonstrated that ‘weak to moderate’ water deficit favoured elevated 3SH concentrations in wines, whereas ‘severe’ water deficit was shown to decrease 3SH concentrations in wines. Both effects were observed to a degree likely to impact sensory properties of the wine. A major effort of this study is the valuable insight, which was obtained regarding the biogenesis of 3SH and its cysteinylated and gluthationylated precursor molecules (cys-S-prc and GSH-S-prc). Upon results from this study, it can be hypothesised, that cys-S-prc accumulated in the time from harvest to pressing, as long as the berries are not crushed. Another observation was that the concentration of cys-S-prc in grape must is highly dependent on GSH-S-prc in grape berries. Therefrom, it could be concluded, that GSH-S-prc formed during grape crushing does not get metabolised into cys-S-prc and does finally not contribute to cys-S-prc concentration in grape must. This observation could be used to influence the generation of cys-S-prc in the berries with consequently controlling the grape musts’ aroma potential regarding 3SH by variation of the post-harvest berry storage times before pressing. The survey over all steps of the wine production also permitted to observe the limits of the analytical methodology. Regarding the analysis of bound monoterpenols and C13-norisoprenoids, results of the aroma potential of these molecules in grapes and must and observed concentrations in finished wines did not comply. For single viticultural treatments the measured aroma potential for linalool underestimated the concentrations of the free compound in the finished wines, whereas it overestimated these for TDN at the same time. This is considered to be due to chemical transformation caused by the harsh conditions of the hydrolysis step for the bound molecules’ analysis. 3) The third goal of this project was to gain deeper insights into Riesling wines’ aroma. In order to decrypt Riesling wines’ aromatic expression, other aroma compounds than pointed

General conclusion and perspectives

173

out in the targeted approach (a) should be identified. The approach of confronting the wine extracts of typical varietal Sauvignon blanc style and Riesling wines in all steps of wine extraction, fractioning and analysis (sensory, GC-Olfactometry and GC-Mass Spectrometry) resulted in the identification of the aroma compound trans-ethyl cinnamate, a well known compound from the esterification of trans-cinnamic acid from grapes. The concentrations of this compound were significantly higher in Riesling wines than in Sauvignon blanc style wines. Surprisingly, there was not much information about the role of trans-ethyl cinnamate on wine aroma available, although it is an aroma compound, which is frequently analysed in wines. Additions to Riesling wines showed that this compound is likely to impact Riesling wines’ typicality, although no correlation between concentrations of this compound and perceived typicality was observed. It can be considered that the effects are highly dependent on the matrix. Results from the wines of the viticultural part of this project demonstrated, that this compound is also affected by viticultural factors and that it increases during bottle aging. Upon these results it is considered that trans-ethyl cinnamate impacts the aroma of bottle aged wines, but that this effect is strongly affected by viticultural conditions during the growing period. In conclusion this work clearly demonstrated that predicted climate change scenarios are likely to impact Riesling wines’ aromatic typicality. This was observed by sensory as well as by instrumental analytical methods. Moreover this work gave new insights into the concept of typicality [french typicité] and applied novel sensory and statistical approaches, which open up new perspectives in aromatic typicality related scientific questions. The importance of perceived citrus fruit, floral, and aging related odours for the Riesling wines’ aromatic concept was demonstrated and the importance of the marker molecules 3-sulfanylhexan-1-ol (3SH), linalool and TDN were pointed out. From the viticultural studies, new questions arise towards environmental impacts on bottle aging behaviour of Riesling wines concerning the development of the ‘kerosene’ odour. This represents a very important aspect to study in the future. Due to this study’s contribution to the increase of knowledge about the biosynthesis of the precursor molecules of 3-sulfanylhexan-1-ol, a new perspective is opened for managing grape musts aromatic potential regarding this highly potent aroma compound. Being able to manage these factors of wine aroma would clearly help winemakers to maintain and improve the production of wines with a high degree of typicality in the future. Moreover, the identification of trans-ethyl cinnamate as an impacting molecule of Riesling wines’ aromatic typicality was realised. This compound showed to be influenced by both, viticultural conditions and bottle storage, leading to the conclusion that it can impart presumably mostly in bottle age aroma of wines. The confrontation of the two varieties Riesling and Sauvignon blanc was highly contributional to the identification of trans-ethyl cinnamate and opens the

General conclusion and perspectives

174

perspective for the identification of more compounds impacting Riesling wines’ as well as Sauvignon blanc style wines’ aroma.

Cited Literature

175

Cited Literature

A Acree, T. E., Braell, P. and Butts, R. M. (1981), The presence of damascenone in cultivars of Vitis vinifera (Linneaus), v. Rotundifolia (Michaux), and v. Labruscana (Baily). Journal of Agricultural and Food Chemistry 29 3 688-690 Adams, D. O. and Liyanage, C. (1993), Glutathione increases in grape berries at the onset of ripening. American Journal of Enology and Viticulture 44 3 333-338 Allen, T., Herbst-Johnstone, M., Girault, M., Butler, P., Logan, G., Jouanneau, S., Nicolau, L. and Kilmartin, P. A. (2011), Influence of grape-harvesting steps on varietal thiol aromas in Sauvignon blanc wines. Journal of Agricultural and Food Chemistry 59 19 10641-10650 Anocibar-Beloqui, A. (1995), Contribution à l'étude des composés soufrés volatiles des vins rouges, Thèse de Doctorat, Université de Bordeaux II, Bordeaux Anonymous (1982) Namen und Herkunft der in Deutschland angebauten Rebsorten, Gesellschaft für Geschichte des Weines, Wiesbaden Antalick, G., Perello, M.-C. and De Revel, G. (2011), Development, validation and application of a specific method for the quantitative determination of wine esters by headspace-solid-phase microextraction - gas chromatography - Mass Spectrometry. Food Chemistry 121 4 1236-1245 Aubry, V., Etiévant, P. X., Giniès, C. and Henry, R. (1997), Quantitative determination of potent flavor compounds in Burgundy Pinot noir wines using a stable isotope dilution assay Journal of Agricultural and Food Chemistry 45 6 2120-2123

B Ballester, J. (2004), Mise en évidence d'un espace sensoriel et caractérisation des marqueurs relatifs à l'arôme des vins issus du cépage Chardonnay, Thèse de Doctorat, Université de Bourgogne and Universitat Politècnica de València, Dijon

Cited Literature

176

Ballester, J., Dacremont, C., Fur, Y. L. and Etiévant, P. (2005), The role of olfaction in the elaboration and use of the Chardonnay wine concept. Food Quality and Preference 16 4 351-359 Ballester, J., Patris, B., Symoneaux, R. and Valentin, D. (2008), Conceptual vs. perceptual wine spaces: Does expertise matter? Food Quality and Preference 19 3 267-276 Bauer, A. (2008), Terroirausprägung bei der Rebsorte Riesling: Korrelation sensorischer, chemischer, bodenkundlicher und klimatischer Parameter, Dissertation, Technische Universität Carolo-Wilhelmina Braunschweig, Braunschweig Baumes, R., Wirth, J., Bureau, S., Gunata, Y. and Razungles, A. (2002), Biogeneration of C13-norisoprenoid compounds: Experiments supportive for an apo-carotenoid pathway in grapevines. Analytica Chimica Acta 458 1 3-14 Belancic, A., Agosin, E., Ibacache, A., Bordeu, E., Baumes, R., Razungles, A. and Bayonove, C. (1997), Influence of sun exposure on the aromatic composition of Chilean Muscat grape cultivars Moscatel de Alexandria and Moscatel rosada. American Journal of Enology and Viticulture 48 2 181-186 Benkwitz, F., Tominaga, T., Kilmartin, P. A., Lund, C., Wohlers, M. and Nicolau, L. (2012), Identifying the chemical composition related to the distinct aroma characteristics of New Zealand Sauvignon blanc wines. American Journal of Enology and Viticulture 63 1 62-72 Beyer, H. and Walter, (1988) Lehrbuch der Organischen Chemie, S. Hirzel Verlag, Stuttgart, pp. 148 BGBl (1994), (WeinG 1994 § 21 (1) Nr. 3) ["Weingesetz in der Fassung der Bekanntmachung vom 18. Januar 2011 (BGBl. I S. 66), das durch Artikel 2 Absatz 13 des Gesetzes vom 22. Dezember 2011 (BGBl. I S. 3044) geändert worden ist"] i.V.m. § 24 (3) WeinV 1995 ["Weinverordnung in der Fassung der Bekanntmachung vom 21. April 2009 (BGBl. I s. 827), die zuletzt durch Artikel 5 der Verordnung vom 29. September 2011 (BGBl. I s. 1996) geändert worden ist"]). Bundesgesetzblatt Bindon, K. A., Dry, P. R. and Loveys, B. R. (2007), Influence of plant water status on the production of C13-norisoprenoid precursors in Vitis vinifera L. cv. Cabernet Sauvignon grape berries. Journal of Agricultural and Food Chemistry 55 11 4493-4500

Cited Literature

177

Bitteur, S., Tesniere, C., Sarris, J., Baumes, R., Bayonove, C. and Flanzy, C. (1992), Carbonic anaerobiosis of Muscat grapes. I. Changes in the profiles of free and bound volatiles. American Journal of Enology and Viticulture 43 1 41-48 Blancher, G., Chollet, S., Kesteloot, R., Hoang, D. N., Cuvelier, G. and Sieffermann, J. M. (2007), French and Vietnamese: How do they describe texture characteristics of the same food? A case study with jellies. Food Quality and Preference 18 3 560-575 Blijlevens, J., Gemser, G. and Mugge, R. (2012), The importance of being 'well-placed': The influence of context on perceived typicality and esthetic appraisal of product appearance. Acta Psychologica 139 1 178-186 Böhm, P., Löhnertz, O. and Muskat, S., (2008) Terroir Hessen. Vielfalt erleben! Gesellschaft für Rheingauer Weinkultur mbH / Weinbauverband Hessische Bergstraße e.V, Eltville am Rhein Bouchilloux, P., Darriet, P., Henry, R., Lavigne-Cruège, V. R. and Dubourdieu, D. (1998), Identification of volatile and powerful odorous thiols in Bordeaux red wine varieties. Journal of Agricultural and Food Chemistry 46 8 3095-3099 Bowen, A. J. (2010), Elucidation of odour-potent compounds and sensory profiles of Vidal blanc and Riesling icewines from the Niagara peninsula: Effect of harvest date and crop level, Doctoral thesis, Brock University St. Catharines - Ontario Breitmeier, (2006) Terpenes - flavors, fragrances, pharmaca, pheromones, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. Britton, G. (2008), Functions of carotenoid metabolites and breakdown products in Carotenoids; Britton, G., et al. (Eds); Birkhäuser Basel, pp. 309-324 Brochet, F. D. R. and Dubourdieu, D. (2001), Wine descriptive language supports cognitive specificity of chemical senses. Brain and Language 77 2 187-196 Buttrose, M. S., Hale, C. R. and Kliewer, W. M. (1971), Effect of temperature on the composition of 'Cabernet Sauvignon' berries. American Journal of Enology and Viticulture 22 2 71-75

Cited Literature

178

C Cadot, Y., Caillé, S., Samson, A., Barbeau, G. and Cheynier, V. (2010), Sensory dimension of wine typicality related to a terroir by quantitative descriptive analysis, Just-About-Right analysis and typicality assessment. Analytica Chimica Acta 660 12 53-62 Cadot, Y., Caillé, S., Samson, A., Barbeau, G. and Cheynier, V. (2012a), Sensory representation of typicality of Cabernet franc wines related to phenolic composition: Impact of ripening stage and maceration time. Analytica Chimica Acta 732 0 91-99 Cadot, Y., Caillé, S., Thiollet-Scholtus, M., Samson, A., Barbeau, G. and Cheynier, V. (2012b), Characterisation of typicality for wines related to terroir by conceptual and by perceptual representations. An application to red wines from the Loire valley. Food Quality and Preference 24 1 48-58 Camacho-Cristóbal, J. and González-Fontes, A. (2007), Boron deficiency decreases plasmalemma H+-ATPase expression and nitrate uptake, and promotes ammonium assimilation into asparagine in tobacco roots. Planta 226 2 443-451 Campo, E., Ferreira, V., Escudero, A. and Cacho, J. (2005), Prediction of the wine sensory properties related to grape variety from dynamic-headspace gas chromatography-olfactometry data. Journal of Agricultural and Food Chemistry 53 14 5682-5690 Campo, E., Do, B. V., Ferreira, V. and Valentin, D. (2008), Aroma properties of young Spanish monovarietal white wines: A study using sorting task, list of terms and frequency of citation. Australian Journal of Grape and Wine Research 14 2 104-115 Campo, E., Ballester, J., Langlois, J., Dacremont, C. and Valentin, D. (2010), Comparison of conventional descriptive analysis and a citation frequency-based descriptive method for odor profiling: An application to Burgundy Pinot noir wines. Food Quality and Preference 21 1 44-55 Candelon, M., Ballester, J., Uscidda, N., Blanquet, J. and Le Fur, Y. (2004), Sensory methodology developed for the the investigation of Sciaccarello wine concept. Journal International des Sciences de la Vigne et du Vin 38 2 147-154

Cited Literature

179

Capone, D. L. and Jeffery, D. W. (2011), Effects of transporting and processing Sauvignon blanc grapes on 3-mercaptohexan-1-ol precursor concentrations. Journal of Agricultural and Food Chemistry 59 9 4659-4667 Capone, D. L., Black, C. A. and Jeffery, D. W. (2012), Effects on 3-mercaptohexan-1-ol precursor concentrations from prolonged storage of Sauvignon blanc grapes prior to crushing and pressing. Journal of Agricultural and Food Chemistry 60 13 3515-3523 Capone, D. L., Sefton, M. A., Hayasaka, Y. and Jeffery, D. W. (2010a), Analysis of precursors to wine odorant 3-mercaptohexan-1-ol using HPLC-MS/MS: Resolution and quantitation of diastereomers of 3-S-cysteinylhexan-1-ol and 3-S-glutathionylhexan-1-ol. Journal of Agricultural and Food Chemistry 58 3 1390-1395 Capone, D. L., Pardon, K. H., Cordente, A. G. and Jeffery, D. W. (2011a), Identification and quantitation of 3-S-cysteinylglycinehexan-1-ol (cysgly-3-MH) in Sauvignon blanc grape juice by HPLC-MS/MS. Journal of Agricultural and Food Chemistry 59 20 11204-11210 Caporale, G., Policastro, S., Carlucci, A. and Monteleone, E. (2006), Consumer expectations for sensory properties in virgin olive oils. Food Quality and Preference 17 12 116-125 Chisholm, M. G., Guiher, L. A., Vonah, T. M. and Beaumont, J. L. (1994), Comparison of some french-american hybrid wines with white Riesling using gas chromatography-olfactometry. American Journal of Enology and Viticulture 45 2 201-212 Choné, X., Van Leeuwen, C., Dubourdieu, D. and Gaudillère, J.-P. (2001), Stem water potential is a sensitive indicator for grapevine water status. Annal of Botany 87 477-483 Choné, X., Lavigne-Cruège, V., Tominaga, T., Van Leeuwen, C., Castagède, C., Saucier, C. and Dubourdieu, D. (2006), Effect of vine nitrogen status on grape aromatic potential: Flavor precursors (S-cysteine conjugates), glutathione and phenolic content in Vitis vinifera L. cv. Sauvignon blanc grape juice. Journal International des Sciences de la Vigne et du Vin 40 1 1-6 Chrea, C., Valentin, D., Sulmont-Rossé, C., Nguyen, D. H. and Abdi, H. (2005), Semantic, typicality and odor representation: A cross-cultural study. Chemical Senses 30 1 37-49

Cited Literature

180

Coombe, B. G. (1987), Influence of temperature on composition and quality of grapes. Acta Horticulturae (ISHS) 206 23-36

D Daniel, M. A., Capone, D. L., Sefton, M. A. and Elsey, G. M. (2009), Riesling acetal is a precursor to 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN) in wine. Australian Journal of Grape and Wine Research 15 1 93-96 Darriet, P., Lavigne, V., Boidron, J.N., Dubourdieu, D. (1991), Caractérisation de l'arôme variétal des vins de Sauvignon par couplage CPG-Olfactométrie. Journal International de la Vigne et du Vin 25 3 167-174 Darriet, P. (1993), L'arôme et les précurseurs d'arôme du Sauvignon, Thèse Doctorat Université de Bordeaux 2, Bordeaux Darriet, P., Thibon, C. and Dubourdieu, D. (2012), Aroma and aroma precursors in grape berry in The biochemistry of the berry; Gerós, H., et al. (Eds); Bentham e books, e-book, pp. 111-136 Darriet, P., Tominaga, T., Lavigne, V., Boidron, J.-N. and Dubourdieu, D. (1995), Identification of a powerful aromatic component of Vitis vinifera L. var. Sauvignon wines: 4-mercapto-4-methylpentan-2-one. Flavour and Fragrance Journal 10 6 385-392 Davis, E. and Croteau, R. (2000), Cyclization enzymes in the biosynthesis of monoterpenes, sesquiterpenes, and diterpenes in Biosynthesis; Leeper, F. and Vederas, J. (Eds); Springer Berlin / Heidelberg, pp. 53-95 Delcroix, A., Günata, Z., Sapis, J.-C., Salmon, J.-M. and Bayonove, C. (1994), Glycosidase activities of three enological yeast strains during winemaking: Effect on the terpenol content of muscat wine. American Journal of Enology and Viticulture 45 3 291-296 Diaz-Maroto, M. C., Schneider, R. and Baumes, R. (2005), Formation pathways of ethyl esters of branched short-chain fatty acids during wine aging. Journal of Agricultural and Food Chemistry 53 9 3503-3509

Cited Literature

181

Dollmann, B., Full, G., Schreier, P., Winterhalter, P., Güntert, M. and Sommer, H. (1995), Synthesis and enantiodifferentiation of Riesling acetals. Phytochemical Analysis 6 2 106-111 Dry, P. R. and Loveys, B. R. (1998), Factors influencing grapevine vigour and the potential for control with partial rootzone drying. Australian Journal of Grape and Wine Research 4 3 140-148 Dry, P. R., Loveys, B. R., Mccarthy and Stoll, M. (2001), Strategic irrigation management in Australian vineyards. Journal International des Sciences de la Vigne et du Vin 35 45 - 61 Du Plessis, C. S. and Augustyn, O. P. H. (1981), Initial study on the guava aroma of Chenin blanc and Colombar wines. South African Journal of Enology and Viticulture 2 2 101-103 Dubourdieu, D. (2012), La valeur du vin entre nature et culture réflexions sur le goût mondial et la typicité des vins in Oeno 2011 - 9e Symposium International d'Oenologie de Bordeaux Darriet, P., et al. (Eds); Dunod - Paris, Bordeaux, pp. 3-6 Dubourdieu, D., Tominaga, T., Masneuf, I., Peyrot Des Gachons, C. and Murat, M. L. (2006), The role of yeasts in grape flavor development during fermentation: The example of Sauvignon blanc. American Journal of Enology and Viticulture 57 1 81-88 Duchêne, E., Legras, J. L., Karst, F., Merdinoglu, D., Claudel, P., Jaegli, N. and Pelsy, F. (2009), Variation of linalool and geraniol content within two pairs of aromatic and non-aromatic grapevine clones. Australian Journal of Grape and Wine Research 15 2 120-130 Duden, (2010) http://www.Duden.De/rechtschreibung/typizitaet. [looked up 05.06.2010] Dukes, B. C. and Butzke, C. E. (1998), Rapid determination of primary amino acids in grape juice using an o-phthaldialdehyde/N-acetyl-L-cysteine spectrophotometric assay. American Journal of Enology and Viticulture 49 2 125-134 DWI (2008) Aroma wheel for German white wines, Deutsches Weininstitut, Mainz DWI, (2011) Vineyard surface in Germany 2010. Deutsches Weininstitut DWI, (2011a) Riesling vineyards world-wide. Deutsches Weininstitut

Cited Literature

182

DWI, (2011b) The major grape-varieties in Germany. Deutsches Weininstitut

E Ebeler, S. E. and Thorngate, J. H. (2009), Wine chemistry and flavor: Looking into the crystal glass. Journal of Agricultural and Food Chemistry 57 18 8098-8108 Eggers, N. J., Bohna, K. and Dooley, B. (2006), Determination of vitispirane in wines by stable isotope dilution assay. American Journal of Enology and Viticulture 57 2 226-232 El-Ansary, D. O., Nakayama, S., Hirano, K. and Okamoto, G. (2005), Response of Muscat of Alexandria table grapes to post-veraison regulated deficit irrigation in Japan. Vitis 44 1 5-9 Escudero, A., Campo, E., Farina, L., Cacho, J. and Ferreira, V. (2007), Analytical characterization of the aroma of five premium red wines. Insights into the role of odor families and the concept of fruitiness of wines. Journal of Agricultural and Food Chemistry 55 11 4501-4510 Escudero, A., Gogorza, B., Melús, M. A., Ortín, N., Cacho, J. and Ferreira, V. (2004), Characterization of the aroma of a wine from Maccabeo. Key role played by compounds with low odor activity values. Journal of Agricultural and Food Chemistry 52 11 3516-3524 Esti, M., Airola, R. L. G., Moneta, E., Paperaio, M. and Sinesio, F. (2010), Qualitative data analysis for an exploratory sensory study of Grechetto wine. Analytica Chimica Acta 660 12 63-67 Etievant, P. X., Issanchou, S. N. and Bayonove, C. L. (1983), The flavour of Muscat wine: The sensory contribution of some volatile compounds. Journal of the Science of Food and Agriculture 34 5 497-504

F Falcao, L. D., Lytra, G., Darriet, P. and Barbe, J. C. (2012), Identification of ethyl 2-hydroxy-4-methylpentanoate in red wines, a compound involved in blackberry aroma. Food Chemistry 132 1 230-236

Cited Literature

183

Fang, Y. and Qian, M. C. (2006), Quantification of selected aroma-active compounds in Pinot noir wines from different grape maturities. Journal of Agricultural and Food Chemistry 54 22 8567-8573 Farquhar, G. D., Ehleringer, J. R. and Hubick, K. T. (1989), Carbon isotope discrimination and photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology 40 1 503-537 Fedrizzi, B., Pardon, K. H., Sefton, M. A., Elsey, G. M. and Jeffery, D. W. (2009), First identification of 4-S-glutathionyl-4-methylpentan-2-one, a potential precursor of 4-mercapto-4-methylpentan-2-one, in Sauvignon blanc juice. Journal of Agricultural and Food Chemistry 57 3 991-995 Fedrizzi, B., Tosi, E., Simonato, B., Finato, F., Cipriani, M., Caramia, G. and Zapparoli, G. (2011), Changes in wine aroma composition according to botrytized berry percentage: A preliminary study on Amarone wine. Food Technology and Biotechnology 49 4 529-535 Ferreira, V., Aznar, M., Lopez, R. and Cacho, J. (2001), Quantitative gas chromatography olfactometry carried out at different dilutions of an extract. Key differences in the odor profiles of four high-quality spanish aged red wines. Journal of Agricultural and Food Chemistry 49 10 4818-4824 Ferreira, V., Hernandez-Orte, P., Escudero, A., Lopez, R. and Cacho, J. (1999), Semipreparative reversed-phase liquid chromatographic fractionation of aroma extracts from wine and other alcoholic beverages. Journal of Chromatography A 864 1 77-88 Fischer, C. and Swoboda (2005) Riesling, Gräfe und Unzer Verlag GmbH, München Fischer, U., Giraud, G. and Serra, J. N., (2005) Development of a typicality assessment methodology. European Union Project FP5 1998-2002 Quality of Life and Management of Living Resources Fischer, U., Roth, D. and Christmann, M. (1999), The impact of geographic origin, vintage and wine estate on sensory properties of Vitis vinifera cv. Riesling wines. Food Quality and Preference 10 4 (5) 281-288

Cited Literature

184

Flanzy, C., Flanzy, M. and Benard, P. (1987) La vinification par maceration carbonique, INRA, Paris Fleischmann, P. and Zorn, H. (2008), Enzymic pathways for formation of carotenoid cleavage products in Carotenoids; Britton, G., et al. (Eds); Birkhäuser Basel, pp. 341-366 Foyer, C. H., Theodoulou, F. L. and Delrot, S. (2001), The functions of inter- and intracellular glutathione transport systems in plants. Trends in Plant Science 6 10 486-492 Fretz, C. B., Luisier, J.-L., Tominaga, T. and Amado, R. (2005), 3-mercaptohexanol: An aroma impact compound of Petite Arvine wine. American Journal of Enology and Viticulture 56 4 407-410

G Galet, P., (1998) Cépages et vignobles de France. ii. L'ampélographie française, Dehan, Montpellier Garrier, G., (2008) Histoire sociale et culturelle du vin, Larousse, Paris Genovese, A., Gambuti, A., Piombino, P. and Moio, L. (2007), Sensory properties and aroma compounds of sweet Fiano wine. Food Chemistry 103 4 1228-1236 Gerdes, S. M., Winterhalter, P. and Ebeler Susan, E. (2001), Effect of sunlight exposure on norisoprenoid formation in White Riesling grapes in Carotenoid-Derived Aroma Compounds; ACS Symposium Series; Volume 802; American Chemical Society, pp. 262-272 Gerhards, D., (2012) Personal communication. Gijs, L., Chevance, F., Jerkovic, V. and Collin, S. (2002), How low pH can intensify β-damascenone and dimethyl trisulfide production through beer aging. Journal of Agricultural and Food Chemistry 50 20 5612-5616

Cited Literature

185

Giorgessi, F., Sansone, L., Tarricone, L., Masi, G., Tamborra, P. and Zanolli, P. (2007), Variation of the vegetative-productive equilibrium of the vine (cv. Sauvignon b.) according to different water conditionings and consequent effects on grape and wine aroma compounds. Rivista di viticoltura e di enologia 60 3 19-36 Gladstone, J. S., (1992) Viticulture and environment, Winetitles, Adelaide, pp. 310 Grant-Preece, P. A., Pardon, K. H., Capone, D. L., Cordente, A. G., Sefton, M. A., Jeffery, D. W. and Elsey, G. M. (2010), Synthesis of wine thiol conjugates and labeled analogues: Fermentation of the glutathione conjugate of 3-mercaptohexan-1-ol yields the corresponding cysteine conjugate and free thiol. Journal of Agricultural and Food Chemistry 58 3 1383-1389 Green, J. A., Parr, W. V., Breitmeyer, J., Valentin, D. and Sherlock, R. (2011), Sensory and chemical characterisation of Sauvignon blanc wine: Influence of source of origin. Food Research International 44 9 2788-2797 Grossmann, M., Rapp, A. and Rieth, W. (1987), Enzymatische Freisetzung gebundener Aromastoffe. Deutsche Lebensmittelrundschau 83 1 7-12 Gruber, B. R. and Schultz, H. R. (2005), Coupling of plant soil water status at different vineyard sites. Acta Hort. (ISHS) 689 381-390 Guinard, J.-X. and Cliff, M. (1987), Descriptive analysis of Pinot noir wines from Carneros, Napa, and Sonoma. American Journal of Enology and Viticulture 38 3 211-215 Günata, Y. Z., Bayonove, C. L., Baumes, R. L. and Cordonnier, R. E. (1985), The aroma of grapes i. Extraction and determination of free and glycosidically bound fractions of some grape aroma components. Journal of Chromatography A 331 0 83-90 Günata, Y. Z., Bayonove, C. L., Baumes, R. L. and Cordonnier, R. E. (1985a), The aroma of grapes. Localisation and evolution of free and bound fractions of some grape aroma components cv. Muscat during first development and maturation. Journal of the Science of Food and Agriculture 36 9 857-862 Guth, H. (1997), Identification of character impact odorants of different white wine varieties. Journal of Agricultural and Food Chemistry 45 8 3022-3026

Cited Literature

186

H Hardie, W. J. and Considine, J. A. (1976), Response of grapes to water-deficit stress in particular stages of development. American Journal of Enology and Viticulture 27 2 55-61 Hardy, P. J. (1970), Changes in volatiles of Muscat grapes during ripening. Phytochemistry 9 4 709-715 Hautus, M. J., Van Hout, D. and Lee, H. S. (2009), Variants of A Not-A and 2AFC tests: Signal detection theory models. Food Quality and Preference 20 3 222-229 Herion, P., Full, G., Winterhalter, P., Schreier, P. and Bicchi, C. (1993), Enantiodifferentiation of isomeric vitispiranes. Phytochemical Analysis 4 5 235-239 Heymann, H. and Noble, A. C. (1987), Descriptive analysis of commercial Cabernet Sauvignon wines from California. American Journal of Enology and Viticulture 38 1 41-44 Hofmann, M. and Schultz, H. R., (2008) Inklim2012IIplus: Einfluss der Klimaentwicklung auf den Wasserhaushalt von Rebstandorten - Möglichkeiten der Modellierung von Steillagenregionen. Forschungsanstalt Geisenheim Hoppe, J.E. (2012) Einfluss von (E)-Zimtsäureethylester auf die Typizität von Rieslingweinen. Bachelor thesis, Hochschule RheinMain, University of Applied Sciences, Wiesbaden Rüsselsheim Geisenheim Howell, K. S., Swiegers, J. H., Elsey, G. M., Siebert, T. E., Bartowsky, E. J., Fleet, G. H., Pretorius, I. S. and De Barros Lopes, M. A. (2004) Variation in 4-mercapto-4-methyl-pentan-2-one release by Saccharomyces cerevisiae commercial wine strains. FEMS Microbiology Letters 240 2 125-129 Huglin, P., (1986) Biologie et ècologie de la vigne, Payot, Lausanne, pp. 371

Cited Literature

187

I ICCP: Field, C. B., et al., Press, C. U., (2012) Managing the risks of extreme events and disasters to advance climate change adaption. A special report of working groups I and II of the intergovernmental panel on climate change. ICCP ICCP: Trenberth, K. E., Jones, P., Ambenje, R., Bojariu, R., Easterling, D., Klein Tank, A., Parker, D., Rahimzadeh, F., Renwick, J. A., Rusticucci, M., Soden, B. and Zhai, P., Solomon, S., et al., Press, C. U., (2007) Observations: Surface and atmospheric climate change. ICCP INAO, (2011) Guide du demandeur d'une Appelation d'Origine Controllée (AOC / AOP). INAO Irmler, S., Raboud, S., Beisert, B., Rauhut, D. and Berthoud, H. L. N. (2008), Cloning and characterization of two Lactobacillus casei genes encoding a cystathionine lyase. Applied and Environmental Microbiology 74 1 99-106 ISO (1977), ISO 3591:1977 Sensory analysis - apparatus - wine-tasting glass, International Organisation of Standardisation, Geneva, Switzerland

J Jackson, M. (1997), Hormones from roots as signals for the shoots of stressed plants. Trends in Plant Science 2 1 22-28 Jaffré, J., Valentin, D., Meunier, J.-M., Siliani, A., Bertuccioli, M. and Le Fur, Y. (2011), The Chardonnay wine olfactory concept revisited: A stable core of volatile compounds, and fuzzy boundaries. Food Research International 44 1 456-464 Jakob, L., (2012) Lexikon der Önologie, Meininger, Neustadt a.d.W. Janusz, A., Capone, D. L., Puglisi, C. J., Perkins, M. V., Elsey, G. M. and Sefton, M. A. (2003), (E)-1-(2,3,6-trimethylphenyl)buta-1,3-diene: A potent grape-derived odorant in wine. Journal of Agricultural and Food Chemistry 51 26 7759-7763

Cited Literature

188

Jones, G., White, M., Cooper, O. and Storchmann, K. (2005), Climate change and global wine quality. Climatic Change 73 3 319-343

K Kalua, C. M. and Boss, P. K. (2010), Comparison of major volatile compounds from Riesling and Cabernet Sauvignon grapes (Vitis vinifera L.) from fruitset to harvest. Australian Journal of Grape and Wine Research 16 2 337-348 Keller, M. (2005), Deficit irrigation and vine mineral nutrition. American Journal of Enology and Viticulture 56 3 267-283 Keller, M. (2010), Managing grapevines to optimise fruit development in a challenging environment: A climate change primer for viticulturists. Australian Journal of Grape and Wine Research 16 56-69 Kennedy, J. A., Matthews, M. A. and Waterhouse, A. L. (2002), Effect of maturity and vine water status on grape skin and wine flavonoids. American Journal of Enology and Viticulture 53 4 268-274 Kerridge, G. and Antcliff, A. (2004) Wine grape varieties, CSIRO, Collingwood Kim, M.-A., Chae, J.-E., Van Hout, D. and Lee, H.-S. (2012), Discriminations of the A Not-A difference test improved when "A" Was familiarized using a brand image. Food Quality and Preference 23 1 3-12 Kobayashi, H., Takase, H., Kaneko, K., Tanzawa, F., Takata, R., Suzuki, S. and Konno, T. (2010), Analysis of S-3-(hexan-1-ol)-glutathione and S-3-(hexan-1-ol)-L-cysteine in Vitis vinifera L. cv. Koshu for aromatic wines. American Journal of Enology and Viticulture 61 2 176-185 Kobayashi, H., Takase, H., Suzuki, S., Tanzawa, F., Takata, R., Fujita, K., Kohno, M., Mochizuki, M., Suzuki, S. and Konno, T. (2011), Environmental stress enhances biosynthesis of flavor precursors, S-3-(hexan-1-ol)-glutathione and S-3-(hexan-1-ol)-cysteine, in grapevine through glutathione S-transferase activation. Journal of Experimental Botany 62 3 1325-1336

Cited Literature

189

Komes, D., Ulrich, D. and Lovric, T. (2006), Characterization of odor-active compounds in Croatian Rhine Riesling wine, subregion Zagorje. European Food Research and Technology 222 1 1-7 Kotseridis, Y., Baumes, R. L. and Skouroumounis, G. K. (1999), Quantitative determination of free and hydrolytically liberated β-damascenone in red grapes and wines using a stable isotope dilution assay. Journal of Chromatography A 849 1 245-254 Koundouras, S., Marinos, V., Gkoulioti, A., Kotseridis, Y. and Van Leeuwen, C. (2006), Influence of vineyard location and vine water status on fruit maturation of non-irrigated cv. Agiorgitiko (Vitis vinifera L.). Effects on wine phenolic and aroma components. Journal of Agricultural and Food Chemistry 54 14 5077-5086 Kovats, E. (1958), Gas chromatographische Charakterisierung organischer Verbindungen. I. Retentions Indices aliphatischer Halogenide, Alkohole, Aldehyde und Ketone. Helvetiae Chimica Acta 41 1932-1951 Kozina, B., Karoglan, M., Herjavec, S., Jeromel, A. and Orlic, S. (2008), Influence of basal leaf removal on the chemical composition of Sauvignon blanc and Riesling wines. Journal of Food, Agriculture & Environment 6 1 28-33 Kwasniewski, M. T., Vanden Heuvel, J. E., Pan, B. S. and Sacks, G. L. (2010), Timing of cluster light environment manipulation during grape development affects C13-norisoprenoid and carotenoid concentrations in Riesling. Journal of Agricultural and Food Chemistry 58 11 6841-6849

L Lacroux, F., Tregoat, O., Van Leeuwen, C., Pons, A., Tominaga, T., Lavigne-Cruège, V. and Dubourdieu, D. (2008), Effect of foliar nitrogen and sulphur application on aromatic expression of Vitis vinifera L. cv. Sauvignon blanc. Journal International des Sciences de la Vigne et du Vin 42 3 125-132 Lavigne, V., Henry, R. and Dubourdieu, D. (1998), Identification et dosage de composés soufrés intervenant dans l'arome 'grillé' des vins. Sciences des Aliments 18 175-191

Cited Literature

190

Lavigne, V., Pons, A. and Dubourdieu, D. (2007), Assay of glutathione in must and wines using capillary electrophoresis and laser-induced fluorescence detection: Changes in concentration in dry white wines during alcoholic fermentation and aging. Journal of Chromatography A 1139 1 130-135 Lawless, H. T. (1984), Flavor description of white wine by "expert" and non-expert wine consumers. Journal of Food Science 49 120-123 Lawless, H. T. and Heymann, (2010) Sensory evaluation of food - principles and practices, Springer, New York, second edition Lê, S., Josse, J. and Husson, F. (2008), FactomineR: An R package for multivariate analysis. Journal of Statistical Software 25 1 2-18 Lebon, E., Dumas, V., Pieri, P. and Schultz, H. R. (2003), Modelling the seasonal dynamics of the soil water balance of vineyards. Functional Plant Biology 30 6 699-710 Lee, H. S., Van Hout, D. and O'mahony, M. (2007a), Sensory difference tests for margarine: A comparison of R-indices derived from ranking and A Not-A methods considering response bias and cognitive strategies. Food Quality and Preference 18 4 675-680 Lee, H. S., Van Hout, D. and Hautus, M. J. (2007b), Comparison of performance in the A Not-A, 2-AFC, and same different tests for the flavor discrimination of margarines: The effect of cognitive decision strategies. Food Quality and Preference 18 6 920-928 Lee, S.-H., Seo, M.-J., Riu, M., Cotta, J. P., Block, D. E., Dokoozlian, N. K. and Ebeler, S. E. (2007), Vine microclimate and norisoprenoid concentration in Cabernet Sauvignon grapes and wines. American Journal of Enology and Viticulture 58 3 291-301 Linsenmeier, A. and Löhnertz, O. (2007), Changes in norisoprenoid levels with long-term nitrogen fertilisation in different vintages of Vitis vinifera var. Riesling wines. South African Journal of Enology and Viticulture 28 1 17-24 Löhnertz, O., Rantz, J. M., (1991) Soil nitrogen and the uptake of nitrogen in grapevines. American Society for Enology and Viticulture

Cited Literature

191

Loken, B. and Ward, J. (1990), Alternative approaches to understanding the determinants of typicality. Journal of Consumer Research 17 111-126 Lorrain, B., Ballester, J., Thomas-Danguin, T., Blanquet, J., Meunier, J. M. and Le Fur, Y. (2006), Selection of potential impact odorants and sensory validation of their importance in typical Chardonnay wines. Journal of Agricultural and Food Chemistry 54 11 3973-3981 Loscos, N., Hernandez-Orte, P., Cacho, J. and Ferreira, V. (2007), Release and formation of varietal aroma compounds during alcoholic fermentation from non-floral grape odorless flavor precursors fractions. Journal of Agricultural and Food Chemistry 55 16 6674-6684 Lott, H., Pfaff, F. and Prior, B. (2010) Taschenbuch der Rebsorten, Fachverlag Dr- Fraund GmbH, Mainz Luisier, J.-L., Buettner, H., Völker, S., Rausis, T. and Frey, U. (2008), Quantification of cysteine S-conjugate of 3-sulfanylhexan-1-ol in must and wine of Petite Arvine vine by stable isotope dilution analysis. Journal of Agricultural and Food Chemistry 56 9 2883-2887 Lund, C. M., Thompson, M. K., Benkwitz, F., Wohler, M. W., Triggs, C. M., Gardner, R., Heymann, H. and Nicolau, L. (2009), New Zealand Sauvignon blanc distinct flavor characteristics: Sensory, chemical, and consumer aspects. American Journal of Enology and Viticulture 60 1 1-12 Lytra, G., Tempere, S., De Revel, G. and Barbe, J.-C. (2012), Distribution and organoleptic impact of ethyl 2-hydroxy-4-methylpentanoate enantiomers in wine. Journal of Agricultural and Food Chemistry 60 6 1503-1509

M Maitre, I., Symoneaux, R., Jourjon, F. and Mehinagic, E. (2010), Sensory typicality of wines: How scientists have recently dealt with this subject. Food Quality and Preference 21 7 726-731 Marais, J. (1983), Terpenes in the aroma of grapes and wine: A review. South African Journal of Enology and Viticulture 4 4 49-58

Cited Literature

192

Marais, J. (1987), Terpene concentration and wine quality of Vitis vinifera L. cv. Gewürztraminer as affected by grape maturity and cellar practices. Vitis 26 231-245 231 Marais, J. (1992b), 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN): A possible degradation product of lutein and beta-carotene. South African Journal of Enology and Viticulture 13 1 52-55 Marais, J., van Wijk, C. J. and Rapp, A. (1991), Carotenoid levels in maturing grapes as affected by climatic regions, sunlight and shade. South African Journal of Enology and Viticulture 12 2 64-69 Marais, J., van Wijk, C. J. and Rapp, A. (1992), Effect of storage time, temperature and region on the levels of 1,1,6-trimethyl-1,2-dihydronaphthalene and other volatiles, and on quality of Weisser Riesling wines. South African Journal of Enology and Viticulture 13 1 33-44 Marais, J., van Wijk, C. J. and Rapp, A. (1992c), Effect of sunlight and shade on norisoprenoid levels in maturing Weisser Riesling and Chenin blanc grapes and Weisser Riesling wines. South African Journal of Enology and Viticulture 13 1 23-32 Marais, J., Versini, G., van Wijk, C. J. and Rapp, A. (1992a), Effect of region on free and bound monoterpene and C13-norisoprenoid concentrations in Weisser Riesling wines. South African Journal of Enology and Viticulture 13 2 71-77 Marchal, A. (2010), Recherches sur les bases moléculaires de la saveur sucrée des vins secs - approches analytique et sensorielle, Thèse de Doctorat, Université Victor Segalen Bordeaux 2, Bordeaux Masneuf-Pomarède, I., Mansour, C., Murat, M.-L., Tominaga, T. and Dubourdieu, D. (2006), Influence of fermentation temperature on volatile thiols concentrations in Sauvignon blanc wines. International Journal of Food Microbiology 108 3 385-390 Masson, G. and Schneider, R. (2009), Key compounds of Provence rosé wine flavor. American Journal of Enology and Viticulture 60 1 116-122 Mateo, J. J. and Jiménez, M. (2000), Monoterpenes in grape juice and wines. Journal of Chromatography A 881 12 557-567

Cited Literature

193

Mathieu, S., Terrier, N., Procureur, J., Bigey, F. and Günata, Z. (2005), A carotenoid cleavage dioxygenase from Vitis vinifera L.: Functional characterization and expression during grape berry development in relation to C13-norisoprenoid accumulation. Journal of Experimental Botany 56 420 2721-2731 Mathieu, S., Wirth, J., Sauvage, F.-X., Lepoutre, J.-P., Baumes, R. and Günata, Z. (2009), Biotransformation of C13-norisoprenoids and monoterpenes by a cell suspension culture of cv. Gamay (Vitis vinifera L.). Plant Cell, Tissue and Organ Culture 97 2 203-213 Matthews, M. A. and Anderson, M. M. (1988), Fruit ripening in Vitis vinifera L.: Responses to seasonal water deficits. American Journal of Enology and Viticulture 39 4 313-320 Matthews, M. A. and Anderson, M. M. (1989), Reproductive development in grape (Vitis vinifera L.): Responses to seasonal water deficits. American Journal of Enology and Viticulture 40 1 52-60 Matthews, M. A., Anderson, M. M. and Schultz, H. R. (1987), Phenologic and growth responses to early and late season water deficits in Cabernet franc. Vitis 24 4 147-160 Mattivi, F., Fedrizzi, B., Zenato, A., Tiefenthaler, P., Tempesta, S., Perenzoni, D., Cantarella, P., Simeoni, F. and Vrhovsek, U. (2012), Development of reliable analytical tools for evaluating the influence of reductive winemaking on the quality of Lugana wines. Analytica Chimica Acta 732 0 194-202 McCloskey, L. P., Sylvan, M. and Arrhenius, S. P. (1996), Descriptive analysis for wine quality experts determining appellations by Chardonnay wine aroma. Journal of Sensory Studies 11 1 49-67 McEwan, J. A. and Schlich, P. (1991), Correspondence analysis in sensory evaluation. Food Quality and Preference 3 1 23-36 Meilgaard, M., Civille, G. V. and Carr, T. B. (2007) Sensory evaluation techniques, CRC Press, Boca Raton, 4th editon Mervis, C. B. and Rosch, E. (1981), Categorization of natural objects. Annual Review of Psychology 32 89-115

Cited Literature

194

Ministère de la Agriculture et de la Pêche (2001), Arreté du 7/12/2011 relatif aux examens analytiques et organoleptiques pour les vins d'AOC. Journal officiel de la République Française Mira De Orduna, R. (2010), Climate change associated effects on grape and wine quality and production. Food Research International 43 7 1844-1855 Moio, L. and Etievant, P. X. (1995), Ethyl anthranilate, ethyl cinnamate, 2,3-dihydrocinnamate, and methyl anthranilate: Four important odorants identified in Pinot noir wines of Burgundy. American Journal of Enology and Viticulture 46 3 392-398 Moio, L., Schlich, P., Issanchou, S. N., Etievant, P. X. and Feuillat, M. (1993), Description de la typicité aromatique des vins de Bourgogne issus du cépage Chardonnay. Journal International des Sciences de la Vigne et du Vin 3 179-189 Murat, M.-L., Tominaga, T. and Dubourdieu, D. (2001), Assessing the aromatic potential of Cabernet Sauvignon and Merlot musts used to produce rose wine by assaying the cysteinylated precursor of 3-mercaptohexan-1-ol. Journal of Agricultural and Food Chemistry 49 11 5412-5417

N Nikolantonaki, M. and Darriet, P. (2011), Identification of ethyl 2-sulfanylacetate as an important off-odor compound in white wines. Journal of Agricultural and Food Chemistry 59 18 10191-10199 Nikolantonaki, M., Thibon, C., Shinoda, K., Tesseidre, P.-L. and Darriet, P. (2012a), Levels and influence of flavan-3-ols must contents on varietal aroma of young Sauvignon blanc wines: Effect of pressing conditions, grape origin and vintage in Oeno 2011 - 9e Symposium International d'Oenologie de Bordeaux Darriet, P., et al. (Eds); Dunod, Bordeaux, pp. 708-715 Nikolantonaki, M., Jourdes, M., Shinoda, K., Teissedre, P.-L., Quideau, S. and Darriet, P. (2012), Identification of adducts between an odoriferous volatile thiol and oxidized grape phenolic compounds: Kinetic study of adduct formation under chemical and enzymatic oxidation conditions. Journal of Agricultural and Food Chemistry 60 10 2647-2656

Cited Literature

195

Noctor, G., Queval, G., Mhamdi, A., Chaouch, S. and Foyer, C. H. (2011), Glutathione. The Arabidopsis Book 9 1-32 Nykänen, L. (1986), Formation and occurrence of flavor compounds in wine and distilled alcoholic beverages. American Journal of Enology and Viticulture 37 1 84-96

O Ojeda, H. N., Andary, C., Kraeva, E., Carbonneau, A. and Deloire, A. (2002), Influence of pre- and postveraison water deficit on synthesis and concentration of skin phenolic compounds during berry growth of Vitis vinifera cv. Shiraz. American Journal of Enology and Viticulture 53 4 261-267 Ormeno, E., Mévy, J. P., Vila, B., Bousquet-Mélou, A., Greff, S., Bonin, G. and Fernandez, C. (2007), Water deficit stress induces different monoterpene and sesquiterpene emission changes in mediterranean species. Relationship between terpene emissions and plant water potential. Chemosphere 67 2 276-284 Ou, C., Du, X., Shellie, K., Ross, C. and Qian, M. C. (2010), Volatile compounds and sensory attributes of wine from cv. Merlot (Vitis vinifera L.) grown under differential levels of water deficit with or without a Kaolin-based, foliar reflectant particle film Journal of Agricultural and Food Chemistry 58 24 12890-12898

P Park, S. K., Morrison, J. C., Adams, D. O. and Noble, A. C. (1991), Distribution of free and glycosidically bound monoterpenes in the skin and mesocarp of Muscat of Alexandria grapes during development. Journal of Agricultural and Food Chemistry 39 3 514-518 Parr, W. V., Heatherbell, D. and White, K. G. (2002), Demystifying wine expertise: Olfactory threshold, perceptual skill and semantic memory in expert and novice judges. Chemical Senses 27 747-755 Parr, W. V., Green, J. A., White, K. G. and Sherlock, R. R. (2007), The distinctive flavour of New Zealand Sauvignon blanc: Sensory characterisation by wine professionals. Food Quality and Preference 18 6 849-861

Cited Literature

196

Parr, W. V., Valentin, D., Green, J. A. and Dacremont, C. (2010), Evaluation of French and New Zealand Sauvignon wines by experienced French wine assessors. Food Quality and Preference 21 1 56-64 Passebois-Ducros, J., Trinquecoste, J.-F. and Viot, C. (2012), De l'influence du jugement de typicalité des étiquettes sur le jugement de typicité d'un vin : Une application aux vins de Bordeaux at 11th International marketing trends conference in Venise, Italy 19-21 January 2012 Perrin, L. (2008), Contribution méthodologique á l'analyse sensorielle du vin, Thèse de Doctorat de l'ENSAR, Ecole Supérieure d'Agriculture - Angers, Angers Perrin, L. and Pagès, J. (2009), A methodology for the analysis of sensory typicality judgments. Journal of Sensory Studies 24 5 749-773 Perrin, L., Symoneaux, R., Maître, I., Asselin, C., Jourjon, F. D. R. and Pages, J. R. M. (2007), Comparison of conventional profiling by a trained tasting panel and free profiling by wine professionals. American Journal of Enology and Viticulture 58 4 508-517 Peyrot des Gachons, C., Tominaga, T. and Dubourdieu, D. (2000), Measuring the aromatic potential of Vitis vinifera L. cv. Sauvignon blanc grapes by assaying S-cysteine conjugates, precursors of the volatile thiols responsible for their varietal aroma. Journal of Agricultural and Food Chemistry 48 8 3387-3391 Peyrot des Gachons, C., Tominaga, T. and Dubourdieu, D. (2002a), Sulfur aroma precursor present in S-glutathione conjugate form: Identification of S-3-(hexan-1-ol)-glutathione in must from Vitis vinifera L. cv. Sauvignon blanc. Journal of Agricultural and Food Chemistry 50 14 4076-4079 Peyrot des Gachons, C., Tominaga, T. and Dubourdieu, D. (2002b), Localization of S-cysteine conjugates in the berry: Effect of skin contact on aromatic potential of Vitis vinifera L. cv. Sauvignon blanc must. American Journal of Enology and Viticulture 53 2 144-146 Peyrot des Gachons, C., van Leeuwen, C., Tominaga, T., Soyer, J.-P., Gaudillère, J.-P. and Dubourdieu, D. (2005), Influence of water and nitrogen deficit on fruit ripening and aroma potential of Vitis vinifera L. cv. Sauvignon blanc in field conditions. Journal International des Sciences de la Vigne et du Vin 85 73-85 73

Cited Literature

197

Pfaffmann, C., Schlossberg, H. and Cornsweet, J. (1954), Variables affecting difference tests in Food acceptance testing methodology; Pilgrim, J. F. and Peterson, M. S. (Eds); Quartermaster Food and Container Institute, Chicago, pp. 4-20 Pineau, B. (2007b), Contribution à l'étude de l'arôme fruité spécifique des vins rouges de Vitis vinifera L. cv. Merlot noir et Cabernet-Sauvignon, Thèse de Doctorat n° 1484, Université Victor Segalen Bordeaux 2, Bordeaux Pineau, B., Barbe, J.-C., van Leeuwen, C. and Dubourdieu, D. (2007), Which impact for β-damascenone on red wines aroma? Journal of Agricultural and Food Chemistry 55 10 4103-4108 Pineau, B., Barbe, J.-C., van Leeuwen, C. and Dubourdieu, D. (2009), Examples of perceptive interactions involved in specific 'red-' and 'black-berry' aromas in red wines. Journal of Agricultural and Food Chemistry 57 9 3702-3708 Pineau, B., Barbe, J.-C., van Leeuwen, C., Dubourdieu, D. and Darriet, P. (2010), Spécifité aromatique des vins rouges de Bordeaux - étude de leurs nuances de fruit rouges et fruits noirs. Revue des Oenologues 134 1 Pineau, B., Trought, M. C. T., Stronge, K., Beresford, M. K., Wohlers, M. W. and Jaeger, S. R. (2011), Influence of fruit ripeness and juice chaptalisation on the sensory properties and degree of typicality expressed by Sauvignon blanc wines from Marlborough, New Zealand. Australian Journal of Grape and Wine Research 17 3 358-367 Pinu, F. R., Jouanneau, S., Nicolau, L., Gardner, R. C. and Villas-Boas, S. G. (2012), Concentrations of the volatile thiol 3-mercaptohexanol in Sauvignon blanc wines: No correlation with juice precursors. American Journal of Enology and Viticulture 63 3 407-412 Piombino, P., Nicklaus, S., Fur, Y. L., Moio, L. and Quara, J.-L. L. (2004), Selection of products presenting given flavor characteristics: An application to wine. American Journal of Enology and Viticulture 55 1 27-34 Pons, A. (2006), Recherches sur l'arôme de vieillissement prèmaturé des vins, Thèse de Doctorat, Université de Bordeaux 2, Bordeaux Pons, A., (2009) Personal communication.

Cited Literature

198

Q Qian, M. C., Fang, Y. and Shellie, K. (2009), Volatile composition of Merlot wine from different vine water status. Journal of Agricultural and Food Chemistry 57 16 7459-7463 Quick, W. P., Chaves, M. M., Wendler, R., David, M., Rodrigues, M. L., Passaharinho, J. A., Pereira, J. S., Adcock, M. D., Leegood, R. C. and Stitt, M. (1992), The effect of water stress on photosynthetic carbon metabolism in four species grown under field conditions. Plant, Cell & Environment 15 1 25-35

R Rapp, A. and Hastrich, H. (1976), Gaschromatische Untersuchungen über die Aromastoffe von Weinbeeren ii. Möglichkeiten der Sortencharakterisierung. Vitis 15 183-192 Rapp, A. and Knipser, W. (1979), 3,7-dimethyl-okta-1,5-dien-3,7-diol - eine neue terpenoide Verbindung des Trauben- und Weinaromas. Vitis 18 229-233 Rapp, A. and Güntert, M. (1985a), Beitrag zur Charakterisierung des Weines der Rebsorte Weißer Riesling ii. Untersuchung der Aromastoffzusammensetzung deutscher Weißweine der Rebsorten Weißer Riesling, Müller-Thurgau und Silvaner. Vitis 24 139-150 Rapp, A. and Mandery, H. (1986), Wine aroma. Experientia 42 873-884 Rapp, A., Mandery, H. and Ullemeyer, H. (1984), Neue Monoterpendiole in Traubenmosten und Weinen und ihre Bedeutung für die Genese einiger cyclischer Monoterpenäther. Vitis 23 84-92 Rapp, A., Güntert, M. and Ullemeyer, H. (1985), Über Veränderungen der Aromastoffe während der Flaschenlagerung von Weißweinen der Rebsorte Riesling. Zeitschrift für Lebensmitteluntersuchung und -Forschung A 180 2 109-116 Rapp, A., Güntert, M. and Heimann, W. (1985b), Beitrag zur Sortencharakterisierung der Rebsorte Weißer Riesling. Zeitschrift für Lebensmitteluntersuchung und -Forschung A 181 5 357-361

Cited Literature

199

Rapp, A., Mandery, H. and Niebergall, H. (1986b), Neue Monoterpenediole in Traubenmost und Wein sowie in Kulturen von Botrytis cinerea. Vitis 25 79-84 Rapp, A., Volkmann, C. and Niebergall, H. (1993), Untersuchung flüchtiger Inhaltsstoffe des Weinaromas: Beitrag zur Sortencharakterisierung von Riesling und Neuzüchtungen mit Riesling-Abstammung. Vitis 32 171-178 Rapp, A., Yavas, I. and Hastrich, H. (1994), Einfache und schnelle Anreicherung ("Kaltronmethode") von Aromastoffen des Weines und deren quantitative Bestimmung mittels Kapillargaschromatographie. Deutsche Lebensmittelrundschau 90 171-174 Rauhut, D., Beisert, B., Berres, M., Gawron-Scibek, M. and Kuerbel, H. (2005), Pulse flame photometric detection: An innovative technique to analyse volatile sulfur compounds in wine and other beverages in State-of-the-art in flavour chemistry and biology; Hofmann, T., et al. (Eds); Deutsche Forschungsanstalt für Lebensmittelchemie, Garching, pp. 363-368 Razungles, A., Bayonove, C. L., Cordonnier, R. E. and Baumes, R. L. (1987), Etude des caroténoides du raisin à maturité. Vitis 26 183-191 Razungles, A., Bayonove, C. L., Cordonnier, R. E. and Sapis, J. C. (1988), Grape carotenoids: Changes during the maturation period and localization in mature berries. American Journal of Enology and Viticulture 39 1 44-48 Reynolds, A. G., Wardle, D. A. and Naylor, A. P. (1996), Impact of training system, vine spacing, and basal leaf removal on Riesling vine performance, berry composition, canopy microclimate, and vineyard labor requirements. American Journal of Enology and Viticulture 47 1 63-76 Reynolds, A. G., Wardle, D. A., Hall, J. W. and Dever, M. (1995), Fruit maturation of four Vitis vinifera cultivars in response to vineyard location and basal leaf removal. American Journal of Enology and Viticulture 46 4 542-558 Ribéreau-Gayon, P., Boidron, J. N. and Terrier, A. (1975), Aroma of Muscat grape varieties. Journal of Agricultural and Food Chemistry 23 6 1042-1047

Cited Literature

200

Ribéreau-Gayon, P., Glories, Y., Dubourdieu, D. and Donèche, B., (2006) Handbook of enology - the chemistry of wine stabilization and treatments, John Wiley & Sons Ltd., Chichester Risvik, E., Colwill, J. S., Mcewan, J. A. and Lyon, D. H. (1992), Multivariate analysis of conventional profiling data: A comparison of a British and a Norwegian trained panel. Journal of Sensory Studies 7 2 97-118 Robert (1993) Le nouveau Petit Robert, Dictionaires Le Robert, Paris, troisiéme édition Roby, G., Harbertson, J. F., Adams, D. A. and Matthews, M. A. (2004), Berry size and vine water deficits as factors in winegrape composition: Anthocyanins and tannins. Australian Journal of Grape and Wine Research 10 2 100-107 Rogiers, S. Y., Hatfield, J. M. and Keller, M. (2004), Irrigation, nitrogen, and rootstock effects on volume loss of berries from pottet Shiraz vines. Vitis 43 1 - 6 Rojas, V., Gil, J. V., Pieaga, F. and Manzanares, P. (2003), Acetate ester formation in wine by mixed cultures in laboratory fermentations. International Journal of Food Microbiology 86 12 181-188 Roland, A., Vialaret, J., Razungles, A., Rigou, P. and Schneider, R. (2010), Evolution of S-cysteinylated and S-glutathionylated thiol precursors during oxidation of Melon B. and Sauvignon blanc musts. Journal of Agricultural and Food Chemistry 58 7 4406-4413 Roland, A. L., Schneider, R. M., Razungles, A. and Cavelier, F. (2011), Varietal thiols in wine: Discovery, analysis and applications. Chemical Reviews 111 11 7355-7376 Rosch, E. and Mervis, C. B. (1975), Family resemblances: Studies in the internal structure of categories. Cognitive Psychology 7 573-605 Ruzicka, L. (1963), Perspektiven der Biogenese und der Chemie der Terpene. Pure Applied Chemistry 6 4 493-524

Cited Literature

201

S Sacks, G. L., Gates, M. J., Ferry, F. X., Lavin, E. H., Kurtz, A. J. and Acree, T. E. (2012), Sensory threshold of 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN) and concentrations in young Riesling and non-Riesling wines. Journal of Agricultural and Food Chemistry 60 12 2998-3004 Salette, J. (1997), La typicité: Une notion nouvelle au service du produit, de ceux qui l'élaborent et de ceux qui le consomment en l'appréciant. Revue des Oenologues 85 11-13 Santosa, M., Hautus, M. and O'mahony, M. (2011), Roc curve analysis to determine effects of repetition on the criteria for same - different and A Not-A tests. Food Quality and Preference 22 1 66-77 Sarrazin, E. (2007b), Recherches sur l'arôme des vins liquoreux de pourriture noble issus des cépages Sémillon et Sauvignon blanc - characterisation de composés clés et étude de leur genese, Thèse de Doctorat, Université Victor Segalen Bordeaux 2, Bordeaux Sarrazin, E., Tominaga, T. and Darriet, P. (2008), Correlation between sensory typicality and aromatic composition in Sauternes botrytised wines in Expression of multidisciplinary flavour science - proceedings of the 12th Weurmann Symposium Interlaken, Switzerland, 2008; Blank, I., et al. (Eds); Zürcher Hochschule für Angewandte Wissenschaften, Winterthur, Zürich, pp. 72-75 Sarrazin, E., Shinkaruk, S., Tominaga, T., Bennetau, B., Frérot, E. and Dubourdieu, D. (2007), Odorous impact of volatile thiols on the aroma of young botrytized sweet wines: Identification and quantification of new sulfanyl alcohols. Journal of Agricultural and Food Chemistry 55 4 1437-1444 Sarrazin, E., Shinkaruk, S., Pons, M., Thibon, C., Bennetau, B. and Darriet, P. (2010), Elucidation of the 1,3-sulfanylalcohol oxidation mechanism: An unusual identification of the disulfide of 3-sulfanylhexanol in Sauternes botrytized wines. Journal of Agricultural and Food Chemistry 58 19 10606-10613

Cited Literature

202

Sauvageot, F. (1994), Le sciences de l'aliment et le concept de typicité ou le chercheur en sciences de la nature a-t-il quelque chose à declarer sur la typicité d'un produit alimentaire ? Sciences des Aliments 14 557-571 Schaller, K., (2000) Praktikum der Bodenkunde und Pflanzenernährung, Gesellschaft zur Förderung der Forschungsanstalt Geisenheim e.V., Geisenheim Schneider, R., Razungles, A., Augier, C. and Baumes, R. (2001), Monoterpenic and norisoprenoidic glycoconjugates of Vitis vinifera L. cv. Melon B. as precursors of odorants in Muscadet wines. Journal of Chromatography A 936 12 145-157 Schneider, R., Charrier, F., Razungles, A. and Baumes, R. (2006), Evidence for an alternative biogenetic pathway leading to 3-mercaptohexanol and 4-mercapto-4-methylpentan-2-one in wines. Analytica Chimica Acta 563 12 58-64 Schneider, R. M., Kotseridis, Y., Ray, J.-L. C., Augier, C. and Baumes, R. (2003), Quantitative determination of sulfur-containing wine odorants at sub parts per billion levels. 2. Development and application of a stable isotope dilution assay. Journal of Agricultural and Food Chemistry 51 11 3243-3248 Schoch, E., Benda, I. and Schreier, P. (1991), Bioconversion of α-damascenone by Botrytis cinerea. Applied and Enironmental Microbiology 57 1 15-18 Scholander, P. F., Hammel, H. T., Bradstreet, E. D. and Hemmingen, E. A. (1965), Sap pressure in vascular plants - negative hydrostatic pressure can be measured in plants. Science 148 339-346 Schreier, P. and Drawert, F. (1974a), Gaschromatographisch-massenspektrometrische Untersuchung flüchtiger Inhaltsstoffe des Weines. Zeitschrift für Lebensmitteluntersuchung und -forschung A 154 5 273-278 Schreier, P., Drawert, F. and Junker, A. (1974b), Identifizierung von 3,7-dimethyl-1,5,7-octatrien-3-ol als flüchtige komponente des Trauben- und Weinaromas. Zeitschrift für Lebensmitteluntersuchung und -forschung A 155 2 98-99 Schreier, P., Drawert, F. and Junker, A. (1976), Identification of volatile constituents from grapes. Journal of Agricultural and Food Chemistry 24 2 331-336

Cited Literature

203

Schultz, H. (2000), Climate change and viticulture: A European perspective on climatology, carbon dioxide and UV-B effects. Australian Journal of Grape and Wine Research 6 1 2-12 Schultz, H. R. and Lebon, E. (2005), Modelling the effets of climate change on grapevine water relations. Acta Hort. (ISHS) 689 71-78 Schultz, H. R., Löhnertz, O., Bettner, W., Balo, B., Linsenmeier, A., Jähnisch, A., Müller, M., Gaubatz, B. and Varadi, G. (1998), Is grape composition affected by current levels of UV-B radiation ? Vitis 37 4 191 - 192 Sefton, M. A. (1998), Hydrolytically-released volatile secondary metabolites from a juice sample of Vitis vinifera grape cv. Merlot and Cabernet Sauvignon. Australian Journal of Grape and Wine Research 4 1 30-38 Sefton, M. A., Francis, I. L. and Williams, P. J. (1993), The volatile composition of Chardonnay juices: A study by flavor precursor analysis. American Journal of Enology and Viticulture 44 4 359-370 Segurel, M. A., Razungles, A. J., Riou, C., Salles, M. and Baumes, R. L. (2004), Contribution of dimethyl sulfide to the aroma of Syrah and Grenache noir wines and estimation of its potential in grapes of these varieties. Journal of Agricultural and Food Chemistry 52 23 7084-7093 Silva Ferreira, A. N. C. S. and Guedes De Pinho, P. (2004), Norisoprenoids profile during port wine ageing - influence of some technological parameters. Analytica Chimica Acta 513 1 169-176 Silva, M. A., (2011) Effet de l'obturateur sur l'évolution de la qualité sensorielle du vin, Thèse de Doctorat, Université de Bordeaux 2, Bordeaux Simpson, R. F. (1979), Aroma and compositional changes in wine with oxidation, storage and ageing. Vitis 17 274-387 Simpson, R. F. and Miller, G. C. (1983), Aroma composition of aged Riesling wine. Vitis 22 51-63

Cited Literature

204

Singleton, V. L., Salgues, M., Zaya, J. and Trousdale, E. (1985), Caftaric acid disappearance and conversion to products of enzymic oxidation in grape must and wine. American Journal of Enology and Viticulture 36 1 50-56 Skinkis, P. A., Bordelon, B. P. and Wood, K. V. (2008), Comparison of monoterpene constituents in Traminette, Gewürztraminer, and Riesling winegrapes. American Journal of Enology and Viticulture 59 4 440-445 Smart, R. and Robinson (1996) Sunlight into wine - a handbook for winegrape canopy management, Winetitles, Adelaide, pp. Smyth (2005), The compositional basis of the aroma of Riesling and unwooded Chardonnay wine, Doctoral thesis, The University of Adelaide, Adelaide Sponholz, W. R. and Hühn, T. (1997), Einfußfaktoren von Klonenematerial und verwendetem Hefestamm auf die Alterung von Riesling Weinen. Vitic. Enol. Sci. 52 2 103-108 Staab, J., (1986) Zur Herkunft der Rieslingrebe - die Geschichte einer Rebsorte, PRO RIESLING, Verein zur Förderung der Riesling-Kultur e.V., Trier, pp. 9 - 11 Stone, H., Sidel, J., Oliviers, S., Woosley, A. and Singleton, R. C. (1974), Sensory evaluation by quantitative descriptive analysis. Food Technol. November 24-28 Strauss, C. R., Dimitriadis, E., Wilson, B. and Williams, P. J. (1986), Studies on the hydrolysis of two megastigma-3,6,9-triols rationalizing the origins of some volatile C13-norisoprenoids of Vitis vinifera grapes. Journal of Agricultural and Food Chemistry 34 1 145-149 Strauss, C. R., Wilson, B., Anderson, R. and Williams, P. J. (1987), Development of precursors of C13-norisoprenoid flavorants in Riesling grapes. American Journal of Enology and Viticulture 38 1 23-27 Subileau, M., Schneider, R. M., Salmon, J.-M. and Degryse, E. (2008), New insights on 3-mercaptohexanol (3MH) biogenesis in Sauvignon blanc wines: Cys-3-MH and (E)-hexen-2-al are not the major precursors. Journal of Agricultural and Food Chemistry 56 19 9230-9235

Cited Literature

205

Swiegers, J. H., Pretorius, I. S., Allen I. Laskin, J. W. B. and Geoffrey, M. G. (2005), Yeast modulation of wine flavor in Advances in applied microbiology; (Eds); Academic Press, pp. 131-175 Symoneaux, R., Perrin, L., Wilson, D., Maitre, I. and Jourjon, F. (2010), La typicité sensorielle des vins d'appellation : Comment la mesurer avec les professionels, comment l'aborder avec les consommateurs ? Qualité et Environment 142 94-101

T Tenenhaus, M., Pages, J. R. M., Ambroisine, L. and Guinot, C. (2005), PLS methodology to study relationships between hedonic judgements and product characteristics. Food Quality and Preference 16 4 315-325 Ter Braak, C. J. F. and Verdonschot, P. F. M. (1995), Canonical Correspondence Analysis and related multivariate methods in aquatic ecology. Aquatic Sciences 57 3 255-289 Thibon, C., Cluzet, S., Mérillon, J. M., Darriet, P. and Dubourdieu, D. (2011), 3-sulfanylhexanol precursor biogenesis in grapevine cells: The stimulating effect of Botrytis cinerea. Journal of Agricultural and Food Chemistry 59 4 1344-1351 Thibon, C., Marullo, P., Claisse, O., Cullin, C., Dubourdieu, D. and Tominaga, T. (2008), Nitrogen catabolic repression controls the release of volatile thiols by Saccharomyces cerevisiae during wine fermentation. FEMS Yeast Research 8 7 1076-1086 Thibon, C., Dubourdieu, D., Darriet, P. and Tominaga, T. (2009), Impact of noble rot on the aroma precursor of 3-sulfanylhexanol content in Vitis vinifera L. cv Sauvignon blanc and Semillon grape juice. Food Chemistry 114 4 1359-1364 Tominaga, T. and Dubourdieu, D. (2006), A novel method for quantification of 2-methyl-3-furanthiol and 2-furanmethanethiol in wines made from Vitis vinifera grape varieties. Journal of Agricultural and Food Chemistry 54 1 29-33 Tominaga, T., Darriet, P. and Dubourdieu, D. (1996), Identification of 3-mercaptohexyl acetate in Sauvignon blanc wine, a powerful aromatic compound exhibiting box-tree odor. Vitis 35 207-210 207

Cited Literature

206

Tominaga, T., Murat, M.-L. and Dubourdieu, D. (1998b), Development of a method for analyzing the volatile thiols involved in the characteristic aroma of wines made from Vitis vinifera L. cv. Sauvignon blanc. Journal of Agricultural and Food Chemistry 46 3 1044-1048 Tominaga, T., Peyrot Des Gachons, C. and Dubourdieu, D. (1998c), A new type of flavor precursors in Vitis vinifera L. cv. Sauvignon blanc: S-cysteine conjugates. Journal of Agricultural and Food Chemistry 46 12 5215-5219 Tominaga, T., Guimbertau, G. and Dubourdieu, D. (2003), Contribution of benzenemethanethiol to smoky aroma of certain Vitis vinifera L. wines. Journal of Agricultural and Food Chemistry 51 5 1373-1376 Tominaga, T., Guimbertau, G. and Dubourdieu, D. (2003b), Role of certain volatile thiols in the bouquet of aged champagne wines. Journal of Agricultural and Food Chemistry 51 4 1016-1020 Tominaga, T., Furrer, A., Henry, R. and Dubourdieu, D. (1998a), Identification of new volatile thiols in the aroma of Vitis vinifera L. var. Sauvignon blanc wines. Flavour and Fragrance Journal 13 3 159-162 Tominaga, T., Baltenweck-Guyot, R., Gachons, C. P. D. and Dubourdieu, D. (2000a), Contribution of volatile thiols to the aromas of white wines made from several Vitis vinifera grape varieties. American Journal of Enology and Viticulture 51 2 178-181 Tominaga, T., Blanchard, L., Darriet, P. and Dubourdieu, D. (2000b), A powerful aromatic volatile thiol, 2-furanmethanethiol, exhibiting roast coffee aroma in wines made from several Vitis vinifera grape varieties. Journal of Agricultural and Food Chemistry 48 5 1799-1802 Tominaga, T., Niclass, Y., Frérot, E. and Dubourdieu, D. (2006a), Stereoisomeric distribution of 3-mercaptohexan-1-ol and 3-mercaptohexyl acetate in dry and sweet white wines made from Vitis vinifera (var. Sauvignon blanc and Semillon). Journal of Agricultural and Food Chemistry 54 19 7251-7255 Torres, A. and Van De Velden, M. (2007), Perceptual mapping of multiple variable batteries by plotting supplementary variables in correspondence analysis of rating data. Food Quality and Preference 18 1 121-129

Cited Literature

207

Trognon, L. (2005), Contribution à l'etude des stratégies de la petite entreprise agro-alimentaire. Strategie de distinction par la construction de la typicité, Thèse de Doctorat, Université Montpellier I, Montpellier Troost, G. (1988) Technologie des Weines, 6. neubearb. Aufl., Ulmer, Stuttgart (Hohenheim)

U Ugliano, M. and Moio, L. (2005), Changes in the concentration of yeast-derived volatile compounds of red wine during malolactic fermentation with four commercial starter cultures of Oenococcus oeni. Journal of Agricultural and Food Chemistry 53 26 10134-10139 Ugliano, M. and Moio, L. (2008), Free and hydrolytically released volatile compounds of Vitis vinifera L. cv. Fiano grapes as odour-active constituents of Fiano wine. Analytica Chimica Acta 621 1 79-85

V van Leeuwen, C. and Seguin, G. (1994), Incidence de l'alimentation en eau de la vigne, appréciée par l'état hydrique du fuillage, sur le développement de l'appereil végétatif et la maturation du raisin (Vitis vinifera variété Cabernet franc, Saint-Emilion, 1990). Journal International des Sciences de la Vigne et du Vin 28 81 - 110 van Leeuwen, C. and Seguin, G. (2006), The concept of terroir in viticulture. Journal of Wine Research 17 1 1-10 van Leeuwen, C., Pieri, P. and Vivin, P. (2010), Comparison of three operational tools for the assessment of vine water status: Stem water potential, carbon isotope discrimination measured on grape sugar and water balance in Methodologies and results in grapevine research; Delrot, S., et al. (Eds); Springer Netherlands, pp. 87-106 van Leeuwen, C., Friant, P., Choné, X., Tregoat, O., Koundouras, S. and Dubourdieu, D. (2004), Influence of climate, soil, and cultivar on terroir. American Journal of Enology and Viticulture 55 3 207-217

Cited Literature

208

van Leeuwen, C., Tregoat, O., Choné, X., Bois, B., Pernet, D. and Gaudillère, J.-P. (2009), Vine water status is a key factor in grape ripening and vintage quality for red Bordeaux wine. How can it be assessed for vineyard management purposes? Journal International des Sciences de la Vigne et du Vin 43 3 121 - 134 van Wyk, C. J., Webb, A. D. and Kepner, R. E. (1967a), Some volatile components of Vitis vinifera variety White Riesling. Journal of Food Science 32 6 660-664 van Wyk, C. J., Kepner, R. E. and Webb, A. D. (1967b), Some volatile components of Vitis vinifera variety White Riesling. 3. Neutral components extracted from wine. Journal of Food Science 32 6 669-674 Vermeulen, C., Gijs, L. and Collin, S. (2005), Sensorial contribution and formation pathways of thiols in foods: A review. Food Reviews International 21 1 69-137 Versini, G., Inama, S. and Sartori, G. (1981), A capillary column gas chromatographic research into the terpenes constituents of 'Riesling Rhenano' (Rhine Riesling) wine from Trentino Alto Adige: Their distribution within berries, their passage into must and their presence in the wine according to different wine-making procedures. Organoleptic considerations. Vini d'Italia XXIII 189-211 Versini, G., Rapp, A., Reniero, F. and Mandery, H. (1991), Structural identification and presence of some p-menth-1-enediols in grape products. Vitis 30 143-149 Versini, G., Rapp, A., Marais, J., Mattivi, F. and Spraul, M. (1996), A new 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN) precursor isolated from Riesling grape products: Partial structure elucidation and possible reaction mechanism. Vitis 35 1 15-21 Versini, G. and Tomasi, T. (1983), Confronto trai componenti volatili dei vini rossi ottenuti con macerazione tradizionale e macerazione carbonica. Importanza differenziale del cinnamato di etile. L'Enotecnico 19 595-600 Vollhardt, K. P. C. (1990) Organische Chemie, VCH Verlagsgesellschaft mbH, Weinheim, pp. 341

Cited Literature

209

W Waldmann, D. and Winterhalter, P. (1992), Identification of a novel vitispirane precursor in Riesling wine. Vitis 31 169-174 Webster, D. R., Edwards, C. G., Spayd, S. E., Peterson, J. C. and Seymour, B. J. (1993), Influence of vineyard nitrogen fertilization on the concentrations of monoterpenes, higher alcohols, and esters in aged Riesling wines. American Journal of Enology and Viticulture 44 3 275-284 Williams, D. W., Williams, L. E., Barnett, W. W., Kelley, K. M. and Mckenry, M. V. (1985), Validation of a model for the growth and development of the Thompson seedless grapevine. I. Vegetative growth and fruit yield. American Journal of Enology and Viticulture 36 4 275-282 Williams, P. J., Strauss, C. R., Wilson, B. and Massy-Westropp, R. A. (1982), Novel monoterpene disaccharide glycosides of Vitis vinifera grapes and wines. Phytochemistry 21 8 2013-2020 Wilson, B., Strauss, C. R. and Williams, P. J. (1986), The distribution of free and glycosidically-bound monoterpenes among skin, juice, and pulp fractions of some white grape varieties. American Journal of Enology and Viticulture 37 2 107-111 Winterhalter, P. (1991), 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN) formation in wine. 1. Studies on the hydrolysis of 2,6,10,10-tetramethyl-1-oxaspiro[4.5]dec-6-ene-2,8-diol rationalizing the origin of TDN and related C13-norisoprenoids in Riesling wine. Journal of Agricultural and Food Chemistry 39 10 1825-1829 Winterhalter, P. and Schreier, P. (1988), Free and bound C13-norisoprenoids in quince (cydonia oblonga, mill.) fruit. Journal of Agricultural and Food Chemistry 36 6 1251-1256 Winterhalter, P. and Rouseff, R. (2001), Carotenoid-Derived Aroma Compounds: An introduction in Carotenoid-Derived Aroma Compounds; ACS Symposium Series; Volume 802; American Chemical Society, pp. 1-17

Cited Literature

210

Winterhalter, P., Sefton, M. A. and Williams, P. J. (1990a), Two-dimensional GC-DCCC analysis of the glycoconjugates of monoterpenes, norisoprenoids, and shikimate-derived metabolites from Riesling wine. Journal of Agricultural and Food Chemistry 38 4 1041-1048 Winterhalter, P., Sefton, M. A. and Williams, P. J. (1990b), Volatile C13-norisoprenoid compounds in Riesling wine are generated from multiple precursors. American Journal of Enology and Viticulture 41 4 277-283 Winton, W., Ough, C. S. and Singleton, V. L. (1975), Relative distinctiveness of varietal wines estimated by the ability of trained panelists to name the grape variety correctly. American Journal of Enology and Viticulture 26 1 5-11

Y Yang, J., Zhang, J., Wang, Z., Zhu, Q. and Liu, L. (2002), Abscisic acid and cytokinins in the root exudates and leaves and their relationship to senescence and remobilization of carbon reserves in rice subjected to water stress during grain filling. Planta 215 4 645-652 Yani, A., Pauly, G., Faye, M., Salin, F. and Gleizes, M. (1993), The effect of a long-term water stress on the metabolism and emission of terpenes of the foliage of cupressus sempervirens. Plant, Cell & Environment 16 8 975-981 Yusop, S. M., O'sullivan, M. G., Kerry, J. F. and Kerry, J. P. (2009), Sensory evaluation of chinese-style marinated chicken by Chinese and European native assessors. Journal of Sensory Studies 24 4 512-533

Z Zamora, M. C. and Guirao, M. (2004), Performance comparison between trained assessors and wine experts using specific sensory attributes. Journal of Sensory Studies 19 6 530-545 Zoecklein, B. W., Marcy, J. E., Williams, J. M. and Jasinski, Y. (1997), Effect of native yeasts and selected strains of Saccharomyces cerevisiae on glycosyl glucose, potential volatile terpenes, and selected aglycones of White Riesling (Vitis vinifera L.) wines. Journal of Food Composition and Analysis 10 1 55-65

Cited Literature

211

Zoecklein, B. W., Hackney, C. H., Duncan, S. E. and Marcy, J. E. (1999), Effect of fermentation, aging and thermal storage on total glycosides, phenol-free glycosides and volatile compounds of White Riesling (Vitis vinifera L.) wines. Journal of Industrial Microbiology & Biotechnology 22 2 100-107

Cited Literature

212

Appendix

i

Appendix Appendix 1a Discriptors applied in the German and French sensory panel in the descriptive task, and their

grouping into descriptor families. ................................................................................................................... iii Appendix 2 Chi-square (χ2) values at a significance level α = 0.05 for the ‘A Not-A test’ derived dry Riesling wines

recognition frequency for each Judge and the whole panel for the German (GH) and the French (BX) respectively .................................................................................................................................................... ix 

Appendix 3 Frequencies on descriptor families for the German panel; RIESLING represents the recognition frequency obtained by the A Not-A procedure. ............................................................................................... x 

Appendix 4 Frequencies on descriptor families for the French panel; RIESLING represents the recognition frequency obtained by the A Not-A procedure. .............................................................................................. xi 

Appendix 5 Pearson correlation table for correlation factor r of the descriptor family frequencies, recognition frequency of Riesling wines (RIESLING) and relative typicality ratings (typicality), determined by the German panel; data set comprising Riesling and non-Riesling wines .......................................................... xiii 

Appendix 7 Pearson correlation table for correlation factor r of the descriptor family frequencies, recognition frequency of Riesling wines (RIESLING) and relative typicality ratings (typicality), determined by the German panel; data set comprising only Riesling wines .............................................................................. xiv 

Appendix 8 Pearson correlation table for correlation factor r of the descriptor family frequencies, recognition frequency of Riesling wines (RIESLING) and relative typicality ratings (typicality), determined by the French panel; data set comprising only Riesling wines ............................................................................................. xv 

Appendix 9 Aroma compound’s concentrations in the wines assessed in sensory studies by the German panel: volatile acids; higher alcohols, acetic acid esters, ethyl esters by means of the “Kaltron method”; value 0 indicates concentration below limit of quantification for the concerning aroma compound .......................... xvi 

Appendix 10 Aroma compound’s concentrations in the wines assessed in sensory studies by the German panel: miscellaneous, monoterpenols, monoterpenol oxides by means of the “Kaltron method”, C13-norisoprenoids by means of SBSE and GC-MS; values 0 indicates concentration below limit of quantification for the concerning aroma compound ...................................................................................................................... xvii 

Appendix 11 Aroma compound’s concentrations in the wines assessed in sensory studies by the German panel: sulfur compounds by means of HS-GC-PFPD1 and according to Tominaga et al., (2006)2; value 0 indicates concentration below limit of quantification for the concerning aroma compound ......................................... xiii 

Appendix 12 Aroma compound’s concentrations in the wines assessed in sensory studies by the French panel: volatile acids; higher alcohols, acetic acid esters, ethyl esters by means of the “Kaltron method”; value 0 indicates concentration below limit of quantification for the concerning aroma compound .........................xxiv 

Appendix 13 Aroma compound’s concentrations in the wines assessed in sensory studies by the French panel: miscellaneous, monoterpenols, monoterpenol oxides by means of the “Kaltron method”, C13-norisoprenoids by means of SBSE and GC-MS; values 0 indicates concentration below limit of quantification for the concerning aroma compound ........................................................................................................................ xx 

Appendix 14 Aroma compound’s concentrations in the wines assessed in sensory studies by the French panel: sulfur compounds by means of GC-PFPD1 and according to Tominaga et al., (2006)2; value 0 indicates concentration below limit of quantification for the concerning aroma compound ......................................... xxi 

Appendix 15 Ions (m/z) used for compounds identification and quantification and calibration curve parameters applied for quantification in the analysis of ‘Kaltron’ extracts by means of GC-MS ..................................... xxii 

Appendix 16 Ions (m/z) used for compounds identification and quantification and calibration curve parameters applied for quantification in the analysis of C13-norisoprenoids by means of GC-MS ................................. xxiii 

Appendix

ii

Appendix 17 Ions (m/z) used for compounds identification and quantification and calibration curve parameters applied for quantification in the analysis of volatile thiols by means of GC-MS ........................................... xxiii 

Appendix 18 Ions (m/z) used for compounds identification and quantification and calibration curve parameters applied for quantification in the analysis of the potential of volatile C13-norisoprenoids and terpenes (PVNT) by means of GC-MS....................................................................................................................................xxiv 

Appendix 20 Calibration curve parameters applied for quantification in the analysis of the low boiling sulfur compounds by means of HS-GC-PFPD ...................................................................................................... xxv 

Appendix 20 Ions (m/z) used for compounds identification, quantification and calibration curve parameters applied for quantification in the analysis of cysteinylated and glutathionylated precurors of 3SH by means of HPLC-HRMS 25 

Appendix 21 Calibration parameters for analysis of N-OPA, [NH4]+, glutathione, and caftaric acid ...................... xxv Appendix 22 List of chemical names used in this thesis – equivalence table .......................................................xxvi 

Appendix

iii

Appendix 1a Discriptors applied in the German and French sensory panel in the descriptive task, and their grouping into descriptor families.

descriptor

family English

(translated) original descriptors

(French) original descriptors

(German)

vegetal vegetal green lawn 

cut gras capsicum green capsicum 

pyrazines 

IBMP stinging nettle vegetable asparagus beetroot 

box tree 

cats piss 

burgeon (cassis) 

thiols 

black currant  

vegetal 

vert 

herbe coupée 

poivron 

pyrazines 

IBMP 

légume 

asperge 

buis 

pipi de chat 

burgeon de cassis 

thiols 

bois de cassis 

vegetativ 

pflanzlich 

grün 

Rasen 

grünes Gras 

Paprika 

grüne Paprika 

Pyrazine 

Brennessel 

Gemüse 

Spargel 

rote Beete 

Katzenurin 

schwarze Johannisbeere 

Thiole 

sauvignon sauvignon 

sauvignon 

Sauvignon 

Appendix

iv

Appendix 1b Discriptors applied in the German and French sensory panel in the descriptive task, and their grouping into descriptor families.

descriptor family

descriptors (English - translated)

original descriptors (French)

original descriptors (German)

spicy spicy 

pepper 

white pepper 

clove 

saffron 

anise 

menthol 

broth  

épicé 

poivre 

Poivre blanc 

clou de girofle 

safran 

anis 

menthe 

bouillon 

fond 

würzig 

Suppe 

Maggi 

Ingwer 

Pfeffer 

Gewürznelke 

Minze 

herbs herbs 

herbaceous 

hay 

herbal 

tea aroma 

pine tree 

herbes 

verveine 

feuilles sèches 

foin‐paille 

kräutig 

krautig 

Heu 

Teearoma 

Pinie 

Ribes ssp. / Sambucus ssp.

blackcurrant 

elderberry 

gooseberry 

schwarze Johannisbeere(Frucht) 

Holunder (Frucht) 

Stachelbeere (Frucht) 

fruity fruity 

strawberry 

fruit syrup 

coconut  

fruité 

fraise 

sirop de fruit 

noix  de coco 

fruchtig 

Erdbeere 

Steinfrucht 

citrus fruit citrus fruit 

grapefruit 

lemon 

lemon zest 

bergamot 

orange 

mandarine  

agrumes 

pamplemousse 

orange 

bergamotte 

citron 

mandarine 

Zitrusfrüchte 

Limette 

Pampelmuse 

Zitrone 

white fruit white fruit 

pear 

apple  

fruits blancs 

poire 

pomme  

weisse Früchte 

Apfel 

grüner Apfel 

Birne 

Appendix

v

Appendix 1c Discriptors applied in the German and French sensory panel in the descriptive task, and their grouping into descriptor families.

descriptor family

descriptors (English - translated)

original descriptors (French)

original descriptors (German)

yellow fruit yellow fruit 

peach 

white peach 

apricot 

nectarine 

quince 

melon   

fruit jaune 

pêche 

pêche blanche

abricot 

nectarine 

coing 

melon 

gelbe Früchte 

Aprikose 

Pfirsich 

Marille  

tropical fruit tropical fruit 

exotic fruit 

exotic 

passion fruit 

lychee 

mango 

guava 

figs  

fruits tropicaux 

fruit exotique 

exotique 

fruit de la passion 

litchi 

mangue 

goyave 

figues 

tropische Früchte 

exotische Früchte 

tropisch 

Passionsfrucht 

Litschi 

Guave   

dried / candied fruit

dried / candied fruit  

candied 

sweet 

quince jelly 

candied quince 

candied grape berries 

candied lemon 

dried apricot 

dry grape berries  

confit de fruit sec 

confit 

sucre 

confit de coing 

pate de coing  

confit de raisin  

confit de citron 

abricot sec 

raisin sec  

getrocknete/ eingemachte Früchte

süß 

Marmelade  

floral floral 

white flowers 

muscaté 

terpenic 

acacia flower 

lily 

jasmine 

rose  

floral 

fleurs blanches

muscate 

terpenique fleur d'acacialis 

jasmin 

rose 

floral 

blumig 

weiße Blüten 

Muskat 

süßliche Blumen

parfümiert 

Rose 

Akazie  

Appendix

vi

Appendix 1d Discriptors applied in the German and French sensory panel in the descriptive task, and their grouping into descriptor families.

descriptor family

descriptors (English - translated)

original descriptors (French)

original descriptors (German)

aging notes aging notes 

evolved 

oxidation 

ripe 

petroleum 

kerosene 

TDN 

ATA 

wax 

caramel   

honey   

notes de vieillissement

evolué 

oxidation 

mûr 

pétrole 

kérosène 

TDN 

UTA 

cire 

caramel   

miel 

Alterungsnoten 

firn 

oxidativ 

Reife 

wachsig 

Schuhcreme 

Petrolton 

TDN 

UTA 

Honig  

minerality mineral 

shell 

oyster shell 

limestone 

wet stone 

dusty 

metallic 

flintstone  

mineral 

coquillage 

coquillage d'huitre

pierre humide 

poussiéreux 

metallique 

silex  

mineralig 

steinig 

nasser Fels 

nasser Stein 

Feuerstein  

fermentative aroma

fermentative aroma 

artificial  

ester 

fruity ester 

soapy 

banana 

pineapple 

candy aroma 

glacial ice candies 

gummy bears   

arômes fermentatifs

savonneux 

banane 

ananas 

bonbon  

fermentatives Aroma

künstlich  

Ester 

Fruchtester 

seifig 

Banane 

Ananas 

Bonbonaroma 

Eisbonbon 

Gummibärchen  

Appendix

vii

Appendix 1e Discriptors applied in the German and French sensory panel in the descriptive task, and their grouping into descriptor families.

descriptor family

descriptors (English - translated)

original descriptors (French)

original descriptors (German)

sulfur sulfur 

reduced 

sulfur off‐flavour 

hydrogen sulfide 

smoky 

grilled 

sweat 

coffee 

sulphidic 

rubber  

soufré 

reduit 

fumée 

grillé 

sueur 

plastique 

caoutchouc 

ailloli  

Schwefel 

reduktiv 

Böckser 

H2S 

rauchig 

Röstaromen 

schweißig 

Kaffee 

sulfidisch 

Gummi 

MLF MLF 

lactic 

rancid 

butter 

milky 

cheesy 

yoghurt  

MLF 

lactique 

beurre 

rance  

BSA 

laktisch 

ranzig 

buttrig 

milchig 

käsig 

Joghurt  

off-flavour off‐flavour 

unpleasant 

unclean 

volatile acidity 

sulphitic 

musty 

cork 

TCA 

brettanomyces 

medical 

glue 

ethylacetate  

défaut 

acescent 

acidité volatile

sulfitique 

liège 

bouchonné 

bouchon 

amylique  

Fehlton 

unangenehm 

unsauber 

flüchtige Säure 

SO2 

muffig 

Kork 

TCA 

Brettanomyces 

medizinisch 

Klebstoffton 

UHU‐Note 

Ethylacetat 

Appendix

viii

Appendix 1f Discriptors applied in the German and French sensory panel in the descriptive task, and their grouping into descriptor families.

descriptor family

descriptors (English - translated)

original descriptors (French)

original descriptors (German)

others others 

wood 

oak 

barrique 

vanilla 

yeast 

bread 

ginger candy 

smoky 

smokehouse 

greasy / bacon 

meaty 

mouldy 

earthy 

wet dog 

dog waste 

plastic 

cardboard 

hand lotion 

scotch whiskey 

botrytis 

spontaneous fermentation caramel crunchy nut 

brittle  

autres 

bois 

levure 

pain d'épice 

biscuit 

noisette 

cacahouète 

champignon 

terre 

tourbé 

saucisse 

 

andere 

Holz 

Eiche 

Barrique 

Vanille 

Hefe 

Brot 

Ingwer Bonbon 

rauchig 

Räucherkammer 

speckig 

fleischig 

schimmelig 

erdig 

nasser Hund 

Hundekot 

Kunststoff 

Karton 

Handlotion 

Scotch Whiskey 

Botrytis 

Spontangärung 

Karamell 

Honig Cornflakes 

crunchy nut 

Krokant  

Appendix

ix

Appendix 2 Chi-square (χ2) values at a significance level = 0.05 for the ‘A Not-A test’ derived dry Riesling wines recognition frequency for each Judge and the whole panel for the German (GH) and the French (BX) respectively

Chi 2 (critical) = 3.84 @ α = 0.05

GH BX Chi 2 (observed) Chi 2 (observed)

Judge A 18.26 Judge A 13.03 Judge B 11.94 Judge B 16.37 Judge C 15.97 Judge C 16.36 Judge D 13.70 Judge D 10.26 Judge E 9.36 Judge E 6.32 Judge F 13.49 Judge F 8.31 Judge G 9.46 Judge G 7.27 Judge H 11.94 Judge H 14.26 Judge I 10.44 Judge I 10.91 Judge J 25.11 Judge J 20.64 Judge K 11.81

TOTAL panel 143.43 TOTAL panel 111.81

mean Chi 2 13.77 mean Chi 2 12.37

stddev 4.40 stddev 4.36 min 9.36 min 6.32 max 25.11 max 20.64

Appendix

x

Appendix 3 Frequencies on descriptor families for the German panel; RIESLING represents the recognition frequency obtained by the A Not-A procedure.

vege

tal

sauv

igno

n

spic

y

herb

s

Rib

es /

Sam

bucu

s

fruity

citru

s fru

it

whi

te fr

uit

yello

w fr

uit

tropi

cal f

ruit

drie

d /

cand

ied

fruit

flora

l

agei

ng n

otes

min

eral

ity

ferm

enta

tive

arom

a

sulfu

r

MLF

off-f

lavo

ur

othe

rs

RIE

SLIN

G

SBSA1 15 22 0 1 4 5 1 0 0 4 0 0 0 0 0 0 0 0 0 0 SBNZ 18 22 0 1 6 2 1 0 0 1 0 0 0 0 0 0 0 0 0 1 SBBX2 16 23 0 0 3 3 2 0 0 1 0 0 0 0 0 2 0 0 0 1 SBSA2 16 23 2 2 5 2 2 0 0 0 0 0 0 0 0 0 0 1 0 0 SBBX 7 22 1 0 9 6 2 0 0 2 2 0 1 0 0 0 0 0 0 4 CHAUS 2 1 1 0 0 1 0 0 0 0 0 1 1 0 3 1 2 0 17 2 CHRCH 2 0 2 0 0 0 0 0 0 0 0 0 0 0 0 3 2 1 13 3 RIMO2 3 1 0 0 0 6 0 0 3 0 0 10 4 0 0 3 1 1 1 8 RIAL8 0 0 0 0 0 5 2 1 1 0 0 0 2 2 0 0 0 14 2 9 RIAL3 0 0 0 0 0 2 0 1 0 1 0 0 5 1 1 1 1 2 1 11 RINA1 2 0 0 2 0 1 0 0 0 0 1 1 2 0 0 4 0 0 4 8 RIRH12 4 1 0 0 0 1 0 0 0 0 0 0 0 3 2 1 0 0 0 5 RIAL7 1 0 0 0 0 2 1 0 0 0 0 0 2 0 0 2 0 11 1 14 RIMO3 0 0 0 0 0 7 1 1 3 1 0 9 0 2 1 2 0 1 2 10 RIRH14 0 0 0 0 0 11 2 2 3 1 0 0 3 1 0 0 5 1 5 15 RIRH13 0 0 1 0 0 6 1 0 3 1 0 1 6 0 1 0 2 0 0 15 RIRH4 0 0 0 1 0 4 0 1 1 0 0 0 1 0 1 0 0 0 3 12 RIPF 0 0 1 1 0 6 1 0 2 0 0 1 1 2 1 1 0 0 0 15 RIRH5 0 0 0 0 0 5 1 2 1 0 0 0 1 1 0 1 0 1 1 16 RINZ 1 0 0 1 0 4 0 0 1 2 0 1 14 1 0 0 0 3 4 14 RIAUT1 0 0 1 0 0 7 0 1 3 1 1 4 2 0 3 0 0 0 1 13 RIRH1 0 0 0 0 0 4 0 1 0 0 0 0 4 2 1 0 0 0 1 13 RIRH2 1 1 0 0 0 7 1 0 3 1 0 0 4 1 1 0 0 0 1 15 RIRH11 0 0 0 0 0 10 4 1 1 1 0 1 6 2 0 0 0 0 1 14 RIRH6 0 1 0 0 0 7 3 1 2 0 0 2 0 3 0 0 0 0 2 17 RIMO1 0 0 1 0 0 9 2 2 1 0 1 2 0 3 3 0 0 0 0 15 RIAUT2 2 0 0 1 0 10 2 2 4 0 0 5 0 3 3 0 0 0 0 15 RIRH9 1 0 0 0 0 10 1 3 5 0 0 0 4 1 0 0 0 0 1 17 RIRH3 1 0 0 0 0 8 3 0 4 0 0 1 6 1 0 0 0 0 0 17 RIRH8 0 0 0 0 0 7 3 0 3 0 1 1 1 2 2 0 0 0 1 14

Appendix

xi

Appendix 4 Frequencies on descriptor families for the French panel; RIESLING represents the recognition frequency obtained by the A Not-A procedure.

vege

tal

sauv

igno

n

spic

y

herb

s

fruity

citru

s

fruit

whi

te

fruit

yello

w

fruit

tropi

cal

fruit

drie

d /

cand

ied

fruit

flora

l

agei

ng n

otes

min

eral

ity

ferm

enta

tive

arom

a

sulfu

r

MLF

off-f

lavo

ur

othe

rs

RIE

SLIN

G

SBBX 5 10 2 4 6 0 2 2 2 0 1 3 1 2 0 0 2 1 2 SBAL 7 4 1 0 14 3 0 7 3 1 1 4 1 0 2 0 0 1 2 SBNZ 18 7 3 2 4 1 0 0 0 2 2 1 2 0 1 0 0 0 0 RIAL4 3 1 0 0 12 7 2 0 0 1 1 2 2 2 2 1 0 3 8 RIAL3 0 0 0 0 12 1 3 3 3 2 2 6 2 0 0 0 1 3 9 GTAL 0 0 0 0 6 0 1 1 1 3 14 3 0 0 2 0 0 2 4 PGAL 0 0 0 0 7 1 0 2 2 1 18 3 0 0 0 0 0 1 4 RINA1 1 0 1 0 11 5 1 2 2 0 2 3 2 0 1 1 1 3 3 RIAL5 2 0 2 0 8 2 2 2 0 1 1 2 4 0 3 1 1 1 6 RIRH4 4 0 0 1 11 5 1 2 2 0 2 1 2 0 1 0 0 0 5 RIAL2 0 0 1 0 15 1 0 5 5 2 5 6 1 0 1 1 1 0 7 RIAL1 0 0 1 0 5 1 1 2 1 0 3 6 5 1 1 1 4 1 7 RIRH1 3 1 0 0 13 8 0 1 2 0 2 2 2 0 0 1 0 1 7 RIRH8 0 0 0 1 9 5 0 3 0 0 8 3 1 0 1 0 2 0 11 RIRH3 0 1 1 0 10 7 0 2 1 0 5 5 1 1 1 0 0 0 9 RINZ 0 0 0 0 9 5 0 1 1 1 3 14 4 0 1 0 0 1 10 RIRH2 1 0 0 0 10 4 2 2 2 0 4 5 2 1 0 0 1 0 2 RIAL7 1 0 1 0 8 4 0 1 3 0 1 3 5 0 4 0 0 4 4 RIAUT2 1 0 1 0 16 5 3 4 2 0 3 1 0 1 0 1 0 1 11 RIRH5 1 0 1 0 13 7 1 2 3 0 1 1 2 0 2 2 0 0 10 RIRH6 0 0 0 0 10 6 1 1 0 1 5 3 5 1 0 0 0 1 10 RIAL6 3 1 0 1 12 6 3 1 1 0 3 5 7 0 1 0 0 1 8 RIMO1 1 1 0 0 8 7 0 0 1 0 10 1 1 1 0 0 0 0 10 RIRH11 0 0 0 0 10 5 1 2 1 0 6 3 1 0 0 0 1 0 12 RIRH9a 1 1 2 0 12 5 1 3 1 1 4 2 0 1 0 0 0 2 11 RIRH7 1 0 4 0 8 3 2 0 0 1 0 10 4 1 1 0 0 0 9 RINA2 2 1 2 2 11 4 1 1 2 0 3 1 1 1 0 0 0 3 10 RIAUT1 0 0 2 0 15 3 0 2 3 5 8 6 0 0 0 0 0 1 11 RIRH10 1 0 1 0 13 4 2 3 3 0 4 9 1 0 0 0 0 0 12

Appendix

xii

Appendix 5 Pearson correlation table for correlation factor r of the descriptor family frequencies, recognition frequency of Riesling wines (RIESLING) and relative typicality ratings (typicality), determined by the German panel; data set comprising Riesling and non-Riesling wines

variables

vege

tal

sauv

igno

n

spic

y

herb

s

Rib

es s

sp. /

Sam

bucu

s s

sp.

fruity

citru

s fru

it

whi

te fr

uit

yello

w fr

uit

tropi

cal f

ruit

drie

d /

cand

ied

fruit

flora

l

agin

g no

tes

min

eral

ity

ferm

enta

tive

arom

a

sulfu

r

MLF

off-f

lavo

ur

othe

rs

RIE

SLIN

G

typi

calit

y

vegetal 1 0.945 0,148 0.428 0.782 -0.357 0.059 -0.379 -0.426 0.412 -0.031 -0.191 -0.335 -0.418 -0.290 -0.008 -0.169 -0.149 -0.186 -0.787 -0.764

sauvignon 0.945 1 0,202 0.344 0.916 -0.257 0.158 -0.352 -0.433 0.509 0.174 -0.232 -0.330 -0.425 -0.319 -0.113 -0.182 -0.150 -0.238 -0.741 -0.755

spicy 0.148 0.202 1 0.156 0.256 -0.253 -0.101 -0.223 -0.208 -0.106 0.235 -0.098 -0.234 -0.276 0.177 0.051 0.194 -0.143 0.349 -0.339 -0.359

herbs 0.428 0.344 0,156 1 0.282 -0.290 -0.151 -0.223 -0.245 0.084 0.000 -0.098 -0.063 -0.224 -0.141 0.102 -0.230 -0.125 -0.085 -0.310 -0.263

Ribes ssp./ Sambucus ssp. 0.782 0.916 0,256 0.282 1 -0.188 0.160 -0.313 -0.398 0.480 0.404 -0.222 -0.280 -0.395 -0.300 -0.179 -0.170 -0.134 -0.229 -0.640 -0.675

fruity -0.357 -0.257 -0,253 -0.290 -0.188 1 0.582 0.606 0.770 0.068 0.102 0.291 0.155 0.437 0.124 -0.497 0.089 -0.179 -0.368 0.651 0.631

citrus fruit 0.059 0.158 -0,101 -0.151 0.160 0.582 1 0.151 0.261 -0.014 0.114 -0.085 -0.092 0.385 -0.160 -0.416 -0.103 0.017 -0.334 0.242 0.257

white fruit -0.379 -0.352 -0,223 -0.223 -0.313 0.606 0.151 1 0.423 -0.215 -0.067 0.075 -0.059 0.420 0.168 -0.301 0.105 -0.023 -0.120 0.520 0.479

yellow fruit -0.426 -0.433 -0,208 -0.245 -0.398 0.770 0.261 0.423 1 -0.172 -0.083 0.444 0.203 0.254 0.172 -0.255 0.102 -0.181 -0.223 0.651 0.658

tropical fruit 0.412 0.509 -0,106 0.084 0.480 0.068 -0.014 -0.215 -0.172 1 0.127 -0.086 0.209 -0.300 -0.216 -0.237 0.023 -0.126 -0.163 -0.310 -0.360

dried / candied fruit -0.031 0.174 0.235 0.000 0.404 0.102 0.114 -0.067 -0.083 0.127 1 0.000 -0.166 -0.078 0.226 -0.013 -0.173 -0.161 -0.120 -0.057 0.010

floral -0.191 -0.232 -0.098 -0.098 -0.222 0.291 -0.085 0.075 0.444 -0.086 0.000 1 -0.038 0.145 0.234 0.318 -0.042 -0.112 -0.071 0.101 0.157

aging notes -0.335 -0.330 -0.234 -0.063 -0.280 0.155 -0.092 -0.059 0.203 0.209 -0.166 -0.038 1 -0.077 -0.232 -0.175 0.082 0.094 -0.026 0.409 0.348

minerality -0.418 -0.425 -0.276 -0.224 -0.395 0.437 0.385 0.420 0.254 -0.300 -0.078 0.145 -0.077 1 0.358 -0.300 -0.218 -0.002 -0.257 0.478 0.595

fermentative aroma -0.290 -0.319 0.177 -0.141 -0.300 0.124 -0.160 0.168 0.172 -0.216 0.226 0.234 -0.232 0.358 1 -0.204 -0.029 -0.249 0.139 0.134 0.277

sulfur -0.008 -0.113 0.051 0.102 -0.179 -0.497 -0.416 -0.301 -0.255 -0.237 -0.013 0.318 -0.175 -0.300 -0.204 1 0.083 0.076 0.322 -0.295 -0.289

MLF -0.169 -0.182 0.194 -0.230 -0.170 0.089 -0.103 0.105 0.102 0.023 -0.173 -0.042 0.082 -0.218 -0.029 0.083 1 -0.057 0.509 -0.009 -0.147

off-flavour -0.149 -0.150 -0.143 -0.125 -0.134 -0.179 0.017 -0.023 -0.181 -0.126 -0.161 -0.112 0.094 -0.002 -0.249 0.076 -0.057 1 0.002 0.038 -0.077

others -0.186 -0.238 0.349 -0.085 -0.229 -0.368 -0.334 -0.120 -0.223 -0.163 -0.120 -0.071 -0.026 -0.257 0.139 0.322 0.509 0.002 1 -0.265 -0.265

RIESLING -0.787 -0.741 -0.339 -0.310 -0.640 0.651 0.242 0.520 0.651 -0.310 -0.057 0.101 0.409 0.478 0.134 -0.295 -0.009 0.038 -0.265 1 0.935

typicality -0.764 -0.755 -0.359 -0.263 -0.675 0.631 0.257 0.479 0.658 -0.360 0.010 0.157 0.348 0.595 0.277 -0.289 -0.147 -0.077 -0.265 0.935 1

significant correlations (a = 0.05) in bold

Appendix

xiii

Appendix 6 Pearson correlation table for correlation factor r of the descriptor family frequencies, recognition frequency of Riesling wines (RIESLING) and relative typicality ratings (typicality), determined by the French panel; data set comprising Riesling and non-Riesling wines

variables

vege

tal

sauv

igno

n

spic

y

herb

s

fruity

citru

s fru

it

whi

te fr

uit

yello

w fr

uit

tropi

cal f

ruit

drie

d /

cand

ied

fruit

flora

l

agin

g no

tes

min

eral

ity

ferm

enta

tive

arom

a

sulfu

r

MLF

off-f

lavo

ur

othe

rs

RIE

SLIN

G

typi

calit

y

vegetal 1 0.723 0.384 0.512 -0.274 -0.192 -0.109 -0.093 -0.178 0.066 -0.309 -0.290 0.011 -0.034 0.106 -0.100 -0.145 -0.129 -0.568 -0.504

sauvignon 0.723 1 0.373 0.806 -0.335 -0.325 -0.030 0.010 -0.068 -0.017 -0.238 -0.202 -0.129 0.334 -0.079 -0.194 0.112 -0.080 -0.520 -0.612

spicy 0.384 0.373 1 0.307 -0.157 -0.339 0.037 -0.070 -0.087 0.215 -0.372 0.076 -0.017 0.181 0.086 -0.004 -0.021 -0.053 -0.106 -0.188

herbs 0.512 0.806 0.307 1 -0.347 -0.300 0.102 -0.146 -0.090 -0.149 -0.187 -0.238 -0.052 0.317 -0.173 -0.249 0.204 -0.026 -0.331 -0.397

fruity -0.274 -0.335 -0.157 -0.347 1 0.416 0.168 0.535 0.553 0.090 -0.212 0.014 -0.250 -0.139 -0.191 0.289 -0.354 0.024 0.438 0.299

citrus fruit -0.192 -0.325 -0.339 -0.300 0.416 1 -0.075 -0.222 -0.191 -0.456 -0.203 -0.163 0.127 0.108 -0.062 0.208 -0.397 -0.078 0.469 0.692

white fruit -0.109 -0.030 0.037 0.102 0.168 -0.075 1 -0.022 -0.103 -0.175 -0.347 0.036 0.231 0.309 -0.132 0.110 0.097 0.149 0.127 0.032

yellow fruit -0.093 0.010 -0.070 -0.146 0.535 -0.222 -0.022 1 0.581 0.036 -0.051 0.038 -0.360 -0.243 -0.026 0.100 0.167 -0.120 -0.024 -0.161

tropical fruit -0.178 -0.068 -0.087 -0.090 0.553 -0.191 -0.103 0.581 1 0.134 -0.026 0.088 -0.295 -0.323 -0.043 0.182 -0.054 0.093 -0.073 -0.127

dried / candied fruit 0.066 -0.017 0.215 -0.149 0.090 -0.456 -0.175 0.036 0.134 1 0.307 0.189 -0.251 -0.238 0.001 -0.162 -0.191 0.093 -0.030 -0.396

floral -0.309 -0.238 -0.372 -0.187 -0.212 -0.203 -0.347 -0.051 -0.026 0.307 1 -0.080 -0.439 -0.190 -0.266 -0.305 -0.086 -0.139 0.060 -0.062

aging notes -0.290 -0.202 0.076 -0.238 0.014 -0.163 0.036 0.038 0.088 0.189 -0.080 1 0.285 -0.132 -0.072 -0.247 0.077 -0.150 0.215 0.160

minerality 0.011 -0.129 -0.017 -0.052 -0.250 0.127 0.231 -0.360 -0.295 -0.251 -0.439 0.285 1 -0.045 0.385 0.061 0.162 0.117 -0.064 0.205

fermentative aroma -0.034 0.334 0.181 0.317 -0.139 0.108 0.309 -0.243 -0.323 -0.238 -0.190 -0.132 -0.045 1 -0.232 -0.004 0.169 0.102 0.050 -0.109

sulfur 0.106 -0.079 0.086 -0.173 -0.191 -0.062 -0.132 -0.026 -0.043 0.001 -0.266 -0.072 0.385 -0.232 1 0.273 -0.041 0.308 -0.334 -0.053

MLF -0.100 -0.194 -0.004 -0.249 0.289 0.208 0.110 0.100 0.182 -0.162 -0.305 -0.247 0.061 -0.004 0.273 1 0.192 0.022 0.049 -0.096

off-flavour -0.145 0.112 -0.021 0.204 -0.354 -0.397 0.097 0.167 -0.054 -0.191 -0.086 0.077 0.162 0.169 -0.041 0.192 1 -0.066 -0.116 -0.305

others -0.129 -0.080 -0.053 -0.026 0.024 -0.078 0.149 -0.120 0.093 0.093 -0.139 -0.150 0.117 0.102 0.308 0.022 -0.066 1 -0.130 -0.113

RIESLING -0.568 -0.520 -0.106 -0.331 0.438 0.469 0.127 -0.024 -0.073 -0.030 0.060 0.215 -0.064 0.050 -0.334 0.049 -0.116 -0.130 1 0.671

typicality -0.504 -0.612 -0.188 -0.397 0.299 0.692 0.032 -0.161 -0.127 -0.396 -0.062 0.160 0.205 -0.109 -0.053 -0.096 -0.305 -0.113 0.671 1

significant correlations (a = 0.05) in bold; descriptor family ‘Ribes ssp. / Sambucus ssp.’ was not applied by the French panel and is therefore missing

Appendix

xiv

Appendix 7 Pearson correlation table for correlation factor r of the descriptor family frequencies, recognition frequency of Riesling wines (RIESLING) and relative typicality ratings (typicality), determined by the German panel; data set comprising only Riesling wines

variables ve

geta

l

sauv

igno

n

spic

y

herb

s

fruity

citru

s fru

it

whi

te fr

uit

yello

w fr

uit

tropi

cal f

ruit

drie

d /

cand

ied

fruit

flora

l

agin

g no

tes

min

eral

ity

ferm

enta

tive

arom

a

sulfu

r

MLF

off-f

lavo

ur

othe

rs

RIE

SLIN

G

typi

calit

y

vegetal 1 0.554 -0.295 0.215 -0.306 -0.324 -0.288 -0.017 -0.230 -0.083 0.249 -0.017 0.017 0.042 0.465 -0.120 -0.076 -0.125 -0.550 -0.275 sauvignon 0.554 1 -0.211 -0.226 -0.128 -0.085 -0.305 0.027 -0.114 -0.211 0.220 -0.142 0.176 -0.053 0.147 -0.059 -0.160 -0.131 -0.274 -0.187 spicy -0.295 -0.211 1 -0.009 0.154 -0.085 -0.040 0.105 0.087 0.395 0.051 -0.105 -0.043 0.501 -0.170 0.046 -0.193 -0.381 0.199 0.085 herbs 0.215 -0.226 -0.009 1 -0.296 -0.300 -0.185 -0.194 -0.050 0.207 -0.035 0.060 -0.164 -0.017 0.384 -0.177 -0.137 0.398 -0.177 -0.104 fruity -0.306 -0.128 0.154 -0.296 1 0.645 0.548 0.747 0.097 -0.007 0.220 -0.039 0.272 0.135 -0.514 0.288 -0.297 -0.093 0.609 0.590 citrus fruit -0.324 -0.085 -0.085 -0.300 0.645 1 0.165 0.312 -0.127 0.013 -0.088 -0.116 0.467 -0.084 -0.416 0.035 0.017 -0.160 0.496 0.556 white fruit -0.288 -0.305 -0.040 -0.185 0.548 0.165 1 0.261 -0.126 -0.040 -0.041 -0.262 0.258 0.120 -0.341 0.163 -0.101 0.057 0.358 0.262 yellow fruit -0.017 0.027 0.105 -0.194 0.747 0.312 0.261 1 0.041 -0.051 0.373 -0.001 -0.037 0.107 -0.290 0.183 -0.320 -0.177 0.462 0.476 tropical fruit -0.230 -0.114 0.087 -0.050 0.097 -0.127 -0.126 0.041 1 -0.114 0.049 0.674 -0.228 -0.061 -0.201 0.312 -0.072 0.359 0.068 -0.068 dried / candied fruit -0.083 -0.211 0.395 0.207 -0.007 0.013 -0.040 -0.051 -0.114 1 0.051 -0.253 -0.043 0.501 0.147 -0.164 -0.193 0.036 -0.092 0.238 floral 0.249 0.220 0.051 -0.035 0.220 -0.088 -0.041 0.373 0.049 0.051 1 -0.161 0.007 0.186 0.404 -0.048 -0.162 -0.096 -0.274 -0.161 aging notes -0.017 -0.142 -0.105 0.060 -0.039 -0.116 -0.262 -0.001 0.674 -0.253 -0.161 1 -0.357 -0.409 -0.172 0.121 0.038 0.249 0.161 0.027 minerality 0.017 0.176 -0.043 -0.164 0.272 0.467 0.258 -0.037 -0.228 -0.043 0.007 -0.357 1 0.363 -0.352 -0.233 -0.104 -0.306 0.052 0.328 fermentative aroma 0.042 -0.053 0.501 -0.017 0.135 -0.084 0.120 0.107 -0.061 0.501 0.186 -0.409 0.363 1 -0.272 -0.186 -0.318 -0.453 -0.061 0.268 sulfur 0.465 0.147 -0.170 0.384 -0.514 -0.416 -0.341 -0.290 -0.201 0.147 0.404 -0.172 -0.352 -0.272 1 -0.068 0.089 0.178 -0.583 -0.588 MLF -0.120 -0.059 0.046 -0.177 0.288 0.035 0.163 0.183 0.312 -0.164 -0.048 0.121 -0.233 -0.186 -0.068 1 -0.060 0.419 0.074 -0.259 off-flavour -0.076 -0.160 -0.193 -0.137 -0.297 0.017 -0.101 -0.320 -0.072 -0.193 -0.162 0.038 -0.104 -0.318 0.089 -0.060 1 0.132 -0.210 -0.459 others -0.125 -0.131 -0.381 0.398 -0.093 -0.160 0.057 -0.177 0.359 0.036 -0.096 0.249 -0.306 -0.453 0.178 0.419 0.132 1 -0.142 -0.253 RIESLING -0.550 -0.274 0.199 -0.177 0.609 0.496 0.358 0.462 0.068 -0.092 -0.274 0.161 0.052 -0.061 -0.583 0.074 -0.210 -0.142 1 0.738 typicality -0.275 -0.187 0.085 -0.104 0.590 0.556 0.262 0.476 -0.068 0.238 -0.161 0.027 0.328 0.268 -0.588 -0.259 -0.459 -0.253 0.738 1

significant correlations (α = 0.05) in bold.

Appendix

xv

Appendix 8 Pearson correlation table for correlation factor r of the descriptor family frequencies, recognition frequency of Riesling wines (RIESLING) and relative typicality ratings (typicality), determined by the French panel; data set comprising only Riesling wines

variables ve

geta

l

sauv

igno

n

spic

y

herb

s

fruity

citru

s fru

it

whi

te fr

uit

yello

w fr

uit

tropi

cal f

ruit

drie

d /

cand

ied

fruit

flora

l

agin

g no

tes

min

eral

ity

ferm

enta

tive

arom

a

sulfu

r

MLF

off-f

lavo

ur

othe

rs

RIE

SLIN

G

typi

calit

y

vegetal 1 0.431 -0.098 0.405 0.131 0.367 0.277 -0.354 -0.123 -0.301 -0.446 -0.411 0.166 0.068 0.188 0.144 -0.357 0.143 -0.341 0.044 sauvignon 0.431 1 -0.077 0.284 0.067 0.558 -0.079 -0.391 -0.247 -0.196 0.109 -0.320 -0.102 0.444 -0.164 -0.102 -0.351 0.181 0.137 0.061 spicy -0.098 -0.077 1 -0.014 -0.041 -0.382 0.063 -0.018 -0.017 0.285 -0.270 0.167 -0.065 0.135 0.137 0.037 -0.093 0.046 0.103 0.008 herbs 0.405 0.284 -0.014 1 -0.012 0.046 0.031 -0.096 -0.075 -0.235 0.037 -0.257 0.025 -0.042 -0.095 -0.278 -0.046 0.108 0.059 0.141 fruity 0.131 0.067 -0.041 -0.012 1 0.131 0.183 0.497 0.572 0.365 0.038 -0.150 -0.525 -0.159 -0.341 0.235 -0.473 -0.010 0.252 -0.120 citrus fruit 0.367 0.558 -0.382 0.046 0.131 1 -0.205 -0.448 -0.319 -0.417 0.112 -0.362 -0.094 0.207 -0.076 0.066 -0.537 -0.107 0.149 0.492 white fruit 0.277 -0.079 0.063 0.031 0.183 -0.205 1 0.081 -0.127 -0.106 -0.449 -0.006 0.204 0.188 -0.100 0.064 -0.023 0.126 -0.019 -0.127 yellow fruit -0.354 -0.391 -0.018 -0.096 0.497 -0.448 0.081 1 0.551 0.152 0.152 -0.016 -0.456 -0.339 -0.186 0.192 0.285 -0.192 0.105 -0.253 tropical fruit -0.123 -0.247 -0.017 -0.075 0.572 -0.319 -0.127 0.551 1 0.258 -0.010 0.047 -0.315 -0.455 -0.049 0.199 -0.091 0.075 -0.149 -0.281 dried / candied fruit -0.301 -0.196 0.285 -0.235 0.365 -0.417 -0.106 0.152 0.258 1 0.218 0.287 -0.191 -0.122 -0.128 -0.108 -0.103 0.084 0.202 -0.371 floral -0.446 0.109 -0.270 0.037 0.038 0.112 -0.449 0.152 -0.010 0.218 1 -0.060 -0.427 -0.012 -0.491 -0.400 0.093 -0.406 0.424 0.142 aging notes -0.411 -0.320 0.167 -0.257 -0.150 -0.362 -0.006 -0.016 0.047 0.287 -0.060 1 0.268 -0.177 -0.080 -0.312 0.071 -0.185 0.121 0.028 minerality 0.166 -0.102 -0.065 0.025 -0.525 -0.094 0.204 -0.456 -0.315 -0.191 -0.427 0.268 1 -0.087 0.467 -0.025 0.162 0.156 -0.356 0.016 fermentative aroma 0.068 0.444 0.135 -0.042 -0.159 0.207 0.188 -0.339 -0.455 -0.122 -0.012 -0.177 -0.087 1 -0.153 -0.016 -0.040 0.123 0.064 -0.129 sulfur 0.188 -0.164 0.137 -0.095 -0.341 -0.076 -0.100 -0.186 -0.049 -0.128 -0.491 -0.080 0.467 -0.153 1 0.324 0.044 0.310 -0.435 -0.022 MLF 0.144 -0.102 0.037 -0.278 0.235 0.066 0.064 0.192 0.199 -0.108 -0.400 -0.312 -0.025 -0.016 0.324 1 0.202 0.015 -0.182 -0.456 off-flavour -0.357 -0.351 -0.093 -0.046 -0.473 -0.537 -0.023 0.285 -0.091 -0.103 0.093 0.071 0.162 -0.040 0.044 0.202 1 -0.075 -0.200 -0.479 others 0.143 0.181 0.046 0.108 -0.010 -0.107 0.126 -0.192 0.075 0.084 -0.406 -0.185 0.156 0.123 0.310 0.015 -0.075 1 -0.260 -0.246 RIESLING -0.341 0.137 0.103 0.059 0.252 0.149 -0.019 0.105 -0.149 0.202 0.424 0.121 -0.356 0.064 -0.435 -0.182 -0.200 -0.260 1 0.338 typicality 0.044 0.061 0.008 0.141 -0.120 0.492 -0.127 -0.253 -0.281 -0.371 0.142 0.028 0.016 -0.129 -0.022 -0.456 -0.479 -0.246 0.338 1

significant correlations (α = 0.05) in bold.

Appendix

xvi

Appendix 9 Aroma compound’s concentrations in the wines assessed in sensory studies by the German panel: volatile acids; higher alcohols, acetic acid esters, ethyl esters by means of the “Kaltron method”; value 0 indicates concentration below limit of quantification for the concerning aroma compound

wine samples 

SBSA

1 

SBNZ 

SBBX

2 

SBSA

2 

SBBX

 

CHAUS 

CHRC

H 

RIMO2 

RIAL8 

RIAL3 

RINA1 

RIRH

12 

RIAL7 

RIMO3 

RIRH

14 

RIRH

13 

RIRH

4 

RIPF

 

RIRH

5 

RINZ 

RIAUT1

 

RIRH

1 

RIRH

2 

RIRH

11 

RIRH

6 

RIMO1 

RIAUT2

 

RIRH

9 

RIRH

3 

RIRH

8 

volatile acids 

caproic acid  [mg/L]  7  7  9  8  5  8  5  6  7  7  6  7  8  6  7  8  8  7  8  9  7  7  7  6  7  8  8  7  7  7 

caprylic acid  [mg/L]  7  5  8  7  5  6  4  6.83  8  7  5  7  9  7  6  7  9  6  6  9  6  6  7  5  6  7  7  6  6  7 

capric acid [mg/L]  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1 

higher alcohols 

iso‐butanol  [mg/L]  27  17  18  26  40  17  37  21  22  31  23  17  33  22  29  15  13  22  22  19  23  15  16  16  25  28  9  23  20  12 

iso‐amyl alcohol  [mg/L]  148  104  110  136  134  116  171  124  127  134  93  107  152  122  152  108  78  114  95  125  139  93  96  53  104  93  72  114  115  92 

tert‐amyl alcohol  [mg/L]  28  19  20  25  34  21  36  33  27  29  27  20  38  27  29  23  14  21  18  18  25  15  19  10  17  18  11  26  32  16 

hexan‐1‐ol  [µg/L]  1766 2178  885  1672  646  1238 1875 1641 752  907  1741 1507 1035 2058 1163  1645 968  1457 1525 966  1452 1017 1749 941  615  1827 1298  751  1233  417 

acetic acid ester 

ethylacetate [mg/L]  116  73  79  151  79  67  59  73  150  34  168  31  246  107  168  140  1  142  43  172  226  132  40  40  101  65  34  157  168  93 

iso‐amylacetate [µg/L]  593  580  1294  237  1  1051 1262 222  92  1  300  743  55  807  1391  161  779  497  141  25  1008 264  127  1  819  555  1143  1011  274  659 

amylacetate [µg/L]  28  23  57  14  1  29  60  20  16  5  36  31  17  43  61  17  45  26  5  4  52  16  12  3  30  26  57  70  25  37 

hexylacetate [µg/L]  91  106  125  58  13  72  95  53  24  17  97  90  24  122  87  41  96  80  40  27  99  55  45  29  41  165  200  95  43  39 

ethylester 

ethyl propanoate [µg/L]  87  81  60  95  93  97  78  76  14  64  61  96  78  87  91  73  29  83  78  118  69  57  63  20  58  41  19  94  82  56 

ethyl iso‐butyrate [µg/L]  80  42  14  105  65  28  37  60  89  77  76  59  93  46  67  90  39  78  61  55  33  43  87  36  33  69  17  41  71  30 

ethyl butyrate  [µg/L]  269  283  331  280  154  333  239  179  264  174  305  185  226  267  314  247  218  302  406  359  345  235  236  214  392  373  294  395  292  256 

ethyl lactate [µg/L]  32  19  19  37  26  30  20  162  111  26  18  67  25  33  98  209  107  18  40  23  15  78  52  71  122  18  11  26  72  12 

ethyl caproate [µg/L]  758  781  1031  817  529  916  498  627  1004 689  787  631  827  636  758  905  771  864  1011 1083 871  651  785  567  877  904  1007  1207  796  726 

diethyl succinate [µg/L]  2240 1378  692  2276  6520 2264 1035 3983 5067 5820 1678 2359 3395 2102 2581  2233 796  1304 2631 3076 1017 1681 2081 2229 1266 1173 478  779  2804  341 

ethyl capryloate  [µg/L]  1312 1125  1460  1484  1079 1286 748  1728 1823 1384 1034 1105 1612 1469 1229  1502 1355 1308 1317 1791 1474 1132 1258 811  1477 1533 1587  1624  1266  1049 

ethyl capriate   [µg/L]  305  327  333  354  311  314  144  351  442  402  288  361  428  541  343  378  435  375  271  519  417  401  427  154  381  401  523  365  258  269 

Appendix

xvii

Appendix 10 Aroma compound’s concentrations in the wines assessed in sensory studies by the German panel: miscellaneous, monoterpenols, monoterpenol oxides by means of the “Kaltron method”, C13-norisoprenoids by means of SBSE and GC-MS; values 0 indicates concentration below limit of quantification for the concerning aroma compound

wine samples 

SBSA

1 

SBNZ 

SBBX

2 

SBSA

2 

SBBX

 

CHAUS 

CHRC

H 

RIMO2 

RIAL8 

RIAL3 

RINA1 

RIRH

12 

RIAL7 

RIMO3 

RIRH

14 

RIRH

13 

RIRH

4 

RIPF

 

RIRH

5 

RINZ 

RIAUT1

 

RIRH

1 

RIRH

2 

RIRH

11 

RIRH

6 

RIMO1 

RIAUT2

 

RIRH

9 

RIRH

3 

RIRH

8 

miscellaneous                 

2‐phenylethanol  [mg/L]  24  14  15  23  33  22  43  187  30  35  16  18  36  91  28  22  18  10  9  16  16  13  19  8  17  14  14  24  19  11 

2‐phenylethylacetate [µg/L]  110  83  150  79  57  152  268  582  65  52  85  102  81  584  280  82  156  51  30  37  69  83  75  25  83  101  158  256  80  80 

monoterpenols                 

linalool [µg/L]  0  3  0  0  6  0  0  147  38  21  13  16  50  68  16  40  15  42  21  39  56  8  24  31  28  60  43  37  26  24 

α‐terpineol [µg/L]  10  11  8  11  14  10  9  124  73  51  23  31  96  58  22  81  20  47  61  29  36  21  66  59  35  53  32  35  94  37 

β‐citronellol [µg/L]  0  1  0  0  0  2  2  0  0  0  0  0  0  0  1  0  0  0  0  0  2  0  0  0  1  0  1  0  0  1 

nerol [µg/L]  2  0  3  2  2  2  0  3  2  2  0  2  0  2  2  2  2  0  0  0  2  0  0  3  0  0  0  0  2  0 

geraniol [µg/L]  10  9  10  8  10  10  10  7  10  8  8  9  9  9  9  9  9  8  12  9  10  9  9  11  9  10  10  9  10  10 

monoterpenol oxides                 

trans‐linalool oxide [µg/L]  0  21  1  0  20  0  0  69  53  55  28  35  0  40  30  67  29  35  56  61  27  39  65  83  40  27  22  31  79  44 

cis‐linalool oxide [µg/L]  0  9  1  0  13  0  0  71  25  23  13  18  0  14  13  31  14  16  28  34  15  21  32  69  27  14  14  19  36  22 

nerol oxide [µg/L]  5  7  4  9  8  3  3  36  24  19  13  18  26  1  14  30  27  14  21  32  15  20  30  43  17  21  10  17  27  21 

C13‐norisoprenoids                 

vitispirane isomers [µg/L]  24  25  20  26  25  20  21  31  33  31  25  21  49  25  26  32  27  29  29  77  28  32  4  30  27  33  27  23  44  30 

TDN [µg/L]  0.5  1.0  0.0  0.5  1.1  0.0  0.0  3.5  5.9  4.8  1.6  1.2  9.4  2.3  1.6  3.3  1.6  3.2  3.3  22.1  1.6  3.4  4.9  3.1  1.7  3.8  2.0  0.9  6.6  2.5 

β‐damascenone [µg/L]  2.0  3.0  3.0  3.0  3.0  2.0  3.0  3.0  3.0  3.0  3.0  2.0  3.0  3.0  4.0  4.0  5.0  3.0  2.0  5.0  4.0  5.0  3.0  2.0  4.0  4.0  3.0  3.0  2.0  3.0 

TPB [µg/L]  0.0  0.0  0.0  0.0  0.0  0.0  0.1  0.0  0.0  0.0  0.0  0.0  0.1  0.0  0.1  0.0  0.0  0.0  0.0  0.1  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0 

α‐ionone [µg/L]  0.5  0.8  0.7  0.9  1.8  0.8  0.8  4.3  1.0  0.5  0.6  0.6  0.0  0.0  0.6  0.5  0.6  0.6  1.1  0.7  0.6  0.7  1.5  0.7  0.7  0.6  0.6  0.7  0.0  0.6 

β‐ionone [µg/L]  0.1  0.2  0.1  0.1  0.1  0.1  0.2  0.1  0.1  0.0  0.1  0.1  0.1  0.0  0.1  0.1  0.2  0.1  0.1  0.1  0.2  0.1  0.2  0.3  0.1  0.1  0.2  0.2  0.1  0.1 

Appendix

xviii

Appendix 11 Aroma compound’s concentrations in the wines assessed in sensory studies by the German panel: sulfur compounds by means of HS-GC-PFPD1 and according to Tominaga et al., (2006)2; value 0 indicates concentration below limit of quantification for the concerning aroma compound

wine samples 

SBSA

1 

SBNZ 

SBBX

2 

SBSA

2 

SBBX

 

CHAUS 

CHRC

H 

RIMO2 

RIAL8 

RIAL3 

RINA1 

RIRH

12 

RIAL7 

RIMO3 

RIRH

14 

RIRH

13 

RIRH

4 

RIPF

 

RIRH

5 

RINZ 

RIAUT1

 

RIRH

1 

RIRH

2 

RIRH

11 

RIRH

6 

RIMO1 

RIAUT2

 

RIRH

9 

RIRH

3 

RIRH

8 

Sulfur compounds                 

hydrogen sulfide [µg/L]  5  18.1  7.7  7.3  13.3  50.4  6.7  9.2  11.9  5.1  9.1  14.7  5.8  11.9  7.9  7.3  6  8  9.9  22.9  13.4  23.9  6.4  4.7  16  8.1  8.2  16.9  24.9  9.5 

methanethiol  [µg/L]  0  1  0  6.1  0  5.5  1  0  0  0  1  6.3  0  5.3  0  1  1  5.5  1  5.3  5.5  4.8  0  0  1  4.6  5.3  1  0  1 

  dimethyl sulfide  [µg/L] 

18.7  34  8.9  29.2  9.7  65.9  19.2  7.2  11.2  8.6  17.2  8.9  7.7  12.1  8.5  17.6  8.9  9.3  18.2  22.5  15.6  12.9  18  8.8  9.4  5.3  17.6  9.9  13.4  7.1 

carbon disulfide  [µg/L]  2  1.6  1.9  1.8  3.2  1.3  2.7  6.1  1.8  2.4  20.1  7.2  6.5  3.2  2.2  2.4  2.3  1.5  2.2  4.6  14.7  1.3  2.4  3  7.2  1  1.6  3.4  3.9  5.4 

 ethyl 2‐sulfanylpropionate [ng/L] 

0  0  0  0  8  0  0  0  0  0  5  0  0  0  0  5  0  5  0  5  5  5  0  0  0  0  5  0  0  0 

4‐methyl‐4‐sulfanylpentan‐2‐one  [ng/L]  0  0  0  0  10  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 

 ethyl 2‐sulfanyl acetate  [ng/L] 

15  10  10  7  65  15  14  15  19  47  0  40  19  22  14  15  13  10  10  31  21  21  11  13  11  15  13  0  21  16 

 ethyl 3‐sulfanylpropionate [ng/L] 

0  27  0  14  75  24  0  23  42  57  14  57  0  29  75  19  19  15  0  35  20  29  38  19  5  15  5  22  27  21 

benzenemethanethiol  [µg/L]  0  0  0  0  10  0  0  0  0  13  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 

3‐sulfanylhexylacetate  [µg/L]  0  0  0  0  0  7  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 

3‐sulfanylhexan‐1‐ol  [µg/L]  697  1898  1529  1125  836  876  779  559  295  261  218  570  519  1189 741  352  686  247  311  546  764  747  450  893  846  339  253  1558  868  1290 

Appendix

xix

Appendix 12 Aroma compound’s concentrations in the wines assessed in sensory studies by the French panel: volatile acids; higher alcohols, acetic acid esters, ethyl esters by means of the “Kaltron method”; value 0 indicates concentration below limit of quantification for the concerning aroma compound

typicity 

SBBX

 

SBAL 

SBNZ 

RIAL4 

RIAL3 

GTA

L 

PGAL 

RINA1 

RIAL5 

RIRH

4 

RIAL2 

RIAL1 

RIRH

1 

RIRH

8 

RIRH

3 

RINZ 

RIRH

2 

RIAL7 

RIAUT2

 

RIRH

5 

RIRH

6 

RIAL6 

RIMO1 

RIRH

11 

RIRH

9 

RIRH

7 

RINA2 

RIAUT1

 

RIRH

10 

volatile acids 

caproic acid  [mg/L]  5  7  7  8 6 8 8 7 8 7 6 6 7 7  7 9 7 8 8 8 7 8 8 7 9 7  7  7  6 

caprylic acid  [mg/L]  5  7  6  7 6 7 7 6 7 8 5 5 4 7  6 9 7 7 8 7 7 7 8 6 8 6  8  7  4 

capric acid [mg/L]  1  2  2  1 2 2 1 1 2 2 1 1 2 2  1 3 2 2 2 2 1 2 2 1 2 2  2  2  1 

higher alcohols 

iso‐butanol  [mg/L]  37  18  21  13 36 14 24 23 33 21 25 35 14 13  20 17 20 25 14 21 23 40 31 16 23 48  19  18  54 

iso‐amyl alcohol  [mg/L]  116  133  114  118 137 106 115 86 129 112 90 124 82 91  107 109 105 121 102 88 93 139 94 54 111 101  89  112  150 

tert‐amyl alcohol  [mg/L]  29  18  21  17 30 18 25 25 29 22 24 32 13 16  29 16 22 29 17 16 15 28 18 10 25 26  19  20  37 

hexan‐1‐ol  [µg/L]  591  2044  2330  451 856 354 291 1685 836 1094 441 996 952 430  1166 899 1715 445 1655 1457 275 1166 1837 969 815 490  1247  1295  422 

acetic acid ester 

ethylacetate [mg/L]  50  148  99  106 62 72 178 122 152 62 256 193 92 80  135 112 90 162 82 18 63 27 64 19 136 38  13  165  87 

iso‐amylacetate [µg/L]  1  526  255  1 1 162 229 54 1 421 1 1 55 265  91 1 1 1 778 1 371 1 267 1 379 1  422  581  339 

amylacetate [µg/L]  1  19  9  1 1 8 27 17 1 20 1 1 5 19  17 1 1 1 40 1 14 1 12 1 30 1  31  31  35 

hexylacetate [µg/L]  10  87  62  13 10 21 19 55 9 50 17 9 33 23  28 18 27 14 133 25 25 18 100 18 46 15  76  66  30 

ethylester 

ethyl propanoate [µg/L]  88  102  92  128  75  104  68  60  76  61  110  1  57  64  83  107  79  62  46  78  59  140  52  20  107  43  33  71  106 

ethyl iso‐butyrate [µg/L]  68  31  47  44  84  54  41  81  84  57  122  87  47  35  73  52  100  81  27  61  37  102  86  36  53  122  49  37  84 

ethyl butyrate  [µg/L]  133  242  246  315  156  265  232  246  200  220  147  191  201  216  261  307  222  166  313  347  316  245  344  177  325  212  137  283  235 

ethyl lactate [µg/L]  25  19  22  138  27  73  22  19  228  148  17  193  76  13  71  22  56  24  14  43  144  34  22  77  33  25  17  15  35 

ethyl caproate [µg/L]  400  658  594  855  486  673  652  611  646  588  452  427  509  515  635  836  663  563  754  817  642  704  764  381  912  545  355  656  448 

diethyl succinate [µg/L]  6205  2197  1670  3520  5989  3023  1661  2215  4985  1019  3389  6612  1597  501  2626  2838  2061  3348  690  2641  1327  5483  1725  1831  1800  3897  322  1248  3380 

ethyl capryloate  [µg/L]  766  1160  818  1279  971  1289  1306  886  1031  1039  839  773  894  745  997  1427  1064  1098  1039  1109  1108  1181  1296  576  1295  1013  733  1043  749 

ethyl capriate   [µg/L]  181  329  218  194  263  363  307  226  227  280  180  206  294  180  195  423  305  281  283  221  278  252  322  107  278  307  192  300  150 

Appendix

xx

Appendix 13 Aroma compound’s concentrations in the wines assessed in sensory studies by the French panel: miscellaneous, monoterpenols, monoterpenol oxides by means of the “Kaltron method”, C13-norisoprenoids by means of SBSE and GC-MS; values 0 indicates concentration below limit of quantification for the concerning aroma compound

typicity 

SBBX

 

SBAL 

SBNZ 

RIAL4 

RIAL3 

GTA

L 

PGAL 

RINA1 

RIAL5 

RIRH

4 

RIAL2 

RIAL1 

RIRH

1 

RIRH

8 

RIRH

3 

RINZ 

RIRH

2 

RIAL7 

RIAUT2

 

RIRH

5 

RIRH

6 

RIAL6 

RIMO1 

RIRH

11 

RIRH

9 

RIRH

7 

RINA2 

RIAUT1

 

RIRH

10 

miscellaneous                 

2‐phenylethanol  [mg/L]  27  14  14  49  27  13  53  14  27  15  20  30  11  10  16  14  16  31  14  8  14  19  12  7  24  36  18  13  18 

2‐phenylethylacetate [µg/L]  31  60  52  72  24  26  103  50  26  81  50  27  47  45  52  24  43  43  104  20  46  33  60  16  134  60  140  42  39 

monoterpenols               

linalool [µg/L]  1  1  1  1  8  26  10  8  11  10  38  19  3  15  18  1  16  26  37  13  17  23  42  16  20  6  13  42  61 

α‐terpineol [µg/L]  12  15  11  31  40  65  12  21  52  18  74  64  18  33  81  25  58  77  30  53  30  70  49  49  40  32  26  34  44 

β‐citronellol [µg/L]  3 5 7 0 2 8 2 12 0 1 0 4 68 55 5 0 6 0 4 4 0 3 10 10 0 1 3 10 0

nerol [µg/L]  0 4 2 2 2 2 3 2 3 2 0 2 10 3 2 2 2 0 3 2 2 3 3 3 2 3 2 3 2

geraniol [µg/L]  11 12 12 0 10 35 9 12 10 12 9 10 90 14 15 12 15 9 15 12 9 12 14 13 11 8 10 16 12

monoterpenol oxides               

trans‐linalool oxide [µg/L]  1  1  1  54  54  67  23  33  68  30  56  65  39  51  78  57  68  1  1  62  40  67  33  91  48  57  35  30  56 

cis‐linalool oxide [µg/L]  1  1  1  21  23  31  15  16  35  19  26  29  19  25  36  30  28  1  1  26  29  30  17  73  23  25  17  15  31 

nerol oxide [µg/L]  5 6 9 20 12 10 10 8 16 9 24 19 6 10 18 15 18 25 9 14 20 15 9 33 11 25 10 11 17

C13‐norisoprenoids               

vitispirane isomers [µg/L]  28 26 29 40 40 25 18 32 51 30 31 55 0 39 75 112 48 41 31 44 26 52 39 39 38 45 25 29 66

TDN [µg/L]  0.6 0.7 0.7 5.0 5.0 0.0 0.0 1.3 4.9 0.9 4.2 5.6 0.0 2.8 5.9 17.2 3.4 4.9 1.3 2.6 1.3 5.9 2.8 2.4 2.1 8.1 1.4 1.4 6.2

β‐damascenone [µg/L]  3.0 2.0 3.0 2.0 3.0 2.0 3.0 3.0 2.0 4.0 2.0 3.0 3.0 3.0 2.0 5.0 3.0 3.0 3.0 2.0 3.0 2.0 3.0 2.0 3.0 4.0 2.0 3.0 3.0

TPB [µg/L]  0.1 0.1 0.1 0.0 0.1 0.0 0.1 0.1 0.0 0.1 0.0 0.1 0.1 0.1 0.1 0.1 0.1 0.0 0.0 0.1 0.0 0.1 0.1 0.0 0.1 0.1 0.0 0.1 0.0

α‐ionone [µg/L]  0.0 1.2 1.7 0.0 0.0 1.3 0.0 1.6 0.0 1.0 0.0 0.0 0.0 1.9 0.7 1.6 1.7 0.4 1.5 1.0 0.8 0.0 2.0 0.0 0.0 0.0 0.0 0.0 1.7

β‐ionone [µg/L]  0.0 1.3 0.2 0.0 0.0 0.2 0.2 0.2 0.1 0.4 0.2 0.0 0.0 0.7 0.0 0.3 0.4 0.3 0.5 0.3 0.2 0.0 0.7 0.0 0.0 0.2 0.0 0.4 0.6

Appendix

xxi

Appendix 14 Aroma compound’s concentrations in the wines assessed in sensory studies by the French panel: sulfur compounds by means of GC-PFPD1 and according to Tominaga et al., (2006)2; value 0 indicates concentration below limit of quantification for the concerning aroma compound

Typicity SB

BX 

SBAL 

SBNZ 

RIAL4 

RIAL3 

GTA

L 

PGAL 

RINA1 

RIAL5 

RIRH

4 

RIAL2 

RIAL1 

RIRH

1 

RIRH

8 

RIRH

3 

RINZ 

RIRH

2 

RIAL7 

RIAUT2

 

RIRH

5 

RIRH

6 

RIAL6 

RIMO1 

RIRH

11 

RIRH

9 

RIRH

7 

RINA2 

RIAUT1

 

RIRH

10 

Sulfur compounds               

hydrogen sulfide [µg/L]  13  14.8  7.7  14  5.4  4.5  3.6  10  10  6  3.6  8.1  20.6  7.7  22.9  14.8  5.8  5.3  5.2  8.9  9.1  7.8  8.6  4.3  19  5.3  3.7  11.5  11.5 

methanethiol  [µg/L]  0  7.1  0  4.5  0  0  0  0  0  0  0  0  0  4.6  5.8  5.6  0  4.3  4.7  0  5.2  4  4.8  0  5.8  0  10.1  0  17.5 

  dimethyl sulfide  [µg/L] 

9.7  48.6  20.6  9.5  7.1  7.4  10.4  14.2  7.2  6.8  7  5.5  8.6  6  11  17.3  15  6.4  12.7  12.3  7  6.1  4.2  6  8.8  7.1  5.1  12.7  6.4 

carbon disulfide  [µg/L]  3.2  3.1  1.9  4  4.2  4.3  8  15.8  2.1  4.2  16.1  25.6  4.5  4.5  5.6  5.3  5  3.8  2  6.1  7.3  1.3  1.8  2.5  4.1  3.1  6.4  18.3  13.8 

 ethyl 2‐sulfanylpropionate [ng/L] 

8  7  0  5  7  7  5  6  0  0  5  11  5  0  0  6  0  0  6  0  5  6  0  0  0  0  0  6  6 

4‐methyl‐4‐sulfanylpentan‐2‐one  [ng/L]  10  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  1  0  0  0  0  0  0  0  0  0  0 

 ethyl 2‐sulfanyl acetate  [ng/L] 

66  26  17  19  48  28  17  0  19  13  37  78  21  17  21  31  11  19  14  10  15  20  15  13  0  14  10  21  22 

 ethyl 3‐sulfanylpropionate [ng/L] 

75  50  40  53  57  37  10  13  0  20  45  69  28  21  27  34  38  15  0  0  20  17  15  18  22  20  13  20  30 

benzenemethanethiol  [µg/L]  0  0  0  0  12  0  0  0  0  0  13  22  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 

3‐sulfanylhexylacetate  [µg/L]  0  0  0  0  0  10  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 

3‐sulfanylhexan‐1‐ol  [µg/L]  837  610  2797  312  261  778  665  218  519  687  306  493  747  1290  868  547  450  254  253  312  507  797  339  893  1559  317  551  764  669 

Appendix

xxii

Appendix 15 Ions (m/z) used for compounds identification and quantification and calibration curve parameters applied for quantification in the analysis of ‘Kaltron’ extracts by means of GC-MS

Compound Ions [m/z] Equation of

calibration curveCoefficient of

determination (R2)Internal

standard Calibration range ethyl acetate 88 / 43 2.4768*x+0.3090 0.9767 IS1 50 mg/L – 198.1 mg/L iso-butanol (2-methylpropan-1-ol) 41 / 74 9.8634*x-0.0147 0.9995 IS1 10.3 mg/L – 80.1 mg/L ethyl propanoate 57 / 102 / 75 1.2105*x+0.0249 0.9918 IS1 41 µg/L – 499 µg/L iso-amyl alcohol (3-methyl-1-butanol) 55 / 70 / 42 38.828*x-0.1136 0.9990 IS1 24.9 mg/L – 199.4 mg/L tert-amyl alcohol (2-methyl-2-butanol) 55 / 70 / 42 45.217*x+0.0324 0.9980 IS1 9.8 mg/L – 97.6 mg/L ethyl iso-butyrate (ethyl 2-methylpropanoate) 43 / 71 / 116 1.7863*x+0.0052 0.9960 IS1 24 µg/L – 298 µg/L ethyl butyrate (ethyl butanoate) 71 / 43 / 88 1.1440*x+0.0200 0.9960 IS1 107 µg/L – 689 µg/L ethyl lactate (ethyl 2-hydroxypropanoate) 45 / 75 19.080*x-0.0983 0.9955 IS1 100.2 mg/L – 250.4 mg/L hexan-1-ol 56 / 69 / 55 / 84 0.3434*x-0.0310 0.9977 IS1 315 µg/L – 3472 µg/L iso-amylacetate (3-methylbutylacetate) 43 / 70 / 61 / 87 2.6426*x+0.6346 0.9860 IS1 293 µg/L – 2518 µg/L amylacetate (2-methylbutylacetate) 43 / 70 / 55 / 73 3.6745*x+0.0320 0.9999 IS1 26 µg/L – 244 µg/L caproic acid (hexanoic acid) 60 / 73 / 87 370.33*x-0.7123 0.9740 IS1 2.2 mg/L – 10.1 mg/L ethyl caproate (ethyl hexanoate) 88 / 99 / 43 1.3669*x+0.1595 0.9897 IS1 204 µg/L – 1506 µg/L hexyl acetate 43 / 56 / 84 3.3751*x-0.0153 0.9931 IS1 20 µg/L – 540 µg/L trans-linalool oxide 59 / 68 / 94 / 111 0.2839*x+0.0001 0.9980 IS1 19 µg/L – 204 µg/L cis-linalool oxide 59 / 68 / 94 / 111 0.2496*x-0.0020 0.9972 IS1 19 µg/L – 204 µg/L linalool 71 / 43 / 93 1.3683*x+0.0048 0.9982 IS1 9 µg/L – 200 µg/L 2-phenylethanol (benzeneethyl) 91 / 92 / 122 81.101*x+0.0114 0.9993 IS1 4.1 mg/L – 66.1 mg/L caprylic acid (octanoic acid) 60 / 73 / 85 145.22*x-1.4624 0.9980 IS1 2 mg/L – 10 mg/L diethyl succinate (diethyl butanedioate) 129 / 128 / 101 / 174 0.4830*x+0.1492 0.9962 IS1 511 µg/L – 5013 µg/L ethyl capryloate (ethyl octanoate) 88 / 101 / 127 1.5247*x+0.2209 0.9832 IS1 198 µg/L – 1506 µg/L α-terpineol 59 / 121 / 136 1.5654*x-0.0047 0.9906 IS1 10 µg/L – 202 µg/L ethyl 2-phenylacetate (ethyl benzeneacetate) 164 / 91 / 92 / 65 0.5914*x-0.0004 0.9943 IS1 10 µg/L – 99 µg/L 2-phenylethylacetate (benzylethyl acetate) 104 / 43 3.7680*x-0.0103 0.9969 IS1 21 µg/L – 400 µg/L decanoic acid (capric acid) 60 / 73 / 129 119.48*x-0.0649 0.9961 IS1 490 µg/L – 4978 µg/L ethyl capriate (ethyl decanoate) 88 / 101 / 73 / 155 1.9798*x+0.0545 0.9928 IS1 49 µg/L – 490 µg/L 2,6-dimethylhept-5-en-2-ol (DMH) (IS) 109 / 124 / 59 IS 1 - - 1151 µg/L cumene (IS) 105 / 120 IS 2 - - 398 µg/L

(IS) indentifies internal standards; m/z in bold identify Ions used for quantification; x: AA/AIS, with A: peak area

Appendix

xxiii

Appendix 16 Ions (m/z) used for compounds identification and quantification and calibration curve parameters applied for quantification in the analysis of C13-norisoprenoids by means of GC-MS

Compound Ions [m/z] Equation of

calibration curve Coefficient of

determination (R2) Internal

standard Calibration range vitispirane isomers 93 / 121 / 136 / 177 / 192 0.0030*x-0.0025 0.9808 IS 2 15 µg/L – 800 µg/L TDN (1,1,6-trimethyl-1,2-dihydronaphthalene) 142 / 157 / 172 0.0263*x-0.0154 0.9869 IS 2 1 µg/L – 75 µg/L β-damacenone 69 / 91 / 105 / 121 / 190 0.0059*x-0.0013 0.9666 IS 2 0.5 µg – 20 µg/L β-ionone 91 / 135 / 177 / 192 0.0273*x-0.0001 0.9814 IS 2 0.5 µg/L – 20 µg/L TPB (1,2,4-trimethylphenyl-3-butadiene) 142 / 157 / 172 0.0216*x+0.0000 0.9943 IS 2 0.05 µg – 10 µg/L trans-ethyl cinnamate 176 / 137 / 136 0.0326*x-0.0013 0.9856 IS 2 0.1 µg/L – 30 µg/L cumol (IS) 105 / 120 IS 1 - - 50 µg/L dichloroaniline (IS) 90 / 161 / 163 IS 2 - - 50 µg/L

(IS) indentifies internal standards; m/z in bold identify Ions used for quantification; x: AA/AIS, with A: peak area

Appendix 17 Ions (m/z) used for compounds identification and quantification and calibration curve parameters applied for quantification in the analysis of volatile thiols by means of GC-MS

Compound Ions [m/z] Equation of

calibration curve Coefficient of

determination (R2) Internal

standard Calibration range benzenemethanethiol 124 / 91 /65 0.0013*x-0.0044 0.9825 IS 1 1 ng/L – 200 ng/L 2-furanylmethanethiol 114 / 81 / 51 0.0011*x+0.0152 0.9483 IS 1 1 ng/L – 200 ng/L 3-sulfanylhexan-1-ol 134 / 116 / 100 0.0040*x+0.0012 0.9885 IS 2 20 ng/L – 2000 ng/L 3-sulfanylhexyl acetate 116 / 101 0.0010*x-0.0019 0.9885 IS 1 5 ng/L – 200 ng/L 4-methyl-4-sulfanylpentan-2-one 132 / 99 / 75 0.0015*x+0.0003 0.9761 IS 1 5 ng/L – 200 ng/L ethyl-2-sulfanyl acetate 134 / 43 / 74 0.0178*x+0.0052 0.9830 IS 1 50 ng/L – 2000 ng/L ethyl-2-sulfanylpropionate 134 / 61 /88 0.0006*x+0.0037 0.9841 IS 1 4 ng/L – 100 ng/L ethyl-3-sulfanylpropionate 134 / 61 / 88 0.0280*x-0.0001 0.9925 IS 1 3 ng/L – 100 ng/L 1-methoxy-3-methylsulfanylbutane (IS) 134 / 100 IS 1 - - 10 ng/L 6-sulfanylhexan-1-ol (IS) 134 / 116 /101 IS 2 - - 100 ng/L

(IS) indentifies internal standards; m/z in bold identify Ions used for quantification; x: AA/AIS, with A: peak area

xxiv

Appendix 18 Ions (m/z) used for compounds identification and quantification and calibration curve parameters applied for quantification in the analysis of the potential of volatile C13-norisoprenoids and terpenes (PVNT) by means of GC-MS

Compound Ions [m/z] Equation of

calibration curve Coefficient of

determination (R2) Internal

standard Calibration range geraniol 139 / 121 / 93 0.00071*x-0.0052 0.9700 IS 1 1 µg/L – 257 µg/L linalool 121 / 93 0.00274*x-0.0049 0.9968 IS 1 2 µg/L – 209 µg/L nerol 139 / 121 / 93, 0.00107*x-0.0021 0.9898 IS 1 9 µg/L – 482 µg/L β-citronellol 138 / 123 / 109 0.00437*x+0.0069 0.9953 IS 1 13 µg/L – 243 µg/L cis-linalool oxide 155 / 137 / 93 / 79 0.00298*x+0.0038 0.9972 IS 1 2 µg/l – 212 µg/L trans-linalool oxide 155 / 137 / 93 / 79 0.00005*x+0.0021 0.9845 IS 1 13 µg/L – 254 µg/L nerol oxide 152 / 109 / 83 0.00360*x-0.0072 0.9696 IS 1 13 µg/L – 855 µg/L α-terpineol 136 / 121 / 93 0.00670*x-0.0494 0.9872 IS 2 15 µg/L – 562 µg/L β-damascenone 190 / 175 / 121 0.02331*x+0.0036 0.9579 IS 2 0.2 µg/L – 10 µg/L α-ionon 192 /136 / 121 / 93 0.00875*x-0.0035 0.9969 IS 2 0.2 µg/L – 50 µg/L β-ionone 91 / 135 / 177 / 192 0.06347*x-0.0051 0.9939 IS 2 0.2 µg/L – 100 µg/L vitispirane isomers 192 / 177 / 136 / 121 / 93 0.00558*x-0.0909 0.9910 IS 2 15 µg/L – 1524 µg/L 1,1,6-trimethyl-1,2-dihydro-naphthalene (TDN) 172 / 157 / 142 0.09653*x+0.0575 0.9958 IS 2 0.8 µg/L – 33 µg/L TPB (1,2,4-trimethylphenyl-3-butadiene) 172 / 157 / 142 0.08469*x-0.0014 0.9890 IS 2 0.1 µg/L – 5 µg/L 3-octanol (IS) 101 / 83 IS 1 - - 100 µg/L 2,6-dimethylhept-5-en-2-ol (DMH) (IS) 109 / 124 IS 2 - - 25 µg/L

(IS) indentifies internal standards; m/z in bold identify Ions used for quantification; x: AA/AIS, with A: peak area

xxv

Appendix 19 Calibration curve parameters applied for quantification in the analysis of the low boiling sulfur compounds by means of HS-GC-PFPD

Compound Ions [m/z] Equation of

calibration curve Coefficient of

determination (R2) Internal

standard Calibration range hydrogen sulfide - 0.68362*x-0.0253 0.9957 IS 2 5 µg/L – 75 µg/L methanethiol - 0.68351*x-0.0969 0.9936 IS 2 2 µg/L – 19 µg/L dimethyl sulfide - 1.10476*x-0.1042 0.9989 IS 2 8 µg/L – 45 µg/L carbon disulfide - 3.65824*x-0.2063 0.9958 IS 2 3 µg/L – 20 µg/L methyl thioacetetate - 0.10127*x+0.0297 0.9997 IS 2 2 µ/L – 56 µg/L dimethyl disulfide - 0.87069*x+0.0272 0.9994 IS 2 1 µg/L – 29 µg/L ethyl thioacetate - 0.11614*x+0.0038 0.9997 IS 2 3 µg/L – 65 µg/L diethyl disulfide - 0.65568*x+0.0374 0.9995 IS 2 2 µg – 38 µg/L dimethyl trisulfide - 0.36474*x-0.0047 0.9996 IS 2 1 µg/L – 26 µg/L butylmethyl sulfide (IS) - IS 1 - - methyl-iso-propyl sulfide (IS) - IS 2 - -

(IS) indentifies internal standards; m/z in bold identify Ions used for quantification; x: AA/AIS, with A: peak area

Appendix 20 Ions (m/z) used for compounds identification, quantification and calibration curve parameters applied for quantification in the analysis of cysteinylated and glutathionylated precurors of 3SH by means of HPLC-HRMS

Compound Ions [m/z]

[M+H]+ Equation of

calibration curve Coefficient of

determination (R2) Internal

standard Calibration range 3-S-cysteinylhexan-1-ol 222.1158407 441.08*x-2.4996 0.9945 IS 10 µg/L – 500 µg/L 3-S-glutathionylhexan-1-ol 408.1798974 160.31*x+0.012 0.9998 IS 10 µg/L – 500 µg/L 3-S-glutathionylhexan-1-ol-d3 (IS) 411.1987276 IS - - 200 µg/L

(IS) indentifies internal standards; m/z in bold identify Ions used for quantification; x: AIs/AA, with A: peak area

Appendix 21 Calibration parameters for analysis of N-OPA, [NH4]+, glutathione, and caftaric acid

Compound Equation of

calibration curve Coefficient of

determination (R2) Internal

standard Calibration range N-OPA (iso-L-leucine) - 0.0075z-0.0023 1 - 28 mg/L – 140 mg/L [NH4]+ - 4862539*x+0.000001 1.0001 - 1 mg/L – 500 mg/L reduced glutathione - 2.8011*x 0.9948 - 1 mg/L – 50 mg/L caffeic acid - 532112*x-0.00002 0.9962 - 1.25 mg/L – 200 mg/L

z: Absorbance; x:peak area

xxvi

Appendix 22 List of chemical names used in this thesis – equivalence table

name used in this thesis systematic name

1-methoxy-3-methylsulfanylbutane 1-methoxy-3-methylsulfanylbutane 2-furanylmethanethiol (2FM) furan-2-ylmethanethiol 3-octanol octan-3-ol 3-sulfanylhexan-1-ol (3SH) 3-sulfanylhexan-1-ol 3-sulfanylhexyl acetate (A3SH) 3-sulfanylhexyl acetate 4-methyl-4-sulfanylpentan-2-one (4MSP) 4-methyl-4-sulfanylpentan-2-one 6-sulfanylhexan-1-ol 6-sulfanylhexan-1-ol alpha-ionone (E)-4-(2,6,6-trimethylcyclohex-2-en-1-yl)but-3-en-2-one alpha-terpineol 2-(4-methylcyclohex-3-en-1-yl)propan-2-ol amylacetate 2-methylbutyl acetate benzeneethyl 2-phenylethanol benzenemethanethiol (BMT) phenylmethanethiol benzylethyl acetate 2-phenylethyl acetate beta-citronellol 3,7-dimethyloct-6-en-1-ol

beta-damascenone (E)-1-(2,6,6-trimethylcyclohexa-1,3-dien-1-yl)but-2-en-1-one

beta-ionone (E)-4-(2,6,6-trimethylcyclohexen-1-yl)but-3-en-2-one butylmethyl sulfide 1-methylsulfanylbutane caffeic acid (E)-3-(3,4-dihydroxyphenyl)prop-2-enoic acid capric acid decanoic acid caproic acid hexanoic acid caprylic acid octanoic acid carbin disulfide methanedithione citric acid 2-hydroxypropane-1,2,3-tricarboxylic acid cumene propan-2-ylbenzene dichloroaniline 2,3-dichloroaniline diethyl disulfide ethyldisulfanylethane diethyl succinate diethyl butanedioate dimethyl trisulfide methylsulfanyldisulfanylmethane 2,6-dimethylhept-5-en-2-ol (DMH) 2,6-dimethylhept-5-en-2-ol dimethyl sulphide (DMS) methylsulfanylmethane ethyl 2-sulfanyl acetate ethyl 2-sulfanyl acetate

xxvii

name used in this thesis systematic name

ethyl 2-sulfanylpropionate (E2SP) ethyl 2-sulfanylpropionate ethyl 3-sulfanylpropionate (E3SP) ethyl 3-sulfanylpropionate ethyl acetate ethyl acetate ethyl benzeneacetat ethyl 2-phenylacetate ethyl butyrate ethyl butanoate ethyl capriate ethyl decanoate ethyl caproate ethyl hexanoate ethyl capryloate ethyl octanoate ethyl iso-butyrate ethyl 2-methylpropanoate ethyl lactate ethyl 2-hydroxypropanoate ethyl propanoate ethyl propanoate ethyl thioacetate S-ethyl ethanethioate formic acid formic acid geraniol (2E)-3,7-dimethylocta-2,6-dien-1-ol hexan-1-ol hexan-1-ol hexyl acetate hexyl acetate hydrochloric acid hydrogen chloride hydrogen sulfide hydrogen sulfide iso-amyl alcohol 3-methyl-1-butanol iso-amylacetate 3-methylbutyl acetate iso-butanol 2-methylpropan-1-ol L-cysteine chloride (2)-2-amino-3-sulfanylpropanoic acid hydrochloride linalool 3,7-dimethylocta-1,6-dien-3-ol methanethiol methanethiol methyldisulfide methyldisulfanylmethane methyl-iso-propyl sulfide 2-methylsulfanylpropane methylthioactetate methylthioactetate monobromobimane 1-(bromomethyl)-2,6,7-trimethylpyrazolo[1,2-

a]pyrazole-3,5-dione nerol (2E)-3,7-dimethylocta-2,6-dien-1-ol nerol oxide 4-methyl-2-(methylprop-1-enyl)-3,6-dihydro-2H-pyran p-hydroxymercuribenzoate sodium salt p-hydroxymercuribenzoate sodium salt S-3-(hexan-1-ol)-glutahione-d5 S-3-(hexan-1-ol)-glutahione-d5

xxviii

name used in this thesis systematic name

S-3-(hexan-1-ol)-glutathione S-3-(hexan-1-ol)-glutathione S-3-(hexan-1-ol)-L-cysteine S-3-(hexan-1-ol)-L-cysteine S-glutathionyl caftaric acid S-glutathionyl caftaric acid sodium sulfate disodium sulfate tartaric acid 2,3-dihydroxybutanedioic acid 1,1,6-trimethyl-1,2-dihydronaphthalene (TDN)

1,1,6-trimethyl-1,2-dihydronaphthalene

tert-amyl alcohol 2-methyl-2-butanol 1,2,4-trimethylphenyl-3-butadiene (TPB) 1,2,4-trimethylphenyl-3-butadiene trans/ cis-linalool oxide 2-(5-ethenyl-5-methyloxolan-2-yl)propan-2-ol trans-ethyl cinnamate ethyl (E)-3-phenylprop-2-enoate Trizma 2-amino-2-(hydroxymethyl)propane-1,3-diol vitispirane 2,6,6-trimethyl-10-methylidene-1-oxaspiro[4.5]dec-8-ene

Publications and Conference communications

Publications and Conference communications Publications Armin Schüttler, Bernd R. Gruber, Cécile Thibon, Magalie Lafontaine, Manfred Stoll, Hans R. Schultz, Doris Rauhut, and Philippe Darriet (2012), Influence of environmental stress on secondary metabolite composition of Vitis vinifera var. Riesling grapes in a cool climate region – water status and sun exposure, in Oeno 2011, Actes de colloques du 9ème Symposium international d’œnologie de Bordeaux, Dunod, Paris, 2012, pp. 65-70 Armin Schüttler, Stefanie Fritsch, Rainer Jung, Doris Rauhut, and Philippe Darriet (2012), Tracing of dry Riesling wines aromatic typicality with sensory and instrumental analytical methods – comparative approach, in Oeno 2011, Actes de colloques du 9ieme Symposium international d’œnologie de Bordeaux, Dunod, Paris, 2012, pp. 814-819 Conference communications 9ème Symposium International d’Œnology – Bordeaux, 15-17 june 2011, Bordeaux : Oral presentation : Influence of environmental stress on secondary metabolite composition of Vitis vinifera var. Riesling grapes in a cool climate region – water status and sun exposure. Armin Schüttler, Bernd R. Gruber, Cécile Thibon, Magalie Lafontaine, Manfred Stoll, Hans R. Schultz, Doris Rauhut, and Philippe Darriet 41. Lebensmittelchemikertag, 10-12 september 2012, Münster : Oral presentation : Identifizierung on Minorkomponeneten in Weinaroma mittels sensorikgeleiteter präparativer Fraktionierung und anschließender Applikation von GC-GC-O/HRMS. Armin Schüttler, Holger Zorn, Doris Rauhut, Philippe Darriet Deutscher Sensoriktag 2012 : 08-09 november 2012, Neustadt a.d. W. : Oral presentation : Typizität von Riesling – Vom Konzept über die Sensorik zur chemischen Grundlage. Armin Schüttler, Stefanie Fritsch, Beata Beisert, Christoph Schüssler, Rainer Jung, Doris Rauhut, Philippe Darriet 9ème Symposium International d’Œnology – Bordeaux, 15-17 june 2011, Bordeaux : Poster presentation and Workshop : Tracing of dry Riesling wines aromatic typicality with sensory and instrumental analytical methods – a comparative approach. Armin Schüttler, Stefanie Fritsch, Rainer Jung, Doris Rauhut, and Philippe Darriet XIII Weurmann Flavour Research Symposium, 27 – 30 september 2011, Zaragossa : Poster presentation and Workshop : Characterising of dry Riesling wines’ aromatic typicality with sensory and instremental analytical methods – comparative approach. Armin Schüttler, Stefanie Fritsch, Rainer Jung, Doris Rauhut, and Philippe Darriet

Declaration of academical honesty

Declaration of academical honesty

Eidesstattliche Erklärung

Ich erkläre: „Ich habe die vorgelegte Dissertation selbstständig und ohne unerlaubte fremde Hilfe und nur mit den Hilfen angefertigt, die ich in der Dissertation angegeben habe. Alle Textstellen, die wörtlich oder sinngemäß aus veröffentlichten Schriften entnommen sind, und alle Angaben, die auf mündlichen Auskünften beruhen, sind als solche kenntlich gemacht. Bei den von mir durchgeführten und in der Dissertation erwähnten Untersuchungen habe ich die Grundsätze guter wissenschaftlicher Praxis, wie sie in der Chartes des thèses der Ecole Doctoral des Sciences de la Vie der Université Segalen Bordeaux 2 und der Satzung der Justus-Liebig-Universität Gießen zur Sicherung guter wissenschaftlicher Praxis niedergelegt sind, eingehalten“. Geisenheim, den 12.11.2012