chemistry of hydroxycinnamate esters and their role as ... · bruxellensis . two classes of berry...

Chemistry of Hydroxycinnamate Esters and

their Role as Precursors to Dekkera Produced

Off-flavour in Wine

A thesis presented in fulfilment of the

requirements for the degree of

Doctor of Philosophy

Josh L. Hixson

BTech (Forens&AnalytChem), BSc (Hons)

School of Agriculture, Food and Wine

March 2012

i

Table of Contents

Abstract ................................................................................................................................ iv

Declaration ......................................................................................................................... vii

Acknowledgements ........................................................................................................... viii

Publications and Symposia ................................................................................................ xi

Abbreviations .................................................................................................................... xiii

Figures, Schemes and Tables ........................................................................................... xvi

Chapter 1: Introduction ...................................................................................................... 1

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

1.2 Dekkera/Brettanomyces bruxellensis ............................................................................... 1

1.3 Volatile Phenols ............................................................................................................... 5

1.3 Introduction to Tartrate Esters ....................................................................................... 11

1.4 Introduction to Glucose Esters ....................................................................................... 16

1.5 Introduction to Ethyl Esters ........................................................................................... 18

1.5 Research Aims ............................................................................................................... 20

Chapter 2: Synthesis of Hydroxycinnamoyl Esters ........................................................ 22

2.1 Synthesis of Hydroxycinnamic Acids and Derivatives ................................................. 22

2.2 Synthesis of Hydroxycinnamoyl Tartrate Esters ........................................................... 24

2.2.1 Introduction to Tartrate Ester Synthesis ............................................................. 24

2.2.2 Synthesis of Hydroxycinnamoyl Tartrate Esters ................................................ 26

2.3 Synthesis of Hydroxycinnamoyl Glucose Esters ........................................................... 34

2.3.1 Introduction to Glucose Ester Synthesis ............................................................. 34

2.3.2 Synthesis of Hydroxycinnamoyl Glucose Esters ................................................ 37

2.4 Conclusions .................................................................................................................... 47

ii

Chapter 3: Acyl Migration of Hydroxycinnamoyl Glucose Esters ............................... 50

3.1 Introduction. ................................................................................................................... 50

3.2 Research Aims ............................................................................................................... 54

3.3 Theoretical Studies into Acyl Migration of Hydroxycinnamoyl Glucoses ................... 55

3.3.1 Thermodynamics of Migration ........................................................................... 55

3.3.2 Kinetics of Migration .......................................................................................... 60

3.4 Liquid Chromatography of Wine. .................................................................................. 67

3.5 Conclusions. ................................................................................................................... 76

Chapter 4: Photoisomerisation of Hydroxycinnamic Acids .......................................... 79

4.1 Introduction. ................................................................................................................... 79

4.1.1 Hydroxycinnamate Photoisomerisation .............................................................. 79

4.1.2 cis-Hydroxycinnamate content in grapes and wine ............................................ 81

4.1.3 Enzymatic Specificity ......................................................................................... 83

4.2 Research Aims ............................................................................................................... 86

4.3 Synthesis of cis-Hydroxycinnamic Acids. ..................................................................... 87

4.4 Theoretical Studies into the Isomerisation of Hydroxycinnamic Acids ........................ 91

4.5 Conclusions .................................................................................................................. 105

Chapter 5: Bioconversion of Hydroxycinnamates by D. bruxellensis ......................... 107

5.1 Bioconversion of trans-Hydroxycinnamate Esters ...................................................... 107

5.1.1 Ethyl Esters ....................................................................................................... 107

5.1.2 Ethyl Esterase Substrate Selectivity ................................................................. 110

5.1.3 Tartrate Esters ................................................................................................... 111

5.1.4 Glucose Esters ................................................................................................... 113

5.1.5 Conclusions for Chapter 5.1 ............................................................................. 114

5.2 Stereoselectivity of D. bruxellensis Enzyme Activities ............................................... 115

5.2.1 Decarboxylase Stereoselectivity ....................................................................... 115

iii

5.2.2 Ethyl Esterase Stereoselectivity ........................................................................ 121

5.2.3 Conclusions for Chapter 5.2 ............................................................................. 125

5.3 Thesis Conclusions and Future Directions .................................................................. 126

Chapter 6: Experimental ................................................................................................. 130

6.1 General Experimental .................................................................................................. 130

6.2 Experimental Procedures for Chapter 2 ....................................................................... 133

6.2.1 Hydroxycinnamoyl Derivatives ........................................................................ 133

6.2.2 Synthesis of Hydroxycinnamoyl Tartrate Esters .............................................. 142

6.2.3 Synthesis of Hydroxycinnamoyl Glucose Esters .............................................. 157




6.5.1 General Procedures for Chapter 5 ..................................................................... 180

6.5.2 Fermentation of trans-Hydroxycinnamate Esters ............................................. 184

6.5.3 Stereoselectivity of D. bruxellensis Enzyme Activities .................................... 184

Appendix 1: Data for Migration Thermodynamics ...................................................... 186

Appendix 2: Data for Migration Kinetics ...................................................................... 188

Appendix 3: Data for Energy Profiles ............................................................................ 190

Appendix 4: Data for Vertical Excitations and HOMO-LUMO Gaps ....................... 192

Appendix 5: Data from Ethylphenol Analyses .............................................................. 196

References ......................................................................................................................... 198

iv

Abstract

The potential for malodour in wine caused by the accumulation of ethylphenols has been

widely studied with respect to the breakdown of the hydroxycinnamic acids, p-coumaric

and ferulic acid, by D. bruxellensis. The presence of esterified hydroxycinnamate

conjugates in grapes and wine is well established and they account for a large proportion of

the hydroxycinnamate content. There exists the possibility that these conjugates could also

provide the potential for spoilage, though they have never been linked to the direct

formation of ethylphenols. The research highlighted within this thesis examines the

potential role of a number of esterified conjugates in the production of ethylphenols by D.

bruxellensis. Two classes of berry derived esters, the tartaric acid and glucose bound

hydroxycinnamates, as well as the vinification formed ethyl esters, were synthesised and

used for model fermentation experiments.

Chapter 2 describes the preparation of a number of protected hydroxycinnamic acid

derivatives that were used in the synthesis of the hydroxycinnamoyl tartrate esters (7 and

8) for the first time. Coupling 1-O-chloroacetyl protected p-coumaric and ferulic acids (21

and 22) with di-tert-butyl-L-tartrate (34) followed by selective hydrolysis of the tert-butyl

esters yielded p-coumaroyl tartrate (7) and feruloyl tartrate (8). Hydroxycinnamoyl glucose

esters (9 and 10) were prepared using the same hydroxycinnamates (21 and 22), esterifying

with a prepared trichloroacetimidate glucosyl donor sequence, though purification of the

glucose esters resulted in undesired chemical transformations. It was found that

photoisomerisation of the glucose esters could be prevented via synthesis under red light,

which gave trans-9 and 10, however migration of the hydroxycinnamoyl moiety around

the glucose ring, which yielded mainly the 2-O-α- and 6-O-α-esters, was a product of

submitting the esters to non-aqueous solvents and could not be avoided.

The acyl migration of the glucose esters that was observed in Chapter 2 has been

researched at a DFT B3LYP 6-31G* theoretical level in Chapter 3 with respect to both the

thermodynamics and kinetics of the transformations. The desired 1-O-β-esters were

thermodynamically favoured only in water, while in any other solvent studied the 2-O-α-

and 6-O-α-esters would prevail. Kinetically, migration to the 3-O-position involved lower

energy barriers which can be equated to a more rapid process, although the ring-flipped

v

conformation needed to achieve the migration would promote subsequent migration to the

6-O-position. Step-wise migration, from the 1-O- to the 2-O-position, was found to be

thermodynamically less favoured than other migrations investigated. This effect can be

rationalised by the formation of a 5-membered cyclic intermediate in comparison to the 6-

membered intermediate produced during 1-O- to 3-O-migration. However, the energy

barriers involved in 1-O-β- to 2-O-β-migration better explain the comparative extent of

migration observed between the p-coumaroyl and feruloyl glucose esters. The possibility

of multiple glucose esters existing in wine was the focus of a brief study, finding two

separate p-coumaroyl glucose esters in red and white wine, while a lesser extent of

migration in feruloyl glucose limited observation to concentrated wine alone. However,

due to co-elution of feruloyl glucose (10) with suspected p-coumaroyl anthocyanin

derivatives in red wine, HPLC-MRM was required to detect it, which is the first report of

this compound in red wine.

Theoretical studies into observed photoisomerisations and the synthesis of cis-

hydroxycinnamates are described in Chapter 4. The cis-ethyl hydroxycinnamates were

isolated and hydrolysed to give a mixture of cis/trans-hydroxycinnamic acids (3 and 4),

which could be separated by flash chromatography, though the pure cis-isomers isomerised

rapidly under ambient conditions and slowly under red light back to the trans-isomers.

Stable isomeric mixtures were achieved by irradiation with ultra-violet light giving

mixtures of 40-50% of the cis-isomer which could be used further in fermentation studies.

Computational evidence suggested that isomerisation of the hydroxycinnamic acids was

favoured with greater resonance throughout the molecule. Those with deprotonated

phenolic moieties possessed the most intramolecular electron movement, decreasing the

HOMO-LUMO gap and promoting photoisomerisation. Smaller solvent and substrate

effects were also noted, though the nature of the phenol and carboxyl clearly played the

most important role in determining stability of each isomer.

Fermentation in the presence of the synthesised trans-hydroxycinnamoyl esters (7-12) and

investigation into the stereospecificity of D. bruxellensis enzyme activities was performed

as detailed in Chapter 5. In Australia, three genetic groups of D. bruxellensis account for

98% of isolates, with the largest of these groups making up 85%. AWRI 1499 is a

representative of the largest genetic group, with AWRI 1608 and AWRI 1613 belonging to

the two remaining significant genetic groups. In the presence of AWRI 1499, the trans-

vi

ethyl esters (11 and 12) were metabolised to varying extents with the preference for

breakdown of ethyl coumarate (11) over ethyl ferulate (12). This selectivity was

investigated further and found to be common for both AWRI 1499 and AWRI 1608, while

AWRI 1613 was unable to breakdown either ester. The preference for formation of 4-

ethylphenol (1) over 4-ethylguaiacol (2) from the ethyl esters could accentuate the ratio of

these compounds as seen in wine, initially thought to be brought about by the relative

concentration of the precursor acids.

Of the berry derived esters, the tartrate esters (7 and 8) were not metabolised by AWRI

1499, and subsequent fermentations with AWRI 1608 and 1613 yielded the same result.

This confirmed that the tartrate esters cannot contribute directly to the formation of

ethylphenols during exposure to D. bruxellensis. The glucose esters were metabolised by

AWRI 1499 to a moderate extent (35% conversion), providing information that these can

contribute to the accumulation of ethylphenols during barrel ageing. Furthermore, the

isomerisation of the glucose esters lead to studies into the stereoselectivity of D.

bruxellensis enzyme activities, whereby the decarboxylase as well as the ethyl esterase

showed selectivity for the trans-isomers and that the cis-hydroxycinnamate content of

grapes and wine are not important in the accumulation of ethylphenols. The experimental

procedures employed throughout Chapters 2-5 are outlined in Chapter 6.

vii

Declaration

This work contains no material which has been accepted for the award of any other degree

or diploma in any university or other tertiary institution and, to the best of my knowledge

and belief, contains no material previously published or written by another person, except

where due reference has been made in the text.

I give consent to this copy of my thesis, when deposited in the University Library, being

available for loan and photocopying, subject to the provisions of the Copyright Act 1968.

I also give permission for the digital version of my thesis to be made available on the web,

via the University’s digital research repository, the Library catalogue, the Australian

Digital Thesis Program (ADTP) and also through web search engines, unless permission

has been granted by the University to restrict access for a period of time.

…………………………………..

Josh L. Hixson

…………………………………..

viii

Acknowledgements

In no particular order, other than chronologically, I would like to thank my supervisors for

their commitment to my learning. Dr. Gordon Elsey…… Gordy, you have been an

absolute inspiration since I met you in 2006 and you are the reason I started this particular

journey. You have pushed me to know more and work harder and have been a constant

source for knowledge outside of the field of chemistry as well as within it. It saddens me

that we didn’t get to finish this journey together, and it saddens me even more that it was

easier for me to dismiss you and carry on without you, rather than help you through some

tough times and for that I am sorry. I will always consider you a friend, regardless of the

past, and I truly believe that you have contributed as much by leaving me to research and

become independent as you have to actively increasing my knowledge.

Dr. Chris Curtin for taking me on during my Ph.D. after seeing my complete lack of

microbiology skills during my honours year and still wanting to get the best out of me and

instill into me as much microbiology knowledge as possible. Also for being a fantastic

outlet when synthetic chemistry became too much and we could discuss fermentation

experiments or I could listen to you get excited about potential enzyme activities that could

be expressed.

When the notion of leaving Flinders University in 2008 arose, the choice to relocate to The

University of Adelaide was made so much easier by being ‘adopted’ by Prof. Dennis

Taylor. Den, thank you for the opportunity of working under you. You have given me such

an insight into the workings of a university always open to discuss which grants you were

applying for and what the outcomes were, when you really didn’t have to. From finding me

a scholarship at very short notice, to offering me roles in the lab to keep my mind off of

what was making me unhappy. You have shown me that there is nothing wrong with

breaking the mold, because it’s not held together that well to start off with.

The final member of the supervision team, who picked up the slack when it was needed,

Dr. Mark Sefton. It has been a pleasure to work with such a fantastic and knowledgeable

flavour and aroma chemist and I sometimes forget how lucky I am. I have honestly been

approached with ‘you work with Dr Sefton? He is a legend, I have read so many of his

ix

papers’ and that was on the other side of the world. Thank you for giving me advice when I

felt like there was nobody else who wanted to give me any.

The members of the original Elsey/AWRI group who moved to Adelaide with me, Natoiya

and Jo, and those who I found when I arrived, Pete and Nicole. You have been good

friends, and have learnt when to leave me alone and when to make me laugh. I have

probably spent as much time complaining to you about all sorts of things than I have

talking to you about science, but by letting me vent, you have definitely helped me get

through.

People that haven’t been on the whole journey, but those that have helped along the way,

Dr. Simon Mathew for advice about theoretical calculations, Dr. Eric Dennis for advice on

anything I needed or just a random message to keep the spirits up, Dr. Dave Jeffery for

being the only person to come and visit me in my red lab in the basement and discuss

deprotection strategies. Ms. Dimi Capone for running ethylphenol analyses and keeping

the instruments running so very well and Dr. Edward Tiekink at the University of Malaysia

for performing X-ray crystallography.

To my family, especially my parents, thank you for understanding why I am doing this and

not out getting a job, and also for pushing me into university when I would have been just

as happy lying on the couch…at the time that is. I am sorry it took me until I started

honours to actually apply myself to do anything, I am hoping you have largely forgotten

my first 21 years of complete laziness and contentment with achieving the minimum

amount to survive.

The biggest thanks of all has to go to my wife and best friend Suey. When you met me I

was a lay about undergraduate student who was about to dropout of uni and find something

more exciting to do, but you have stuck with me, and supported me throughout my

extended university stay, financially, emotionally and physically. Thank you for being

there when I have needed you, and for knowing when I have needed support, even if I was

too tired or grumpy or hungry to figure it out. Thank you for putting up with long days,

long nights, long sleeps and long ramblings about my project, without you I wouldn’t have

made it through this. Also for giving me the ultimate inspiration to get this done and start

x

my working life, I can’t wait to meet Googy and finally take my place as the provider for

my new family.

xi

Publications and Symposia

Publications:

Hixson, J. L.; Sleep, N. R.; Capone, D. L.; Elsey, G. M.; Curtin, C. D.; Sefton, M. A.;

Taylor, D.K. Hydroxycinnamic acid ethyl esters as precursors to ethylphenols in wine. J.

Agric. Food Chem. Accepted 12/02/2012.

Hixson, J. L.; Curtin, C. D.; Sefton, M. A.; Taylor, D. K. Stereospecificity of D.

bruxellensis in the production of ethylphenol off-flavour in wine. Proceedings of the 13th

Weurman Flavour Research Symposium. In press.

Hixson, J. L.; Taylor, D. K.; Ng, S. W.; Tiekink, E. R. T. Di-tert-butyl (2R,3R)-2-({(2E)-3-

[4-(acetyloxy)-3-methoxyphenyl]prop-2-enoyl}oxy)-3-hydroxybutanedioate. Acta

Crystallographica, Section E 2012, 68 (3), o509-o510.

Hixson, J. L.; Taylor, D. K.; Ng, S. W.; Tiekink, E. R. T. Di-tert-butyl (2R,3R)-2-({(2E)-3-

[4-(acetyloxy)phenyl]prop-2-enoyloxy)-3-hydroxybutanedioate. Acta Crystallographica,

Section E 2012, 68 (2), o568-o569.

Hixson, J. L.; Elsey, G. M.; Curtin, C. D.; Sefton, M. A.; Taylor, D.K. Hydroxycinnamoyl

glucose and tartrate esters and their role in the formation of ethylphenols in wine. J. Agric.

Food Chem. In draft.

Symposia:

Hixson, J. L.; Curtin, C. D.; Taylor, D. K.; Elsey, G. M. Mapping the Metabolic Inputs of

‘Brett’ Taint. Poster presented at the 2009 YPD conference (Meeting of the Australasian

Yeast Group).

Hixson, J. L.; Curtin, C. D.; Taylor, D. K. Stereospecificity of the Decarboxylase Enzyme

of D. bruxellensis. Poster presented at the 14th Wine Industry Technical Conference, 2010.

xii

Hixson, J. L.; Elsey, G. M.; Curtin, C. D.; Taylor, D. K. Isomerisation of the

Hydroxycinnamic Acids and their Role in the Production of Wine Off-aroma. Seminar

presented at the 2010 Adelaide Synthetic Chemistry Symposium.

Hixson, J. L.; Curtin, C. D.; Sefton, M. A.; Taylor, D. K. Determination of Alternative

Precursors to Brettanomyces/Dekkera Produced Off-flavour. Seminar presented at the 13th

Weurman Flavour Research Symposium, 2011.

xiii

Abbreviations

4-EG 4-Ethylguaiacol

4-EP 4-Ethylphenol

Å Angstroms

Ac Acetyl

AcCl Chloroacetyl

AcCN Acetonitrile

app. d Apparent doublet

Ar Aromatic

AWRI Australian Wine Research Institute

Bn Benzyl

br Broad

COSY Correlation spectroscopy

d Doublet

DAD Diode array detector

DCM Dichloromethane

dd Doublet of doublets

ddd Doublet of doublet of doublets

DFT Density functional theory

EIC Extracted ion chromatogram

ESI Electrospray ionization

Et Ethyl

Et2O Diethyl ether

EtOAc Ethyl acetate

g Grams

GC Gas chromatography

Glc Glucose

HCA Hydroxycinnamic acid

HMBC Heteronuclear multiple bond correlation

HMQC Heteronuclear multiple quantum coherence

HOMO Highest occupied molecular orbital

HPLC High-performance liquid chromatography

HRMS High resolution mass spectroscopy

xiv

Hz Hertz

hν Light

J Coupling constant

kJ Kilojoules

L Litre

LC Liquid chromatography

Lit. Literature

LUMO Lowest unoccupied molecular orbital

m Multiplet

M Molar (moles/litre)

m/z Mass to charge ratio

mg Milligrams

MgSO4 Magnesium sulphate

MHz Megahertz

ML Megalitre

mL Millilitre

MMFF Merck Molecular Force Field

mmol Millimoles

mol Moles

m.p. Melting point

MRM Multiple reaction monitoring

MS Mass spectrometry

MYPG Malt, yeast extract, peptone, glucose

nm Nanometres

NMR Nuclear magnetic resonance

p Para

Ph Phenyl

ppb Parts per billion

ppm Parts per million

q Quartet

Rf Retension factor

rpm Revolutions per minute

s Singlet

S0 Singlet ground state

xv

S1 Singlet first excited state

t Triplet

T1 Triplet first excited state

tert Tertiary

THF Tetrahydrofuran

TIC Total ion chromatogram

TLC Thin layer chromatography

TMS Tetramethyl silane

UV Ultra-violet

Vis Visible

VNBC Viable but non-culturable

X4 Hexane fraction

YNB Yeast extract, nitrogen, base

YPD Yeast extract, peptone, dextrose

δ Chemical shift

µ Micro

xvi

Figures, Schemes and Tables

List of Figures:

Figure 1.1: Ethylphenols produced by D. bruxellensis in red wine. ...................................... 5

Figure 1.2: Enzymatic conversion of hydroxycinnamic acids to volatile phenols. ............... 6

Figure 1.3: L-Tartaric acid esters of p-coumaric acid (7) and ferulic acid (8). ................... 11

Figure 1.4: 1-O-β-D-Glucose esters of p-coumaric acid (9) and ferulic acid (10). ............. 16

Figure 1.5: Ethyl hydroxycinnamates. ................................................................................. 18

Figure 1.6: Evolution of ethyl coumarate in Shiraz wine. ................................................... 19

Figure 1.7: Hydroxycinnamoyl tartrate (7 and 8), glucose (9 and 10) and ethyl esters (11

and 12) to be synthesised and used in these studies. ........................................................... 20

Figure 2.1: Molecular structure and crystallographic numbering scheme for 35. ............... 31

Figure 2.2: Molecular structure and crystallographic numbering scheme for 36. ............... 32

Figure 2.3: 1H proton NMR spectrum of the chloroacetyl protons in 2,3,4,6-O-

tetrachloroacetyl-β-D-glucopyranosyl cinnamate (48). ....................................................... 40

Figure 2.4: NMR spectrum of isomerised glucose esters. a) cis/trans-Feruloyl glucose (10).

b) cis/trans-Cinnamoyl glucose (53). .................................................................................. 42

Figure 2.5: Hydroxycinnamate esters to be used in fermentation experiments. .................. 48

Figure 2.6: Dominant equilibria in hydroxycinnamoyl glucose ester mixtures to be used in

fermentation experiments. ................................................................................................... 48

Figure 3.1: Acyl migration in p-coumaroyl glucose. ........................................................... 50

Figure 3.2: Initial silica catalysed 3-S- to 6-O-migration observed by Whistler et al. ........ 51

Figure 3.3: Migration intermediates. a) Base-catalysed 1-O-β- to 2-O-β-migration

intermediate proposed by Iddon et al. b) Acid-catalysed 4-O-α- to 6-O-α-migration

intermediate proposed by Horrobin et al. ............................................................................ 51

Figure 3.4: Proposed migration of mono-O-chloroacetyl derivatives to the 6-O-position. . 53

Figure 3.5: Twenty possible esters of p-coumaroyl glucose (9) and feruloyl glucose (10). 55

Figure 3.6: Energy of p-coumaroyl and feruloyl glucose esters in water, relative to the 1-O-

β-esters.. ............................................................................................................................... 56

Figure 3.7: Energy of p-coumaroyl and feruloyl glucose esters in dichloromethane, relative

to the 1-O-β-esters.. ............................................................................................................. 57

Figure 3.8: Energy of p-coumaroyl and feruloyl glucose esters in ethanol, relative to the 1-

O-β-esters.. ........................................................................................................................... 58

xvii

Figure 3.9: Energy of p-coumaroyl and feruloyl glucose esters in toluene, relative to the 1-

O-β-esters.. ........................................................................................................................... 58

Figure 3.10: p-Coumaroyl glucose (9) ester energies calculated in changing solvents,

relative to the 1-O-α-esters.. ................................................................................................ 59

Figure 3.11: Key intermediates (Int. 1-4) for the acid-catalysed 1-O-β- to 2-O-β- acyl

migration of p-coumaroyl glucose (9). ................................................................................ 62

Figure 3.12: Energy of the intermediates in 1-O-β- to 2-O-β-p-coumaroyl glucose

migration, relative to intermediate 1.. .................................................................................. 63

Figure 3.13: Energy of the intermediates in 1-O-β- to 2-O-β-feruloyl glucose migration,

relative to intermediate 1.. ................................................................................................... 63









Figure 3.18: Glucose ring-flip to facilitate 1-O- to 3-O-migration and 1-O- to 6-O-

migration. ............................................................................................................................. 67

Figure 3.19: p-Coumaroyl glucose. a) Extracted ion chromatogram of m/z 325. b) Mass

spectrum at 29.6 to 29.8 minutes. ........................................................................................ 69

Figure 3.20: Feruloyl glucose. a) Extracted ion chromatogram of m/z 355. b) Mass

spectrum at 36.5 to 36.6 minutes. ........................................................................................ 69

Figure 3.21: Concentrated white wine, extracted ion chromatogram of m/z 325. ............... 70

Figure 3.22: Mass spectra of compounds identified in extracted ion chromatogram of m/z

325. ...................................................................................................................................... 71

Figure 3.23: Concentrated white wine, extracted ion chromatogram of m/z 355. ............... 72


355. ...................................................................................................................................... 72

Figure 3.25: Red wine chromatogram (DAD). .................................................................... 73

Figure 3.26: Concentrated red wine, extracted ion chromatogram of m/z 325. ................... 73

Figure 3.27: Concentrated red wine, extracted ion chromatogram of m/z 355. ................... 74

xviii

Figure 3.28: HPLC-MRM traces (aglycone - blue, aglycone minus water - red) of

hydroxycinnamoyl glucose esters. a) Pure glucose esters. b) Neat white wine. c)

Concentrated white wine. d) Concentrated red wine. .......................................................... 75

Figure 4.1: Photoisomerisation of the hydroxycinnamoyl glucose esters. .......................... 79

Figure 4.2: Electron configuration of π bonding and anti-bonding molecular orbitals in

ground and excited states. .................................................................................................... 80

Figure 4.3: Compounds investigated in decarboxylation studies. ....................................... 85

Figure 4.4: Proposed resonance assisted conversion of cis-p-coumaric acid to trans-p-

coumaric acid. ...................................................................................................................... 88

Figure 4.5: Intended effect of metal coordination on cis-hydroxycinnamates. ................... 91

Figure 4.6: Frontier molecular orbital diagrams of trans-p-coumaric acid (3). a) HOMO of

S0 trans-p-coumaric acid. b) LUMO of S0 trans-p-coumaric acid. c) HOMO of T1 trans-p-

coumaric acid. ...................................................................................................................... 92

Figure 4.7: Electron spin density in T1 trans-p-coumaric acid. ........................................... 93

Figure 4.8: Energy profile of p-coumaric acid (3).. ............................................................. 94

Figure 4.9: Energy profile produced from forward and reverse dynamic, and manual

constraint of ethyl coumarate (11).. ..................................................................................... 95

Figure 4.10: Pyramidilised alkene resulting from rotation of the dihedral angle from 180o

to 0o in ethyl coumarate (11). ............................................................................................... 95

Figure 4.11: cis-Ethyl coumarate conformers produced by: a) drawing trans-ethyl

coumarate and rotating the dihedral; and b) drawing cis-ethyl coumarate. ......................... 96

Figure 4.12: Energy profile for p-coumaroyl glucose (9), relative to S0 trans-isomer.. ...... 96

Figure 4.13: a) T1 energy profile for p-coumaric acid (3) in water, relative to the S0 trans-

acid. b) T1 energy profile for p-coumaroyl glucose (9) water, relative to the S0 trans-

isomer. .................................................................................................................................. 97

Figure 4.14: S0-T1 vertical excitation energy for trans-p-coumaric acid (3) and trans-p-

coumaroyl glucose (9). . ...................................................................................................... 98

Figure 4.15: HOMO-LUMO gap for trans-p-coumaric acid and trans-p-coumaroyl

glucose.. ............................................................................................................................... 99

Figure 4.16: a) Vertical excitation energies (S0-T1) of trans-p-coumaroyl glucose phenolate

in solvents of differing polarity. b) HOMO-LUMO gap.. ................................................... 99

Figure 4.17: a) HOMO-LUMO gap of trans-hydroxycinnamates. b) HOMO-LUMO gap of

cis-hydroxycinnamates. ..................................................................................................... 101

xix

Figure 4.18: HOMO-LUMO gaps of cis-hydroxycinnamates during base-catalysed ester

hydrolysis. . ........................................................................................................................ 102

Figure 4.19: Numbering of oxygen atoms in hydroxycinnamate skeleton. ....................... 103

Figure 4.20: HOMO-LUMO gap of p-coumaric acid carboxylate.. .................................. 103

Figure 4.21: HOMO-LUMO gaps of hydroxycinnamate derivatives against ratio of charge

between oxygen 1 and oxygen 3.. ...................................................................................... 104

Figure 4.22: Relationship between HOMO-LUMO gap and double bond length in

hydroxycinnamate derivatives.. ......................................................................................... 105

Figure 5.1: Ethyl coumarate (11) and ethyl ferulate (12). ................................................. 107

Figure 5.2: Percentage of the theoretical maximum conversion of ethyl esters (11 and 12)

to ethylphenols. .................................................................................................................. 108

Figure 5.3: Percentage recovery of coumarates in fermentations. ..................................... 109

Figure 5.4: Percentage recovery of ferulates in fermentations. ......................................... 109

Figure 5.5: Percentage of the theoretical maximum conversion from ethyl coumarate (11)

and ethyl ferulate (12) to ethylphenols by different strains of D. bruxellensis. ................ 110

Figure 5.6: p-Coumaroyl L-tartrate (7) and feruloyl L-tartrate (8). ................................... 111

Figure 5.7: p-Coumaroyl glucose (9) and feruloyl glucose (10). ...................................... 113

Figure 5.8: Percentage of the theoretical maximum conversion of hydroxycinnamoyl

glucose esters (9 and 10) to ethylphenols. ......................................................................... 113

Figure 5.9: Percentage of the theoretical maximum conversion to 4-ethylguaiacol for the

trans- and cis/trans- fermentations. ................................................................................... 116

Figure 5.10: Evolution of 4-ethylguaiacol in cis/trans-fermentations as a percentage of

maximum conversion observed in trans-fermentations. .................................................... 117

Figure 5.11: Compounds by percentage in end-point fermentation samples. ................... 119

Figure 5.12: Percentage of the theoretical maximum conversion to 4-ethylphenol in trans-

and cis/trans- fermentations. .............................................................................................. 120

Figure 5.13: Evolution of 4-ethyphenol in cis/trans-fermentations as a percentage of

maximum conversion observed in trans-fermentations. .................................................... 120

Figure 5.14: Percentage of the theoretical maximum ethylphenol conversion from cis-ethyl

esters. ................................................................................................................................. 122

Figure 5.15: Total coumarate recovery from cis-fermentations. ....................................... 124

Figure 5.16: Total ferulate recovery from cis-fermentations. ............................................ 124

Figure 5.17: Breakdown of ethyl ferulate in a single fermentation. .................................. 125

xx

List of Schemes:

Scheme 2.1: Synthesis of hydroxycinnamic acid derivatives. ............................................. 22

Scheme 2.2: Literature syntheses of mono-esters of tartaric acid. ...................................... 25

Scheme 2.3: Literature syntheses of chicoric acid. .............................................................. 26

Scheme 2.4: Synthesis of benzylated hydroxycinnamoyl tartrate esters. ............................ 27

Scheme 2.5: Attempted debenzylation procedures. ............................................................. 28

Scheme 2.6: Synthesis of di-tert-butyl tartrate. ................................................................... 29

Scheme 2.7: Esterification of hydroycinnamic acids and di-tert-butyl tartrate. .................. 30

Scheme 2.8: Synthesis of hydroxycinnamoyl tartrate esters. .............................................. 34

Scheme 2.9: Modified Koenigs-Knorr reaction conditions employed within this research

group. ................................................................................................................................... 35

Scheme 2.10: Glycosylation reactions of Ziegler. ............................................................... 36

Scheme 2.11: Glycosylation method described by Galland. ............................................... 37

Scheme 2.12: Synthesis of 1-O-benzyl hydroxycinnamoyl glucopyranoses. ...................... 39

Scheme 2.13: Synthesis of glucose esters with free hydroxycinnamic acids. ..................... 41

Scheme 2.14: Glycosylation with 1-O-acetyl hydroxycinnamic acids and partial

deacetylation using XAD-8 resin. ........................................................................................ 44

Scheme 2.15: Glycosylation of 1-O-chloroacetyl hydroxycinnamates, and migration of the

free glucose esters. ............................................................................................................... 46

Scheme 3.1: Mechanism for acid catalysed 1-O-β- to 2-O-β- acyl migration of p-

coumaroyl glucose (9). ........................................................................................................ 61

Scheme 4.1: Attempted synthesis of cis-hydroxycinnamic acids. ....................................... 88

List of Tables:

Table 1.1: Hydroxycinnamoyl tartrate ester concentrations in different grape varieties. .... 12

Table 1.2: Changes in p-coumaroyl tartrate concentration during malolactic fermentation.

............................................................................................................................................. 13

Table 1.3: Changes in p-coumaroyl tartrate concentration during wine storage. ................ 14

Table 2.1: 1H NMR shifts for migrated hydroxycinnamoyl glucose esters. ........................ 47

Table 4.1: Content of cis- and trans-p-coumaroyl tartrate in the skin and juice of red and

white grapes. ........................................................................................................................ 82

Table 4.2: Isomeric ratio of p-coumaric acid (3) under different storage conditions. ......... 90

Table 4.3: Solvent polarities and ET30 values. ..................................................................... 98

xxi

Table 5.1: Ethylphenol content in tartrate ester fermentation experiments. ...................... 112

Table 5.2: Concentration of cis- and trans-ferulic acid in end-point fermentation samples.

........................................................................................................................................... 119

Table 5.3: Final trans-ethyl ester content in cis-ethyl ester fermentations. ....................... 123

“Success is the ability to go from one failure to another with no loss of enthusiasm” Sir Winston Churchill

“I’m a great believer in luck, and I find the harder I work the more I have of it”

Thomas Jefferson

Chapter 1: Introduction

1

Chapter 1: Introduction.

1.1 General Introduction.

The global history of wine production spans back many thousands of years, supported by

the discovery of wine vessels that have been dated to as early as circa 5400-5000 B.C., as

well as there being numerous biblical references to wine.1 The history of Australian wine,

however, begins with European settlement in 1788 when the Lady Penrhyn,2 one of the

eleven ships in the First Fleet, arrived carrying on it vine cuttings and seeds from the

species Vitis vinifera obtained en route from Brazil and the Cape of Good Hope.3 Within

days of landing, the vines had been planted in the governor’s garden in Sydney Cove but

failed to thrive due to the humid coastal climate. Late in 1788 the first inland farming

settlement was established at Rose Hill (now Parramatta), some 24 km further inland. This

region had a much drier climate allowing for cultivation of a vineyard and by 1791 boasted

3 acres consisting of 8000 vines and the following year produced approximately 150 kg of

table grapes.3 In 1795 Philip Schaffer became the colony’s first vigneron when he

produced some 90 gallons, or approximately 400 litres of wine.2

From humble beginnings, the Australian wine industry has developed extensively,4 now

consisting of 2477 wineries, crushing 1,603,000 tonnes of grapes and producing 1,533 ML

of wine,5 of which, 777 ML is exported worldwide, with a value of A$2,167,200,000.6

South Australia contributes 48% of the total volume of Australian wine production, which

is a product of crushing 689,000 tonnes of grapes.7

Even though the world of wine has largely moved past the fortuitous fermentation of grape

juice caused by indigenous yeast, and into an industry of more controlled and predictable

fermentations with cultured or purchased yeasts,8-9 there still remains uncertainties that can

be brought about through the presence of unwanted microrganisms.10

1.2 Dekkera/Brettanomyces bruxellensis.

A constant issue encountered in wine making throughout the world,11 and of interest in

Australia since the first reported occurrence in 1986, is that of contamination caused by


2

yeast of the Brettanomyces and Dekkera genera,12 through their potential to cause a prolific

economic impact on the wine industry.13-14

The name Brettanomyces was originally introduced in 1904 as the characteristics produced

by this yeast were similar to the English beers of the time, with the prefix ‘Brettano’ a

reference to the British brewing industry. The first report of this yeast in wine came much

later, when in 1930 Mycotorula intermedia was isolated from a French must, later to be

reclassified as Brettanomyces. The early 1950’s saw the first report of this genus in bottled

wine, which was closely followed by the discovery of a sporulating form of

Brettanomyces,15 which was categorised into the new genus, Dekkera.16

Since then species within Brettanomyces and Dekkera have been classified and re-

classified numerous times,17 with the work of Smith and Poot18 and Boekhout19 being

largely responsible for the current taxonomy of these genera. There consists five species of

Brettanomyces (bruxellensis, anomolus, naardenensis, custersianus, nanus), of which

bruxellensis and anomolus have teleomorphs in the genus Dekkera.17 However, due to the

difficulties associated with characterising yeasts into either the Brettanomyces or Dekkera

genera on the basis of sporulation,20 these two names are often used interchangeably.11 The

current preference is for Dekkera on the basis of molecular identification.21

While there have been numerous wine-related studies focusing on different strains and

species of Brettanomyces and Dekkera, recent attempts to characterise grape, wine and

winery isolates have failed to identify a Dekkera species other than D. bruxellensis.22-27

Also, early research may have used outdated yeast classifications describing species that

have since been re-classified. As such, reference to previous literature will reflect the

author’s original classifications, but wine related instances of these yeasts will be referred

to as D. bruxellensis.

Many studies into Dekkera and Brettanomyces yeasts have shown great variety between

strains within the same species for easily observable characteristics including growth,

nutritional requirements and metabolic output,23-24, 28-36 but the development of genetic

characterisation has allowed for a more in depth study of these yeasts. Conterno et al.

obtained 47 wine strains of B. bruxellensis from around the world, of which 35 isolates

were studied in great detail. While no two strains displayed the same characteristics in


3

terms of growth, temperature dependence, ethanol tolerance, sulfite tolerance and

metabolic output, the genetic characteristics of these strains could link them with their

geographic origin, vintage year and wine variety.23

With respect to Australian winemaking, Curtin et al. characterised 244 D. bruxellensis

isolates from wine making regions across the country and showed that all isolates could be

placed into closely related genetic groups. One group dominated, accounting for 85% of all

the isolates, with another two genetic groups contributing 6% and 7% respectively.22 The

second two groups, while less represented in Australia, were shown to be closely related to

reference isolates from France and California, which indicates that these secondary D.

bruxellensis groups in Australia may be representative of internationally isolated strains.

While Dekkera strains can be classified into distinct genetic clusters, it is the more generic

characteristics that lead to yeast being classified into the Brettanomyces or Dekkera genera.

Prior to the work of Renouf and Lovaud-Funel,37 the origin of these yeasts in the winery

was not completely understood, but through the development of an enrichment media

specifically for growth of B. bruxellensis, detection on grape berries and the vineyard

origin was inferred. These wild yeasts encounter the grapes by becoming airborne and

settling on them in the vineyard, or can be spread by fruit fly and bees with traces of these

yeasts having being found in the feeding and breeding areas of both insects, as well as on

their legs, bodies and wings.15 Once present in the winery, D. bruxellensis can become

established in any area with which the affected wine comes in contact.11, 15

Even though D. bruxellensis can be found throughout the winery, and has been isolated as

early in the winemaking process as the completion of alcoholic fermentation,33 malolactic

fermentation is an important period in the development of D. bruxellensis due to the low

sulphur dioxide levels needed for growth of lactic acid bacteria, as well as the residual

sugar still remaining in the wine.38 The ability to grow and thrive with few nutrients,39-40

along with the increased tolerance for high ethanol concentrations make the later stages of

winemaking ideal for D. bruxellensis, with little competition from other wine

microorganisms.20 This allows for survival throughout vinification until conditions are

more encouraging for growth,40 which is why the most common place of Dekkera infection

is in the barrel, during wine ageing.15, 41-42 D. bruxellensis can live solely off of cellobiose,

a product of cellulose degradation, allowing it to remain in empty barrels and contaminate


4

subsequent wine additions.11, 15, 23 Once a barrel has become contaminated, sterilisation can

be attempted through shaving, toasting, steaming or hot water treatments,15, 43 though there

is no guaranteed method by which to eradicate Dekkera, with prevention being the

recommended course of action.43

While re-use of barrels can increase the chance of further infection by D. bruxellensis,44-45

the use of new barrels can provide the yeast with additional sugar and oxygen in which to

establish themselves,20 and techniques such as micro-oxygenation which is used to

accelerate wine ageing can provide a more favourable environment for growth.44-45 This is

not to say that lack of oxygen is a great inhibitor, Dekkera can adapt to conditions of low

oxygen, with anaerobiosis only restricting growth and not preventing it.44, 46

In the suspended volume of wine, sulphur dioxide has proven relatively successful in

controlling D. bruxellensis growth.47 Along with the use of barrels during red wine

production,15, 42 the increased sulphur dioxide efficacy at the lower pH of white wine is

probably the main reason why this yeast is more commonly found in red wines.20

However, due to the ability of wine to penetrate deep into the barrel, Dekkera can be

carried deep into cracks, between staves, and around the bung, proliferating away from the

dissolved sulphur dioxide.14 Furthermore, the use of sulphur dioxide may induce a viable

but non-culturable state (VBNC) whereby the yeast is no longer active but can become

viable again given favourable conditions. In a VBNC state Dekkera cells can shrink from

an average 5-8 x 3-4 µm to be small enough to pass through a 0.45 µm filter and then

proliferate in the ‘filtered’ wine,48-50 with the potential to become the major organism in

bottled wine.51 Alternatively trialled methods for controlling Dekkera include

dimethyldicarbonate,52-53 sorbic acid,54 increased temperatures,55-56 low-voltage electric

current,57 ozone,38 and high-power ultra sonic radiation.58

Furthermore, detection of D. bruxellensis in wine has proven to be difficult due to its

comparatively slow growing nature and limited carbon dioxide production.11 Agar plate

smears can be used for the detection of yeasts in winemaking, but D. bruxellensis is often

overgrown by other yeasts that are present and missed, or can even develop long after the

agar plate has been disposed of.15 Current techniques for detection revolve around

additions of compounds inhibitory to the growth of other organisms, allowing Dekkera to


5

be the sole occupant and more easily identified,59 in a similar manner to that previously

described for detection on grapes.37

If able to grow in wine, D. bruxellensis is associated with formation of several “spoilage”

compounds. It has been connected with the production of acetic acid in wines,60 and

though this effect is lessened with decreasing amounts of oxygen, it can still be produced

under full anaerobiosis.44-46, 60 D. bruxellensis has also been directly linked with the

production of ‘mousy’ aromas in wine from tetrahydropyridines,12, 29 the formation of

isovaleric acid,15 which has been described as rancid or cheesy,43 and the production of

volatile phenols.61-62

1.3 Volatile Phenols.

Of the spoilage compounds produced by D. bruxellensis, those of greatest interest,

especially in red wine, are 4-ethylphenol (1) and 4-ethylguaiacol (2). These compounds,

their presence in wine, and their link to D. bruxellensis has been extensively researched

over the past 25 years,15, 41-43, 61-67 and they are produced almost exclusively by Dekkera

under oenological conditions.61-62 Only trace amounts of 4-ethylphenol and 4-ethylguaiacol

have been identified in grape musts, with very little present at the conclusion of malolactic

fermentation. The resulting wines can have much higher concentrations of ethylphenols,

with the greatest increase usually occurring during barrel ageing,41 where Dekkera

proliferates.

OH OH

OCH3

1 2

Figure 1.1: Ethylphenols produced by D. bruxellensis in red wine.

Phenols, 1 and 2, are formed via the activity of two enzymes that are active towards the

hydroxycinnamic acids, p-coumaric acid (3) and ferulic acid (4). The first of these

activities, a decarboxylase, converts compounds 3 or 4 into 4-vinylphenol or 4-


6

vinylguaiacol (5 or 6), respectively, by removing the carboxylic acid moiety and releasing

carbon dioxide.68 The second activity is a vinyl reductase which acts by reducing the C-C

double bond, generating ethylphenols (1 or 2).68-69

OH OH OH

COOHDecarboxylase Vinyl Reductase

3 R = H4 R = OCH3

1 R = H2 R = OCH3

5 R = H6 R = OCH3

R R R

Figure 1.2: Enzymatic conversion of hydroxycinnamic acids to volatile phenols.

In studies of B. bruxellensis and B. anamolus, decarboxylase activity towards caffeic acid

as well as p-coumaric and ferulic acids (3 and 4) has been shown.68, 70 However, it has not

been until recently that the metabolite of caffeic acid, 4-ethylcatechol, has been quantified

in wine.71 While caffeic acid concentrations in wine can exceed that of p-coumaric and

ferulic acid,28, 72-73 4-ethylcatechol concentrations are much lower than 4-ethylphenol and

are closer to that of 4-ethylguaiacol.71 The sensory threshold of 4-ethylcatechol is yet to be

adequately determined in wine. Initial reports suggest a detection threshold around 50

µg/L,74 though unpublished investigations by the Australian Wine Research Institute

(AWRI) suggest that it is much higher. A recent study into the detection thresholds in cider

also support a significantly increased threshold when compared to that of 4-ethylphenol,75

implying that 4-ethylcatechol is of little importance in the production of volatile phenol

off-flavour in wine.

Other microbes present during winemaking possess the necessary enzymatic abilities to

breakdown p-coumaric and ferulic acids and can produce varying amounts of volatile

phenols. However, unlike Dekkera, the activity of the Saccharomyces cerevisiae

decarboxylase is inhibited by the presence of polyphenols in red wines.41, 76 As such, S.

cerevisiae is able to perform decarboxylation and contribute to the accumulation of

vinylphenols in white wine alone,76 although does not possess the ability to subsequently

reduce the vinylphenols to the ethyl analogues.77-78


7

This is also the case for lactic acid bacteria. For many that possess the ability to

decarboxylate hydroxycinnamic acids and produce vinylphenols, the activity is inhibited

by the presence of polyphenols, though in situations where decarboxylation can be

performed, subsequent formation of ethylphenols is hindered by limited vinyl reductase

activity.61-62 Specifically, Oenococcus oeni, the organism largely responsible for malolactic

fermentation, displays limited decarboxylase activity even when uninhibited.61, 79 For those

that do possess decarboxylase activity it has been found that it needs to be induced,

whereby the bacteria need to grow in the presence of the hydroxycinnamic acids for it to be

activated.80

Chatonnet et al. investigated numerous bacteria from the genera Leuconostoc,

Lactobacillus, Pediococcus and Acetobacter, and yeasts Candida, Hanseniaspora,

Metchnikovia, Pichia, Hansenula, Kluyveromyces, Torulaspora and Zigosaccharomyces

with respect to production of volatile phenols.41 Other non-wine related microorganisms

that have been studied include Lactobacillus plantarum, Lactobacillus hilgardii,

Lactobacillus brevis, Pediococcus pentosaceus, Pediococcus damnosus,61 Klebsiella

oxytoca,81-82 Erwinia uredovora,82 Aerobacter aerogens,83 Cladosporium phlei,84

Polyporus circinata,85 Bacillus subtilis86 and Pichia guilliermondii.87-88

Of these Lactobacillus plantarum can produce ethylphenols in synthetic media (2.55%

conversion, compared with 68.6% by D. bruxellensis), but like most other organisms is

inhibited by the presence of polyphenolic compounds.61 P. guilliermondii can also produce

ethylphenols but only in grape juice, and as such can only contribute to volatile phenol

accumulation prior to alcoholic fermentation.88 Furthermore, this organism has only been

associated with wine prior to alcoholic fermentation, having been isolated from grapes,

grape juice and winery equipment, but not from wine itself.87 Other than L. plantarum and

P. guilliermondii, the remaining organisms were either studied with respect to

decarboxylase activity and the ability to produce vinylphenols, or did not possess the

necessary enzymatic abilities to produce ethylphenols.

Thus, while winemaking micro-organisms other than D. bruxellensis can possess the

enzymatic abilities to breakdown the hydroxycinnamic acids, they do not produce

ethylphenols in the quantities seen by D. bruxellensis in wine,61-62 due to either inhibition

of the necessary activities by polyphenolic compounds present in red wine or poor survival


8

under oenological conditions. Organisms that do possess the ability to produce volatile

phenols are not present during barrel ageing where the majority of the spoilage occurs.41

Studies into the sensory impact of the volatile phenols found that 4-ethylphenol has a

detection threshold of 605 µg/L and a rejection threshold of 620 µg/L in a red wine, while

4-ethylguaiacol was detected at 110 µg/L with the wine being rejected at 140 µg/L.41

However, the amounts of these compounds found in wine differs greatly depending on

wine variety, with ratios of 4-ethylphenol:4-ethylguaiacol in Australian red wine varying

from 10:1 in Cabernet Sauvignon to 3.5:1 in Pinot Noir and with an average ratio of

approximately 8:1,63 which is said to be a result of the relative amounts of precursors

present in the grape.41 Also, differences in yeast nutrients, winemaking practices, D.

bruxellensis strains, temperature and use of oak can contribute to altering the ratios and

concentration of ethylphenols in finished wine.43

At an 4-ethylphenol:4-ethylguaiacol ratio of 10:1, which has become known as the

Bordeaux ratio, due to first being determined in Bordeaux red wine by Chatonnet and Pons

(1988, cited in Romano et al. 2009),89 a combined 4-ethylphenol and 4-ethylguaiacol

detection threshold of 369 µg/L and a rejection threshold of 426 µg/L was determined.41 A

study of the ethylphenol concentration of Australian red wines reported a combined

concentration of ethylphenols in excess of 2500 µg/L in three wines, and across the entire

survey an average concentration of 795 µg/L for 4-ethylphenol and 99 µg/L for 4-

ethylguaiacol, with approximately 60% of the wines possessing combined concentrations

in excess of 426 µg/L.63 The threshold of these compounds relies heavily on the wine

variety in which they are found, with a lighter wine being spoiled at a much lower

threshold than a full-bodied wine.43 The descriptors used for each compound are spicy,

phenolic, medicinal, wet horse, woody and smoky for 4-ethylphenol, and smoky or clove-

like for 4-ethylguaiacol.62, 66, 90

A more recent study into the sensory properties of the ethylphenols has linked both

isovaleric acid and isobutyric acid with masking effects.89 This study indicates that the

presence of these two acids can increase the detection thresholds of the ethylphenols by as

much as four times. This would have a similar effect as the wine style, whereby a wine

with higher isovaleric and isobutyric acid concentrations could contain slightly more


9

ethylphenols than a wine with lower concentrations of these acids and exhibit the same

sensory properties.

While the ethylphenols are the main contributors to Dekkera related spoilage in red

wines,31 the vinylphenols, while less thoroughly researched, are of greater importance in

white wines.76 Their relative scarcity in red wine is brought about by the efficacy of the

vinyl reductase that D. bruxellensis possesses, converting most to give the ethyl

analogues,41 as well as the potential incorporation of vinylphenols into

pyranoanthocyanins.91 As mentioned previously, the vinylphenols in white wines are due

to the decarboxylase ability of S. cerevisiae when not impeded by polyphenolics, and when

these vinylphenols are present in wine, a 1:1 mixture (4-vinylphenol:4-vinylguaiacol)

imparts pharmaceutical or phenolic nuances at concentrations above 770 µg/L.76

The production of volatile phenols in wine is proportional to the Dekkera population,41

with 4-ethylphenol able to be used as a marker for growth,11 while yeast growth is

proportional to the sugar concentration. Sugar at a level of 300 mg/L allows for

proliferation of 1 x 103 cells/mL, enough to yield 600 µg/L of 4-ethylphenol. In wine that

has completed malolactic fermentation, up to 1 g/L or more of residual sugar can be

found.61

Research towards the removal of ethylphenols has shown promise in lowering

concentrations, with reverse-osmosis found to reduce the concentrations of 4-ethyphenol

and 4-ethylguaiacol in wine, but the reduction in the volatile phenols was matched with

losses to other desirable aroma compounds.92 Experimentation with lyophilised yeast as an

adsorbent for 4-ethylphenol also resulted in reductions of desirable compounds, in this case

the loss of anthocyanins produced a reduction in wine colour.93

With current efforts at ethylphenol removal resulting in concomitant reductions in wine

quality, one effective way of avoiding spoilage, apart from limiting Dekkera growth, is to

minimise the initial concentration of precursors in the wine.64 As such, the role of

hydroxycinnamic acids in the production of volatile phenols should come under further

scrutiny.


10

There have been conflicting reports as to the presence of p-coumaric and ferulic acids in

grapes with some studies identifying these compounds in grapes, juice or must,73, 94-97

while others have failed to do so.72, 98-99 This could in part be due to insufficient extraction

from the berry, as those that could identify them, found large concentrations in the skin.94-

96 Reported concentrations in juice or must generally range from not present, or not

detected, to around 0.2 mg/L73, 97 which is seen to increase throughout vinification with

changes observed due to skin contact, alcoholic fermentation, malolactic fermentation and

ageing or storage.97, 99-105 A comprehensive study of 547 red wines from multiple countries

and wine regions found p-coumaric acid in concentrations ranging from not detected

through to 6.7 mg/L,106 and although this study did not quantify ferulic acid, it is often

present in lower concentrations than p-coumaric acid,72-73, 101, 103 hence lower 4-

ethylguaiacol concentrations following breakdown by D. bruxellensis. The increase of free

hydroxycinnamic acids during vinification is largely attributed to the release from

conjugated forms,99-100 with the hydroxycinnamic acids having been found as tartaric acid

esters, glucose esters, glucosides, ethyl esters, bound to anthocyanins, or in combinations

of these.107-108 The hydroxycinnamic acids are known to possess antimicrobial

properties109-110 and can be stored in the grape in an inert form until needed to fight off

unwanted organisms, also they may be conjugated to assist in both solubility and

transport.111 Therefore, before p-coumaric and ferulic acids can be decarboxylated and

reduced, yielding ethylphenols, they must be freed from the conjugated forms in which

they are found in grapes.

A common oenological technique is to employ commercial enzyme preparations during

maceration to aid in the release of phenolic compounds from the grape berries. Those

containing undesired cinnamoyl esterase capabilities can be effective at hydrolysing bound

forms, which then leaves the free hydroxycinnamic acids available for conversion to the

associated volatile phenols. As such, enzyme preparations that do possess cinnamoyl

esterase activity are not recommended for use in winemaking as they can increase the

chance of spoilage.64, 112 Furthermore, Dekkera has been shown to be active in the

formation and degradation of ethyl esters,113 and in the release of aglycones from

glycosidically bound forms.114 However, the study of the direct volatile phenol production

from bound hydroxycinnamates by Dekkera has not been studied.


11

1.3 Introduction to Tartrate Esters.

The first report of L-tartaric acid conjugates of the hydroxycinnamic acids in grapes was by

Ribereau-Gayon in 1965, from paper chromatography of black grape skin extracts and of

red wine, identifying the caffeoyl tartaric acid ester along with the p-coumaroyl tartaric

acid ester (7) and feruloyl tartaric acid ester (8). This study failed to identify the quinic

acid esters of the hydroxycinnamic acids, which is the form in which they exist in many

other plants, and were believed to exist in grapes.115 Since this discovery, much research

has been done on the tartaric acid esters of hydroxycinnamic acids, including the discovery

of them in the whole grape berry and not just the skin,116 and showing that they are largely

found in the juice of the grape.117 These esters are also the main phenolic constituent of

fresh grape juice and it has been confirmed that they exist mainly in their trans-form, with

the cis-isomer being present at lower levels.118 The hydroxycinnamoyl tartrates have no

odour, but can add to the taste and astringency of wine, possessing a bitter taste above a

concentration of 10 ppm.119

O O

OH

HOOC

COOH

OH

O O

OH

HOOC

COOH

OH

OCH3

7 8

Figure 1.3: L-Tartaric acid esters of p-coumaric acid (7) and ferulic acid (8).

The highest concentration of hydroxycinnamoyl tartrates are found in immature berries and

they decrease as the berry ripens,94-96, 120-122 which is accentuated by an increase in berry

volume, though in most cases the weight of compound per berry also shows a decline as

the berry matures.95, 121 One study observed an initial drop during ripening followed by a

slow re-accumulation to the original amounts as the berries matured.122

The identification and quantification of the tartrate esters in grapes,72, 94-96, 122-125 skin,98, 126-

127 juice73, 120, 127 and must97, 128-129 has been performed across both red and white grape


12

varieties. The literature data has been collected from different grape varieties from around

the world, and the concentrations determined using a number of analytical techniques

which quantify varied parts of the grape berry or during different stages of winemaking. As

such, a summary of the tartrate ester content in grapes and wine would, at best, only be an

average of many different analyses which possess significant variation. Also, feruloyl

tartrate can be present at low enough concentrations that some analytical methods, if it can

be detected, may give unreliable results.128 The table below shows the results achieved by

Ong and Nagel from analysing different grape varieties for the presence of p-coumaroyl

and feruloyl tartrate in the berry. 122

Table 1.1: Hydroxycinnamoyl tartrate ester concentrations in different grape varieties.122

The tartrate ester content continues to change during winemaking due to the effects of skin

contact,97, 102, 129 alcoholic fermentation,97, 99, 101-102, 104, 129 malolactic fermentation,97, 101-103

and storage or ageing.99-100, 102, 104, 129-130 Some studies have focused simply on the

difference between either grape berry, juice or must concentration, and that of finished

wine, with significant reductions in concentration noted.72-73, 128

In separate studies Gil-Munoz et al.,102 Somers et al.,99 and Nagel and Wulf129 monitored

the changes in p-coumaroyl tartrate throughout fermentations of Monastrell, Chardonnay,

and Cabernet Sauvignon and Merlot wines, respectively. All authors reported an initial

increase in concentration, peaking during alcoholic fermentation, followed by a large

decrease throughout either malolactic fermentation and storage for the red varieties or just

storage for Chardonnay (3-5 fold reductions). Any reductions witnessed in the early stages

NOTE: This table is included on page 12 of the print copy of the thesis held in the University of Adelaide Library.


13

of winemaking have been associated with enzymatic cleavage of the tartrate esters

resulting in liberation of the free hydroxycinnamic acids.99

Observing malolactic fermentation alone (Table 1.2), Hernandez et al. employed different

lactic acid bacteria as well as studying spontaneous malolactic fermentation reporting

increases and decreases in p-coumaroyl tartrate concentration, depending on the

conditions.103 From grapes that were crushed at ambient temperature, Gil-Munoz et al.

witnessed close to a 3-fold reduction of p-coumaroyl tartrate concentration during

malolactic fermentation of Monastrell wine, compared with an increase in wine made with

grapes crushed at 10 oC.102

Table 1.2: Changes in p-coumaroyl tartrate concentration during malolactic

fermentation.102-103

Initial (mg/L) Final (mg/L) Variety Details Reference

8.68 Spontaneous17.98 O. oeni-1813.84 O. oeni-1598.98 L. plantarum-5112.77 L. plantarum-39

397 148 Normal temp.163 255 Low temp.

Tempranillo13.75

Monastrell

Hernandez et al. (2007)

Gil-Munoz et al. (1999)

Ageing or storage of wines, both red and white, results in a reduction of the tartrate ester

concentration over different lengths of storage and under a number of conditions (Table

1.3). The bottle ageing of red wine has been studied from 8 months to 26 months, with

large reductions seen in Monastrell wine over shorter storage times,130 small reductions

observed in Cabernet Sauvignon and Merlot over longer periods,129 and differences

between wine varieties with Monagas reporting a dramatic loss of p-coumaroyl tartrate in

Tempranillo, but only slight fluctuations in Graciano and Cabernet Sauvignon.100 In the

study of Gil-Munoz mentioned above, a 2-fold reduction through 210 days of ageing is

described.


14

Table 1.3: Changes in p-coumaroyl tartrate concentration during wine storage.99-100, 102, 104,

129-130

Initial Final Variety Storage Length Reference

1.9 0.111.9 0.132.1 0.151.9 0.17

15 11.8 Cab. Sav. 10.9 6.6 Merlot

0.77 0.14 Tempranillo0.9 0.8 Graciano1.12 0.9 Cab. Sav.

148 88 Monastrell 210 days Gil-Munoz et al. (1999)

0.8 Steel - 94 days1 Oak - 94 days

4.5 3.43.8 3.33.2 3.21.7 1.5

2.6

Bautista-Ortin et al. (2007)

Nagel and Wulf (1979)

Chardonnay

Pinot Blanc 11 months

Somers et al. (1987)

Vrhovsek and Wendelin (1998)

Monagas et al. (2005)

Concentration (mg/L)

Monastrell 8 months

192 days

18.5 months

These results were mimicked in white wine with different vinification treatments of Pinot

Blanc showing only slight reductions of tartrate esters over 11 months of ageing,104 while

Chardonnay after 125 days contained around one-third of the p-coumaroyl tartrate seen

before storage and experienced no loss in feruloyl tartrate.99

Though there are variations between studies regarding the concentration changes, with

different varieties and techniques utilised, the overall observation is for a reduction in the

tartrate ester concentration from grape through to wine, which is mirrored in simpler

studies comparing grape content with wine content alone.72-73, 128 However, with respect to

D. bruxellensis, the interest lies in the hydroxycinnamoyl tartrate concentration in red

wines from the completion of alcoholic fermentation, throughout storage when this yeast

would be in contact with the wine.

Very few studies have quantified feruloyl tartrate in wine as well as p-coumaroyl tartrate.

A study by Nagel et al. monitored the changes in concentration of both compounds from

must to wine in 3 white and 3 red varieties, observing feruloyl tartrate at concentrations of


15

1.4, 1.9 and 1.2 ppm in Cabernet Sauvignon, Merlot and Pinot Noir, respectively,

compared with p-coumaroyl tartrate at 5.2, 3.1 and 4.7 ppm.128

In other studies, p-coumaroyl tartrate has been found in wine that has undergone malolactic

fermentation in concentrations ranging from less than 1 mg/L101 up to greater than 20

mg/L,103 and in wines with different lengths of ageing approximately 1 mg/L100, 131 through

to around 10 mg/L.97, 129 With typical p-coumaric concentrations in red wines ranging from

trace to 6.7 mg/L, as reported by Goldberg et al.,106 the role of the hydroxycinnamoyl

tartaric acid esters in the formation of ethylphenols has the potential to exceed that of the

free acids. However, due to the relative molecular weights of the tartrate esters (7 and 8)

compared with the acids (3 and 4), breakdown of an equal concentration of each substrate

will result in a greater amount of ethylphenols in the case of the acids.

While the decrease in hydroxycinnamoyl tartrate ester concentration during winemaking

has been linked to typical winemaking practices such as fermentation and the yeast

Saccharomyces cerevisiae,97, 99, 101-102, 104, 129, 132 malolactic fermentation and the bacteria

Oenococcus oeni,97, 101-104, 129 as well as wine ageing,99-100, 102, 104-105, 129-130 the use of

commercial enzyme preparations has also been studied.64, 112 The breakdown of the tartrate

esters, followed by fermentation with either S. cerevisiae or B. bruxellensis resulted in an

increase in the volatile phenol concentration, linking the use of enzyme preparations to the

liberation of free hydroxycinnamic acids, which can be further modified by yeast.

There has also been a single study linking the p-coumaroyl tartrate to D. bruxellensis,28 but

this study failed to link any observed losses of the tartrate ester with an increase in volatile

phenol production. Instead, only p-coumaroyl tartrate and p-coumaric acid were quantified

in wine having undergone fermentation with multiple strains of D. bruxellensis, with both

compounds exhibiting inconsistent changes in concentration. In some cases the

concentration of p-coumaric acid in the samples inoculated with D. bruxellensis were

higher than in the uninoculated control, and did not exhibit a corresponding loss in the

tartrate ester, leaving it unclear as to whether p-coumaroyl tartrate could be enzymatically

hydrolysed by D. bruxellensis and subsequently broken down to yield volatile phenols.


16

1.4 Introduction to Glucose Esters.

The presence of glucose esters of hydroxycinnamic acids was first postulated in 1978 by

Ong and Nagel who tentatively identified them during a study on the constituents of grapes

using high pressure liquid chromatography.120 Soon after this discovery, Herrmann and

Reschke analysed a number of different fruits to isolate and determine their constituents,133

including white grapes, from which the authors confirmed the presence of p-coumaroyl

glucose (9) and feruloyl glucose (10) (Figure 1.4). More recently Baderschneider and

Winterhalter isolated both the p-coumaroyl and feruloyl glucose esters from a German

Riesling,108 which was the first reported occurrence of these two compounds in wine. The

first report of the isolation of the p-coumaroyl glucose ester from red wine was in 2004 by

Monagas,107 though no feruloyl glucose was detected.

OO

OH

OO

OH

H3CO

9 10

O O

OH

HOHO

OH

OH

HOHO

OH

Figure 1.4: 1-O-β-D-Glucose esters of p-coumaric acid (9) and ferulic acid (10).

Very little research has been performed with regard to grape and wine concentrations of 9

and 10. Through extraction of Riesling wine and purification, Baderschneider was able to

isolate 24.2 mg of p-coumaroyl glucose and 23.9 mg of feruloyl glucose, although in this

publication there is no mention of the scale of the extraction.108 In a later report the co-

author, Winterhalter, states that the earlier study was performed on 100 litres of wine,134

which would indicate an approximate glucose ester concentration of approximately 0.24

mg/L, without taking into account losses during extraction and purification.

The identification of p-coumaroyl glucose in four monovarietal red wines via liquid-liquid

extraction was described by Monagas et al.131 Two compounds with identical

fragmentation patterns were observed, denoted ‘Hexose ester of trans-p-coumaric acid (1)’


17

and ‘Hexose ester of trans-p-coumaric acid (2)’, which were rationalised by attachment to

glucose via different glucose hydroxyls. Both esters were found in the four wines at similar

concentrations with combined amounts ranging from 0.46 mg/L in Merlot to 0.71 mg/L in

Tempranillo, which were determined after 1.5 months of bottle ageing. An extended study

by the same author investigated the concentration changes in the two p-coumaroyl glucose

ester conjugates through 26 months of ageing in the bottle for three of the four varieties

used above.100 From the figures in their publication, which utilise two x-axes without

specification of which axis the data corresponds with, it is initially unclear as to the

concentration of the glucose esters, although the initial data points correspond with those

described in their previous publication,131 and can therefore be established. In all three

varieties studied, the glucose ester concentrations only fluctuate mildly over 26 months.

Using a similar extraction method, Hernandez et al. studied polyphenolic compounds in

red wine during malolactic fermentation, and again, two p-coumaroyl hexose esters were

observed.103 Initial concentrations in wine were 1.23 and 1.51 mg/L, with increases seen

throughout both spontaneous and inoculated malolactic fermentation with different species

and strains. Final concentrations of p-coumaroyl glucose (1) were between 1.39 and 7.16

mg/L, and 1.95 to 2.63 mg/L for p-coumaroyl glucose (2). While there appears little reason

for such a disproportionate increase, this study infers the potential for either chemical or

enzymatic formation of the glucose esters during vinification.

More recently, results of Boido et al. mirror those detailed above with a solid-phase

extraction resulting in identification of two p-coumaroyl glucose esters in Tannat red wine,

both in near identical concentrations.98 Along with quantification in wine, the content of

the two esters in skin was monitored during grape ripening with concentrations of 1.8 and

1.2 mg/kg 20 days after veraison, rising to 2.4 and 1.4 mg/kg 10 days before harvest and

finally 2.6 and 1.7 mg/kg at harvest.

All of the studies mentioned above used p-coumaric acid to quantify the p-coumaroyl

glucose esters, and employ extraction techniques that can result in loss of analyte with little

means of determining the extraction efficiency due to a lack of a pure standard. However,

it can be concluded that the glucose esters which accumulate during grape ripening,

continue to increase during malolactic fermentation, and are stable throughout wine

ageing.


18

Though there exists very little information on the breakdown of the glucose esters,

potential pathways to degradation during the vinification process include the use of

enzyme preparations containing the appropriate esterases,64, 112 or by the activity of

microbiological esterases, either by those of intended wine microflora or by those of

unwanted microorganisms.135-136 Finally, the acidic environment of the wine could promote

acid-catalysed hydrolysis,135 though the results of Monagas imply that the glucose esters

are stable, or that any hydrolysis is in equilibrium with re-formation.100

To date, no research has been performed on this class of compounds with respect to the

potential for breakdown by D. bruxellensis into volatile phenols.

1.5 Introduction to Ethyl Esters.

While numerous grape derived hydroxycinnamoyl derivatives have been identified, there

also exist those that are a product of the winemaking process. This is the case for ethyl

coumarate (11) which was identified by Somers et al. and assumed to be the product of the

free hydroxycinnamic acid and ethanol during alcoholic fermentation.137

O O

OH

O O

OH

OCH3

11 12

Figure 1.5: Ethyl hydroxycinnamates.

The quantities observed in wine are presumed to be dependent on the original

concentration of hydroxycinnamic acid, and the ester synthesising abilities of the yeast

employed for alcoholic fermentation. Somers et al. reported that ethyl coumarate was not

present in the must of commercial Chardonnay wine, but increased throughout the course

of alcoholic fermentation to finally be observed at 2.7 mg/L at day 31.99 This increase in


19

ethyl coumarate was preceded by an initial increase in p-coumaric acid concentration,

followed by a steady decline as ethyl coumarate was formed.

Compared with ethyl coumarate (11), ethyl ferulate (12) has been less frequently observed

in wine, which is most likely to be a product of the relative quantities present, though it has

previously been identified using both HPLC and GC in red wine,138 and also in Riesling

wine.139-140

Both ethyl coumarate and ethyl ferulate were quantified in 3 red wines and 3 white wines

in 2003 when Sleep developed a method for analysis using the deuterated analogues.141

Both esters were identified in all 6 wines analysed, with ethyl coumarate found at 0.35 to

1.02 mg/L and ethyl ferulate at 0.01 to 0.14 mg/L. Furthermore, the ethyl esters were not

observed in standard ratios, and wines with the highest ethyl ferulate did not necessarily

correspond to those with the highest ethyl coumarate concentrations.

Using the above SIDA method, the evolution of ethyl coumarate was monitored in a Shiraz

wine throughout vinification and ageing (AWRI unpublished results) with a sigmoidal

increase during alcoholic fermentation, which was followed by a slow accumulation during

ageing, peaking at 3600 ppb after 300 days (Figure 1.6).

Figure 1.6: Evolution of ethyl coumarate in Shiraz wine.

Previously described evidence indicated that the activity of alcoholic fermentation has the

ability to break down the tartrate esters to give the free acids, which, as shown above, can

then be esterified again to give the ethyl hydroxycinnamates. However, there has been no


20

evidence as to the ability of D. bruxellensis to then breakdown the ethyl esters to give

volatile phenols.

1.5 Research Aims.

With D. bruxellensis reported to express esterase activites,113 it is surprising that to date no

studies have actively measured the volatile phenol output when fermentation is conducted

in the presence of common and known esters of hydroxycinnamic acids. The

hydroxycinnamoyl tartrate and glucose esters, which are present in the grape berry, survive

throughout the winemaking process. Additionally, the evolution of free acids through

cleavage of esterified forms, and subsequent esterification to yield the ethyl esters results

in a number of hydroxycinnamoyl derivatives that are present during barrel ageing. These

esterified forms could contribute to the accumulation of ethylphenols, but are yet to be

examined.

As such, this study aims to synthesise the aforementioned hydroxycinnamoyl esters (Figure

1.7) and examine the ability of D. bruxellensis to metabolise these compounds and

determine whether they can act as direct precursors to the volatile phenols, or if enzyme

preparations containing cinnamoyl esterase are indeed required for the release of the free

hydroxycinnamic acids before this conversion can take place.

O O

OH

HOOC

COOH

OH

R

O O

OH

R

O O

OH

R

7 R = H8 R = OCH3

9 R = H10 R = OCH3

11 R = H12 R = OCH3

OHO

OH

HOOH

Figure 1.7: Hydroxycinnamoyl tartrate (7 and 8), glucose (9 and 10) and ethyl esters (11

and 12) to be synthesised and used in these studies.


21

As previously mentioned, there exist three significant genetic groups of D. bruxellensis in

Australia, with one of these accounting for 85% of isolates. By including the

hydroxycinnmamoyl esters in fermentation studies with a representative strain (AWRI

1499) from the most common group, the results will be largely representative of what

could be expected from Australian isolates of D. bruxellensis.22 As such, small scale D.

bruxellensis fermentations will be conducted in the presence of each of the esters with the

result determined by analysis for the production of 4-ethylphenol and 4-ethylguaiacol as

previously described by Pollnitz et al.63

Chapter 2: Synthesis of Hydroxycinnamoyl Esters

22

Chapter 2: Synthesis of Hydroxycinnamoyl Esters.

2.1 Synthesis of Hydroxycinnamic Acids and Derivatives.

The synthesis of hydroxycinnamic acids and protected derivatives is a matter of simple

organic transformation. More importantly, the hydroxycinnamate derivative must be

compatible with the conditions required for esterification with glucose or tartaric acid and

subsequent deprotection so that the desired product can be achieved. As such, for use in the

synthesis of glucose and tartrate esters, a variety of protected hydroxycinnamic acid

derivatives were synthesised, as outlined below (Scheme 2.1).

O

O

HO

O

HO

R R

O

O

BnO

R

OH

O

BnO

R

OH

O

HO

R

OH

O

AcO

ROH

O

ClAcO

R

13 R = H 14 R = OCH3

11 R = H (93%)12 R = OCH3 (72%)

19 R = H (83%)20 R = OCH3 (79%)

21 R = H (35%)22 R = OCH3 (65%)

3 R = H (97%)4 R = OCH3 (96%)

15 R = H (93%)16 R = OCH3 (85%)

17 R = H (74%)18 R = OCH3 (69%)

i

ii iii

ivv ii

i) (carbethoxymethylene)triphenylphosphorane. ii) potassium hydroxide. iii) anhydrous potassium

carbonate, benzyl bromide. iv) pyridine, acetic anhydride. v) sodium hydroxide, chloroacetyl

chloride.

Scheme 2.1: Synthesis of hydroxycinnamic acid derivatives.


23

p-Hydroxybenzaldehyde (13) and vanillin (14) underwent Wittig olefination to afford the

ethyl esters (11 and 12) in good yields, initially being achieved at room temperature in the

presence of 2-3 equivalences of the appropriate stabilised ylide over a number of weeks.

Alternatively, microwave assisted synthesis proved more facile as a method of synthesis

and also afforded 12 with only a small excess of ylide required, but could only be

performed on small scales with our system due to the energetic nature of the process, with

reaction volumes limited to 15 mL. An attempted scale-up, though minor, for the synthesis

of 11 proved unsuccessful given the small volume of the vessel and much of the reaction

mixture was lost. As such, the original method at ambient temperature over a number of

weeks was the method of choice.

The synthesis of the 1-O-benzyl hydroxycinmamic acids (15 and 16) was achieved by

benzylation of 11 or 12 to give 17 or 18 in 74 and 69% yield respectively, followed by

base-catalysed hydrolysis to furnish 15 and 16 in 93 and 85% yield. Previously, the

benzylation of caffeic acid was described by Galland et al. and was originally achieved via

benzylation of the free acid to give the benzylic ester as well as the ether, followed by

hydrolysis of the ester to yield the di-phenolic protected analogue.142 However, in this case,

the production of 11 and 12 allowed for benzylation of the phenol alone.

Unlike in the preparation of 15 and 16, the phenolic protecting groups with ester

functionality were not installed directly onto 11 and 12 due to the potential for removal of

the protecting group during hydrolysis of the ethyl ester. Hence hydrolysis of 11 and 12, in

an analogous method to that used for production of 15 and 16, afforded the free

hydroxycinnamic acids (3 and 4) which were subsequently protected.

1-O-Acetyl protection to furnish 19 and 20 in good yields (83 and 79%) was achieved by

allowing 3 or 4 to react with acetic anhydride in pyridine.143 The successful incorporation

of the acetyl group was confirmed by the inclusion of a 3-proton singlet in each of the 1H

NMR spectra corresponding to the acetyl methyl, and the downfield shift of the ring proton

signals caused by the electron-withdrawing nature of the protecting group. However, when

an analogous process was attempted for preparation of the 1-O-chloroacetyl derivatives (21

and 22), it proved unsuccessful. After 16 hours of reaction the mixture had solidified and

by dissolving in methanol, analysis by TLC showed no desired product. Instead, these

derivatives were prepared using 2M sodium hydroxide solution and chloroacetyl


24

chloride.144 This process had to be optimised as the original procedure employed a large

volume of sodium hydroxide solution which was found to promote degradation of the

chloroacetyl chloride. Therefore it was determined that the hydroxycinnamic acid should

be dissolved in a minimal amount of sodium hydroxide solution, aided with sonication,

followed by addition of chloroacetyl chloride which afforded 21 in 35% yield and 22 in

65% yield. The yields were diminished by the inability to successfully separate 21 or 22

from 3 or 4, as only the initial recrystallisation yielded pure product with subsequent

recrystallisation attempts yielding mixtures of the product and the free acid (3 or 4).

Characterisation of 21 and 22 by 1H NMR showed the inclusion of a 2-proton singlet at

approximately 4.6 ppm in accordance with the literature data.144-145

These aforementioned processes gave four different derivatives for each hydroxycinnamic

acid (free acids 3 and 4, 1-O-benzyl protected 15 and 16, 1-O-acetyl protected 19 and 20

and 1-O-chloroacetyl protected 21 and 22) which would be used in the synthesis of the

target glucose and tartrate esters, as well as 11 and 12 which could be used directly in later

fermentation experiments.

2.2 Synthesis of Hydroxycinnamoyl Tartrate Esters.

2.2.1 Introduction to Tartrate Ester Synthesis

While the existence of tartrate esters in grapes and wines has been widely researched, the

synthesis of tartrate conjugates has received less attention. Synthetically, they are mostly

built as di-conjugates but because of the prevalence of mono-esterified tartrates in grapes it

has become desirable to synthesise mono-esters.146-147 The synthesis of mono-esterified

tartaric acid conjugates has been reported on only a few occasions, but all follow a similar

path. The esterification of dibenzyl tartrate with either the acid in the presence of

trifluoroacetic anhydride, 148-149 or DMAP and DCC,150-151 or with an acid chloride,152 was

followed by removal of the benzyl groups via hydrogenolysis (Scheme 2.2).


25

HO

COOBnHO

COOBnCOOH

OR

OR

DMAP, DCCor

TFAA

HO

COOBnO

COOBn

R = H, Me, Pri

OOR

OR

H2, Pd/C

HO

COOHO

COOHOOR

OR

HO

COOBnHO

COOBn COCl

OMe

OMe

Et3N, DMAP

Scheme 2.2: Literature syntheses of mono-esters of tartaric acid.

The only reported syntheses of hydroxycinnamoyl tartrates are that of chicoric acid, or

dicaffeoyl tartrate (Scheme 2.3). Scarpati and Oriente managed the preparation of chicoric

acid through reaction of tartaric acid and the acid chloride of caffeic acid protected as a

cyclic carbonate (Pathway A),153 and many other methods for esterification directly from

tartaric acid have been utilised.147 However, the polarity of the products would provide

limited methods by which the mono- and di-esters could be separated. In the preparation of

di-esters the issue of purification can be simplified by reacting with an excess of the

hydroxycinnamate and avoiding the formation of the mono-ester, this would not be the

case when attempting to produce mono-esters alone. As such, protection of the tartaric acid

moiety is required to simplify the handling of the esterified products.

Zhao and Burke described the preparation of chicoric acid by reaction of diacetyl caffeoyl

acid chloride and di-tert-butyl tartrate, followed by treatment with trifluoroacetic acid to

remove the tert-butyl groups, and mild acid hydrolysis of the acetyl groups to yield the

desired product (Pathway B).143 However, the synthesis described by Lamidey et al.

employs a synthetically simpler starting tartrate derivative and a single protecting group,

removing the need for a two-step deprotection.154

Esterification of dibenzyl caffeic acid with dibenzyl tartrate in the presence of 1-(3-

dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride and DMAP yielded the

hexabenzyl-protected chicoric acid (Pathway C). Removal of the benzyl ethers and esters

was achieved in a single step via hydrogenolysis in the presence of triethylsilane and


26

triethylamine. While these examples describe the preparation of a di-ester, and the exact

procedures may not be useful in synthesising mono-esters, at the very least they provide

evidence as to the conditions under which the hydroxycinnamoyl esters are, and will be,

stable.

HO

COOBnHO

COOBn

O

OBn

OBn

R'O COOBn

COOBnR'O

R'''OH Et3SiH, Pd(OAc)2, Et3N

HO

COOHHO

COOH

O

O

O

O

R''' =

O

OH

OHR =

R' =

R'ClR''O

R''O

O

O

O

80% Acetic acid

RO COOH

COOHRO

HO

COOButHO

COOBut R'''O COOBut

COOButR'''O

1) TFAA

R''Cl

O

OAc

OAcR'' =

2) HClPathway B)

Pathway A)

Pathway C)

Chicoric acid

Scheme 2.3: Literature syntheses of chicoric acid.

2.2.2 Synthesis of Hydroxycinnamoyl Tartrate Esters

The method by which all of the mono-esters have been created, utilising dibenzyl tartrate,

has also been employed in the production of chicoric acid, and was utilised here in the

synthesis of requisite hydroxycinnamic acid esters (7 and 8). Protection of L-tartaric acid

(23) with benzyl alcohol in the presence of p-toluenesulphonic acid proceeded smoothly in

excellent yields (94%). Esterification of dibenzyl L-tartrate (24) with acids 15 or 16

catalysed by trifluoroacetic anhydride yielded the tri-benzyl protected L-tartrate esters (25

and 26). Furthermore, flash chromatography gave two fractions, one of the pure mono-


27

esters and a second consisting largely of the di-esters (27 and 28) with only minor co-

elution of the mono-esters.

Formation of the mono-esters was confirmed by the unsymmetrical nature of the tartrate

proton shifts and the downfield shift of H2’, while the undesired di-esters were identified

by a symmetrical tartrate as outlined in the literature,143, 154 a downfield shift of the tartrate

H2’ and H3’ shifts from 4.63 ppm observed in 24 to approximately 5.9 ppm corresponding

to formation of the desired ester functionality, and a 1:1 ratio between hydroxycinnamoyl

and tartrate shifts. The di-esters were not fully purified, with 1H NMR data extracted from

the crude mixture and simply compared with that for benzyl protected chicoric acid as

reported by Lamidey,154 to explain the reduced formation of the mono-esters.

OH

O

BnO

R

15 R = H16 R = OCH3

O

O

BnO

R

COOBn

COOBn

OHR'O

BnOOC COOBn

OR'

25 R = H (47%)26 R = OCH3 (50%)

27 R' = 15 28 R' = 16

+

HO

COOHHO

COOH

BnOH/p-TsOH

HO

COOBnHO

COOBn

23

24 (94%)

+

TFAA

Scheme 2.4: Synthesis of benzylated hydroxycinnamoyl tartrate esters.

Removal of the benzyl groups by hydrogenolysis was first attempted as described by

Galland et al. for 1-O-benzyl-caffeoyl glucose,142 using 1,4-cyclohexadiene as the proton

donor, but when attempted on 26 this method resulted in reduction of the double bond in


28

preference to debenzylation to give 29, as determined by the loss of the alkene proton

shifts in the proton spectrum and the appearance of saturated alkane shifts between 2.84

and 2.48 ppm. The characterisation of multiple reduced hydroxycinnamate derivatives

show all signals for the alkane protons observed in the region of 2.5 to 3.1 ppm.155-158 In

addition, minor shifts due to this second tartaric acid derivative were observed, moving

upfield from those corresponding to 26. Furthermore, limited debenzylation was suspected

as the entire reaction mixture was soluble in chloroform-d, which is not expected for the

desired products (7 and 8).

A subsequent attempt using triethylsilane as a more mild hydrogen donor,154 again resulted

in reduction of the double bond, though concomitant debenzylation also occurred affording

30, which was characterised by a further downfield shift of the tartrate proton shifts and the

loss of the signals corresponding to the benzyl moieties. Altering the stoichiometry of the

reactants showed evidence of the desired product, but only in very small quantities, and not

consistently. In conclusion, the use of benzyl protection was an insufficient method for

consistent and reproducible production of the desired hydroxycinnamoyl tartrates, thus an

alternative approach was investigated.

O

O

BnO

H3CO

COOBn

COOBn

OH

1) 1,4-cyclohexadiene, Pd/C2) Et3SiH, Pd/C

O

O

HO

H3CO

COOH

COOH

OH

29 30

+

26

O

O

BnO

H3CO

COOBn

COOBn

OH

Scheme 2.5: Attempted debenzylation procedures.

In a similar fashion to Zhao or Lamidey et al. the synthesis of the tartrate esters was

attempted using di-tert-butyl tartrate (34). However, due to the cost of the starting material

it was synthesised rather than purchased.


29

The use of 34 was described by Uray and Lindner including details of the conversion of di-

O-acetyl tartaric acid (32) to 34 using an undesirable reagent, isobutylene.159 The use of

gaseous isobutylene was avoided by Wright et al., who reported a method for the in-situ

formation of isobutylene from the acid-catalysed dehydration of tert-butanol.160 A review

of tartrate synthesis by Syndoradzki et al. outlined the procedure for synthesis of 32, which

proceeded through di-O-acetyl tartaric anhydride (31).146

O

AcO

AcO

O

O

COOH

COOH

AcO

AcO

COOH

COOH

HO

HO

COOBut

COOBut

AcO

AcO

COOBut

COOBut

HO

HO

SOCl2, t-BuOH

23 31 (82%) 32 (99%) 33 (49%)

34 (29%)

AcCl H2O

KOH

Scheme 2.6: Synthesis of di-tert-butyl tartrate.

Procedures for the preparation of 31 are well known,161-163 though under these conditions,

one study noted the formation of 32 directly. Barros et al. reported the formation of 32

directly, without first yielding 31, and furthermore was characterised in deuterated

chloroform, in which 32 was found to be insoluble, unlike the anhydride.162 Here, 31 was

achieved in 82% yield, but the simplicity of the proton NMR led to ambiguity. The 1H

NMR spectra of 31 and 32 (run in acetone-d6) both display only a shift for the CH protons

at either 6.17 and 5.72 ppm and a shift for the acetyl protons at 2.19 or 2.11 ppm,

respectively, and differ slightly between solvents.161-163 As such, the formation of 31 was

confirmed via determination of the melting point, with the literature value differing by

around 20 oC to that of 32. Preparation of 32 was achieved through stirring with 2

equivalences of water in acetone, though recommendations are for the use of water alone

to ring-open the anhydride,146 it was discovered however, that reacting in an organic

solvent with minimal water lead to a more simple work-up. Concentration followed by

trituration with hexane gave 32 as a white solid in 99% yield after removing excess water

in vacuo.


30

Installation of the tert-butyl ester functionality was successful on small scales using the

method of Wright, but yields diminished as the scale of the reaction was increased.

Nonetheless, this proved superior to another method which involved first producing the di-

acid chloride and reacting with tert-butanol,161-162 which tended to cause dehydration of 32

with concomitant formation of 31. In the preparation of 33, mono-O-acetyl di-tert-butyl

tartrate was produced as a minor product, which could also be used in the subsequent

deacetylation. Initially, the two products were separated for characterisation, though a

subsequent synthesis used a mixture of the two for the following reaction.

Uray describes the use of potassium hydroxide to deacetylate 33, although the amount of

potassium hydroxide required proved less than reported.159 The literature procedure

involves a larger amount of potassium hydroxide and a very quick reaction time, though

the reaction was much more controllable with smaller amounts of potassium hydroxide for

a longer period. Minor amounts of tert-butyl methyl tartrate was isolated during

purification and characterised by 1H NMR, with the data corresponding to that previously

reported.159 Presumably upon addition of potassium hydroxide to the reaction mixture a

portion of the methanol was deprotonated and the resulting methoxide was responsible for

the replacement of a tert-butyl group for a methyl ester.

Synthesis of 35 or 36 was achieved by converting the hydroxycinnamate to the acid

chloride, and then reacting with 34, but this could only be achieved for hydroxycinnamates

with phenolic protection (19-22).

OH

O

AcO

RO

O

AcO

1) SOCl2

19 R = H20 R = OCH3

R

2) Pyridine, 34

35 R = H (31%)36 R = OCH3 (48%)

COOBut

COOBut

OH

Scheme 2.7: Esterification of hydroycinnamic acids and di-tert-butyl tartrate.

Preparation of 35 and 36 was achieved in relatively good yields, with minor formation of

the di-ester, which could be lessened by using a greater excess of 34. The unreacted 34 was

easily recovered during flash chromatography of the reaction mixture, eluting with 10%


31

EtOAc/X4 while 35 and 36, with similar Rf values, eluted with 20-30% EtOAc/X4.

Interestingly, the di-esters displayed higher Rf values by TLC, but eluted after the mono-

esters from flash chromatography, which is presumably an effect caused by the differences

in the silica gel used during flash chromatography and that on the TLC plates. However,

purification with flash chromatography provided adequate separation of 35 and 36 from the

di-esters to obtain pure samples, with the mixtures analysed by 1H NMR to confirm the

presence the di-esters. While the data is not shown, the di-esters exhibited the same

hydroxycinnamoyl 1H shifts as 35 and 36, a symmetrical tartrate lacking a free hydroxyl

and integration of all signals confirmed two hydroxycinnamoyl moieties for every tartrate,

as described by Zhao.143

Recrystallisation of 35 and 36 from ethyl acetate/X4 yielded hygroscopic white crystals

which were submitted to X-ray crystallography (Figures 2.1 and 2.2), which confirm

retention of the (R,R)-stereochemistry during preparation of 34 and esterification to afford

35 and 36.

Figure 2.1: Molecular structure and crystallographic numbering scheme for 35.


32

Figure 2.2: Molecular structure and crystallographic numbering scheme for 36.

The 1H and 13C shifts of the tert-butyl groups within 35 and 36 were assigned using the

shifts of the previously synthesised tert-butyl derivatives. The 1H tert-butyl shifts of 33

appear at approximately 1.44 ppm, whereas they appear at 1.52 ppm in 34. It was assumed

that the tert-butyl group adjacent to the esterified hydroxyl would possess the upfield shift,

which was confirmed with the tert-butyl shifts for the di-esters appearing at 1.44 ppm.

Attempted deprotection of 35 using a previously described method which employed

trifluoroacetic acid to remove the tert-butyl groups followed directly by acid promoted

deacetylation,143 resulted in the desired product (7), which by 1H NMR analysis was shown

to exist in a mixture with other p-coumaroyl tartrate derivatives. Four components were

observed all possessing overlapping signals corresponding to a p-coumaroyl moiety bound

to tartaric acid, which was rationalised by the formation of tartaric acid sodium salts during

work-up. Purification using reverse-phase chromatography was unsuccessful as all

components co-eluted, and in an attempt to protonate the sodium salts the mixture was

taken up in methanol, the pH adjusted to 1 (2M HCl solution) and stirred at room

temperature for 16 hours, which was unsuccessful as shown by 1H NMR analysis.

Furthermore, the use of acetyl protection in the synthesis of hydroxycinnamoyl tartrates

was discontinued when it was discovered that removal of these groups within the glucose

esters (54 and 55, described in a later section), which was performed concurrently, was not

possible while retaining the hydroxycinnamoyl ester linkage.

Utilising a more labile phenolic protection strategy the 1-O-chloroacetyl derivatives, 21

and 22, were esterified with 34, yielding the chloroacetylated products 37 and 38. During


33

purification of 38 with flash chromatography, a second compound (of lower Rf) was

isolated and identified as 40, having undergone phenolic deprotection under the reaction

conditions. To test the efficacy of the dechloroacetylation, the p-coumaroyl analogue (37)

was dissolved in 1:1 pyridine/benzene and stirred at room temperature for 24 hours.

Concentration and separation by flash chromatography, gave an approximate 60:40 ratio of

39:37.

For the following esterification attempts reaction times were increased, which gave 39 and

40 as the major products (19 and 28% yields), with some 37 and 38 remaining (11 and

6%). This reaction was not further optimised, as it gave the desired products in acceptable

yields, and further deprotection of the chloroacetylated compounds under the same

conditions could be achieved.

Recrystallisation of 39 and 40 in an analogous manner to 35 and 36 yielded white solids,

which were not suitable for X-ray crystallography, and though the microanalysis of 40 was

within acceptable limits, the hygroscopic nature of the compounds resulted in incorrect

microanalytical data for 39. Nonetheless, the crystallographic data and optical rotations of

35 and 36 and the microanalysis of 40 achieved under the same conditions, along with the

similar 1H and 13C spectra, optical rotations and expected high resolution accurate mass

determination strongly support the structure of 39 and 40.

Deprotection of 39 and 40 was achieved using trifluoroacetic acid, concentrated and

purified with reversed-phase chromatography.143 The crude product was dissolved in

methanol and loaded onto a pre-packed 4 g C18 cartridge which was washed with

water/formic acid (99:1) and 7 and 8 eluted with water/acetonitrile/formic acid (69:30:1).

Concentration of the fractions containing only pure product yielded 7 and 8 which could be

used directly in fermentation experiments, described in Chapter 5.


34

OH

O

ClAcO

R

O

O

ClAcO

1) SOCl22) Pyridine, 34

21 R = H22 R = OCH3

R

37 R = H (30%)38 R = OCH3 (34%)

COOBut

COOBut

OH

O

O

HO

R

40 R = OCH3 (24%)

COOBut

COOBut

OH

+

O

O

HO

R

7 R = H (82%)8 R = OCH3 (41%)

COOH

COOH

OH

TFA, C18-RP chromatography

37 R = H (11%)38 R = OCH3 (6%)

39 R = H (19%)40 R = OCH3 (28%)

Reacted for 16 hours

Reacted for 45 hours

Scheme 2.8: Synthesis of hydroxycinnamoyl tartrate esters.

2.3 Synthesis of Hydroxycinnamoyl Glucose Esters.

2.3.1 Introduction to Glucose Ester Synthesis

Glycosylation methods have been extensively studied, and also widely reviewed in books

and journal articles,164-170 including specific articles discussing effects that control

glycosylation such as the anomeric effect.171


35

Previous attempts in this research group to synthesise glucose conjugates, have taken

advantage of a modified Koenigs-Knorr reaction. Again, since the original publication

which employed glucopyranosyl bromide, and silver carbonate in the presence of the

alcohol to achieve glycosylation,172 this reaction has been modified and improved, and

these advancements have also been reviewed numerous times.173-175

Glucosylation has been achieved via pivaloyl or acetyl protection of glucose followed by

activation of the glucosyl donor via installation of anomeric bromine (Scheme 2.9).175-177

The glucose ether linkage was created using silver triflate as a catalyst, followed by

removal of glucosyl protecting groups under basic conditions. However, the synthesis of a

glucose ester differs slightly to that of a glucoside due to the ease of hydrolysis of the ester

linkage under the basic conditions required to remove the glucosyl protection.

OR1OR1O

R1OBr

OR1

OR1OR1O

R1O

OR1

OR2

R2OH, 2,6-lutidine

AgOTf, CH2Cl2

R1 = Ac, Piv

Scheme 2.9: Modified Koenigs-Knorr reaction conditions employed within this research

group.

This procedure was employed for glycosylation of hydroxycinnamic acids by the current

author using acetyl glucose protection, which could not be removed with the glucose ester

linkage remaining intact,178 even though Birkofer et al. has reported the use of acetyl

protection on both the glucosyl hydroxyls and on the phenol to achieve hydroxycinnamoyl

glucose esters.179 To ensure consistent and adequate yields for the target esters, a different

protecting group was needed for the glycosyl protection in order to achieve synthesis of

hydroxycinnamoyl glucose esters.

The use of the chloroacetyl group in carbohydrate synthesis was first described by

Bertolini and Glaudemans which provides an alternative to the acetyl group in that the

former can be removed under more mildly nucleophilic conditions due to the decreased

electron density of the carbonyl carbon caused by the electron withdrawing nature of the

chlorine.180 Chloroacetyl protection was utilised by Ziegler and Pantkowski to prepare

cinnamic and hydroxycinnamic acid esters via numerous modified Koenigs-Knorr


36

glycosylations (Scheme 2.10).144 The glycosylation attempts, with either cinnamic acid (i)

or 3,4,5-trimethoxybenzioc acid (ii ), gave 51-58% of a mixture of α- and β-glucose ester

(77-81% α-ester) using a glucosyl fluoride in the presence of boron trifluoride etherate

(Method 1). Purification of the crude mixtures afforded 39% of the α-anomer and 27% of

the α-anomer for reactions i and ii , respectively. Using chloroacetyl protected

galactopyranosyl bromide and the silver carboxylate yielded 70% of the β-galactose ester

(Method 2). However, improving on either selectivity for the β-anomer (over method 1) or

simplicity of reaction (eliminating light sensitive reagents in method 2), the use of a

glucopyranosyl trichloroacetimidate in the presence of cinnamic acid and trimethylsilyl

trifluoromethanesulfonate (TMSOTf) resulted in a 75% yield with a ratio of 92:8 β:α-

glucose ester (Method 3).

O

OAcCl

ClAcOClAcO

ClAcO O CCl3

NH

O

OAcCl

ClAcOClAcO

ClAcOF

O

ClAcO

BrClAcO

ClAcO

OAcCl

O

OAcCl

ClAcOClAcO

ClAcO

O

OAcCl

ClAcOClAcO

ClAcO

O

ClAcO

OC(O)RClAcO

ClAcO

OAcCl

O

OAcCl

ClAcOClAcO

ClAcO

O

OAcCl

ClAcOClAcO

ClAcO

O

ClAcO

ClAcO

ClAcO

OAcCl

OC(O)R

OC(O)R

OC(O)R

OC(O)R

OC(O)R

RCO2H

RCO2H

RCO2Ag

75% 8:92 (α:β)

i) R = cinnamic acid, 58% crude, 81:19 (α:β)ii) R = 3,4,5-trimethoxybenzoic acid, 51% crude, 77:23 (α:β)

70% 0:100 (α:β)

Method 1)

Method 2)

Method 3)

Scheme 2.10: Glycosylation reactions of Ziegler.

The trichloroacetimidate glycosylation method was developed by Schmidt and involved

preparation from trichloroacetonitrile under basic conditions.181 This method has been

explored in great detail appearing as book chapters authored or co-authored by Schmidt.168-

169 The trichloroacetimidate method, along with chloroacetyl protection was implemented

by Galland to synthesise caffeoyl glucose esters which were used for co-pigmentation

studies in wines (Scheme 2.11).142


37

Sieves, AgOTf

HO O

OBn

OBn

O

OAcCl

ClAcOClAcO

ClAcO

O

O

OBn

OBn

1) Pyridine, water

2) 1,4-cyclohexadiene, Pd/C

O

OH

HOHO

OH

O

O

OH

OH

O

OAcCl

ClAcOClAcO

ClAcOO CCl3

NH

Scheme 2.11: Glycosylation method described by Galland.

While Galland employed silver triflate to catalyse the glycosylation, Ziegler utilised

trimethylsilyl triflate, as is outlined by numerous examples in Preparative Carbohydrate

Chemistry,169 which removes the need for the reaction to be carried out in the dark. This

minor change aside, preparation on the glucosyl donor via chloroacetylation, anomeric

deprotection and activation through preparation of trichloroacetimidate via previously

described methods,142, 144, 180-181 along with the phenolic protection strategies of Galland to

give the protected glucose esters, deprotection was expected to yield p-coumaroyl and

feruloyl glucose esters (9 and 10).

2.3.2 Synthesis of Hydroxycinnamoyl Glucose Esters

Formation of 41 was achieved by addition of chloroacetylchloride (46) to D-glucose

yielding an approximate 55:45 mixture of α-:β-41 after purification and in excellent yields

(91%). Subsequent preparation of 42 gave some residual unreacted 41 solely as the α-

anomer, from which it could be seen by 1H NMR that the shifts for the chloroacetyl groups

are not singlets, as previously reported,142, 182 but distorted AB quartets. This effect is

displayed as a large central signal representing the overlapping inner lines, with very minor

outer lines which when obscured leaves only the single signal corresponding to the inner

lines.


38

One chloroacetyl group in α-41 showed separated inner lines, appearing as two singlets, as

the satellite peaks were unresolved from other shifts and the coupling in the distorted

doublets cannot be determined. This effect is difficult to observe in a spectrum of both

anomers, possibly why the previous reports give a two proton singlet for each chloroacetyl

group.142, 182 In the experimental section (Chapter 6) the chloroacetyl shifts are referred to

as apparent singlets (app. s) as they are technically highly distorted pairs of doublets where

the coupling cannot be sufficiently determined.

In the purification of 41 or 42 both anomers largely co-eluted on column chromatography,

and appeared as one spot by TLC. However, the α-anomer did elute slightly before the β-

anomer, meaning that discarded fractions containing one anomer and an impurity could

alter the anomeric ratio which, if important to the synthesis, should be determined prior to

the final purification.

Preparation of 42 gave a 55% yield with an anomeric ratio of 70:30 (α:β) as expected,182

but subsequent synthesis of 43 did not give the pure α-anomer as was reported by

Galland.142 Instead 43 was isolated in a 2:1 ratio (α:β), supporting the findings of Ziegler,

having isolated a 2.6:1 mixture of α:β-anomer.144 The anomers of 43 possess different Rf

values by both TLC and flash chromatography, and as such the β-anomer, eluting second,

could potentially be mistaken for a by-product and discarded. Furthermore, 43 undergoes

hydrolysis, as shown by Skouroumounis,174 and it is was found that the β-anomer was

hydrolysed preferentially to the α-anomer, with the hydrolysis product eluting between the

α- and β-43. From column chromatography the order of elution is α-43, the hydrolysis

product, then β-43.


39

O

OAcCl

ClAcOClAcO

ClAcO

O

OAcCl

ClAcOClAcO

ClAcO

O

OAcCl

ClAcOClAcO

ClAcO

Hydrazine acetate

Cl3CCN, DBU

41 (91%, 55:45 α:β)42 (55%, 70:30 α:β)

43 (75%, 67:33 α:β)

Sieves, TMSOTf

O

OH

HOHO

HO OHOAcClOH

O CCl3

NH

OHO

R1

O

OAcCl

ClAcOClAcO

ClAcO

O

O

R1

R2

ClAcCl, pyridine

15 R1 = OBn, R2 = H16 R1 = OBn, R2 = OCH347 R1 = H, R2 = H

R2

+

44 R1 = OBn, R2 = H (52%)45 R1 = OBn, R2 = OCH3 (54%)48 R1 = H, R2 = H (54%)

Scheme 2.12: Synthesis of 1-O-benzyl hydroxycinnamoyl glucopyranoses.

The glycosylation procedure was initially trialled with cinnamic acid (47) with the desired

product, 48, isolated in 54% yield and possessing a distinct 1H NMR shift for the anomeric

proton, with a coupling constant of 8.2 Hz supporting production of the β-ester and

matching the literature data.144 While synthesis of 48 proved useful in confirming the

efficacy of the glycosylation technique and determining the stoichiometric conditions, the

lack of phenol functionality meant that it was not useful in confirming the entire synthetic

pathway due to the inability to investigate phenolic deprotection strategies. However,

synthesis of a single anomer of 48 allowed for confirmation of the AB quartet nature of

some chloroacetyl protons (Figure 2.3). The inner lines at 4.03 and 4.04 ppm corresponded

with satellite peaks at 3.99 and 4.08 ppm, giving 4.05 and 4.02 ppm doublets with 14.5 Hz

coupling.


40

Figure 2.3: 1H proton NMR spectrum of the chloroacetyl protons in 2,3,4,6-O-

tetrachloroacetyl-β-D-glucopyranosyl cinnamate (48).

Synthesis of 44 and 45 gave very similar yields to the cinnamate (48), and the formation of

the β-esters was confirmed by the shift of the anomeric proton signal from 6.60 ppm (3.7

Hz) and 5.94 ppm (7.6 Hz) observed for 43, to a single signal at 5.91 ppm possessing an

8.2 Hz coupling constant corresponding to the 1,2-trans conformation of the sugar.

However, debenzylation in the presence of the α,β-double bond was investigated

concurrently during the synthesis of the hydroxycinnamoyl tartrate esters (7 and 8,

described above) and due to those findings 44 and 45 were not used further in the

production of the glucose esters.

Glucosylation using 3 and 4 gave inadequate yields of 49 and 50 which co-eluted with a

sugar impurity from column chromatography. In one instance 50 was prepared as a pure

compound, though the outcome of the reactions were inconsistent with 49 only being

achieved as a mixture of compounds identified based on the 1H NMR spectra, of which the

data for 50 matched that later reported by Zhu and Ralph.183

Regardless, impure 49, as well as 48 and 50, were deprotected followed by attempted

purification on XAD-2, which proved insufficient. Although multiple purification attempts


41

using XAD-8 resin provided separation of the desired products and the 6-O-chloroacetyl-

β-D-glucopyranose derivatives (51 and 52) which were identified by the 1H NMR

characterisation of 51 alone.

43, Sieves, TMSOTf

OHO

OH

R

O

OAcCl

ClAcOClAcO

ClAcO

O

O

OH

R

3 R = H4 R = OCH3

49 R = H (29% crude)50 R = OCH3 (39%)

Pyridine, water

O

OH

HOHO

OH

O

O

OH

R

O

OH

HOHO

OH

O

O

OH

R

9 R = H10 R = OCH3

cis-9 R = H (33% both isomers)cis-10 R = OCH3 (20% both isomers)

XAD-8

O

OAcCl

HOHO

OH

O

O

OH

R

51 R = H (25%)52 R = OCH3

+

O

OH

HOHO

OH

O

O

OH

R

trans-9 R = Htrans-10 R = OCH3

+

Scheme 2.13: Synthesis of glucose esters with free hydroxycinnamic acids.

Compounds 9, 10 and 53 were isolated with minor impurities where the excess signals

were attributed to only a single compound as indicated by 1H NMR integration. By

identifying the distinct ring coupling present in 10, the presence of a β-anomeric proton

signal consistent with a glucose ester, and coupling constants of 13.0 Hz for the α,β-

doublets, led to identification of the impurity as cis-10. Upon investigation of the literature,


42

the changes in 1H signals between trans-10 and the impurity were consistent with that

observed between cis- and trans-ferulic acid (4)184 and between cis- and trans-ethyl

ferulate (12)185 supporting the presence of the cis-isomer, which was also found to be the

case for 9 and 53 (Figure 2.4).

From the work of Kahnt,186-187 isomerisation was determined to be a result of exposure to

ultra-violet radiation during purification. However, the nature of ambient light conditions

that 9, 10 and 53 were exposed to indicated that photoisomerisation was not limited to

ultra-violet light and exposure to laboratory lighting would also induce this effect. Isomeric

ratios differed slightly between glucose esters, but were in the vicinity of 4:1 (trans:cis).

Figure 2.4: NMR spectrum of isomerised glucose esters. a) cis/trans-Feruloyl glucose (10).

b) cis/trans-Cinnamoyl glucose (53).

Furthermore, photoisomerisation was limited to 9, 10 and 53, and not experienced during

the synthesis of any other hydroxycinnamate derivative. As a precaution further synthetic

attempts towards 9 and 10 were conducted under red light while further investigation into

the photoisomerisation is outlined in Chapter 4. While 9 and 10 could be synthesised

directly from 3 and 4 without employing phenolic protection, the yields were lower than

desired and unpredictable enough to warrant further investigation into the use of phenolic

protection to develop a reproducible and reliable method of synthesis.


43

The use of 19 and 20 in glucosylation gave good yields of 54 and 55, but removal of the

acetyl group could not be achieved with retention of glucose ester functionality.

Previously, attempted removal of glucosyl acetates resulted in hydrolysis of the

hydroxycinnamoyl glucose linkage,178 though the use of phenolic acetyl protection was

attempted based on the documented successful removal in the presence of a

hydroxycinnamoyl ester.143, 179

Removal of the glycosyl chloroacetyl groups from 54 in 1:1 pyridine/water failed to

remove the phenolic acetyl group, yielding 56 as evidenced by crude 1H NMR analysis.

Attempted purification on XAD-8 resulted in the appearance of another compound with a

lower Rf value than 56. Partial separation by flash chromatography gave a small amount of

the lower Rf compound which was determined to be 9 by 1H NMR, as well as a mixture of

9 and the original 56. While 56 could only be isolated as a mixture, with subsequent

purification attempts removing some of the residual 9 but also giving slight deacetylation

producing additional 9, the 1H NMR could be assigned by ignoring those signals known to

belong to 9. Removal of all chloroacetyl groups was evident by the glucosyl proton signals

all appearing between 3.86 and 3.38 ppm, indicating the deprotected species, but also

possessing a 3-proton signal at 2.29 ppm and ring proton signals slightly downfield from

those observed in 9, which is consistent with the presence of a phenolic acetyl group.


44

43, Sieves, TMSOTf

OHO

OAc

R

O

OAcCl

ClAcOClAcO

ClAcO

O

O

OAc

R

19 R = H20 R = OCH3

54 R = H (40%)55 R = OCH3 (44%)

XAD-8

O

OH

HOHO

OH

O

O

OAc

1:1 Pyridine/water

O

OAcCl

ClAcOClAcO

ClAcO

O

O

OAc

54

56 (20%)

O

OH

HOHO

OH

O

O

OH

+

9 (4%)

Scheme 2.14: Glycosylation with 1-O-acetyl hydroxycinnamic acids and partial

deacetylation using XAD-8 resin.

While XAD-8 gave minor deprotection of the acetyl group, it was not enough to be an

effective method in which to synthesise 9 and 10, with only trace amounts produced.

Alternative methods for deacetylation included stirring with Amberlite IR-120 (H+

form),188 Amberlite IR-4B (-OH form) and catalytic amounts of triethylamine, which failed

to yield 9, with increasing hydrolysis of the glucose ester bond as the methods became less

mild. As such, preparation of 54 (and 55) was not a viable pathway to 9 (and 10), giving

very small yields through accidental deprotection on XAD-8.

The use of 21 and 22 for glycosylation has been attempted by Ziegler using a glucosyl

fluoride in the presence of boron trifluoride etherate, which gave a moderate yield (51%)

and largely the α-glucose ester (Scheme 2.10). Purification of the mixture gave 27% of the

α-glucose ester, though the use of similar glycosyl acceptors and varying glucosyl donors


45

suggests that the preference for production of the α-anomer is a factor of the glucosyl

donor, and the use of 21 with the trichloroacetimidate method might therefore yield largely

the β-anomer, as was seen when used with cinnamic acid. Though the deprotection of this

compound is not reported,144 the phenolic chloroacetyl group was not labile under

glycosylation conditions and was expected to be labile under the same deprotection

conditions used to remove the glucosyl chloroacetates.

Compounds 57 and 58 were prepared in 48 and 64% yield respectively, and were

characterised fully, following which the data reported for the synthesis of 58 by Zhu and

Ralph matches that obtained here.183 Deprotection of 57 and 58 required reaction times that

were 50% longer (6 hours) to remove all five chloroacetyl groups than previously

experienced for the removal of four groups (4 hours, 48-50, 54). Again, a large proportion

of the incomplete reaction mixture existed as the 6-O-chloroacetyl protected species.

Given the large proportion of mono-protected species, and co-elution from XAD-8,

alternative purification attempts included flash chromatography using 1% formic acid/ethyl

acetate, 5% methanol/ethyl acetate and 10% methanol/dichloromethane. The two former

solvent systems resulted in co-elution, the later afforded pure fractions of each, from which

re-reaction of the mono-protected analogues yielded the desired products, 9 and 10. Given

full conversion to 9 and 10, purification by flash chromatography using 5% methanol in

ethyl acetate is a convenient and quick method, but fails to remove the mono-protected

species.

Furthermore, TLC of 9 and 10 in aqueous solvent systems utilised heat to evaporate excess

solvent between applications, which resulted in three spots, which were found to

correspond to the desired product (9 or 10), D-glucose and the aglycone (based on

corresponding Rf values of TLC standards) in varying ratios depending on the amount of

heat applied. The decomposition of the compound on silica by heat could have resulted in

previously purified samples being deemed impure and further purifications attempted. As

such, TLC of the glucose esters should be performed without the use of a heat gun, with

the excess solvent allowed to evaporate at room temperature.

Following an adequate method of purification, a previously pure sample of glucose ester

contained impurities that did not show up by TLC, but as seen by 1H NMR possessed the


46

correct shifts to be a hydroxycinnamoyl glucose derivative with ester functionality. The

ratio of the hydroxycinnamate to glucose shifts remained 1:1, but various shifts that were

consistent with anomeric protons in different environments. After investigation of the

literature, the appearance of the extra proton shifts were determined to be caused by acyl

migration, or movement of the hydroxycinnamate onto different glucose hydroxyls, which

was observed to occur to a greater extent in p-coumaroyl glucose than in feruloyl glucose.

In labelling the migrated structures, the attachment to glucose is designated by the

hydroxyl number (1, 2, 3, 4 or 6) and the orientation of the anomeric hydroxyl designated

either α or β.

43, Sieves, TMSOTf

OHO

OAcCl

R

O

OAcCl

ClAcOClAcO

ClAcO

O

O

OAcCl

R

21 R = H22 R = OCH3

57 R = H (48%)58 R = OCH3 (64%)

Pyridine, water

O

OH

HOHO

OH

O

OH

O

OR2

HOHO

R1OOH

R1 = HCA, R2 = H, 2-O-α-esterR1 = H, R2 = HCA, 6-O-α-ester

9 R = H (43% all esters)10 R = OCH3 (20% all esters)

Migration

R

O

HCA = O

OH

R

Scheme 2.15: Glycosylation of 1-O-chloroacetyl hydroxycinnamates, and migration of the

free glucose esters.


47

The shifts for the migrated esters (Table 2.1) were assigned using data from Brecker who

studied the migration of glucosyl acetate and formate esters,189 and the main esters

produced, apart from the 1-O-β-esters, are the 2-O-α-esters and the 6-O-α-esters.

However, shifts could be assigned for the 1-O-α-ester and the 3-O-α-ester, as well as

minor production of the corresponding 2/3/6-O-β-esters. In addition to altered positions

and coupling constants for the anomeric protons, the shifts for the proton at the point of

attachment appeared further downfield. Coupling constants of 3.7 Hz corresponded to α-

anomeric hydroxyls, with 7.8-8 Hz for β-anomeric hydroxyls, with the nature of the

aglycone (p-coumaroyl or feruloyl) having no impact on the shifts of the glucose protons.

Table 2.1: 1H NMR shifts for migrated hydroxycinnamoyl glucose esters in CD3OD.

Ester H1' (ppm) H2' (ppm) H3' (ppm) H6a' (ppm) H6b' (ppm)

1-O-αααα 6.21 (d, 3.7 Hz)

1-O-ββββ 5.6 (d, 7.8 Hz)

2-O-αααα 5.33 (d, 3.7 Hz) 4.67 (dd, 10.0 and 3.7 Hz)

2-O-ββββ 4.69 (d, 8.0 Hz) 4.77 (dd, 9.6 and 8.0 Hz)

3-O-αααα 5.14 (d, 3.7 Hz) 4.97 (dd)

3-O-ββββ 4.6 (d, 7.8 Hz) 5.04 (dd)

6-O-αααα 5.08 (d, 3.7 Hz) 4.5 (dd, 11.8 and 2.1 Hz) 4.31 (dd, 11.8 and 5.4 Hz)

6-O-ββββ 4.48 (d, 8.0 Hz) 4.45 (dd, 11.8 and 2.2 Hz) 4.29 (dd, 11.8 and 6.0 Hz)

When the migrated mixtures were submitted to wine-like conditions (10% ethanol, pH of

3.5) the 1-O-β-esters predominated, but attempts at isolation of these gave a mixture of

esters, as determined by 1H NMR. The migration of these esters was studied further

(Chapter 3), although a mixture of esters was used in fermentation experiments (Chapter

5). While the desired 1-O-β-esters should proliferate in the fermentation media, the

evolution of 4-ethylphenol and 4-ethylguaiacol would require metabolism of a

hydroxycinnamoyl glucose ester, regardless of its position of attachment to glucose.

2.4 Conclusions.

Along with ethyl esters (11 and 12), which could be used directly in fermentation

experiments, the synthesised hydroxycinnamoyl derivatives (3, 4, 15, 16, 19-22) were used

to explore a new method for successful synthesis of 7 and 8 for the first time, which can


48

now be investigated directly with respect to ethylphenol formation in the presence of D.

bruxellensis.

O

O COOH

COOH

OHHO

RO

O

HO

R

11 R = H12 R = OCH3

7 R = H8 R = OCH3

Figure 2.5: Hydroxycinnamate esters to be used in fermentation experiments.

The synthesis of 9 and 10 was achieved, following which the synthetic methodology was

further confirmed by Zhu and Ralph,183 who published the synthesis of feruloyl glucose

with only slight differences to that detailed above. However, the transformations that were

observed for 9 and 10 have not been reported by any other author, and will be investigated

in greater detail. The speed of photoisomerisation of 9 and 10 during synthesis was not

observed for any other hydroxycinnamoyl derivative, and as such should be investigated to

determine what factors contributed to this phenomenon, with the ultimate aim of achieving

synthesis of pure trans-glucose esters without having to handle them exclusively under red

light. Additionally, if the ratio of cis:trans-9 and 10 that was observed in the laboratory is

seen in the grape berry, the contribution of the cis-isomers to the production of

ethylphenols, or otherwise, could have a large effect on the organoleptic properties of

wine.

O

OR

HOHO

HOOH

O

OH

HOHO

ROOH

O

OH

HOHO

OH

OR

R = trans-p-coumaroyl or trans-feruloyl

Figure 2.6: Dominant equilibria in hydroxycinnamoyl glucose ester mixtures to be used in

fermentation experiments.

The migration of the glucose esters could not be controlled under experimental conditions,

though the 1-O-β-esters were observed to predominate in a wine-like environment. The

potential for migration in wine should be investigated to determine if esters other than the


49

1-O-β-esters isolated by Baderschneider108 could be present, and if the two p-coumaroyl

hexose esters observed by Monagas100, 131 and Hernandez103 are indeed present in wine and

not products of the analysis, or if additional esters are present that were not observed.

Regardless, 9 and 10 can be submitted to D. bruxellensis to determine their role in the

production of ethylphenols, given that under fermentation conditions they should exist

mainly as the 1-O-β-esters.

Chapter 3: Acyl Migration of Hydroxycinnamoyl Glucose Esters

50

Chapter 3: Acyl Migration of Hydroxycinnamoyl Glucose

Esters.

3.1 Introduction.

Previously (Chapter 2) it was observed that the prepared 1-O-β-hydroxycinnamoyl glucose

esters (9 and 10) underwent acyl migration under acidic conditions to give a range of

undesired mono-hydroxycinnamoyl esters which exhibited alternative glycosyl attachment,

affording largely the 2-O-α- and 6-O-α-esters, in addition to the initial 1-O-β-esters. While

the extent of migration and the ratio of the resulting esters differed between synthetic

attempts, a common factor for production was flash chromatography on silica gel, with no

migration observed during alternative purification attempts.

O

OH

HOHO

OH

O

OH

O

OR2

HOHO

R1OOH

R1 = HCA, R2 = H, 2-O-α-esterR1 = H, R2 = HCA, 6-O-α-ester

O

H+

123

46

123

46

HCA = O

OH

Figure 3.1: Acyl migration in p-coumaroyl glucose.

A similar phenomenon was observed for an S- to O-acetyl group migration occurring for a

furanosyl derivative in the presence of silica (Figure 3.2), which the authors rationalised by

the formation of a pseudo hydrogen bond between the silica and the carbonyl oxygen,

which decreased the electron density around the carbonyl carbon, promoting nucleophilic

attack and allowed for easy migration to the neighbouring oxygen.190 The extent of


51

migration was lessened with decreasing activities of silica, which was achieved by

hydrating the silica, reducing the electron accepting abilities. While it was found not to

occur for an O- to O-acetyl migration, this report confirms that silica can promote acyl

migration, which appeared to be the case for 9 and 10.

O

OOSAc

OHHO

O

OOSH

OHAcO

Silica

Figure 3.2: Initial silica catalysed 3-S- to 6-O-migration observed by Whistler et al.

Iddon et al. studied the migration of phenyl acetic acid glucosides under basic conditions,

describing cyclic transition states, and suggested that the speed of migration was largely

related to the stability of the intermediates.191 Migration from the 4-O- to that 6-O-position

proceeded through a 6-membered cyclic intermediate and was found to be much favoured

over those created by nucleophilic attack of the neighbouring hydroxyl groups, which

proceed through a 5-membered intermediate. Similar work by Horrobin et al. investigated

4-O- to 6-O-migrations under acid-catalysed conditions which resulted in the proposal of a

mechanism that proceeded through a cationic cyclic intermediate (Figure 3.3). Again, the

speed of migration was governed by the stability of the intermediate, but the relative ratios

of the products were purely under thermodynamic control.192 Furthermore, the speed of the

4-O- to 6-O-migrations shown by both Horrobin and Iddon rationalise the lack of 4-O-

glucose ester observed in the migrated mixtures of 9 and 10, which can quickly migrate to

the 6-O-position.

OHO

OH

HO

O

OO

R

O

AcO

AcOOAc

O

O

OH

a) b)

H

Figure 3.3: Migration intermediates. a) Base-catalysed 1-O-β- to 2-O-β-migration

intermediate proposed by Iddon et al. b) Acid-catalysed 4-O-α- to 6-O-α-migration

intermediate proposed by Horrobin et al.


52

Yoshimoto and Tsuda monitored migrations to and from multiple positions in glucose and

discovered that a direct 1-O- to 6-O-migration was not possible, but occurred in a step-

wise fashion via first a 1-O- to 3-O-, then a 3-O- to 6-O-migration.193 By protecting the 3-

O-position, the previously observed 1-O- to 6-O-migration was inhibited, indicating the

involvement of the 3-O-position in the process. They also documented that the 4-O- to 6-

O-migration occurred rapidly, as did a 1-O-α- to 2-O-transformation, while 2-O- to 3-O-,

3-O- to 4-O- and 4-O- to 2-O-migrations occurred at much slower rates.194 These findings

support the assigned composition of the mixtures for the hydroxycinnamoyl glucose esters

(9 and 10), with rapid migration away from the 4-O-position limiting its presence, while

the ease of 3-O- to 6-O-migration can explain both the prevalence of the 6-O-esters and

limited production of the 3-O-esters. However, Yoshimoto and Tsuda failed to induce a 1-

O-β- to 2-O-migration and claimed that the 1-O-β-ester was stable to migration, in contrast

to the hydroxycinnamoyl glucose esters.194 Furthermore, most of the studies into glucose

acyl migrations have found that migrations to yield the preferred 6-O-esters are largely

irreversible,194-195 which again, is not the case for hydroxycinnamoyl esters, with a

migrated mixture returning to the 1-O-β-ester under wine-like conditions.

As studied in sucrose migrations, Mollinier et al. discovered that basic conditions favoured

the formation of the 6-O-ester, though in acidic conditions the 6-O- as well as the 3-O- and

2-O-esters were produced,196 mimicking that seen for the hydroxycinnamoyl glucoses (9

and 10) which migrated to give the 6-O- and 2-O-esters, with traces of the 3-O-esters

detected. Furthermore, the reported preference for the 6-O-ester under basic conditions

may explain the occurrence of 51 and 52, the 6-O-chloroacetyl derivatives, which were

formed during the deprotection of 49 and 50 to produce 9 and 10, as explained in Chapter

2. It may not be that the 6-O-chloroacetyl is the last group removed, but instead, when a

single chloroacetyl group remains, regardless of the position, migration to the 6-O-position

is promoted under the influence of basic conditions (Figure 3.4).


53

OHOHO

OH

OAcCl

OR

OHOClAcO

OH

OH

OR

OClAcOHO

OH

OH

OR

OHOHO

OAcCl

OH

OR

B-

4-O- to 6-O- 3-O- to 6-O-

B-

B- 2-O- to 6-O-

Figure 3.4: Proposed migration of mono-O-chloroacetyl derivatives to the 6-O-position.

The above studies confirm that acyl migrations within glucose derivatives are common,

and can be achieved under acidic or basic conditions, with the conditions playing an

important role in the final products. Largely, the ratios of the products formed are under

thermodynamic control, with the stabilities of the intermediates only determining the speed

at which the final products are formed. If these processes are controlled completely by

thermodynamics, then by mapping the relative stabilities of each possible ester one should

be able to gain an indication as to the potential for the formation of each ester under given

conditions.

A density functional study of acyl migration in formyl nucleosides determined that

mapping a step-wise mechanism was a more valid pathway than a concerted migration, and

that the geometry of the products should be optimised in the desired solvent rather than a

geometry optimisation in a vacuum, followed by a single-point energy calculation in

solvent.197 The use of a step-wise mechanism to study migration supports those proposed

by both Horrobin and Iddon, using a number of cyclic intermediates.191-192

In addition, it was discovered that the energy of migration was lowered by increasing the

polarity of the solvent,197 most likely by stabilising the charge of the intermediates.

However, for the glucose esters (9 and 10) a greater extent of migration away from the 1-

O-β-ester was observed in less polar solvents. Whether this was for kinetic or

thermodynamic reason was initially unclear.


54

3.2 Research Aims.

Acyl migrations can occur under acidic, basic and neutral conditions198 which can lead to

the formation of different products.196 The hydroxycinnamoyl glucose esters (9 and 10)

were seen to migrate away from the 1-O-β-esters under acidic non-aqueous conditions and

the resulting mixtures migrated back to the 1-O-β-esters in an acidic aqueous environment

(Chapter 2). The role of thermodynamics has been identified in the literature as the main

factor in determining both if migration will occur and to what extent it will occur under

given conditions,192, 196 while the kinetics of each migration relied on the stability of the

intermediates.191-192

Initially the thermodynamics of migration were studied here in an attempt to determine

why migration occurred, and to justify the ratios of esters observed, including the

propensity for increased migration for the p-coumaroyl glucose system compared with the

feruloyl derivative. Following the findings of Rangelov et al., the equilibrium geometry of

each ester was optimised in the desired solvents, rather than optimising the geometry in a

vacuum and then performing single-point energy calculations in each solvent using the

vacuum optimised geometry common to all.197

Additionally, by applying an analogous mechanism to that described by Horrobin192 the

energy of key intermediates was expected to provide insight into how quickly each

transformation can occur. This should indicate the likely pathway of migration and

whether the ratios obtained experimentally are purely under thermodynamic control, or

whether the kinetics do contribute.

Once the nature of migration is determined, protocols for avoiding, minimising, controlling

or even predicting migration can be developed so that synthesis can be achieved more

simply and without having to characterise and utilise mixtures.

Finally, several analyses have identified two separate hexose esters of p-coumaric acid

without designation of which esters were present.100, 103, 131 If the formation of multiple

esters is found to be possible in wine, then the nature of the esters identified could be

determined, also providing information as to whether multiple esters will need to be

considered during quantification in grapes and wine.


55

3.3 Theoretical Studies into Acyl Migration of Hydroxycinnamoyl Glucoses.

3.3.1 Thermodynamics of Migration

Experimentally the 1-O-β-, 2-O-α- and 6-O-α-esters were produced preferentially, though

each of the twenty possible esters (α- and β-anomers for 1-O-, 2-O-, 3-O-, 4-O- and 6-O-

esters, Figure 3.5) need to be considered, not only explain the presence of those that were

seen, but also why others were absent from the “migrated” mixtures.

O

OH

HOHO

OH

O

OH

O

O

OH

HOHO

HOO

O

OH

O

OH

HOHO

O

OH

O

OH

HOHO

OOH

O

O

OH

HOO

OH

OH

O

OH

HOO

HOOH

O O

HO HO

O

OH

OHO

OH

OH

O

OH

OHO

HOOH

OO

HO HO

O

O

HOHO

OH

OH

O

O

HOHO

HOOH

O O

HO HO

O

HO HO

ββββ-D-glucopyranosyl αααα-D-glucopyranosyl

1-O-

2-O-

3-O-

4-O-

6-O-

R

R

RR

R R

R R

R R

9 R = H10 R=OCH3

Figure 3.5: Twenty possible esters of p-coumaroyl glucose (9) and feruloyl glucose (10).


56

The equilibrium geometries of each of the ten p-coumaroyl and ten feruloyl esters were

determined in water using the DFT B3LYP level of theory and a 6-31G* basis set, and the

energy of the optimised structures compared to the energy of the desired 1-O-β-ester

(Figure 3.6).

En

erg

y (

kJ/m

ol)

-est

ers

αααα1-

O--e

ster

s

ββββ2-

O--e

ster

s

αααα2-

O--e

ster

s

ββββ3-

O--e

ster

s

αααα3-

O--e

ster

s

ββββ4-

O--e

ster

s

αααα4-

O--e

ster

s

ββββ6-

O--e

ster

s

αααα6-

O-

0

20

40

60

80p-Coumaroyl glucoseFeruloyl glucose

Figure 3.6: Energy of p-coumaroyl and feruloyl glucose esters in water, relative to the 1-O-

β-esters. See Appendix 1, Table A1.1 for ground state energies and relative differences.

The relative energy of each of the glucose esters as calculated in water indicate precisely

what was observed during the synthesis of the glucose esters; that the 1-O-β-esters are

thermodynamically favoured, and when exposed to conditions conducive to migration in

an aqueous environment, the 1-O-β-esters should prevail. As such, when migrated

mixtures of 9 and 10 were subjected to storage in acidic aqueous conditions, the 1-O-β-

esters could be recovered as a result of thermodynamic influences. Notably, the relative

energies of the 2-O-α- and 6-O-α-esters are lower than the remaining seven esters

suggesting that given migration away from the 1-O-β-ester in water, the formation of these

two species would be favoured.

Experimentally, migration was observed during flash chromatography on silica employing

solvent systems consisting of a small fraction of methanol in dichloromethane. To examine

the effect of the solvent system the energy of each ester was calculated in dichloromethane

with the results indicating that thermodynamically, the 6-O-α- and the 2-O-α-esters are

more favoured than the 1-O-β-ester (Figure 3.7). Already it can be seen why migrated


57

mixtures of the glucose esters possessed the ratios that they did, with the preferred product

in water, migrating in dichloromethane to give a mixture heavily favouring the formation

of the 2-O-α- and 6-O-α-esters. While very little regarding the relative extent of migration

observed between p-coumaroyl glucose and feruloyl glucose can be explained by the

thermodynamic influences, Figures 3.6 and 3.7 explain the occurrence of different esters

under changing solvent conditions. However, these results suggest that in dichloromethane

complete migration to the 6-O-position would eventuate, and also that the ratios observed

experimentally were under kinetic as well as thermodynamic control. The length of

exposure to dichloromethane, or the kinetics of migration determined the extent of

migration from initially being exposed to dichloromethane to the time of characterisation,

as equilibrium has not yet been established.

En

erg

y (

kJ/m

ol)

-est

ers

αααα1-

O--e

ster

s

ββββ2-

O--e

ster

s

αααα2-

O--e

ster

s

ββββ3-

O--e

ster

s

αααα3-

O--e

ster

s

ββββ4-

O--e

ster

s

αααα4-

O--e

ster

s

ββββ6-

O--e

ster

s

αααα6-

O-

-20

0

20

40


Figure 3.7: Energy of p-coumaroyl and feruloyl glucose esters in dichloromethane, relative

to the 1-O-β-esters. See Appendix 1, Table A1.2 for ground state energies and relative

differences.

To further investigate the role of solvents in determining the thermodynamics of migration

for both 9 and 10, the energy of each of the esters was determined in ethanol (Figure 3.8)

and toluene (Figure 3.9), representing polar non-aqueous and non-polar solvents

respectively, to investigate whether a trend in ester energy with respect to solvent

properties could be established.


58

En

erg

y (

kJ/m

ol)

-est

ers

αααα1-

O--e

ster

s

ββββ2-

O--e

ster

s

αααα2-

O--e

ster

s

ββββ3-

O--e

ster

s

αααα3-

O--e

ster

s

ββββ4-

O--e

ster

s

αααα4-

O--e

ster

s

ββββ6-

O--e

ster

s

αααα6-

O-

-20

0

20


Figure 3.8: Energy of p-coumaroyl and feruloyl glucose esters in ethanol, relative to the 1-

O-β-esters. See Appendix 1, Table A1.3 for ground state energies and relative differences.

En

erg

y (

kJ/m

ol)

-est

ers

αααα1-

O--e

ster

s

ββββ2-

O--e

ster

s

αααα2-

O--e

ster

s

ββββ3-

O--e

ster

s

αααα3-

O--e

ster

s

ββββ4-

O--e

ster

s

αααα4-

O--e

ster

s

ββββ6-

O--e

ster

s

αααα6-

O-

-20

0

20

40


Figure 3.9: Energy of p-coumaroyl and feruloyl glucose esters in toluene, relative to the 1-

O-β-esters. See Appendix 1, Table A1.4 for ground state energies and relative differences.

The thermodynamic preference for the 1-O-β-esters only occurred for water, while the

other three solvents studied theoretically show a preference for formation of the 6-O-α-

esters with the 2-O-α-esters closely following. The relative energies in ethanol show not

only a strong preference for the 2-O-α- and 6-O-α-esters, but also suggest that the 3-O-α-

and β-esters are as likely to occur as the 1-O-β-esters. NMR characterisation of the glucose

esters was performed in d4-methanol, and while the effect of methanol on the energies of

the esters could not be studied due to the restraints of the program, it is expected that a

similar trend would be experienced to that calculated for ethanol. As such, any migration

that resulted from exposure to silica in dichloromethane could be compounded by exposure


59

to methanol during characterisation. As such, the limited exposure to methanol could have

increased the thermodynamic preference for the 3-O-esters, as well as the 6-O-and 2-O-

esters, and allowed for minor formation of these esters which could be detected in minor

quantities by 1H NMR analysis.

The relative energies shown in Figures 3.6-3.9 show large changes in the energy of every

ester with changing solvent, though this is only a product of plotting the data relative to the

1-O-β-esters. Extracting the data for the p-coumaroyl glucose esters in all four solvents and

plotting the energies relative to the less important 1-O-α-esters furnishes Figure 3.10. The

largest changes in energy between solvents is present for the 1-O-β-esters, indicating that

in different environments, the energies of the other esters does not change to a great extent,

but really only relative to the 1-O-β-esters. The four energies obtained for the 1-O-β-esters

(relative to the 1-O-α-ester) show the greatest variation across the four solvents with

energies of -59.44, -45.53, -34.66 and -51.77 kJ/mol producing a standard deviation of 10.5

kJ/mol. By performing the same analysis for the remaining esters, the 2-O-α-esters

produce a standard deviation of 4.5 kJ/mol, 4.2 kJ/mol for the 2-O-β-esters, 3.9 kJ/mol for

the 4-O-β-esters and the deviations of the remaining esters between 2.2 and 2.9 kJ/mol. A

similar trend can be observed when the data is plotted relative to any ester other than the 1-

O-β, and shows that as the environment moves away from aqueous, the preference for the

1-O-β-esters decreases as opposed to preference for the 2-O-α- or 6-O-α-esters increasing.

En

erg

y (k

J/m

ol)

ββββ1-

O- ββββ2-

O- αααα2-

O- ββββ3-

O- αααα3-

O- ββββ4-

O- αααα4-

O- ββββ6-

O- αααα6-

O-

-60

-40

-20

0DichloromethaneWaterEthanolToluene

Figure 3.10: p-Coumaroyl glucose (9) ester energies calculated in changing solvents,

relative to the 1-O-α-esters. See Appendix 1, Table A1.5 for relative energies and standard

deviations.


60

By studying ethanol and toluene in addition to water and dichloromethane, it can be seen

that preference for the 1-O-β-esters is not a product of solvent polarity and these appear to

only be favoured in aqueous environments, with migration to the 2-O-α- and 6-O-α-esters

likely to occur in any solvent other than water. Though formation of other esters may be

favoured, the occurrence of mixtures with differing ratios indicates that equilibrium has not

yet been achieved, and the thermodynamic products will prevail once equilibrium has been

reached. However, instantaneous determination of which esters are present will rely on

both the thermodynamics to determine which esters are being formed and the kinetics to

describe to what extent it has occurred at the time of measurement.

3.3.2 Kinetics of Migration

While base-catalysed migration under some reaction conditions might be expected to a

certain extent, the observed migrations as well as any expected to occur in wine, must be

acid-catalysed, and the lack of free acid seen experimentally in the migrated mixtures

indicate that this is not a case of ester hydrolysis followed by re-attachment, but that the

process is one of intramoleular transesterification. By combining the observations

proposed by Horrobin et al.,192 and that known for an acid-catalysed transesterification, the

mechanism of 1-O-β- to 2-O-β-migration for the p-coumaroyl glucose analogue (9) can be

hypothesised (Scheme 3.1).


61

O

O O

HO

O

O

OOH

H

O

OH

OHOH

O

O

OOH

OH

H

O OH

OH

H

O

O O

HO

OH

H

O

OH

O O

OH

Scheme 3.1: Mechanism for acid catalysed 1-O-β- to 2-O-β- acyl migration of p-

coumaroyl glucose (9).

In an acidic environment a pH dependant equilibrium will exist between the protonated

and unprotonated carboxyl oxygen, with the energy difference between the two forms

being of little consequence to the kinetics of the migration. As such, the four cationic

intermediates of interest in mapping the kinetics of this particular migration are given

below (Figure 3.11). While migration has been mapped for the 1-O-β- to 2-O-migration in

p-coumaroyl glucose, a similar mechanism is expected for all those investigated.


62

O

O O

OH

O

O

OOH

H

O

OH

OH

HO

OH

HOHO

OH

HOHO

OH

HO

OH

O

O

OOH

OH

HOHO

OH

H

O OH

OH

H

Int. 1 Int. 2

Int. 3 Int. 4

Figure 3.11: Key intermediates (Int. 1-4) for the acid-catalysed 1-O-β- to 2-O-β- acyl

migration of p-coumaroyl glucose (9).

The transformation shown (Scheme 3.1) involves formation of the 2-O-β-ester, though

experimentally the 2-O-α-ester was favoured with very little of the 2-O-β-ester able to be

detected. Ignoring any conversion between α- and β-glucose that could possibly occur

during migration (in an attempt not to complicate the study with concurrent reactions),

migration from the 1-O-β-ester to form the 2-O-α-ester must occur either via mutarotation

before or after migration. Due to mechanistic constraints in mutarotation between the 1-O-

β- and the 1-O-α-ester it is assumed that the migration occurs first, from the 1-O- to the 2-

O-position, followed by mutarotation to convert the 2-O-β- to the 2-O-α-ester. Each four

intermediates (Int. 1-4) were optimised at the B3LYP 6-31G* level and the calculated

energy compared with that of intermediate 1.


63

En

erg

y (

kJ/m

ol)

Int. 1 Int. 2 Int. 3 Int. 4

-50

0

50

100

150

DichloromethaneVacuum

Water


migration, relative to intermediate 1. See Appendix 2, Table A2.1 for ground state energies

and relative differences.

Unlike the glucose ester energies calculated in the previous section, the energies of the

migration intermediates were calculated for water and dichloromethane alone, rather than

in all four solvents as were employed above (removing ethanol and toluene).

Experimentally, water and dichloromethane were the most common solvents that the

glucose esters (9 and 10) experienced, having been allowed to react in an aqueous system

and purified in either water/methanol or dichloromethane/methanol. Analogous

intermediates to those used for p-coumaroyl glucose (as shown in Figure 3.11), were used

in calculation of the 1-O-β- to 2-O-β-feruloyl glucose migration (Figure 3.13).

En

erg

y (

kJ/m

ol)


50

100

150

200WaterDichloromethaneVacuum


relative to intermediate 1. See Appendix 2, Table A2.2 for ground state energies and

relative differences.

The relative energies of the intermediates for the p-coumaroyl and feruloyl glucose esters

show similar patterns with intermediates 2 and 3 being highest in energy, which is

accentuated in water. Therefore more energy is required for migration in water than

dichloromethane, which is the opposite effect to that seen by Rangelov,197 whereby more


64

polar solvents reduced the energy barriers to migration. However, in this case all four

intermediates investigated are cationic, with intermediate 1 expected to exist in equilibrium

with the neutral species under acidic conditions. Therefore, the energy involved in the

formation of the cationic intermediates is not considered and there is no expected effect

between the intermediates based on charge stabilisation.

Not only is the formation of the 2-O-α-esters more favoured in solvents other than water,

but the energy of migration is lessened in dichloromethane also. When the glucose esters

were synthesised and purified by column chromatography in a dichloromethane based

solvent system, migration away from the 1-O-β-esters under acidic conditions was

thermodynamically favoured and kinetically more favoured also.

Furthermore, the energy to migration in feruloyl glucose (10) compared with p-coumaroyl

glucose (9) in dichloromethane is almost twice as much. For feruloyl glucose the migration

is limited by intermediate 2, lying some 60 kJ/mol higher in energy than intermediate 1,

whereas the same transition for p-coumaroyl glucose requires only 35 kJ/mol.

Experimentally, it was seen that migration occurred faster, or to a greater extent for p-

coumaroyl glucose than for feruloyl glucose, which can be explained by the kinetics of

migration. With a faster process occurring for p-coumaroyl glucose, migration had

occurred to a greater extent at the time of characterisation than for feruloyl glucose.

As mentioned previously, it is expected that by performing the NMR characterisation of

the glucose esters in methanol, migration would continue beyond that caused simply by

purification. As a result, the equilibrium of the mixtures may not have been achieved over

the course of 2-3 hours needed for purification and characterisation, with the kinetically

favoured process (migration in p-coumaroyl glucose) having occurred to a greater extent.

While it would have been relatively simple to allow the mixture to equilibrate over hours,

or even days, the key outcome of this study was to obtain pure 1-O-β-esters and as such,

migration was not encouraged.

Additionally, intermediate 4 was lower in energy than intermediate 1 for p-coumaroyl

glucose under all the conditions investigated, showing a preference for formation of the

cationic 2-O-β-ester over the cationic 1-O-β-ester, which is not the case for feruloyl


65

glucose, indicating that kinetically, migration to the 2-O-position within p-coumaroyl

glucose is comparatively favoured.

While step-wise migration around the glucose ring has been proposed,189 migration to the

2-O-position is not the only path of migration. Yoshimoto dismissed direct 1-O- to 6-O-

migration, though their observations differ from those seen in these studies enough for it

not to be completely discounted.193 Calculating the energy barriers for migration to the 6-

O-position should support or eliminate direct migration, as such analogous intermediates to

the 1-O- to 2-O-migration were optimised and the energies calculated, the same

mechanism was assumed, with the attacking nucleophile changed to the hydroxyl of the

final ester.

En

erg

y (

kJ/m

ol)


-50

0

50

100





En

erg

y (

kJ/m

ol)


-50

0

50

100






66

Figures 3.14 and 3.15 show that direct migration to the 6-O-β-esters is as likely as

migration to the 2-O-β-esters, with very similar energy barriers to overcome. In a similar

fashion to the 2-O-β-p-coumaroyl glucose migration in dichloromethane (Figure 3.12), the

protonated 6-O-β-esters are more favoured than the protonated 1-O-β-esters under all

conditions for both 9 and 10, suggesting that in an aqueous acidic environment direct

migration to the 6-O-position would be favoured over migration to the 2-O-position.

Both transitions involve intermediates of unfavourable conformations, with the 1-O- to 2-

O-migration involving a 5-membered cyclic intermediate, and the 1-O- to 6-O-migration a

7-membered transition state. Given the evidence of Iddon,191 the migration to the 3-O-

position involving a 6-membered cyclic intermediate should be more favourable than both

of the previously calculated transformations.

En

erg

y (

kJ/m

ol)


-40

-20

0

20

40





En

erg

y (

kJ/m

ol)


-40

-20

0

20






67

As expected, the migration to the 3-O-position involves smaller energy barriers, and can be

expected to occur more quickly than migration to the 2-O- and 6-O-positions. However,

thermodynamics suggests a preference for the formation of the 2-O- and 6-O-esters, which

would be a case of subsequent migration.

The optimised intermediates 2 and 3 produced in studying both the 1-O- to 3-O- and 1-O-

to 6-O-migrations involve a glucose ring-flip, with the result being that for the 1-O-β-ester

the 1-OH, the 3-OH and the 6-OH are all axial and on the same face above the ring,

leading to a less hindered migration.

OOORHO

HO

OH

OH

OR

OH

OH

OH

OH

Ring-flip

Figure 3.18: Glucose ring-flip to facilitate 1-O- to 3-O-migration and 1-O- to 6-O-

migration.

This effect is seen more greatly for the 1-O- to 3-O-migration where very small energy

barriers are seen. The unfavoured ring-flipped conformation is stabilised by the formation

of the bicyclic intermediates, or more accurately that the formation of the cyclic

intermediate in the ring-flipped conformer is more favourable than in the original

conformation. A glucose ring-flip would explain the findings of Yoshimoto, that 1-O- to 3-

O- to 6-O-migrations were occurring rapidly, and potentially why they did not experience

1-O-β- to 2-O-migration, with these groups being too far removed for the transformation to

take place.193

3.4 Liquid Chromatography of Wine.

The ease of migration in non-aqueous solvents provides an explanation as to why two p-

coumaroyl hexose esters were seen by both Monagas and Hernendez,100, 103, 131 and brings

into question whether one or both esters observed are an artifact of the extraction being

performed in non-aqueous solvents, rather than a grape or wine product. A study by Perez-


68

Magarino looked at the ability of different resins to absorb and retain phenolic compounds

by loading, then washing with water, ether and finally ethyl acetate.199 From a single resin

one p-coumaroyl glucose ester was eluted with water, but when non-aqueous solvents were

used, two p-coumaroyl glucose esters were observed. It is unlikely that the difference in

chemical properties of the two p-coumaroyl glucose esters would result in a single ester

eluting while the other is retained and further supports the theory that the 1-O-β-esters are

solely found in wine, and any other esters are produced upon exposure to non-aqueous

solvents.

To investigate the presence of multiple hydroxycinnamoyl glucose esters in wine, the

extraction method of Monagas was employed to extract red and white wine spiked with

both p-coumaroyl and feruloyl glucose ester (9 and 10). By comparing the amount

remaining in the “spiked” extraction to that of a pure sample, the liquid-liquid extraction

efficiency was approximately 20% for both compounds, and thus considered inadequate

for detecting small quantities in wine.

Solid-phase extraction by loading spiked wine onto XAD-8, elution with 25%, 50% and

75% methanol in water, followed by HPLC analysis indicated that the majority of the

glucose ester content was found in the 50 and 75% methanol fractions and that the

extraction efficiency was approximately 60%. When analysed by LC-MS, the pure glucose

ester standards, which were not subjected to extraction, consisted largely of single esters,

while the extracted wines contained two glucose esters, the most predominant being the 1-

O-β-ester, with the minor peak expected to be either the 2-O-α- and 6-O-α-esters,

indicating that the spiked glucose esters were migrating under the extraction conditions.

With liquid-liquid extraction giving poor extraction efficiencies, the nature of the solvents

likely to yield migration, and the solid-phase extraction also resulting in migration, neat

red and white wine along with concentrated samples (5 times concentrated under reduced

pressure) were submitted to analysis by LC-MS. The extracted ion chromatograms of the

pure esters again showed largely single esters (Figures 3.19a and 3.20a) which suggests

that migration is not an effect of the HPLC method. The fragmentation pattern of p-

coumaroyl glucose matches literature data,131 and feruloyl glucose fragmented in an

analogous manner, as reported200 (Figures 3.19b and 3.20b).


69

Figure 3.19: p-Coumaroyl glucose. a) Extracted ion chromatogram of m/z 325. b) Mass

spectrum at 29.6 to 29.8 minutes.

Figure 3.20: Feruloyl glucose. a) Extracted ion chromatogram of m/z 355. b) Mass

spectrum at 36.5 to 36.6 minutes.


70

Injection of neat white wine, as well as concentrated white wine gave the same results,

with 5 main peaks appearing in the area of interest in the extracted ion chromatogram for

m/z 325 (Figure 3.21). Based on the fragmentations, the first peak observed in the EIC for

m/z 325 corresponds to the p-coumaroyl glucoside (A), lacking the fragmentation which

corresponds to the loss of water from the aglycone seen in the glucose ester fragmentation,

the second two (B and C), with aglycone peaks matching ferulic acid are likely to be

feruloyl tartrate derivatives,131 and the fourth and fifth peaks (D and E) are the two p-

coumaroyl glucose esters (Figure 3.22).

Figure 3.21: Concentrated white wine, extracted ion chromatogram of m/z 325.

A B

C

D E


71


325.

The fragmentations of the two peaks identified in the extracted ion chromatogram of m/z

355 imply that the first peak (A) is the feruloyl glucoside and the second (B) is the feruloyl

glucose ester (Figure 3.24).

The observation that the white wine, when analysed neat, or concentrated appears to

possess two p-coumaroyl glucose esters, but only single feruloyl glucose ester, is

consistent with the reluctance of feruloyl glucose ester to migrate which was observed

during synthesis and supported by theoretical studies.

A)

B)

C)

D)

E)


72

Figure 3.23: Concentrated white wine, extracted ion chromatogram of m/z 355.


355.

With little difference in the results from neat and concentrated white wine, the analysis of

red wine was only repeated with concentrated red wine (Figure 3.25), with the extracted

ion chromatogram for m/z 325 possessing additional peaks, which are tentatively identified

A)

B)

A

B


73

as p-coumaroyl anthocyanin derivatives (Figure 3.26). However, the presence of the extra

peaks (at 32.6, 34.3, 36.7 and 40.0 minutes) in the same region of the chromatogram as

feruloyl glucose (approximately 36.6 minutes) led to decreased resolution of the m/z 355

extracted ion chromatogram (Figure 3.27), with no fragmentations able to be found

matching that of the reference sample.

Figure 3.25: Red wine chromatogram (DAD).

Figure 3.26: Concentrated red wine, extracted ion chromatogram of m/z 325.


74

Figure 3.27: Concentrated red wine, extracted ion chromatogram of m/z 355.

Due to decreased sensitivity for feruloyl glucose resulting from the presence of additional

peaks in red wine, the samples were submitted to HPLC-MRM, with the fragmentation

from the parent ions to the aglycone and the aglycone-water being monitored. For p-

coumaroyl glucose, p-coumaric acid (m/z 163, red line) and p-coumaric acid minus water

(m/z 145, blue line) fragmentations are shown on the left of Figure 3.28, and for feruloyl

glucose, ferulic acid (m/z 193, red line) and ferulic acid minus water (m/z 175, blue line)

are shown on the right of Figure 3.28.


75

Figure 3.28: HPLC-MRM traces (aglycone - blue, aglycone minus water - red) of

hydroxycinnamoyl glucose esters. a) Pure glucose esters. b) Neat white wine. c)

Concentrated white wine. d) Concentrated red wine.

From the chromatograms in Figure 3.28, the presence of multiple p-coumaroyl glucose

esters in white and red wine can be observed, but the presence of a second glucose ester of

ferulic acid is not immediately obvious in white wine. Although the concentration of

feruloyl glucose in concentrated red wine is somewhat lower than in white wine, evidence

p-Coumaroyl glucose Feruloyl glucose

a)

b)

c)

d)


76

of a second peak in the concentrated red wine sample can be seen. Even though this data

confirms the finding of Monangas, Hernandez and Perez-Magarino that there are multiple

glucose esters in wine, it has also shown that the extraction method can contribute to the

extent of migration, with pure glucose esters undergoing migration during solid-phase

extraction. These results also indicate why the feruloyl glucose ester was not observed in

previous studies, as the presence of what are likely to be p-coumaroyl anthocyanin

derivatives in red wine co-elute, and prevent identification and quantification. Feruloyl

glucose could only be seen in this case by determining the retention time and comparing

fragmentations with the pure reference compound.

3.5 Conclusions.

Previous studies have evaluated the migrations to and from multiple positions of glucose,

and the reverse processes. However, in this study, only the migrations away from the 1-O-

β-esters were examined as the destruction and formation of these esters were specifically

of interest. Migrations involving 5-, 6- and 7-membered cyclic intermediates have been

studied with kinetic preference for migration through the 6-membered intermediate, to an

ester that is thermodynamically unfavourable.191-192 Migration between neighbouring

hydroxyls, in this case 1-O- to 2-O-migration, occurs through a 5-membered intermediate,

while migration to the 3-O-position, which was found to be kinetically more favoured,

involves an extra carbon atom, resulting in a 6-membered intermediate. Whereas migration

from the 1-O- to the 6-O-position requires formation of a 7-membered intermediate, and

was found to be kinetically as favoured as formation through a 5-membered intermediate.

Studying the kinetics of migration away from the 1-O-β-esters is valuable in determining

the likelihood of migration, and it can now be rationalised why previous authors193 have

seen migration through the 3-O-position to get to the 6-O- and 2-O-position. However, the

fact that the 3-O-esters do not accumulate in the migrated glucose ester mixtures

strengthens the argument that these migrations are ultimately under thermodynamic

control. Thus, while in water the energy barriers for migration to the 3-O-position are quite

favourable, the energies of the end products suggest that this wouldn’t be the case,

indicating that the 3-O-esters are intermediates and the secondary migration to the 6-O-

position is extremely facile.


77

However, these results do indicate that in wine like conditions (largely aqueous) that the 1-

O-β-esters will be thermodynamically favoured, and that most migrations are suppressed

due to higher energy barriers to migration, with the only exception being the migration to

the 3-O-position which requires a ring flip of the glucose ring before the migration can

take place.

It can be seen that the migrations observed in the hydroxycinnamoyl glucose esters are

largely under thermodynamic control and that under conditions that favour products other

than the 1-O-β-ester, migration will most likely occur rapidly. This study did not

investigate the base-catalysed migration, as it would be of little importance in acidic wine

medium and the 1-O-β-esters appeared stable under mildly basic conditions. There is the

potential that synthetically, basic conditions should also be avoided, not the least because

of the greater likelihood of hydrolysis.

While the kinetics would suggest that migration away from the 1-O-β-esters should

proceed through the 3-O-position, the relative ratios of the different glucose esters

observed during synthesis showed the p-coumaroyl moiety migrating to a greater extent

than the feruloyl, which suggests that migration of p-coumaroyl glucose should be

kinetically more favoured. This effect is only predicted for a 1-O- to 2-O-migration

(Figures 3.12 and 3.13) with the energy required for migration much higher in the feruloyl

derivative than in the p-coumaroyl. Furthermore, if migration to the 3-O-position is

experienced, the ring-flip required for this transformation to occur would most likely

promote further migration to the 6-O-position, resulting in formation of the

thermodynamically more stable esters.

In future synthetic attempts, the use of solvents other than water should be limited,

especially under conditions conducive to migration. If organic solvents are employed, they

should be done so under neutral conditions, or preferably in the presence of a buffer. In the

event of migration away from the desired 1-O-β-esters, it has been shown that storage

under aqueous acidic conditions will again yield the desired esters.

Furthermore, it can be expected that in wine, and wine-like environments such as model

fermentations, that the 1-O-β-esters will predominate and other esters would be a product


78

of the conditions that the compounds are exposed to. An analytical method using a liquid-

liquid or solid-phase extraction could promote formation of other esters and render the

quantification inaccurate. As such, to quantify these compounds in wine, care must be

taken to ensure that migration hasn’t occurred, and that the compounds seen are products

of the grape or wine conditions, rather than artifacts of the methodology. If extraction and

concentration is required, analysis of a neat wine sample might assist in determining if any

migration has occurred as a result of the processing. The differences in the ratio of esters

observed between model wine, where the 1-O-β-ester is predominant, and red wine, where

multiple esters were observed, are most likely a product of the matrices and effected by

such factors as pH and dielectric constant.

In addition, this study also describes the identification of feruloyl glucose (10) for the first

time in red wine, which, along with p-coumaroyl glucose can exist as multiple esters in

both red and white wine.

Chapter 4: Photoisomerisation of Hydroxycinnamic Acids

79

Chapter 4: Photoisomerisation of Hydroxycinnamic Acids.

4.1 Introduction.

4.1.1 Hydroxycinnamate Photoisomerisation

In addition to the acyl migrations that were observed (Chapter 2) and studied (Chapter 3),

trans-p-coumaroyl glucose (9) and trans-feruloyl glucose (10) were found to undergo

photoisomerisation resulting in formation of cis-analogues (Figure 4.1). Interconversion

between trans- and cis-hydroxycinnamic acids has been known since Kahnt reported and

investigated the nature of the conversion and found the equilibrium to be effected by the

solvent in which the transition occurred, the concentration of the hydroxycinnamic acids,

and also the nature of the compounds, with caffeic acid existing in different isomeric ratios

to an esterified analogue.187

O

OH

HOHO

OH

O

O

OH

R

O

OH

HOHO

OH

O

O

OH

R9 R = H10 R = OCH3

cis-9 R = Hcis-10 R = OCH3

hν

Figure 4.1: Photoisomerisation of the hydroxycinnamoyl glucose esters.

A further investigation by Kahnt measured the pH dependence of the photoisomerisation

and observed that maximum conversion to the cis-acids was achieved within a pH range of

5-7.186 Subsequent studies have shown there to be changes in equilibria due to: additional

substituents, with TMS ethers producing different isomeric ratios than the free

hydroxycinnamic acids, although the direction of the equilibrium change was not

consistent;201 subtle differences in conversion to the cis-analogue in differing solvents;202

as well as due to the wavelength of incident light.203 Longer wavelengths of ultra-violet

light induce a slower isomerisation, while shorter wavelengths induce a quicker and more

complete isomerisation but can also result in degradation after prolonged exposure.203


80

Under irradiation, the majority of the hydroxycinnamic acids exist at isomeric ratios

favouring the trans-isomer, with the cis-acid contributing between 40 and 49%.204 While

many studies have shown the effect of ultra-violet light on isomerisation, other reports

suggest that it can be promoted under more innocuous light conditions, providing the

recommendation that the hydroxycinnamic acids should be handled in the dark to avoid

any potential isomerisation.205-207

p-Coumaric acid (3) plays an important role in bacteria, being the basis for the

chromophore for the photoactive yellow protein or PYP, and as such the isomerisation of

p-coumaric acid, and derivatives, have been extensively studied at various theoretical

levels.208-213 Kort et al. first investigated the photoisomerisation in the PYP and found that

the thioester of p-coumaric acid is in equilibrium between the cis- and trans-forms, but

also concluded that not only can the cis-isomer be photochemically converted to the trans-

isomer, but that the cis- to trans-isomerisation could be facilitated thermally, which is not

the case for the reverse process.214 In the case of the true PYP, the photoisomerisation can

be induced by light up to 430 nm, although many investigations into the PYP however,

have begun with p-coumaric acid as a model system.

Common to these studies is the concept that the isomerisation proceeds via an excited

electronic state through promotion of an electron from the alkene π-bond into an anti-

bonding orbital, allowing for free bond rotation with the product determined by the nature

of the preferred conformation in the excited state. There still remains conjecture as to

whether the photoisomerisation of p-coumaric acid proceeds through a singlet excited state

(S1) with paired electron spin, or a triplet excited state (T1) with unpaired electron spin

(Figure 4.2).

π

π∗

S0 S1 T1

Figure 4.2: Electron configuration of π bonding and anti-bonding molecular orbitals in

ground and excited states.


81

Li and Fang studied the excited states of trans-p-coumaric acid with respect to the nature

of the excitation, finding that the lowest singlet transition (S0-S1) possessed more n-π*

character which resulted from excitation of the unpaired electrons of the carbonyl oxygen,

but that the S0-T1 transition was dominated by a π - π* electron promotion. The S0-T1

transition resulted in T1 p-coumaric acid possessing the lowest energy conformation with

the alkene p-orbitals at a 90o dihedral angle, compared with 180o seen in the S0 state.

Relaxation from T1 at a 90o conformation to S0 accounted for the formation of cis- and

trans-isomers with either able to be formed.208

Furthermore, Sergi et al. compared the excitation energies of p-coumaric acid against the

phenolate anion (the form in which it is found in the PYP) and found that the vertical

excitation energy of the anion was 20% lower than that of the protonated form,211 though

in acidic wine-like conditions (pH 3.5) it is extremely unlikely that the phenolate anion

would be found.

4.1.2 cis-Hydroxycinnamate content in grapes and wine

There are few examples of grape or wine quantifications that include both the cis- and

trans-hydroxycinnamates, with the majority of these studies focusing on p-coumaroyl

tartrate,72-73, 96, 101, 103, 118, 123-127, 131 although feruloyl tartrate72 and p-coumaric acid have

also been considered.101, 103 Of the studies that do consider hydroxycinnamate

stereochemistry, it is sometimes only specified for some compounds and not others,28, 72, 98,

101, 104 or only the trans-isomer is (or can be) quantified,73, 99-100, 103, 123-124, 126, 131 while other

studies do not consider the stereochemistry at all.97, 102, 105-106, 119, 121-122, 128-130, 215-216 For

some studies it is unclear whether the quantification techniques fail to distinguish the two

forms, or if the cis-isomers are present in concentrations lower than the detection

threshold. However, it is understandable that quantifying both cis- and trans-isomers has

been achieved most regularly for the most prevalent form, p-coumaroyl tartrate.

Both isomers of feruloyl tartrate were quantified and expressed as molar percentages in

Cencibel grapes and the resulting wine, with an initial cis-content of 26.3%, dropping to

9.4% in wine.72 The effect of malolactic fermentation on cis- and trans-p-coumaric acid

was studied, with an initial isomeric ratio of 54.8% cis-p-coumaric acid observed; resulting

in 25.2% of the cis-isomer after malolactic had been conducted in steel, against 32.2% in a


82

barrel.101 In the same study, red wine was aged for 14 months resulting in p-coumaric acid

content consisting of 6.5% of the cis-isomer. The same author performed malolactic

fermentation with several lactic acid bacteria and found that the p-coumaric acid content

changed from 48.1% of the cis-isomer to between 3.1 and 38.1% after malolactic

fermentation.103

The effect of ultra-violet light on cis/trans-ratios in grapes can be observed by compiling

the cis- and trans-p-coumaroyl tartrate content in red and white skins of Montealegra et al., 126 and the content in red and white juices as determined by Singleton et al. (Table 4.1).127

Table 4.1: Content of cis- and trans-p-coumaroyl tartrate in the skin and juice of red and

white grapes.126-127

trans-p -coumaroyl tartrate cis-p-coumaroyl tartrate % cis-isomer

Skin content (mg/kg) White (n = 6) 7.63 2.92 27.65Red (n = 4) 6.28 1.79 22.15

Juice content (mg/L) White (n = 19) 15.32 3.11 16.87Red (n = 21) 18.38 3.43 15.73

Average Concentration

The highest cis-content is in the skin of white grapes (27.7%), where increased exposure to

ultra-violet light is expected, followed by the skin of red grapes (22.2%) where

pigmentation can provide some relief from ultra-violet radiation. The juice content of cis-

p-coumaroyl tartrate is lower than observed in the skins likely caused by the absorbance of

radiation by compounds in the skins protecting the “juice” hydroxycinnamates, again with

white grapes (16.9%) having a higher cis-content than red grapes (15.7%). While this data

has come from two separate sources, the approximate ratio of cis- to trans-p-coumaroyl

tartrate in the entire berry decreases with decreasing exposure to ultra-violet radiation. In

red or white grapes the approximate cis:trans ratio agrees with those observed during

synthesis of the glucose esters, which were observed to exist as approximately 20-25% of

the cis-isomer.

The data obtained by Lee and Jaworski in Pinot Blanc grapes supports that shown in Table

4.1. They observed ratios of 16.6 and 21.8% of cis-p-coumaroyl tartrate at harvest across

two separate vintages,96 as well as Gomez-Alonso reporting 21% of the cis-isomer in

Cencibel grapes.72 Meanwhile, other studies have described much different ratios. Betes-

Saura et al. analysed free run juice of 3 white grape varieties and found an average cis-


83

content of 47.6 %,73 while a further study by Lee and Jaworski found cis-contents ranging

from 10 – 67% across 21 white cultivars.124

In wines, the isomeric ratio of p-coumaroyl tartrate also varies, with one study by Monagas

et al. observing ratios between 10.9 to 55.3% of the cis-isomer in four different red wine

varieties,131 while other studies reported contents anywhere between 6.5 and 47.4% of the

cis-isomer.72-73, 101, 103

Of the hydroxycinnamates of interest in this study, p-coumaroyl tartrate is the analogue to

have been most thoroughly quantified, although the presence of other cis-

hydroxycinnamates in wine have been documented. Baderschnieder108 investigated the

phenolic content of a Riesling wine identifying a cis-isomer for most trans-

hydroxycinnamate species identified (excluding the glucose esters of p-coumaric and

ferulic acid), including p-coumaric acid, p-coumaric and ferulic-4-O-glucosides, and p-

coumaroyl and feruloyl tartrate esters. With the extent of isomerisation observed for the

glucose esters in the laboratory (described in Chapter 2), it is possible that a more sensitive

technique might have detected cis-glucose esters also.

The hydroxycinnamates, a class of compound that are widely researched because of the

possibility that they contribute to spoilage during winemaking, have not been considered in

microbial breakdown with respect to stereochemistry. The cis-hydroxycinnamates which,

as shown above, can contribute to around 20% of the hydroxycinnamate content of the

berry, have not been specifically evaluated with respect to metabolism by D. bruxellensis

and whether these can contribute to the accumulation of ethylphenols in wine during barrel

ageing.

4.1.3 Enzymatic Specificity

The breakdown of hydroxycinnamate esters to yield ethylphenols involves two potentially

stereospecific enzymes, an esterase and a decarboxylase. Submitting cis-esters to D.

bruxellensis would simultaneously test both of these enzymes in the same experiment,

potentially leading to ambiguous results. Furthermore, once Dekkera has expressed

decarboxylase activity, the product of the cis- and trans-acids, the vinylphenols (5 and 6),

do not possess differing alkene stereochemistry and as such the subsequent vinyl reductase


84

will not be a factor in stereospecific metabolism. Formation of ethylphenols from acids as

well as esters must both proceed via decarboxylation, and as such, this stage of

hydroxycinnamate metabolism is a key step at which the stereospecificity should be

scrutinised.

A purified p-coumarate decarboxylase from B. bruxellensis has been tested for substrate

specificity with p-coumaric acid, caffeic acid, ferulic acid and m-coumaric acid.68 The

decarboxylase was active towards caffeic, p-coumaric and ferulic acids in that order of

preference, and inactive towards m-coumaric acid. Although, it showed that a para-

hydroxyl group was required, and by shifting it to the meta position, decarboxylation was

retarded, the specificity within this study towards cis-acids was not tested.

Similar results were seen from a purified hydroxycinnamate decarboxylase enzyme from

B. anomalus when tested towards a number of similar acids.70 Again, caffeic acid was

preferentially decarboxylated before p-coumaric and ferulic acids with relative activities of

37.5 and 31.3%, but the decarboxylase was inactive towards cinnamic acid, sinapic acid,

hydrocaffeic acid, o-coumaric acid, m-coumaric acid, p-methoxycinnamic acid, p-

hydroxybenzoic acid, iso-ferulic acid, 5-hydroxyferulic acid, 3,4-methylenedioxycinnamic

acid, phenylalanine and pyruvic acid. This provides further information that a para-

methoxy group is insufficient to facilitate decarboxylation, but again, no cis-acids were

tested.

Gramatica et al. showed the ability of Saccharomyces cerevisiae to decarboxylate p-

coumaric acid, p-methoxycinnamic acid, ferulic acid and 3,4-dimethoxycinnamic acid, but

found that it was not active towards cinnamic acid, caffeic acid and

methylenedioxycinnamic acid.217 cis-3,4-Dimethoxycinnamic acid was then tested and the

decarboxylase was not active towards it, indicating an inability to decarboxylate this

particular cis-acid. Unlike the previous reports, this study shows the ability for a

decarboxylation to occur for a compound possessing a para-methoxy group, though

possessing a para-hydroxyl group doesn’t appear to guarantee decarboxylation with caffeic

acid not being affected.


85

HO

COOH

HO

COOH

HO

COOH

COOH COOH

H3CO

COOH

COOH

HO H3CO

COOH

HO

COOH

H3CO HO

HO

OH

HOH3CO

OH

COOHO

O

COOH

NH2

COOH

O

COOH

O

HO

COOH

OHCOOH

COOH

HO

COOH

HO

COOH

HO

COOHHO

HO

COOHH3CO

HO

COOHH3CO

OCH3

COOH

O

COOH

O

p-Coumaric acid

m-Coumaric acid o-Coumaric acid

Ferulic acid Caffeic acid

p-Methoxycinnamic acid

p-Hydroxybenzoic acid iso-Ferulic acid 5-Hydroxyferulic acid

3,4-methylenedioxycinnamic acid Phenyl alanine Pyruvic acid Hydroxypyruvic acid

Mandelic acid Hydrocinnamic acid 4-Hydroxyphenyl acetic acid

4-Phenyl but-3-enoic acid Phloretic acid Dihydrocaffeic acid

Dihydroferulic acid Sinapic acid Acrylic acid Crotonic acid

Figure 4.3: Compounds investigated in decarboxylation studies.

Goodey and Tubb studied S. cerevisiae in relation to the gene then designated as POF1

(now known as PAD1218), which is responsible for the decarboxylation of

hydroxycinnamic acids and resulting production of phenolic off-flavour as observed in

beer.219 Those strains possessing the ability to decarboxylate were designated Pof+. S.

cerevisiae strains possessing the Pof+ phenotype showed the ability to decarboxylate

ferulic acid, cinnamic acid and p-coumaric acid, in that order of preference, while caffeic

acid, 4-hydroxypyruvic acid, mandelic acid, hydrocinnamic acid, 4-hydroxyphenylacetic


86

acid, 4-hydroxybenzoate and 4-phenylbut-3-enoate were unaffected by the S. cerevisiae

decarboxylase. Again, compounds possessing a para-hydroxyl group weren’t necessarily

decarboxylated, and additionally, cinnamic acid that has no aromatic ring substitution was

able to be decarboxylated.

Harada and Mino studied the substrate specificity of the decarboxylase activity of

Cladosporium phlei.84 It was active towards cis-p-coumaric acid, trans-p-coumaric acid,

caffeic acid and ferulic acid, in that order of preference, and was inactive towards cinnamic

acid, m-coumaric acid, o-coumaric acid, phloretic acid, p-methoxycinnamic acid,

dihydrocaffeic acid, dihydroferulic acid, sinapic acid, acrylic acid and crotonic acid.

With these examples of decarboxylases, one fungus which showed otherwise similar

substrate specificity to Dekkera did possess the ability to metabolise cis-acids, while

brewing yeasts, which showed very different substrate specificity lacked the ability to

metabolise cis-acids. However, there are no reports as to the stereospecificity of the

decarboxylase of D. bruxellensis and whether cis-hydroxycinnamates, in addition to the

trans-isomers, can be metabolised and contribute to the accumulation of ethylphenols in

wine.

4.2 Research Aims.

Following the photoisomerisation of the hydroxycinnamoyl glucose esters (9 and 10)

which resulted in partial conversion to furnish cis-analogues, and the ease by which it

occurred, the role of the cis-hydroxycinnamates in the production of ethylphenols has

become of importance to this research. Investigation into the known cis-hydroxycinnamate

content of grapes has indicated that for most trans-hydroxycinnamates there exists a

corresponding cis-isomer, and of those quantified, the cis-content is in the vicinity of 20%

of the total hydroxycinnamates. The ability of D. bruxellensis to metabolise the cis-

isomers, or otherwise, could have an impact on the production of ethylphenols by as much

as 20% within wine.

The decarboxylase of D. bruxellensis is active in the bioconversion of the free acids and

esters to ethylphenols and will be tested for stereospecificity by conducting fermentation in


87

the presence of both the cis- and trans-hydroxycinnamic acids. The trans-hydroxycinnamic

acids have already been synthesised, being isolated during preparation of the esters

(Chapter 2), with the cis-hydroxycinnamic acids now requiring synthesis.

In addition to synthesis of cis-p-coumaric and cis-ferulic acids, the photoisomerisation of

the hydroxycinnamic acids will be investigated at a theoretical level beyond that of the free

acids to explain the rapid isomerisation experienced for the glucose esters, which was not

observed for any other hydroxycinnamate derivative. Using the existing theoretical studies

of p-coumaric acid, the procedures will be extended to the hydroxycinnamate esters in an

effort to determine the energy barriers associated with photoisomerisation, and develop

protocols to increase the ease of synthesis of the glucose esters (9 and 10) while

maintaining stereoisomeric purity.

4.3 Synthesis of cis-Hydroxycinnamic Acids.

By isolating the minor products from the Wittig reaction (as described in Chapter 2) via

column chromatography, pure cis-11 and cis-12 could be achieved, though these were only

produced in limited quantities as the trans-products are thermodynamically favoured. The

formation of ethyl coumarate (11) affords approximately 90% trans-11 and 10% cis-11,

while the same reaction for the production of ethyl ferulate (12) yields around 70% trans-

12 and 30% cis-12, which can largely be separated by flash chromatography with only

minor co-elution of isomers. While cis-11 and cis-12 could be isolated and characterised

under ambient light conditions, attempted base-catalysed ester hydrolysis to yield the cis-

acids (cis-3 and cis-4) proved unsuccessful, resulting in isomeric mixtures. Hydrolysis of

cis-11 gave a mixture of cis- and trans-3 in a ratio of 20:80, while hydrolysis of cis-12

afforded a 35:65 mixture of cis- and trans-4.


88

O H

OH

R

OH

R

O OEt

OH

R

O

OEt

OH

R

O

OH

13 R = H14 R = OCH3

cis/trans-11 R = H cis/trans-12 R = OCH3

cis-3 R = Hcis-4 R = OCH3

+

Wittig

KOH

Scheme 4.1: Attempted synthesis of cis-hydroxycinnamic acids.

Initial conclusions for the production of the isomeric mixtures of 3 and 4 were that the

basic reaction conditions caused deprotonation of the phenolic hydroxyl, generating a

phenolate resonance structure that interrupted the α,β-unsaturated double bond and causing

a conversion back to the thermodynamically more favoured trans-isomer (Figure 4.4).

OH

O OEt

O-

O

OEt

O

-O

OEt

O

-O OEt

H+

Figure 4.4: Proposed resonance assisted conversion of cis-p-coumaric acid to trans-p-

coumaric acid.

Separation of cis- and trans-4 was achieved through flash column chromatography using

10% methanol in dichloromethane, yielding the pure cis-4 acid as indicated by TLC. After

standing for 16 hours in solution, analysis by NMR showed a 2:1 mixture of cis- to trans-4,


89

contradicting the previous hypothesis that isomerisation to the trans-acid was facilitated by

formation of the phenolate.

Though the trans-acids proved stable under ambient light conditions during synthesis, this

was not the case for the cis-acids, undergoing isomerisation under laboratory lighting. As

such, isolation of the pure cis-acids by chromatography was then performed under red light

in an attempt to minimise photoisomerisation, although under these conditions the cis-acids

still underwent a slow conversion back to the trans-acids. This indicated that spiking a

fermentation with pure cis-acids in the dark would result in some conversion back to the

trans-isomer which D. bruxellensis could metabolise and yield ethylphenols. With the cis-

acids (cis-3 and cis-4) found to isomerise to give trans-acids even under conditions of low

light exposure, attempted synthesis and isolation of the pure cis-3 and cis-4 was

discontinued.

As the decarboxylase of D. bruxellensis is active towards only a few hydroxycinnamic

acids, a more stable alternative cis-substrate could not be utilised in the investigation of

stereoselective metabolism. Furthermore, an alternative substrate would not provide

adequate information as to the ability of D. bruxellensis to produce the ethylphenols of

interest to this study. Therefore it was required that fermentation experiments be performed

on isomeric mixtures of cis- and trans-hydroxycinnamic acids. These could be produced

through ultra-violet irradiation with literature observations indicating that the

hydroxycinnamates, upon irradiation, exist in approximately 40-49% in the cis-form.204

Photoisomerisation of p-coumaric acid (trans-3 to cis-3) was initially performed under 254

nm ultra-violet light (a readily available lamp in organic laboratories used to view TLC

plates) on small scale in NMR tubes for simple analysis requiring no sample preparation.

After 66 hours of irradiation a stable 52:48 ratio of trans- to cis-3 was produced as

indicated by 1H NMR analysis. Concurrently trans-ethyl coumarate (11) was irradiated to

investigate the thermodynamic effects of the isomerisation, assuming that cis-11 would be

produced to a greater extent given the stability comparative to cis-3 during synthesis. After

66 hours a 60:40 mixture of trans- and cis-11 was observed suggesting that the stability of

cis-11 during synthesis and chromatography is not due to thermodynamic stabilities of the

cis-isomers, but more likely a result of the higher energy required to facilitate the

conversion.


90

One study, of o-coumaroyl glucoside isomerisation, reported a quicker trans- to cis-

conversion with a shorter wavelength of ultra-violet light, but noted that prolonged

exposure of the glucosides to 254 nm light gave significant degradation of the product,

with only 40% remaining after 1 hour of exposure. The isomerisation induced by using a

365 nm lamp was found to be slower, but produced the cis-isomer to greater extents and

resulted in very little degradation of the substrate.203

Isomerisation using 365 nm light was performed for both hydroxycinnamic acids (trans-3

and trans-4), with the percentage conversion to the cis-isomer higher in 4. The extent of

isomerisation for 3 was less than experienced under 254 nm irradiation, however 365 nm

light was used in preference to avoid any potential and unnecessary degradation of the

acids. The isomeric ratios of the cis:trans-mixtures were determined by integrating the

signal for H8 of each isomer as these signals not only show a large change in chemical shift

between isomers, but are removed from other signals thus avoiding overlapping shifts.

The stability of the cis/trans-mixtures were tested by storing a small amount of cis/trans-p-

coumaric acid (3) under different conditions. A 61:39 mixture of trans:cis-3 was stored in

acetone or as a solid under the conditions outlined in Table 4.2, with the final ratios

determined after two weeks of storage.

Table 4.2: Isomeric ratio of p-coumaric acid (3) under different storage conditions.

trans cis

Solid 63 37In acetone 62 38



Solid 62 38In acetone 63 37-20 oC

Condition% of isomer

Room temperature, Dark

Room temperature, Light

4 oC

Slight deviations in the observed isomeric ratios are most likely a product of the variability

of the analytical method (NMR), otherwise the mixtures of cis/trans-3 proved stable,

indicating that the mixtures produced by irradiation would remain constant over the course

of the planned fermentation studies.


91

Additionally, in an attempt to promote formation of cis-ethyl coumarate (cis-11) by

binding transition metals between the aromatic ring and carbonyl, and effectively

“holding” the molecule in the cis-configuration (Figure 4.5), palladium and platinum were

stirred in the presence of 11. Both metals were found to retard photoisomerisation in an

undesired manner, and had the effect of yielding trans-11 from the cis-isomer.

O

O

OH

M

Figure 4.5: Intended effect of metal coordination on cis-hydroxycinnamates.

Submitting trans-ethyl coumarate (11) to ultra-violet radiation in the presence of either

palladium or platinum on activated carbon, isomerisation to cis-11 was inhibited by

approximately the same amount as the molar ratio in which the metal was present.

Additionally, by submitting cis-ethyl coumarate (cis-11) to palladium acetate, which was

utilised instead of 10% palladium on carbon to increase the number of moles of metal per

mass of reagent, had a large effect of converting cis-11 to the trans-isomer. With an excess

of palladium, 100% conversion from cis-11 to trans-11 was achieved under ambient light

conditions. So, not only did complexation of metals fail in promoting formation of cis-11,

but assisted in achieving the opposite, production of trans-11, which is consistent with

literature reports.220 However, future use of transition metals to maintain isomeric purity,

or convert unwanted cis-analogues to the trans-isomer may be of use. In the case of

undesired formation of cis-glucose esters (cis-9 and cis-10), future synthetic attempts may

involve producing isomeric mixtures under ambient light conditions, followed by recovery

of the trans-isomers by exposure to palladium or platinum.

4.4 Theoretical Studies into the Isomerisation of Hydroxycinnamic Acids.

Ab initio studies into the photoisomerisation of p-coumaric acid have been well

documented,208-213 but this study aimed to explain the differences in isomerisation observed


92

between hydroxycinnamate derivatives in the laboratory, namely: the ease of trans- to cis-

isomerisation for the glucose esters (9 and 10) that has not been observed previously; the

instability of the cis-hydroxycinnamic acids (cis-3 and cis-4) relative to the cis-ethyl

hydroxycinnamates (cis-11 and cis-12); and the rapid formation of the trans-acids seen

during the base-catalysed ester hydrolysis of the cis-ethyl esters (cis-11 and cis-12).

Using a DFT B3LYP 6-31G* level of theory, trans-p-coumaric acid underwent an

equilibrium geometry optimisation at the ground state (S0) and then again at the first

excited state. While there remains some discrepancy in the literature regarding whether the

isomerisation proceeds via excitation to the S1 or T1 state,208 optimisation of the ‘first

excited state’ of trans-p-coumaric acid identifies this as the T1 state. As the S1 and T1 state

occupy the same molecular orbital, with the HOMO of each theoretically corresponding

with the LUMO of the S0 state, the only difference in energy should arise from the

opposing spin of the promoted electron, which should be constant across all the geometries

and of little consequence to this study.

The HOMO and LUMO of S0 trans-p-coumaric acid (3), along with the HOMO of T1

trans-p-coumaric acid were calculated for the optimised structures and are shown below

(Figure 4.6). These support that promotion of an electron from S0 trans-p-coumaric acid

will result in a molecular orbital corresponding to T1 trans-p-coumaric acid.

Figure 4.6: Frontier molecular orbital diagrams of trans-p-coumaric acid (3). a) HOMO of

S0 trans-p-coumaric acid. b) LUMO of S0 trans-p-coumaric acid. c) HOMO of T1 trans-p-

coumaric acid.


93

Also, the spin density of T1 trans-p-coumaric acid (Figure 4.7) is largely concentrated

around the α,β-unsaturated double bond, suggesting that the alkene is highly affected by

the excitation into the T1 state.

Figure 4.7: Electron spin density in T1 trans-p-coumaric acid.

The energy profile of p-coumaric acid (3) resulted from constraining the dihedral angle of

trans-p-coumaric acid around the α,β-unsaturated double bond to 180o, optimising the

geometry at a DFT B3LYP 6-31G* level and calculating the energy of the conformation.

Using a dynamic constraint, the dihedral angle was rotated from 180o to 0o through 19

possibilities, optimising the geometry and calculating the energy every 10 degrees. This

process was applied to the singlet ground state (S0), then repeating for the triplet state (T1)

representing the photoisomerisation of p-coumaric acid in a vacuum (Figure 4.8).


94

Dihedral Angle

En

erg

y (

kJ/m

ol)

0 30 60 90 120 150 1800

100

200

300

S0

T1

Figure 4.8: Energy profile of p-coumaric acid (3). See Appendix 3, Table A3.1 for

calculated energies and relative differences.

The potential for isomerisation of the hydroxycinnamic acids, or compounds possessing a

similar molecular backbone, can be observed in the energy profile of p-coumaric acid.

Excitation of S0 trans-p-coumaric acid to the T1 state results in the preferred conformation

at a 90o dihedral angle, which corresponds to the two p-orbitals sitting orthogonal.

Relaxation back to S0 state can give either trans- or the cis-p-coumaric acid, which

corresponds to literature evidence.213

Repeating the analysis for ethyl coumarate (11) the S1 profile was generated with

questionable results (Figure 4.9). Rotation of the dihedral from 180o results in geometries

of increasing energy until the 70o conformation whereby calculations for the 60o and 50o

conformations fail to converge, and expected values are achieved for the remaining

conformations (0-40o). This effect also observed for calculations using a dynamic dihedral

constraint and for calculation using individually drawn structures (manual constraint),

though the manually drawn structures are a product of the preceding conformation (170o

drawn from the 180o, and so on). As the dihedral angle of the carbon skeleton is altered,

the alkene protons remain at 180o to each other producing structures with increasing

degrees of pyramidilisation and energy (50 and 60o conformations) until the calculations

fail to converge, whereby the protons are forced to a 0o dihedral angle (0-40o

conformations) (Figure 4.10). As such, the S0 profile was achieved by dynamically rotating

the dihedral from 180o to 90o (forwards) and from 0o towards 90o (backwards), producing

improved results.


95

Dihedral Angle

En

erg

y (

kJ/m

ol)

0 30 60 90 120 150 1800

100

200

300Manual S0 ForwardsDynamic S0 BackwardsDynamic S0 FowardsT1

Figure 4.9: Energy profile produced from forward and reverse dynamic, and manual

constraint of ethyl coumarate (11). See Appendix 3, Table A3.2 for calculated energies and


H

RH

RHH

R

H

R

RH

R R

H

R

H

H

R

R

H

180o 140o 30o60o90o

Figure 4.10: Pyramidilised alkene resulting from rotation of the dihedral angle from 180o

to 0o in ethyl coumarate (11).

The dynamic S0 profiles (forwards and backwards) differ for the 0-40o conformations by

approximately 17 kJ/mol, a product of optimising different ethyl coumarate conformers.

Beginning with trans-ethyl coumarate and rotating the alkene to give cis-ethyl coumarate

produces a different conformer to simply drawing cis-ethyl coumarate, which differ by

rotation around the C8-C9 bond (Figure 4.11). Both configurations, when optimised,

produce geometries that are in their own right in potential energy wells of the surface. As

such, the optimisation of the higher energy conformer determines a local energy minimum

which does not correspond to the global energy minimum. Calculations started from an

initial MMFF geometry, gave higher energy structures than those started from the MMFF

conformer, which identifies the global energy minimum.


96

EtO O

OH OH

EtO

O

OH

O

OEt

a) b)

OR

C8

C9

Figure 4.11: cis-Ethyl coumarate conformers produced by: a) drawing trans-ethyl

coumarate and rotating the dihedral; and b) drawing cis-ethyl coumarate.

The energy profile of p-coumaroyl glucose (9) (Figure 4.12) incorporates first identifying

the MMFF conformer before calculating the S0 profile from both the trans- and cis-isomers

towards the 90o dihedral. The results suggest that different p-coumarate substrates do not

appear to have a large enough effect on the excitation energy to heavily effect the extent or

speed of isomerisation. The energy profiles of p-coumaric acid (3), ethyl coumarate (11)

and p-coumaroyl glucose (9) display similar characteristics and energy barriers.

Dihedral Angle

En

erg

y (

kJ/m

ol)

0 30 60 90 120 150 1800

100

200

300

S0 ForwardsS0 BackwardsT1

Figure 4.12: Energy profile for p-coumaroyl glucose (9), relative to S0 trans-isomer. See

Appendix 3, Table A3.3 for calculated energies and relative differences.

The potential for the different solvent conditions that the glucose esters encountered during

synthesis (aqueous) compared with the other hydroxycinnamates (organic) to have an

effect on the isomerisation exists, as noted by Kahnt.187 Repeating the theoretical analysis

for p-coumaric acid (3) and p-coumaroyl glucose (9) in water (Figure 4.13) yields near

identical energy profiles as shown by the T1 energy profiles which mimic that observed

respectively within Figures 4.8 and 4.12, which were calculated in a vacuum.


97

a)

Dihedral Angle

En

erg

y (

kJ/m

ol)

0 30 60 90 120 150 180180

200

220

240

260

b)

Dihedral Angle

En

erg

y (

kJ/m

ol)

0 30 60 90 120 150 180200

220

240

260

280

Figure 4.13: a) T1 energy profile for p-coumaric acid (3) in water, relative to the S0 trans-

acid. b) T1 energy profile for p-coumaroyl glucose (9) water, relative to the S0 trans-

isomer. See Appendix 3, Table A3.4 for calculated energies and relative differences.

The solvated and vacuum calculated T1 profiles appear to have similar energies, and all

display diabatic profiles, whereby the lowest energy conformation in the T1 state is at a 90o

dihedral angle, which will relax to the S0 state at the highest energy point and potentially

produce either isomer. The shape of the T1 and S0 profiles, and the energy differences

between the S0 and T1 states (the vertical excitation energies) do not change excessively as

a result of either solvation or substrate. Furthermore, optimising the geometry of the T1

state may lead to incorrect vertical excitation energies as the S0 geometry of a molecule is

excited directly to the T1 state, rather than excited to a different, optimised T1 geometry.

Therefore, the vertical excitation energies should be a result of an optimised geometry for

the S0 state, followed by a single point energy calculation for that geometry at the T1 state.

In addition to the S0-T1 vertical excitation energy, the HOMO-LUMO gap was determined

for trans-p-coumaric acid (3) and trans-p-coumaroyl glucose (9) under the effect of

solvation (Figure 4.14). The vertical excitation energy, along with the HOMO-LUMO gap

should decrease with increasing solvent polarity for a π- π* transition due to the

comparative stabilising effect on the π* orbital compared with the π orbital.221

Determination of solvation effects was investigated using solvent ET30 values as shown in

Table 4.3, with a solvent of higher polarity possessing a greater ET30 value.


98

Table 4.3: Solvent polarities and ET30 values.

Solvent Dielectric Constant ET30 Value

Water 80 63.1

Ethanol 24.6 51.9Acetone 21 42.2DCM 9.1 40.7THF 7.5 37.4Ether 4.3 34.5

ET30

En

erg

y (

kJ/m

ol)

30 35 40 45 50 55 60 65245

250

255

260

265trans-p-Coumaric Acidtrans-p-CoumaroylGlucose

Figure 4.14: S0-T1 vertical excitation energy for trans-p-coumaric acid (3) and trans-p-

coumaroyl glucose (9). See Appendix 4, Table A4.1 for calculated vertical excitation

energies.

The vertical excitation energies of trans-p-coumaric acid and trans-p-coumaroyl glucose in

numerous solvents (Figure 4.14) differ by approximately 7 kJ/mol throughout the solvents

tested, and it is unlikely that such a small difference in vertical excitation energy between

the substrates would explain the vast difference in the ease of isomerisation that was

observed during synthesis. However, the HOMO-LUMO gaps of the S0 compounds in each

of the solvents (Figure 4.15) indicate a similar change in energy as displayed for the

vertical excitation energies, without having to calculate the energy of the T1 state,

providing a more rapid indication of the vertical excitation energies.


99

ET30

En

erg

y (

kJ/m

ol)

30 35 40 45 50 55 60 65380

390

400

410

420trans-p-Coumaric Acidtrans-p-CoumaroylGlucose

Figure 4.15: HOMO-LUMO gap for trans-p-coumaric acid and trans-p-coumaroyl

glucose. See Appendix 4, Table A4.1 for calculated HOMO-LUMO gaps.

The effects of solvent and substrate on isomerisation as described by Kahnt187 have not

yielded justification to the changes in isomerisation observed during synthesis, although

Kahnt also identified pH as another determining factor in the isomeric equilibrium.186 In

addition, Sergi et al. found a decrease in excitation energy of p-coumaric acid upon

formation of the phenolate,211 and during synthesis, the glucose esters were believed to

have been deprotonated on XAD-2 resin. Slight deprotonation of the glucose esters on

XAD-8 resin could facilitate a more rapid isomerisation which was experienced for these

compounds alone as only they were submitted to XAD resins. Repeating the vertical

excitation energy and HOMO-LUMO gap calculations for the trans-p-coumaroyl glucose

phenolate afforded Figure 4.16.

a)

ET30

En

erg

y (

kJ/m

ol)

30 35 40 45 50 55 60 65180

185

190

195

200

Figure 4.16: a) Vertical excitation energies (S0-T1) of trans-p-coumaroyl glucose phenolate

in solvents of differing polarity. b) HOMO-LUMO gap. See Appendix 4, Table A4.1 for

calculated vertical excitation energies and HOMO-LUMO gaps.

b)

ET30

En

erg

y (

kJ/m

ol)

30 40 50 60310

312

314

316

318

320


100

As seen by Sergi, a significant reduction of the vertical excitation energy is observed for

the trans-p-coumaroyl glucose phenolate to between 185 and 195 kJ/mol from 250-255

kJ/mol calculated for the protonated form. Again, evidence of this trend is observed in the

reduction of the HOMO-LUMO gap, indicating that phenolic deprotonation facilitates the

isomerisation away from the trans-isomer. Furthermore, if deprotonation of p-coumaroyl

glucose significantly effects the energy barrier to isomerisation, then the earlier hypothesis

regarding the base-catalysed ester hydrolysis inducing isomerisation from the cis-

hydroxycinnamates back to the trans-isomers, may in part be justified.

By comparing the HOMO-LUMO gaps of the cis- and trans-hydroxycinnamates examined

throughout this study (p-coumaric acid, ethyl coumarate and p-coumaroyl glucose), similar

conclusions can be made, that the trans-glucose esters are slightly more prone to

isomerisation away from the trans-isomer, and the energy barrier to isomerisation is

lowered with increasing solvent polarity (Figure 4.17). Additionally, the HOMO-LUMO

gaps of the cis-isomers gave the same trend with regard to solvent polarity, again with only

minor differences observed between substrates.


101

a)

ET30

Ene

rgy

(kJ/

mol

)30 40 50 60

380

390

400

410

420trans-Ethyl Coumaratetrans-p-Coumaric Acidtrans-p-CoumaroylGlucose

b)

ET30

En

erg

y (k

J/m

ol)

30 40 50 60390

395

400

405

410cis-Ethyl Coumaratecis-p-Coumaric Acidcis-p-Coumaroyl Glucose

Figure 4.17: a) HOMO-LUMO gap of trans-hydroxycinnamates. b) HOMO-LUMO gap of

cis-hydroxycinnamates. See Appendix 4, Table A4.2 and A4.3 for orbital energies HOMO-

LUMO gaps.

The relative stability of the cis-ethyl esters (cis-11 and cis-12) compared with the cis-acids

(cis-3 and cis-4) can only be attributed to the different solvents that the compounds

experienced, with the ethyl esters being synthesised and purified with less polar solvents

(i.e. dichloromethane), while the acids were prepared in ethanol and water, with this study

providing no other explanation as to the ease with which the acids isomerise.

In order to assess the hypothesis that during the attempted cis-hydroxycinnamic acid

synthesis, base-catalysed ester hydrolysis was the main contributing factor to

isomerisation, the nature of the compounds that existed under the reaction conditions have

to be determined. Literature pKa values for p-coumaric acid are 4.35 and 8.80 which

correspond to the carboxyl group and the phenol group, respectively.222 With a pKa of

4.35, the carboxylate would be heavily deprotonated at the pH needed to form the

phenolate, as such, the phenolate needs only be considered if the carboxylate anion is also

taken into account. At the pH of reaction, determined to be 13 as shown by pH strips, cis-


102

ethyl coumarate (cis-11) would exist as the phenolate. Ester hydrolysis yields cis-p-

coumaric acid dianion, which is protonated fully during acidic work-up (to pH 3) yielding

cis-p-coumaric acid (cis-3). The HOMO-LUMO gaps of each of the two anionic structures

present during reaction have been calculated under different solvent conditions, and

compared with cis-ethyl coumarate and cis-p-coumaric acid (Figure 4.18).

ET30

En

erg

y (k

J/m

ol)

30 35 40 45 50 55 60 65320

340

360

380

400

420cis-Ethyl Coumaratecis-Ethyl Coumarateanioncis-p-Coumaric Aciddianioncis-p-Coumaric Acid

Figure 4.18: HOMO-LUMO gaps of cis-hydroxycinnamates during base-catalysed ester

hydrolysis. See Appendix 4, Table A4.4 for orbital energies and HOMO-LUMO gaps.

The HOMO-LUMO gap of the cis-ethyl coumarate phenolate is considerably lower than

for the protonated form, implying that the greatest potential for cis- to trans-isomerisation

during this reaction is observed for the phenolate. While the cis-p-coumaric dianion has a

reduced HOMO-LUMO gap compared with the protonated forms, and would be more

likely to isomerise back to the trans-isomer. Thus, the cis-ethyl coumarate phenolate must

be largely responsible.

Additionally, those species existing as phenolic anions would be expected to have a much

greater electron donating character, leading to increased resonance forms and hence

reducing the double bond character of the alkene. It can only be assumed that this effect is

not as great for the p-coumaric dianion, as the formation of the carboxylate anion reduces

the electron withdrawing character of the carboxyl and retarding electron movement

throughout the molecule. If this is indeed the case, then calculation of the p-coumaric

carboxylate (with a protonated phenol) HOMO-LUMO would result in an increase in

HOMO-LUMO gap due to a similar electron donating character of the phenol as observed

for p-coumaric acid, but with reduced electron withdrawing character at the top of the

molecule, as determined by the charge on oxygen 3 (Figure 4.19).


103

O O

R

OR'

1

23

Figure 4.19: Numbering of oxygen atoms in hydroxycinnamate skeleton.

ET30

En

erg

y (k

J/m

ol)

30 40 50 60

380

400

420

440trans-p-Coumariccarboxylatecis-p-Coumaric carboxylate

Figure 4.20: HOMO-LUMO gap of p-coumaric acid carboxylate. See Appendix 4, Table

A4.5 for HOMO-LUMO gaps.

The HOMO-LUMO gap of the p-coumaric acid carboxylate in less polar solvents is similar

to those observed for p-coumaric acid, but in more polar solvents, there is a great increase

in the HOMO-LUMO gap of the carboxylate. This supports the theory that the energies of

isomerisation within the hydroxycinnamates are largely dependent on the nature of

intramolecular electronics, and the ability to reduce the double bond character of the

alkene.

The natural charge on the phenolic oxygen (Oxygen 1) and on the single-bonded

carboxylic oxygen (Oxygen 3) was determined for all of the hydroxycinnamates

investigated throughout this study (p-coumaric acid, p-coumaric acid carboxylate, p-

coumaric acid dianion, ethyl coumarate, ethyl coumarate phenolate, p-coumaroyl glucose

and p-coumaroyl glucose phenolate). The charge ratio (O1/O3) was then calculated in

order to provide an indication of the nature of the electronics of each group and what effect

they will have directly on the HOMO-LUMO gap (Figure 4.21). With good linear


104

correlation observed between oxygen charge ratio and HOMO-LUMO gap, additional

substrates were investigated including 1-O-acetyl p-coumaric acid (19) and 1-O-

chloroacetyl p-coumaric acid (21), with an expected reduction in electron donating

character of the ring giving a higher HOMO-LUMO gap, and the carboxylate anions of 19

and 21, in which the effect should be intensified.

HOMO-LUMO gap

Oxy

gen

Ch

arg

e R

atio

300 350 400 4500.0

0.5

1.0

1.5

2.0

2.5

Figure 4.21: HOMO-LUMO gaps of hydroxycinnamate derivatives against ratio of charge

between oxygen 1 and oxygen 3. See Appendix 4, Table A4.6 and A4.7 for charges on the

oxygen atoms and calculated ratios.

The HOMO-LUMO gap shows a rough linear relationship with the charge ratio of the

compounds investigated. Those with a large charge on oxygen 3 compared with oxygen 1

(lower ratio) are expected to have a lesser effect on the double bond and display an

increased HOMO-LUMO gap, where as those with large negative phenolic oxygen charges

and less negative charge on oxygen 3 are expected to have a much larger effect on the

double bond. Increased electron movement within the molecule decreases the HOMO-

LUMO gap and suggest that they will be more prone to isomerisation, which is also seen

by mapping the HOMO-LUMO gap against double bond length (Figure 4.22). The linear

correlation between the bond length or oxygen charge ratio and the HOMO-LUMO gap

indicates that increased resonance of the substrates increases the chance of

photoisomerisation.


105

HOMO-LUMO gapB

on

d L

eng

th (

Å)

300 350 400 4501.34

1.35

1.36

1.37

1.38

Figure 4.22: Relationship between HOMO-LUMO gap and double bond length in

hydroxycinnamate derivatives. See Appendix 4, Table A4.6 and A4.7 for alkene bond

length.

4.5 Conclusions.

The synthesis and attempted isolation of the cis-hydroxycinnamic acids (cis-3 and cis-4)

resulted in isomeric mixtures, which proved stable and could be maintained under

fermentation conditions. The cis-ethyl esters (cis-11 and cis-12) could be isolated and

handled under ambient light conditions without noticeable isomerisation back to the trans-

isomers. As such, fermentation studies into the stereospecificity of the D. bruxellensis

decarboxylase activity will need to be limited to spikes of isomeric mixtures of cis/trans-

hydroxycinnamic acids, and be fermented against pure trans-acids with the difference in

ethylphenol formation between them examined (Chapter 5).

Base-catalysed ester hydrolysis of the cis-ethyl esters involved formation of cis-ethyl ester

phenolates which were found to have a much lower HOMO-LUMO gap that could be

largely implemented in the conversion back to the trans-isomers. The relative stabilities of

the ethyl esters compared with the hydroxycinnamic acids could only be attributed to a

lowering of the HOMO-LUMO gap in solvents of increasing polarity, with the

environments needed to handle (synthesise or analyse) each compound contributing to

isomerisation. Additionally, the use of group 10 metals in the presence of cis- or trans-

ethyl coumarate was found to encourage formation of the trans-isomer. The use of

transition metals may have future applications in stereochemical control of the

hydroxycinnamic acids in synthetic attempts.


106

The energy required to excite p-coumaroyl glucose (9) to the T1 state and facilitate

cis/trans-isomerisation was lower than for p-coumaric acid (3), but was further lowered by

formation of the phenolate, of which there was evidence of occurring during contact with

XAD resins. As such, the use of XAD resins with the hydroxycinnamic acids should be

performed under strictly acidic conditions, preventing phenolate formation, or it should be

performed under light conditions of lower energy. In this study, red light proved useful in

preventing isomerisation.

Further studies into the isomerisation of hydroxycinnamates observed during synthesis

showed a relationship between the electronic make-up of the molecule and the energy

needed to facilitate photoisomerisation, with compounds allowing increased electron

movement, having reduced HOMO-LUMO gaps. This result indicates that even if the

glucose esters had not been completely deprotonated on the XAD resin, any extent of

hydrogen bonding to the phenol that would increase the electron donating character of the

phenolic oxygen would also decrease the energy required for photoisomerisation to occur.

Chapter 5: Bioconversion of Hydroxycinnamates by D. bruxellensis

107

Chapter 5: Bioconversion of Hydroxycinnamates by D.

bruxellensis.

5.1 Bioconversion of trans-Hydroxycinnamate Esters.

5.1.1 Ethyl Esters

O O

OH

O O

OH

OCH3

11 12

Figure 5.1: Ethyl coumarate (11) and ethyl ferulate (12).

Bioconversion of the ethyl hydroxycinnamates (11 and 12) by D. bruxellensis strain AWRI

1499, a representative of the predominant strain grouping in Australian winemaking,22 was

studied and the outcome determined by the production of 4-ethylphenol and 4-

ethylguaiacol. The self-anaerobic fermentations were conducted to maximise the

conversion from precursors, through vinylphenols to ethylphenols.21 Yeast biomass peaked

at day 6 and the fermentations were conducted for a further 3 days, concluding shortly after

the yeast entered stationary phase to maximise the potential metabolism of the ethyl esters.


108

Fermentation Progress (Days)

Per

cen

tag

e C

on

vers

ion

2 4 6 8 100

20

40

604-Ethylphenol4-Ethylguaiacol

Figure 5.2: Percentage of the theoretical maximum conversion of ethyl esters (11 and 12)

to ethylphenols.

AWRI 1499 displayed a much greater affinity for metabolism of ethyl coumarate (51.4%

conversion to 4-ethylphenol), over ethyl ferulate (4.0% conversion to 4-ethylguaiacol).

This appears to be a product of the esterase activity as such a preference is not observed for

the decarboxylase during the metabolism of the free hydroxycinnamic acids, as seen in a

later experiment and in literature reports.

Furthermore, Godoy et al. studied a purified p-coumarate decarboxylase enzyme from B.

bruxellensis and tested for substrate specificity with p-coumaric acid, caffeic acid and

ferulic acid. The decarboxylase was effective in metabolism of all three substrates with an

activity of 120 and 80% for caffeic and ferulic acids relative to that of p-coumaric acid.68

Similar results were observed by Edlin et al. for a hydroxycinnamate decarboxylase from

B. anomalus, although preferential breakdown was witnessed for caffeic acid, followed by

p-coumaric and ferulic with relative activities of 37.5 and 31.3%, respectively.70

The preferential breakdown of p-coumaric acid compared with that for ferulic acid, as

detailed in these two studies, occurred with relative activities of 1.25 and 1.20, rendering

unlikely the possibility that in this instance the decarboxylase could account for a favoured

formation of 4-ethylphenol over 4-ethylguaiacol by a factor of 12.85.

Therefore the substrate selectivity is presumably a product of the esterase activity of

AWRI 1499 which could be substantiated by the recovery of the remaining ethyl esters in

the fermentation samples using the GC-MS SIDA method as described by Sleep.141 This

method has been validated for quantification up to 10 mg/L in wine, and as such could be


109

employed without dilution of the fermentation samples. Combining the recovery of the

initially spiked ethyl esters with the evolution of ethylphenols provides a total recovery

(Figures 5.3 and 5.4).


Per

cen

t R

eco

vere

d

2 4 6 8 100

20

40

60

80

100 4-EthylphenolEthyl CoumarateTotal Recovery

Figure 5.3: Percentage recovery of coumarates in fermentations.


Per

cen

t R

eco

vere

d

2 4 6 8 100

20

40

60

80

100 4-EthylguaiacolEthyl FerulateTotal Recovery

Figure 5.4: Percentage recovery of ferulates in fermentations.

The total recovery at the conclusion of fermentation was approximately 80% for both the

ferulate series and the coumarate series. In addition to minor contributions by the acids and

vinylphenols, slight losses are expected through adsorption onto the yeast,93, 223 as well as

through loss of the volatile ethylphenols through the gas-lock. Although, with the

significant amount of remaining ethyl ferulate in the ferments, it can be concluded that, for

AWRI 1499, uptake and metabolism of ethyl coumarate is preferential over ethyl ferulate.

This suggests that the ethyl esterase activity of AWRI 1499 exhibits a substrate selectivity.


110

5.1.2 Ethyl Esterase Substrate Selectivity

As well as establishing the potential for spoilage caused by the breakdown of ethyl

coumarate (11), the ethyl ester substrate selectivity shown for AWRI 1499 has the potential

to accentuate the ratio of 4-ethylphenol:4-ethylguaiacol in wine. To determine whether this

selectivity is strain dependant or is common throughout D. bruxellensis, fermentation

experiments with strains representing the two remaining significant genetic groups (AWRI

1608 and AWRI 1613), in the presence of ethyl coumarate (11) and ethyl ferulate (12)

were conducted, along with a repeat fermentation with AWRI 1499. By using D.

bruxellensis strains AWRI 1499, 1608 and 1613, representatives of the three genetic

groups that contribute to 98% of Australian wine isolates were studied. End-point analyses

of 4-ethylphenol and 4-ethylguaiacol were conducted for all fermentations (Figure 5.5).

Per

cen

tag

e C

on

vers

ion

AWRI 1

499

AWRI 1

608

AWRI 1

613

0

20

40

60

804-Ethylphenol4-Ethylguaiacol

Figure 5.5: Percentage of the theoretical maximum conversion from ethyl coumarate (11)

and ethyl ferulate (12) to ethylphenols by different strains of D. bruxellensis.

The preference shown in the previous experiment by AWRI 1499 was observed again, with

ethyl coumarate metabolised over ethyl ferulate by a factor of 8.75. In this instance the

fermentations were conducted over a longer period as the evolution of ethylphenols was

still showing an upward trend after 9 days in the previous experiment (Figure 5.2). AWRI

1608 also displayed a preferential metabolism for ethyl coumarate, however in this case by

a factor of 18.75, due largely to lesser production of 4-ethylguaiacol.

D. bruxellensis AWRI 1613 did not convert either ethyl ester to the respective ethylphenol.

Subsequent analysis of the end-point fermentation samples by HPLC using ethyl

coumarate and ethyl ferulate as external standards indicated that both esters remained at


111

the same concentration at which they were spiked. The experiment with AWRI 1613 was

repeated to confirm these findings, and again did not metabolise the ethyl esters, indicating

an inability to express the necessary esterase activity to facilitate this breakdown.

Greater concentrations of 4-ethylphenol over 4-ethylguaiacol in wine is well

documented,41, 63 and has been attributed to the relative amounts of precursors present in

the grapes. However, these results show that not only will this ratio be defined by the

relative amounts of free hydroxycinnamic acids present in the berry, but also by the

relative amounts of ethyl esters produced during vinification. A 10:1 ratio of 4-

ethylphenol:4-ethylguaiacol could imply an initial 10:1 ratio of p-coumaric acid:ferulic

acid, but it could also imply a 1:1 ratio of ethyl coumarate:ethyl ferulate and a 10:1

selectivity in the metabolism of the esters, or a combination of these two effects.

As the ethyl esters are able to be metabolised by some strains of D. bruxellesis and

contribute to the accumulation of ethylphenols, the esterification of the free acids during

alcoholic fermentation is not a means to avoiding spoilage caused by D. bruxellensis.

Instead, this confirms that additional factors need to be considered when assessing a wine

for the potential production of ethylphenols, such as the extent of formation of esters as

well as the strain of D. bruxellensis that proliferates, given that in the presence of AWRI

1613 the ethyl esters are benign in the accumulation of ethylphenols.

5.1.3 Tartrate Esters

O O

OH

HOOC

COOH

OH

O O

OH

HOOC

COOH

OH

OCH3

7 8

Figure 5.6: p-Coumaroyl L-tartrate (7) and feruloyl L-tartrate (8).


112

Fermentation experiments with AWRI 1499 in the presence of p-coumaroyl and feruloyl

tartrate (7 and 8) yielded no conversion to 4-ethylphenol (4-EP) or 4-ethylguaiacol (4-EG)

after 10 days of fermentation, and to ensure the legitimacy of the result the experiment was

repeated, and identical results observed. As such, the ability of different strains of D.

bruxellensis to metabolise the tartrate esters were tested in fermentation experiments with

AWRI 1608 and AWRI 1613. Analysis for ethylphenol content in the end-point

fermentation samples again failed to detect either 4-ethylphenol (4-EP) or 4-ethylguaiacol

(4-EG) for both strains examined (Table 5.1).

Given, the dilutions used prior to analysis, and the limit of detection for the methodology

at 10 µg/L, the minimum detectable concentration of ethylphenols in the AWRI 1499

fermentations corresponds to 1.25% conversion from the tartrate esters. Detection in the

AWRI 1608 and AWRI 1613 fermentations was limited at around 0.7% conversion, having

employed a smaller dilution factor in preparation of the samples for analysis, due to the

lack of ethylphenol production observed in the initial experiments.

Table 5.1: Ethylphenol content in tartrate ester fermentation experiments.

4-EP 4-EGAWRI 1499 rep. 1 N.D. N.D.AWRI 1499 rep. 2 N.D. N.D.

AWRI 1608 N.D. N.D.AWRI 1613 N.D. N.D.

As outlined in Chapter 1, with the abundance of the tartrate esters, they are often the major

hydroxycinnamates found in grape juice, although the potential for them to contribute to

ethylphenol spoilage is now understood to be somewhat limited and unlike the ethyl esters

appears not to be determined by the yeast strain.

While D. bruxellensis appears to lack the capability to hydrolyse the tartrate esters, and

hence they are unlikely to contribute to the accumulation of the ethylphenols during barrel

ageing, the tartrate ester content of the berries still remains a source of the free

hydroxycinnamic acids (3 and 4). The previously mentioned studies of Dugelay and

Gerbeaux showed that the tartrates could be hydrolysed using commercial enzyme

preparations,64, 112 and there is a loss observed during vinification,97, 99-105, 129-130, 132 as such

the hydroxycinnamoyl tartrate esters (7 and 8) cannot be completely ignored during the

entire vinification process.


113

5.1.4 Glucose Esters

OO

OH

OO

OH

H3CO

9 10

O O

OH

HOHO

OH

OH

HOHO

OH

Figure 5.7: p-Coumaroyl glucose (9) and feruloyl glucose (10).

The synthetic samples of the glucose esters (9 and 10) were initially spiked as mixtures of

multiple esters, with the majority existing as the 1-O-β-, 2-O-α- and 6-O-α-esters, though

the 1-O-β-esters were found to be thermodynamically favoured in wine-like environments,

as shown by theoretical studies (Chapter 3) and by their prevalence during storage in

fermentation-like conditions (Chapter 2). Regardless of the ratio of esters present at the

time of fermentation, the release of the free hydroxycinnamic acids must be achieved by

hydrolysis of a hydroxycinnamoyl glucose ester. Fermentation studies with AWRI 1499

were concluded after 16 days, and all samples that were taken throughout the experiment

were analysed for content of the ethylphenols (Figure 5.8).


Per

cen

tag

e C

on

vers

ion

4 8 12 160

10

20

30

40 4-Ethylphenol4-Ethylguaiacol

Figure 5.8: Percentage of the theoretical maximum conversion of hydroxycinnamoyl

glucose esters (9 and 10) to ethylphenols.


114

At the conclusion of fermentation, 4-ethylphenol and 4-ethylguaiacol were detected at

levels corresponding to 35% conversion from the glucose esters (9 and 10), and remained

relatively constant from day 10 onwards. Unlike the ethyl esters, p-coumaroyl glucose and

feruloyl glucose were metabolised to a similar extent (Figure 5.8).

Due to the previously described stability of the glucose esters in fermentation-like

conditions, both experimentally observed in Chapter 2 and reported in the literature,100 it

can be concluded that the evolution of the ethylphenols is not a product of chemical

hydrolysis of the glucose esters followed by the expected metabolism of the acids. The

moderate conversion of the glucose esters to ethylphenols, along with the observed glucose

ester concentrations in wine (outlined in Chapter 1), indicate that while they can contribute

to the accumulation of ethylphenols during barrel ageing, the metabolism of the glucose

esters alone would not have major effects on the organoleptic properties of the resulting

wine.

5.1.5 Conclusions for Chapter 5.1

For the first time, this work has shown the ability of D. bruxellensis to metabolise

esterified hydroxycinnamic acids directly to ethylphenols. Furthermore, the differences in

breakdown observed between different classes of esters implies the presence of multiple

pathways, or enzyme activities, involved in the release of free hydroxycinnamic acids from

an esterified form. The formation of 4-ethylphenol and 4-ethylguaiacol from the ethyl

esters (11 and 12) shows an overall preference for the breakdown of ethyl coumarate,

though the reasons for stereoselective bioconversion remain to be identified. The high

proportions of ethyl ferulate in the fermentation samples could be a result of a decrease in

transport into the yeast cell for an intracellular esterase, equivalent transport into the cell

but decreased conversion due to the nature of the enzyme, or selective activity of an

extracellular esterase.

The inability of D. bruxellensis to metabolise the tartrate esters (7 and 8), and the moderate

conversion observed for the glucose esters, mean that apart from the metabolism already

known for p-coumaric and ferulic acids, ethyl coumarate has the largest potential to

contribute to wine spoilage, having been found at concentrations high enough to generate

sufficient ethylphenols to affect wine aroma and flavour. However, the formation and


115

metabolism of ethyl coumarate depends on many variables which would need to be

considered. Initially, the concentration of p-coumaric acid in the wine will be determined

both by that present in the berry, as well as the release from p-coumaroyl tartrate through

enzymatic or chemical hydrolyses. Following this, the formation of ethyl coumarate from

p-coumaric acid via esterification with ethanol could be effected by the enzymatic abilities

of the wine microflora (transferases), or by conditions affecting the equilibrium of a

chemical esterification, or trans-esterification, such as pH and ethanol concentration.

Finally, as shown above (Figure 5.5), the strain of D. bruxellensis can determine the

conversion of ethyl coumarate to 4-ethylphenol, if any.

The glucose esters could contribute to the accumulation of ethylphenols in barrels, though

the concentrations in which they are present they are unlikely to be able to produce

sufficient ethylphenols to cause wine spoilage on their own. However, in addition to the

contribution from the ethyl esters and the free acids, the glucose esters could be the

determining factor in whether the ethylphenol content is above or below the perception

threshold.

It can now be seen that not only the free acids contribute to the accumulation of

ethylphenols, but that certain hydroxycinnamate esters are also potential sources of

spoilage, although this cannot be assumed for all hydroxycinnamates. With the extent of

bioconversion differing between substrate, class of ester and strain of D. bruxellensis, the

breakdown of each esterified hydroxycinnamate must be tested individually and strain

dependencies must be considered.

5.2 Stereoselectivity of D. bruxellensis Enzyme Activities.

5.2.1 Decarboxylase Stereoselectivity

Following the photoisomerisation of the hydroxycinnamoyl glucose esters during

synthesis, the ability of D. bruxellensis to produce ethylphenols through the metabolism of

cis-hydroxycinnamates was of interest. As the breakdown of all hydroxycinnamates to

form ethylphenols must proceed via decarboxylation of the acid, synthesis of the cis-acids

was attempted (Chapter 4). The inability to synthesise a pure sample of cis-ferulic acid


116

resulted in the use of an isomeric mixture of known stable ratio was chosen. In the case of

ferulic acid (4) a thermodynamically stable 50:50 mixture of the cis- and trans-isomers

was achieved, whereas p-coumaric acid (3) existed as a 39:61 mixture of the cis- and trans-

isomers. The analytical technique most readily available for real-time determination of the

isomeric ratio during fermentation was NMR. Due to overlap of crucial proton shifts

between p-coumaric and ferulic, the decarboxylase stereoselectivity could not be examined

for both acids concurrently. As ferulic acid was obtained with an equal cis:trans-ratio, it

was tested prior to p-coumaric acid, and the isomeric mixture was submitted to

fermentation studies with AWRI 1499 (cis/trans-fermentations) and compared with

fermentations spiked with pure trans-ferulic acid (trans-fermentations).

The isomeric ratio during the fermentation experiments was monitored by extracting

uninoculated controls and analysing by NMR. Enough of each acid needed to be present in

the uninoculated controls to ensure that adequate spectra could be obtained to determine

the isomeric ratio. Thus, allowing for losses during extraction, 10 mg of the mixture was

required in each control (200 mL), which equated to initial spiked concentrations of 50

mg/L. As Kahnt described the changes in isomerisation equilibrium with changing

substrate concentrations,187 to avoid any effects of concentration on the isomeric ratio the

inoculated fermentations were also spiked at 50 mg/L, five times higher than previously

used for the trans-hydroxycinnamate esters. Throughout fermentation the ratio of

cis:trans-ferulic acid in the uninoculated controls remained stable, and the results for 4-

ethylguaiacol analysis of samples taken throughout fermentation are given below (Figure

5.9).


Per

cen

tag

e C

on

vers

ion

2 4 6 8 100

20

40

60

80 trans-Fermentationscis/trans-Fermentations

Figure 5.9: Percentage of the theoretical maximum conversion to 4-ethylguaiacol for the

trans- and cis/trans- fermentations.


117

The fermentations supplemented with a mixture of 50:50 cis:trans-ferulic acid produced

approximately half as much 4-ethylguaiacol as the fermentations containing only trans-

ferulic acid. The evolution of 4-ethylguaiacol in the cis/trans-fermentations as a percentage

of the maximum conversion observed in the trans-fermentations is displayed in Figure

5.10, and shows more clearly the 50% reduction of 4-ethylguaiacol produced during the

cis/trans-fermentations.


Per

cen

tag

e o

ftra

ns-F

erm

ent

2 4 6 8 100

10

20

30

40

50

Figure 5.10: Evolution of 4-ethylguaiacol in cis/trans-fermentations as a percentage of

maximum conversion observed in trans-fermentations.

While these results strongly suggest that cis-ferulic acid is not at all metabolised by D.

bruxellensis, quantification of the remaining ferulic acid in the fermentation samples would

confirm the inability to metabolise cis-ferulic acid.

Ferulic acid quantification techniques as outlined in a review by Barberousse et al. largely

employ reverse-phase HPLC, methanol-water-acid ternary solvent systems and run times

in excess of 20 minutes.224 However, a method previously developed by the AWRI

describes quantification of p-coumaric and ferulic acid using ion-exchange HPLC,

resulting in retention times of around 3 minutes for trans-ferulic acid (unpublished

method). Although analysis times are reduced using ion-exchange HPLC, the previous

analysis did not consider both isomers of ferulic acid and required optimisation to achieve

resolution of the cis- and trans-ferulic acid.

Using formic acid in water (0.1:99.9, solvent A) and formic acid in acetonitrile (0.1:99.9,

solvent B), the AWRI method (45% B, isocratic elution) gave very little separation of the


118

isomers. This method was further refined, requiring less of the organic solvent (A), with an

isocratic profile of 30% B resulting in similar separation to less organic systems, but with

significantly shorter run times.

To adequately quantify both trans- and cis-ferulic acid, a calibration curve for each had to

be produced. Preparing trans-ferulic acid samples in the concentration range of 1-75 mg/L

gave calibration curves with correlation coefficients of 0.9998 and 0.9973 for absorbance

at 280 and 320 nm, respectively. To overcome the instability of pure cis-ferulic acid,

cis/trans-ferulic acid mixtures of differing, yet known, concentrations and ratios were

produced and analysed. A maximum concentration of 75 mg/L was used with a 50:50

isomeric ratio equating to a maximum cis-ferulic acid content of 37.5 mg/L. Using the

trans-ferulic acid calibration curves the trans-isomer could be quantified in the mixture,

and the remaining cis-ferulic acid in the sample determined, giving calibration curves with

correlation coefficients of 0.9968 and 0.9937 for 280 and 320 nm, respectively. While a

number of different wavelengths were originally used for detection (280, 320, 353, 370 and

520 nm), only 4 of these gave reliable calibration curves for trans-ferulic acid (280, 320,

353 and 370 nm), and only two of those gave reliable calibration curves for cis-ferulic

acid, 280 and 320 nm, the most common wavelengths used to quantify ferulic acid.224

With initially spiked concentrations of 50 mg/L for trans-ferulic acid and 25 mg/L for cis-

ferulic acid (50% of a 50 mg/L spike), preparation of the calibration curves from 1-75

mg/L and 0.5-37.5 mg/L, respectively allowed for the fermentation samples to be analysed

without dilution. After the first 6 samples, overlap of the cis- and trans-ferulic acid peaks

occurred, caused by the drift in retention time of cis-ferulic acid. After refreshing the

column, the analysis could be resumed, though after analysis of the calibration samples and

only a few additional fermentation samples, peak overlap was experienced again. Ion-

exchange HPLC was not an adequate method of analysis for multiple samples due to the

limited number of samples that could be run before the column required refreshing.

Fortuitously, the original 6 samples analysed were the final (triplicate) samples of each of

the trans- and cis/trans-fermentations, as shown in Table 5.2.


119

Table 5.2: Concentration of cis- and trans-ferulic acid in end-point fermentation samples.

trans-Ferulic acid cis-Ferulic acid trans-Ferulic acid cis-Ferulic acid

trans-Ferments 1.02 ± 0.05 2.23 ± 0.11 50.0 0.0cis/trans-Ferments 0.44 ± 0.07 12.43 ± 0.07 25.0 25.0

Concentration (mg/L) Initial spike (mg/L)

The final ferulic acid concentrations (expressed as the percentage remaining), along with

the initial 4-ethylguaiacol quantifications (expressed as percentage conversion) are

combined to give total recoveries of approximately 70% (Figure 5.11). For the trans-

fermentations, most of the spiked acid is converted to 4-ethylguaiacol (70%) with around

3% remaining as ferulic acid, existing as 2% trans- and 1% cis-ferulic acid. As expected,

metabolism of trans-ferulic acid had occurred to a great extent leaving very little acid,

though some had isomerised to give a small cis-content. The cis/trans-fermentations

contain 4-ethylguaiacol concentrations corresponding to 40% conversion from the spiked

acid, with around 30% remaining as ferulic acid (25% cis-isomer and 5% trans-isomer).

Per

cen

tag

e R

eco

very

trans-

Ferm

ents

cis/tr

ans-

Ferm

ents

0

20

40

60

804-Ethylguaiacoltrans-Ferulic Acidcis-Ferulic Acid

Figure 5.11: Compounds by percentage in end-point fermentation samples.

Recovery of cis-ferulic acid was approximately 50% of the spiked 25 mg/L indicating a

potential for uptake by the yeast, but an inability to decarboxylate it. However, as

incomplete recovery of cis-ferulic acid was observed, the inability of AWRI 1499 to

metabolise cis-ferulic acid cannot be definitively confirmed.

The above fermentation experiment was repeated under the same conditions with cis- and

trans-p-coumaric acid (3) to determine whether the same effect was observed, fermenting


120

D. bruxellensis in the presence of a 39:61 cis:trans-mixture alongside trans-p-coumaric

acid. The evolution of 4-ethylphenol in both sets of fermentations was monitored

throughout the experiment (Figure 5.12), also the isomeric ratio in the blanks were

observed to be stable by NMR.


Per

cen

tag

e C

on

vers

ion

to

4-E

P

2 4 6 8 10 120

20

40

60

80trans-Fermentscis/trans-Ferments

Figure 5.12: Percentage of the theoretical maximum conversion to 4-ethylphenol in trans-

and cis/trans- fermentations.

In an analogous fashion to the ferulic acid fermentations, conversion to 4-ethylphenol in

the cis/trans-p-coumaric acid fermentations corresponded to the trans-content, giving final

4-ethylphenol concentrations 39% lower than observed in the trans-fermentations (Figure

5.13).


Per

cen

tag

e o

ftr

ans-

Fer

men

t

4 8 120

20

40

60

Figure 5.13: Evolution of 4-ethyphenol in cis/trans-fermentations as a percentage of

maximum conversion observed in trans-fermentations.


121

Slight deviations in 4-ethylphenol concentration may arise from a) partial isomerisation of

cis-p-coumaric acid back to the trans-isomer, b) from slight loss of 4-ethylphenol in the

trans-ferments due to the high concentrations produced, or c) from slight variations in the

amount of p-coumaric acid spiked into, or remaining in the fermentations produced from

solubility issues, caused by performing these fermentation experiments at higher

concentrations than employed previously.

The quantification of ferulic acid in the previous experiments did not provide additional

evidence as to the role of the cis-acid, with the final conclusions based purely on the

isomeric ratio of the spikes compared with the ratio of produced 4-ethylguaiacol. As such,

quantification of the cis- and trans-p-coumaric acids was not performed for these

fermentations. However, these results confirm that the metabolism of p-coumaric acid

occurs to a much greater extent for the trans-isomer and the role of the cis-acids in the

production of ethylphenols is not significant.

5.2.2 Ethyl Esterase Stereoselectivity

During the synthesis of the cis-hydroxycinnamic acids (3 and 4) detailed in Chapter 4, the

cis-ethyl esters were found to be stable under ambient light conditions. As the ability of D.

bruxellensis to metabolise the ethyl esters had been shown, if the cis-ethyl esters could also

be enzymatically hydrolysed, they were a potential source of pure cis-hydroxycinnamic

acids. Thus, cis-ethyl coumarate (11) and cis-ethyl ferulate (12) were simultaneously

submitted to fermentation experiments with AWRI 1499 and the evolution of ethylphenols

monitored (Figure 5.14).


122


Per

cen

tag

e C

on

vers

ion

2 4 6 8 100.0

0.2

0.4

0.6

0.8

1.04-Ethylphenol4-Ethylguaiacol

L.o.Q.

Figure 5.14: Percentage of the theoretical maximum ethylphenol conversion from cis-ethyl

esters.

The amounts of 4-ethylphenol and 4-ethylguaiacol produced from metabolism of the cis-

ethyl esters were so low that some data points were below the limit of quantification

(L.o.Q.) for the analytical method (10 µg/L). The values that lay under the L.o.Q. are

approximated only but do indicate an upward trend that continues to above the limit of

detection.

Conversion to the ethylphenols under the same conditions (9 days of fermentation, Figure

5.2) was observed at 51 and 4% from trans-ethyl coumarate and trans-ethyl ferulate,

respectively. The production of ethylphenols observed during the cis-ethyl ester

fermentations, though minimal, could be explained by either the esterase being active

towards the cis-ethyl esters, or by partial isomerisation to yield trans-ethyl esters. For

production of ethylphenols via metabolism of the cis-ethyl esters, the resulting cis-acids

must isomerise and the trans-acids then metabolised. Otherwise the cis-ethyl esters

partially isomerised during the experiment to give a small amount of the trans-ethyl esters,

which were then broken down, or the samples of cis-ethyl esters used to spike the

fermentations contained trace impurities of the trans-isomers. The difference in conversion

observed between the cis-ethyl coumarate and cis-ethyl ferulate would suggest the latter,

brought about by the esterase substrate selectivity established earlier. Using the

quantification method of Sleep, the trans-ethyl esters in the fermentation samples could be

determined141 and the method was also applied in an attempt to quantify the cis-ethyl

esters.


123

Table 5.3: Final trans-ethyl ester content in cis-ethyl ester fermentations.

Initial cis-esterInoculated Uninoculated spike (mg/L)

Ethyl coumarate 1.1 ± 0.1 1.5 ± 0.1 10.0Ethyl ferulate 1.8 ± 0.1 1.8 ± 0.1 10.0

trans-Ester concentration (mg/L)

The uninoculated samples contain a minor but significant proportion of the trans-ethyl

esters which is not expected to occur purely over the course of the fermentation

experiments, and is most likely a result of the length of storage (due to instrument

availability the fermentation samples had to be stored for 9 months prior to analysis of the

ethyl esters content). If conversion from cis-ethyl coumarate to trans-ethyl coumarate had

occurred during fermentation a much larger concentration of 4-ethylphenol would be

expected, based on the conversions observed in previous experiments.

However, the minor differences in the amounts of trans-ethyl coumarate between the

inoculated and uninoculated fermentations suggest that minor conversion could have

occured during fermentation, followed by metabolism of the resulting trans-ethyl

coumarate to yield 4-ethylphenol.

The cis-ethyl esters could not be accurately quantified using the same method as the trans-

esters due to differences in both the extraction efficiencies and the mass spectral responses.

Mixtures of known cis- and trans-ethyl ester ratios were analysed by GCMS to determine

the differences in ionisation potentials, then mixtures were extracted and analysed to

determine the differences in responses, which were then factored into the quantification to

afford a more accurate indication of the cis-ethyl ester concentration. When combined with

the trans-ethyl ester and the ethylphenol quantifications, the total content at the end of

fermentation could be determined (Figures 5.15 and 5.16).


124

Per

cen

tag

e R

eco

very

Inocu

lated

Uninocu

lated

0

20

40

60

80

100 4-Ethylphenoltrans-Ethyl coumaratecis-Ethyl coumarate

Figure 5.15: Total coumarate recovery from cis-fermentations.

Per

cen

tag

e R

eco

very

Inocu

lated

Uninocu

lated

0

20

40

604-Ethylguaiacoltrans-Ethyl ferulatecis-Ethyl ferulate

Figure 5.16: Total ferulate recovery from cis-fermentations.

While the recovery for the coumarates were good, this was not the case for the ferulates.

Rough GCMS quantifications of ethyl ferulate (not specific to either isomer) from a single

fermentation throughout the experiment were performed using an abbreviated SIDA

method (Figure 5.17).


125

Fermentation Progress (Days)E

thyl

Fer

ula

te (

mg

/L)

0 2 4 6 8 100

2

4

6

8

10

Figure 5.17: Breakdown of ethyl ferulate in a single fermentation.

The spiked cis-ethyl ferulate within the inoculated fermentation samples was slowly

degraded, and as the final concentration of cis-ethyl ferulate was reduced in both

inoculated and uninoculated ferments it can be assumed a chemical breakdown rather than

a microbiological breakdown took place. With many potential causes, it was not further

investigated. In any case the role of cis-ethyl ferulate in the production of 4-ethylguaiacol

is negligible. Even allowing for a 50% reduction in cis-ethyl ferulate in the fermentation,

the amount of 4-ethylguaiacol produced would correspond to a conversion of no more than

0.5%.

5.2.3 Conclusions for Chapter 5.2

The results of the first part of this study, limited by the reduced recovery of the spiked cis-

acids, indicate that the decarboxylase of D. bruxellensis is specific to metabolism of the

trans-hydroxycinnamic acids (3 and 4). However, from the isomeric ratio of the spiked

acids and the subsequent production of ethylphenols, the decarboxylase of D. bruxellensis

has little or no activity towards the cis-acids. For the cis-acids to be completely disregarded

as precursors to the ethylphenols one of two things must be achieved: either the

experimental conditions must be improved to allow for a pure cis-acid spike and it remains

isomerically pure; or the full recovery of the cis-acids from isomeric mixtures needs to be

achieved. The nature of the photoisomerisation does not allow for use of a pure cis-acid

spike, and an attempt to use one would most likely result in slow conversion back to the

trans-acid, giving a dynamic mixture that would be difficult to characterise over the course

of these experiments. The incomplete recovery of the cis-acids is most likely a result of

adsorption onto the yeast cell wall or by uptake, which could be overcome by examining


126

the enzyme directly, through isolation and purification, as opposed to examining the

indirect products of the activity of the decarboxylase enzyme which has been done in this

study.

Even though the cis-hydroxycinnamic acids cannot be completely eliminated as precursors

to ethylphenols, the conversions observed from the isomeric mixtures strongly suggest that

they are not metabolised by D. bruxellensis, and implies that the decarboxylase possesses

stereospecificity.

In reviewing the content of the cis-hydroxycinnamates in grapes and wine (Chapter 4) it

was noted that the tartrate esters exist in both isomeric forms, with the cis-esters

contributing around 20%. While it has been shown that the tartrate esters are not

metabolised by D. bruxellensis, rendering them of little relevance to the direct

stereoselective breakdown, the tartrate esters are hydrolysed either enzymatically and/or

chemically prior to barrel ageing (described in Chapter 1). Assuming that the cis-tartrate

esters also undergo the same hydrolyses, then the production of cis-acids would not lead to

further metabolism by D. bruxellensis.

By displaying stereoselectivity in decarboxylase activity, the cis-hydroxycinnamate

content of wine, esterified or otherwise, can be ignored in relation to the build-up of

ethylphenols during barrel ageing, as cis-hydroxycinnamate esters could either be

enzymatically hydrolysed, which may be limited, or chemically hydrolysed yielding cis-

acids which will not contribute to the production of ethylphenols (unless enough time is

allowed between hydrolysis and D. bruxellensis growth for some trans-acid to form).

5.3 Thesis Conclusions and Future Directions.

This study has investigated the synthesis and chemical transformations of

hydroxycinnamate esters, their role in the production of ethylphenols in wine by D.

bruxellensis and the stereochemical factors contributing to the enzymatic breakdown of

hydroxycinnamates in wine.


127

In Chapter 2, the synthesis of p-coumaroyl and feruloyl tartrate (7 and 8) has been

described for the first time, via the coupling of the 1-O-chloroacetyl protected acids and di-

tert-butyl tartrate. This has allowed them to be dismissed in the accumulation of

ethylphenols by D. bruxellensis unless first hydrolysed via an alternative pathway.

Moreover, the synthetic methodology developed can now be applied for the synthesis of

other hydroxycinnamoyl tartrate derivatives such as caffeoyl tartrate ester or the grape

reaction product.225 Furthermore, incorporating isotopically labelled hydroxycinnamate

moieties into the synthesis would allow for accurate and convenient quantification of these

compounds in grapes and wine, an addition that could also be applied to the glucose esters.

Synthesis of the glucose esters from the 1-O-chloroacetyl protected acids and a

trichloroacetimidate glucosyl donor and deprotection in pyridine/water resulted in both

photoisomerisation and acyl migration, which have not previously been detailed for the

hydroxycinnamoyl glucose esters. During the course of this study the synthesis of feruloyl

glucose (10) using a similar synthetic pathway was described by Zhu,183 although no

migration or photoisomerisation was described. Theoretical studies into the migration of

the glucose esters identified the role of non-aqueous conditions in altering the

thermodynamic preference for different esters, and in changing kinetic aspects of the

migration away from the desired 1-O-β-esters (Chapter 3). Investigation into the migration

during previously reported analytical procedures showed that the extraction process can

have an effect on the ratios of esters present, but that wine naturally contains multiple

esters for both p-coumaroyl glucose and feruloyl glucose. During these studies, feruloyl

glucose was identified for the first time in red wine, which is rationalised to have gone

unnoticed by other authors due to coelution with p-coumaroyl derivatives tentatively

identified as anthocyanin derivatives.

As outlined in Chapter 4, photoisomerisation of the glucose esters under ambient light

conditions was found to be accelerated in comparison to other synthesised

hydroxycinnamate derivatives, resulting in production of cis-glucose esters in roughly 1:4

ratio with the trans-isomers. For the duration of synthesis this was subdued by working

under red light. Theoretical investigations into the photoisomerisation found that while

solvent and substrate do play a minor role in the ease of isomerisation, the nature of the

phenol and carboxyl contribute to a much larger extent. By increasing the electron

donating character of the phenol or increasing the electron withdrawing character of the


128

carboxyl, electron movement within the molecule was increased resulting in a decrease in

the α,β-double bond length. This reduced the HOMO-LUMO gap of the molecule,

affecting the S0-T1 vertical excitation energy and altering the energy required for

photoisomerisation. To avoid photoisomerisation, production of the phenolate should be

prevented either by using phenolic protection as far into a synthetic route as possible, or by

preventing exposure to basic conditions.

The ease by which the glucose esters isomerised encouraged the investigation into the

ability of D. bruxellensis to metabolise cis-hydroxycinnamates in addition to the trans-

isomers. Attempts at synthesising pure cis-p-coumaric and ferulic acids were unsuccessful,

with pure cis-ethyl esters undergoing isomerisation during base-catalysed ester hydrolysis,

assisted by formation of the phenolate of the ethyl esters. Instead stable isomeric mixtures

were produced with ferulic acid existing in a 50:50 ratio of cis:trans-ferulic acid and p-

coumaric as a 39:61 ratio, which could be used in fermentation experiments, along with the

pure cis-ethyl esters that were synthesised and remained pure under ambient light

conditions.

Fermentation of the ethyl esters with multiple strains of D. bruxellensis showed that a

substrate selectivity exists in some strains for the preferential breakdown of ethyl

coumarate over ethyl ferulate, which was observed in two strains representing nearly 92%

of Australian D. bruxellensis isolates, while a third strain was unable to metabolise either.

This work showed the ability of D. bruxellensis to form ethylphenols directly from

esterified hydroxycinnamates for the first time, also identifying a substrate preference that

could contribute to the 4-ethylphenol:4-ethylguaiacol ratio seen in red wines. The

bioconversion of ethyl coumarate and subsequent production of 4-ethylphenol has also

been established.

The tartrate esters, while abundant enough to have a large impact on the ethylphenol

content of wine, were unable to be metabolised by D. bruxellensis, showing that for this

compound to be of significance it must first be hydrolysed by alternative enzymatic

methods, or chemically. The inability of D. bruxellensis to breakdown the tartrates was

common to all three strains tested.


129

Metabolism of migrated mixtures of p-coumaroyl glucose and feruloyl glucose by D.

bruxellensis showed a moderate 35% conversion to the ethylphenols. Given the previous

quantifications in wine, the glucose esters have the ability to contribute to the production

of ethylphenols, but are not abundant enough to solely spoil wine.

In fermentation experiments to examine the stereoselectivity of the decarboxylase activity,

metabolism of cis-ferulic and cis-p-coumaric acids to yield 4-ethylguaiacol and 4-

ethylphenol, was at most limited. The inability to recover all of the spiked cis-acid in

conjunction with the nature of the isomerisation did not allow determination of whether the

decarboxylase activity towards the cis-acids was non-existent or just very small. A similar

result was achieved in testing the ethyl esterase for stereoselectivity, indicating that the cis-

ethyl esters were not broken down, with minor production of ethylphenols attributed to

isomerisation or the presence of trace impurities. Regardless, both decarboxylase and ethyl

esterase were more active towards the trans-isomers, rendering the cis-hydroxycinnamates

unimportant in the accumulation of ethylphenols in wine.

Following the findings that hydroxycinnamates other than the free acids can contribute to

ethylphenol accumulation, further conjugates should be tested in order to determine all

metabolic inputs in the production of 4-ethylphenol and 4-ethylguaiacol. In addition to

those tested in this study, other derivatives have been identified which have the potential to

contribute to a greater extent than those shown here. During the LCMS studies of red wine,

a peak attributed to the glucoside of p-coumaric acid was identified, a compound which

has been identified in red wine and isolated from white wine previously,108, 131 and

appeared to more abundant than the glucose esters. Furthermore, p-coumaroyl anthocyanin

derivatives have been identified in red wine,107 which could also be examined as potential

precursors to ethylphenols.

Regardless of the hydroxycinnamate under investigation, the work here has shown that

metabolism of each substrate by D. bruxellensis must be tested individually, as the result

can be dependent on the substrate, the stereochemistry, and the strain used. While much

has been reported regarding the breakdown of trans-p-coumaric and trans-ferulic acid, and

this thesis has detailed the role of esterified precursors to ethylphenols and the

stereoselective metabolism by D. bruxellensis, the potential for spoilage of this species in

wine still requires further investigation.

Chapter 6: Experimental

130

Chapter 6: Experimental.

6.1 General Experimental.

Solvents and reagents

Dry organic solvents were purchased and dispensed using a Puresolv™ solvent purification

system. Pyridine was dried by storage on 4Å molecular sieves and tert-butanol was

distilled from CaH2 onto 4Å molecular sieves. General organic solvents were obtained and

distilled where needed. Reagents other than those synthesised were purchased from Sigma-

Aldrich Chemical Company Ltd. and used without further purification.

Naming of synthesised compounds

Compounds are named using common nomenclature as they would appear in literature,

followed by the IUPAC name as generated using ACD/Labs 12.0 software. Where alkene

stereochemistry is not denoted the trans-isomer was produced. For the assignment of NMR

shifts, the numbering systems are shown below.

O OH

OH

O

HO

O

OH

O

OH

O

OH

HOHO

OH

O

12

34

5

6

78

9

1'2'3'

4' 5'

6'1'

2'

3'4'

Chromatography

Column chromatography was performed using Davisil 40-63 µm silica gel. Thin layer

chromatography was performed using Merck silica gel 60 F254 alumina sheets (20 x 20 cm)

and viewed under UV light.


131

Melting points

Melting points were obtained using a Buchi B-540 melting point apparatus and are

uncorrected.

Optical Rotation

Rotation was measured with a polAAr 21 polarimeter and referenced to the sodium D line

(589 nm) at 20 oC in a cell with 1 dm path length. The concentration is specified in g/100

mL and in the solvent as reported.

Infrared spectroscopy

Spectra were acquired with a Perkin Elmer Spectrum One FT-IR spectrometer using neat

samples.

High-resolution mass spectrometry

Accurate mass determination was performed by The Organic Mass Spectrometry Facility,

University of Tasmania. Where an appropriate spectrum was obtained prior to accurate

mass determination, other significant fragmentations are quoted.

Mass spectra of compounds 9 and 10

Spectra were obtained during LCMS studies of 9 and 10 in wine, as detailed in Chapter 6.3

(Experimental Procedures for Chapter 3).

Elemental analysis

Analysis was performed at the University of Otago, New Zealand.

X-ray crystallography

Crystallographic data was performed by Dr Edward R. T. Tiekink at the Department of

Chemistry, University of Malaysia.


132

NMR spectroscopy

The 1H and 13C spectra were acquired with either a Bruker Ultrashield Plus 400 MHz

Spectrometer or a Bruker Ultrashield Plus 600 MHz Spectrometer, where indicated.

Spectra were recorded in the specified solvent, and referenced as described by Gottlieb et

al.226 In cases of overlapping solvent and compound shifts, the spectra were referenced to

the TMS peak at 0.00 ppm. Chemical shifts (δ ) are reported in ppm and coupling constants

(J) in Hz. 13C assignments were made using 2D correlation experiments HMQC and

HMBC.

Photoisomerisation

Was carried out using either a 365 nm or 254 nm ultra-violet lamp, as specified.

Computational Chemistry

Was performed using Spartan ’08 software package, with final calculations employing

density functional theory (DFT) and the supplied B3LYP 6-31G* basis set.


133

6.2 Experimental Procedures for Chapter 2.

6.2.1 Hydroxycinnamoyl Derivatives

(Carbethoxymethylene)triphenylphosphorane

Ethyl bromoacetate (17 mL, 0.15 mmol) and triphenylphosphine (40.31 g, 0.15 mmol)

were heated under reflux in toluene (150 mL) for 15 hours. The resulting precipitate was

filtered and washed with toluene (3 x 50 mL). The prepared salt was then stirred with

sodium hydroxide (11.88 g, 0.30 mmol) in water (500 mL) for 1 hour. The product was

extracted with ethyl acetate, dried (Na2SO4) and concentrated to give

(carbethoxymethylene)triphenylphosphorane as a beige solid (44.50 g, 91%); m.p. 117-122 oC (lit. m.p. 126-127 oC).227 1H NMR: (400 MHz, CDCl3) δ: 7.69-7.64 (m, 6H, ArH), 7.57-7.52 (m, 3H, ArH), 7.48-

7.44 (m, 6H, ArH), 3.96 (q, 2H, J = 7.1 Hz, OCH2CH3), 2.80 (br. s, 1H, CH), 1.03 (t, 3H, J

= 7.1 Hz, OCH2CH3). Physical and spectral properties were as previously reported.227

Ethyl coumarate (11)

Ethyl 3-(4-hydroxyphenyl)prop-2-enoate

(Carbethoxymethylene)triphenylphosphorane (10.07 g, 28.91 mmol) and p-

hydroxybenzaldehyde (3.34 g, 27.37 mmol) were stirred in dry dichloromethane (70 mL)

under nitrogen at ambient temperature. After 10 days a further portion of

(carbethoxymethylene)triphenylphosphorane (8.01 g, 23.01 mmol) was added and the

mixture stirred for a further 12 days. The reaction mixture was concentrated and purified

using column chromatography (30% EtOAc/X4) which gave partial separation of isomers,

yielding 4.87 g (93%) of trans-ethyl coumarate as a white solid and 0.20 g (4%) of a

mixture of cis/trans-ethyl coumarate (9:1) as a colourless oil.

trans-Ethyl coumarate


134

Ethyl (2E)-3-(4-hydroxyphenyl)prop-2-enoate

m.p. 72.1-73.0 oC (lit. m.p. 73 oC).185

Rf (50% EtOAc/X4): 0.47 1H NMR: (400 MHz, CDCl3) δ: 7.63 (d, 1H, J = 15.9 Hz, H7), 7.42 (app. d, 2H, J = 8.6 Hz,

H3,5), 6.85 (app. d, 2H, J = 8.6 Hz, H2,6), 6.29 (d, 1H, J = 15.9 Hz, H8), 4.27 (q, 2H, J = 7.1

Hz, OCH2CH3), 1.33 (t, 3H, J = 7.1 Hz, OCH2CH3).

cis-Ethyl coumarate

Ethyl (2Z)-3-(4-hydroxyphenyl)prop-2-enoate

Rf (50% EtOAc/X4): 0.50 1H NMR: (400 MHz, CDCl3) δ: 7.63 (app. d, 2H, J = 8.6 Hz, H3,5), 6.85 (d, 1H, J = 12.7

Hz, H7), 6.80 (app. d, 2H, J = 8.6 Hz, H2,6), 5.83 (d, 1H, J = 12.7 Hz, H8), 4.21 (q, 2H, J =

7.1 Hz, OCH2CH3), 1.29 (t, 3H, J = 7.1 Hz, OCH2CH3). Data was extracted from the

mixture of isomers.

For both isomers, all physical and chemical properties were as previously reported. 185, 228

Ethyl ferulate (12)

Ethyl 3-(4-hydroxy-3-methoxyphenyl)prop-2-enoate

Reaction of vanillin (3.56 g, 23.41 mmol) using the same procedure as described for 11

(above), yielded 3.73 g (72%) of trans-ethyl ferulate as a pale yellow solid and 0.89 g

(17%) of a mixture of cis/trans-ethyl ferulate (8:2) as a yellow oil.

trans-Ethyl ferulate

Ethyl (2E)-3-(4-hydroxy-3-methoxyphenyl)prop-2-enoate

m.p. 39.6-41.2 oC (lit. m.p. 39 oC).185

Rf (50% EtOAc/X4): 0.45 1H NMR: (400 MHz, CDCl3) δ: 7.61 (d, 1H, J = 15.9 Hz, H7), 7.07 (dd, 1H, J = 8.2 and

1.9 Hz, H5), 7.03 (d, 1H, J = 1.9 Hz, H3), 6.91 (d, 1H, J = 8.2 Hz, H6), 6.28 (d, 1H, J = 15.9

Hz, H8), 4.27 (q, 2H, J = 7.1 Hz, OCH2CH3), 3.92 (s, 3H, OCH3), 1.33 (t, 3H, J = 7.1 Hz,

OCH2CH3).

cis-Ethyl ferulate


135

Ethyl (2Z)-3-(4-hydroxy-3-methoxyphenyl)prop-2-enoate

Rf (50% EtOAc/X4): 0.53 1H NMR: (400 MHz, CDCl3) δ: 7.77 (d, 1H, J = 1.9 Hz, H3), 7.11 (dd, 1H, J = 8.5 and 1.9

Hz, H5), 6.88 (d, 1H, J = 8.5 Hz, H6), 6.79 (d, 1H, J = 12.9 Hz, H7), 5.81 (d, 1H, J = 12.9


OCH2CH3). Data was extracted from the mixture of isomers.

For both isomers, all physical and chemical properties were as previously reported.185, 229

Microwave synthesis of ethyl ferulate (12)

Vanillin (0.52 g, 3.44 mmol) and (carbethoxymethylene)triphenylphosphorane (1.17 g,

3.36 mmol) were added to dry dichloromethane (15 mL) in a 30 mL reaction vessel and

heated at 50 oC under microwave radiation (CEM discover microwave reactor) for 5

minute intervals. After 20 minutes the reaction mixture showed no further change by TLC

and was concentrated and purified using column chromatography (20% EtOAc/X4) which

yielded 0.46 g (60%) of trans-ethyl ferulate and 0.13 g (17%) of a mixture of trans- and

cis-ethyl ferulate (15:85).

Attempted microwave synthesis of ethyl coumarate (11)

(Carbethoxymethylene)triphenylphosphorane (2.00 g, 5.73 mmol) and p-

hydroxybenzaldehyde (0.59 g, 4.79 mmol) were added to dry dichloromethane (15 mL) in

a 30 mL reaction vessel and heated at 50 oC under microwave radiation for 5 minutes.

After 5 minutes most of the reaction mixture had leaked into the microwave reactor, and a

repeat procedure of this scale provided the same result.


136

1-O-Benzyl p-coumaroyl ethyl ester (17)

Ethyl (2E)-3-[4-(benzyloxy)phenyl]prop-2-enoate

Anhydrous potassium carbonate (0.38 g, 2.71 mmol) and benzyl bromide (0.47 mL, 3.95

mmol) were added to a mixture of ethyl coumarate (11) (0.50 g, 2.60 mmol) in dry

dichloromethane (10 mL) and the reaction mixture heated under reflux. After 20 hours,

further benzyl bromide (0.47 mL, 3.95 mmol) was added. After 4 days, the reaction was

quenched with water, the organics extracted with ethyl acetate (3 x 10 mL), dried (MgSO4)

and concentrated. Purification by column chromatography (10% EtOAc/X4) gave 0.54 g

(74%) of 17 ester as a white solid. m.p. 64.8-65.7 oC (lit. m.p. 65-67 oC).230

Rf (50% EtOAc/X4): 0.61 1H NMR: (400 MHz, CDCl3) δ: 7.64 (d, 1H, J = 16.0 Hz, H7) 7.47 (app. d, 2H, J = 8.8 Hz,

H3,5), 7.44-7.34 (m, 5H, ArH), 6.98 (app. d, 2H, J = 8.8 Hz, H2,6), 6.31 (d, 1H, J = 16.0 Hz,

H8), 5.10 (s, 2H, OCH2Ph), 4.28 (q, 2H, J = 7.1 Hz, OCH2CH3), 1.34 (t, 3H, J = 7.1 Hz,

OCH2CH3). All physical and chemical properties were as previously reported.230

1-O-Benzyl feruloyl ethyl ester (18)

Ethyl (2E)-3-[4-(benzyloxy)-3-methoxyphenyl]prop-2-enoate

Using the same reaction conditions as described for 17 (above), ethyl ferulate (12) (1.00 g,

4.50 mmol) gave 0.96 g (69%) of 18 as a white solid. m.p. 67.7-68.4 oC (lit. m.p. 64 oC).231

Rf (50% EtOAc/X4): 0.69 1H NMR: (600 MHz, CDCl3) δ: 7.61 (d, 1H, J = 15.9 Hz, H7), 7.44-7.42 (m, 2H, ArH),

7.38-7.36 (m, 2H, ArH), 7.33-7.30 (m, 1H, ArH), 7.07 (d, 1H, J = 2.0 Hz, H3), 7.03 (dd,

1H, J = 8.4 and 2.0 Hz, H5), 6.87 (d, 1H, J = 8.4 Hz, H6), 6.30 (d, 1H, J = 15.9 Hz, H8),

5.19 (s, 2H, OCH2Ph), 4.26 (q, 2H, J = 7.1 Hz, OCH2CH3), 3.92 (s, 3H, OCH3), 1.33 (t,

3H, J = 7.1 Hz, OCH2CH3). Physical and spectral properties were as previously

reported.231-232


137

1-O-Benzyl p-coumaric acid (15)

(2E)-3-[4-(Benzyloxy)phenyl]prop-2-enoic acid

Potassium hydroxide (0.10 g, 1.71 mmol) was dissolved in water (5 mL) and added to a

mixture of 1-O-benzyl coumaroyl ethyl ester (17) (0.21 g, 0.73 mmol) in ethanol (5 mL).

The reaction mixture was stirred at room temperature for 20 hours then concentrated under

reduced pressure. The residue was taken up in water (10 mL), acidifed to pH 3 with 10%

HCl solution and caused precipitation of a white solid, which was extracted into ethyl

acetate (3 x 10 mL), washed with saturated brine solution (2 x 10 mL), dried (MgSO4) and

concentrated. This gave 0.17 g (93%) of 15 as a white solid. m.p. 200.2-201.8 oC (lit. m.p.

198-201 oC).233

Rf (50% EtOAc/X4): 0.27 1H NMR: (400 MHz, CDCl3) δ: 7.74 (d, 1H, J = 15.9 Hz, H7), 7.50 (app. d, 2H, J = 8.7 Hz,


H8), 5.11 (s, 2H, OCH2Ph). 1H NMR: (600 MHz, Acetone-d6) δ: 7.65-7.62 (m, 3H, H3,5,7), 7.50-7.49 (m, 2H, ArH),

7.42-7.39 (m, 2H, ArH), 7.36-7.34 (m, 1H, ArH), 7.08 (app. d, 2H, J = 8.8 Hz, H2,6), 6.39

(d, 1H, J = 15.9 Hz, H8), 5.20 (s, 2H, OCH2Ph). Physical properties and spectral properties

in acetone were as previously reported.233

1-O-Benzyl ferulic acid (16)

(2E)-3-[4-(Benzyloxy)-3-methoxyphenyl]prop-2-enoic acid

1-O-Benzyl feruloyl ethyl ester (18) (0.31 g, 1.00 mmol) was submitted to the same

procedure as described for 15 (above). This gave 0.24 g (85%) of 16 as a pale yellow solid.

m.p. 189.6-190.3 oC (lit. m.p. 191 oC).231

Rf (50% EtOAc/X4): 0.14


138

1H NMR: (400 MHz, CDCl3) δ: 7.70 (d, 1H, J = 15.9 Hz, H7), 7.43 (app. d, 2H, J = 7.1 Hz,

ArH), 7.38 (app. t, 2H, J = 7.3 Hz, ArH), 7.31 (app. t, 1H, J = 7.2 Hz, ArH), 7.09 (d, 1H, J

= 2.0 Hz, H3), 7.06 (dd, 1H, J = 8.3 and 2.0 Hz, H5), 6.89 (d, 1H, J = 8.3 Hz, H6), 6.31 (d,

1H, J = 15.9 Hz, H8), 5.20 (s, 2H, OCH2Ph), 3.93 (s, 3H, OCH3). 1H NMR: (600 MHz, Acetone-d6) δ: 7.61 (d, 1H, J = 15.9 Hz, H7), 7.51-7.49 (m, 2H,

ArH), 7.41-7.39 (m, 2H, ArH), 7.36 (d, 1H, J = 2.0 Hz, H3), 7.35-7.32 (m, 1H, ArH), 7.19

(dd, 1H, J = 8.3 and 2.0 Hz, H5), 7.08 (d, 1H, J = 8.3 Hz, H6), 6.42 (d, 1H, J = 15.9 Hz,

H8), 5.19 (s, 2H, CH2Ph), 3.91 (s, 3H, OCH3). Physical properties and spectral properties

in chloroform were as previously reported.231, 234

p-Coumaric acid (3)

(2E)-3-(4-Hydroxyphenyl)prop-2-enoic acid

trans-Ethyl coumarate (11) (1.00 g, 5.20 mmol) was dissolved in 1:1 aqueous ethanol (v/v,

20 mL) followed by the addition of potassium hydroxide (0.87 g, 15.52 mmol), then the

reaction mixture was stirred at room temperature for 3 days. The mixture was then diluted

with water (10 mL), unwanted organics extracted with diethyl ether (2 x 20 mL), the

aqueous layer acidified to pH 3 with 2 M hydrochloric acid solution and extracted with

ethyl acetate (2 x 20 mL). Concentration at reduced pressure gave 0.83 g (97%) of 3 as an

off-white solid. m.p. 208.7-209.8 oC (lit. m.p. 214-216 oC).235

Rf (10% MeOH/DCM): 0.31 1H NMR: (600 MHz, Acetone-d6) δ: 7.62 (d, 1H, J = 16.0 Hz, H7), 7.55 (app. d, 2H, J =

8.6 Hz, H3,5), 6.90 (app. d, 2H, J = 8.6 Hz, H2,6), 6.34 (d, 1H, J = 16.0 Hz, H8).

Physical and chemical properties were as previously reported.214, 235


139

Ferulic acid (4)

(2E)-3-(4-Hydroxy-3-methoxyphenyl)prop-2-enoic acid

With the same hydrolysis conditions as used above (for 3), reaction of trans-ethyl ferulate

(12) (0.67 g, 3.01 mmol) gave 0.56 g (96%) of 4 as a yellow solid. m.p. 169.1-170.2 oC (lit.

m.p. 168-169 oC).236

Rf (10 % MeOH/DCM): 0.33 1H NMR: (600 MHz, Acetone-d6) δ: 7.60 (d, 1H, J = 15.9 Hz, H7), 7.33 (d, 1H, J = 2.0 Hz,

H3), 7.14 (dd, 1H, J = 8.1 and 2.0 Hz, H5), 6.87 (d, 1H, J = 8.1 Hz, H6), 6.38 (d, 1H, J =

15.9 Hz, H8), 3.92 (s, 3H, OCH3). Physical and chemical properties were as previously

reported.184, 235-236

1-O-Acetyl p-coumaric acid (19)

(2E)-3-[4-(Acetyloxy)phenyl]prop-2-enoic acid

p-Coumaric acid (3) (0.68 g, 4.17 mmol) was dissolved in dry pyridine (5 mL) followed by

the addition of acetic anhydride (2 mL, 21.10 mmol) and the mixture stirred at room

temperature under a nitrogen atmosphere. After 52 hours the mixture was concentrated and

the crude solid recrystallised from ethanol to furnish 0.71 g (83%) of 19 as colourless

needles. m.p. 203.3-204.5 oC (lit. m.p. 200-205 oC).237

Rf (10% MeOH/DCM): 0.33 1H NMR: (400 MHz, DMSO-d6) δ: 7.74 (app. d, 2H, J = 8.6 Hz, H3,5), 7.59 (d, 1H, J =

16.0 Hz, H7), 7.18 (app. d, 2H, J = 8.6 Hz, H2,6), 6.51 (d, 1H, J = 16.0 Hz, H8), 2.28 (s, 3H,

OCOCH3). 1H NMR: (400 MHz, Acetone-d6) δ: 7.73 (app. d, 2H, J = 8.7 Hz, H3,5), 7.68 (d, 1H, J =


OCOCH3). Spectral properties correspond with that previously reported.238


140

1-O-Acetyl ferulic acid (20)

(2E)-3-[4-(Acetyloxy)-3-methoxyphenyl]prop-2-enoic acid

Ferulic acid (4) (1.06 g, 5.46 mmol) was reacted using the same conditions described for

the preparation of 19 (above). Recrystallisation of the crude solid from ethanol afforded

1.02 g (79%) of 20 as an off-white solid. m.p. 196.3-197.2 oC (lit. m.p. 197-200 oC).239

Rf (10% MeOH/DCM): 0.28 1H NMR: (400 MHz, DMSO-d6) δ: 7.57 (d, 1H, J = 16.0 Hz, H7), 7.48 (d, 1H, J = 1.8 Hz,


16.0 Hz, H8), 3.82 (s, 3H, OCH3), 2.26 (s, 3H, OCOCH3). 1H NMR: (600 MHz, Acetone-d6) δ: 7.65 (d, 1H, J = 16.0 Hz, H7), 7.47 (d, 1H, J = 2.0 Hz,


16.0 Hz, H8), 3.91 (s, 3H, OCH3), 2.25 (s, 3H, OCOCH3). Spectral properties in DMSO

were as previously reported.239

Attempted synthesis of 1-O-chloroacetyl coumaric acid (21)

(2E)-3-{4-[(Chloroacetyl)oxy]phenyl}prop-2-enoic acid

p-Coumaric acid (3) (151.5 mg, 0.92 mmol) was dissolved in dry pyridine (5 mL) followed

by the addition of chloroacetic anhydride (0.73g, 4.26 mmol) and the mixture stirred at

ambient temperature. After 16 hours the solid reaction mixture was dissolved in methanol

(10 mL) and analysed by TLC which indicated no formation of the desired product.


141

1-O-Chloroacetyl p-coumaric acid (21)

(2E)-3-{4-[(Chloroacetyl)oxy]phenyl}prop-2-enoic acid

Chloroacetyl chloride (0.90 mL, 11.30 mmol) was added to p-coumaric acid (3) (0.60 g,

3.68 mmol) dissolved in a minimal volume of 2M sodium hydroxide solution (8 mL) at 0 oC. The resulting suspension was stirred at room temperature for 5 minutes before being

acidified with 2M hydrochloric acid solution. The precipitate was filtered, washed with

cold water, dried and recrystallised from acetone/X4 to give 0.31 g (35%) of 21 as a

crystalline white solid. m.p. 182.9-184.5 oC (lit. m.p. 186-187 oC).144

Rf (10% MeOH/DCM): 0.41 1H NMR: (400 MHz, Acetone-d6) δ: 7.78 (app. d, 2H, J = 8.7 Hz, H3,5), 7.69 (d, 1H, J =


CH2Cl). Physical and chemical properties were as previously reported.144

1-O-Chloroacetyl ferulic acid (22)

(2E)-3-{4-[(Chloroacetyl)oxy]-3-methoxyphenyl}prop-2-enoic acid

As described for 21 (above), reaction of ferulic acid (4) (0.60 g, 3.07 mmol) gave a crude

mixture which was recrystallised from ethanol/water to give 0.54 g (65%) of 22 as a pale

yellow crystalline solid. m.p. 148.3-149.6 oC (lit. m.p. 146-148 oC).145

Rf (10% MeOH/DCM): 0.40 1H NMR: (400 MHz, Acetone-d6) δ: 7.66 (d, 1H, J = 16.0 Hz, H7), 7.51 (d, 1H, J = 1.9

Hz, H3), 7.30 (dd, 1H, J = 8.2 and 1.9 Hz, H5), 7.19 (d, 1H, J = 8.2 Hz, H6), 6.56 (d, 1H, J

= 16.0 Hz, H8), 4.58 (s, 2H, CH2Cl), 3.93 (s, 3H, OCH3). NMR data assignments made

based on those of 21, and physical properties were as previously reported.145


142

6.2.2 Synthesis of Hydroxycinnamoyl Tartrate Esters

Dibenzyl L-tartaric acid (24)

Dibenzyl (2R,3R)-2,3-dihydroxybutanedioate

L-Tartaric acid (3.01 g, 20.04 mmol) was added to a mixture of benzyl alcohol (6.3 mL,

60.88 mmol) and p-toluene sulphonic acid (0.36 g, 2.06 mmol) in toluene (40 mL).

Employing a Dean-Stark apparatus, the mixture was heated under reflux for 3 hours, in

which 0.72 mL (100 % theoretical) of water was collected. The mixture was allowed to

cool to room temperature, diluted with diethyl ether (30 mL), poured into saturated sodium

bicarbonate solution (60 mL), and the product extracted with diethyl ether (3 x 30 mL),

dried (MgSO4) and concentrated. Trituration with X4/EtOAc (20:1) gave 6.24 g (94%) of

24 as a white solid. m.p. 49.8-51.0 oC (lit. m.p. 49-50 oC).152

Rf (50% EtOAc/X4): 0.42

= +11.4o (c 1.01, acetone) (lit. = +10.1)148

1H NMR: (400 MHz, CDCl3) δ: 7.38-7.36 (m, 10H, ArH), 5.30 (d, 2H, J = 12.1 Hz,

CHaHbPh), 5.25 (d, 2H, J = 12.1 Hz, CHaHbPh), 4.63 (d, 2H, J = 7.7 Hz, CH), 3.35 (d, 2H,

J = 7.7 Hz, OH). All physical and chemical properties were as previously reported.148, 151-

152

General procedure for esterification with dibenzyl-L-tartrate

The hydroxycinnamate (15 or 16) (0.50 mmol) and 24 (0.60 mmol) were dissolved in dry

dichloromethane (15 mL) followed by the addition of trifluoroacetic anhydride (0.60

mmol) at 0oC. The mixture was then stirred at ambient temperature for 6 hours before

being poured onto saturated sodium bicarbonate solution (20 mL), extracted with

dichloromethane (3 x 15 mL), washed with water (3 x 15 mL), dried (MgSO4) and

concentrated in vacuo. Purification by column chromatography (DCM-3% Et2O/DCM)

furnished the desired products (25 or 26), as well as a mixture of the mono- and di-ester.


143

1-O-Benzyl p-coumaroyl dibenzyl L-tartaric acid (25)

Dibenzyl (2R,3R)-2-({(2E)-3-[4-(benzyloxy)phenyl]prop-2-enoyl}oxy)-3-

hydroxybutanedioate

From 15 (101.6 mg, 0.40 mmol), afforded 107.0 mg (47%) of 25 as a colourless gum, as

well as 49.1 mg of a mixture of 25 and 27 (1:9).

Rf (50% EtOAc/X4): 0.56

= +12.6o (c 0.7, chloroform) 1H NMR: (400 MHz, CDCl3) δ: 7.60 (d, 1H, J = 15.9 Hz, H7), 7.45-7.15 (m, 17H, ArH and

H3,5), 6.98 (app. d, 2H, J = 8.8 Hz, H2,6), 6.21 (d, 1H, J = 15.9 Hz, H8), 5.63 (d, 1H, J = 2.3

Hz, H2’), 5.27-5.02 (m, 6H, 3 x CH2Ph,), 4.86 (dd, 1H, J = 2.3 and 7.5 Hz, H3’), 3.28 (d,

1H, J = 7.5 Hz, OH). 13C NMR: (400 MHz, CDCl3) δ: 170.7 (C4’), 166.8 (C1’), 165.7 (C9), 161.0 (C1), 146.2

(C7), 136.5, 135.1, 134.6 (3 x Ar), 130.0 (C3,5), 130.3-127.0 (Ar), 115.2 (C2,6), 113.2 (C8),

73.7 (CH2Ph), 72.8 (C2’). 70.9 (C3’), 70.2, 67.7 (2 x CH2Ph).

IR (neat) ν: 3475, 2955, 1718, 1145, 732, 695.

LRP (+EI) m/z (%): 566 (M+, 1), 475 (1), 325 (2), 254 (5), 237 (9), 181 (2), 164 (2), 147

(3), 107 (3), 91 (100), 65 (7).

HRMS calculated for C34H30O8 [M] + 566.1941, found 566.1948.

Di(1-O-benzyl coumaroyl)dibenzyl L-tartaric acid (27)

Dibenzyl (2R,3R)-2,3-bis({(2E)-3-[4-(benzyloxy)phenyl]prop-2-enoyl}oxy)butanedioate

Isolated from the synthesis and purification of 25.


144

Rf (50% EtOAc/X4): 0.69 1H NMR: (600 MHz, CDCl3) δ: 7.64 (d, 1H, J = 16.0 Hz, H7), 7.45-7.15 (m, 12H, ArH and

H3,5), 6.99 (app. d, 2H, J = 8.8 Hz, H2,6), 6.25 (d, 1H, J = 16.0 Hz, H8), 5.91 (s, 1H, CH),

5.24 (d, 1H, J = 12.2 Hz, CHaHbPh), 5.14 (d, 1H, J = 12.2 Hz, CHaHbPh), 5.11 (app. s, 2H,

CH2Ph). Identification and characterisation based on reported data for the caffeoyl

derivative.154

1-O-Benzyl feruloyl dibenzyl L-tartaric acid (26)

Dibenzyl (2R,3R)-2-({(2E)-3-[4-(benzyloxy)-3-methoxyphenyl]prop-2-enoyl}oxy)-3-

hydroxybutanedioate

From 16 (185.9 mg, 0.65 mmol), yielded 197.0 mg (50%) of 26 as a colourless gum, as

well as 67.5 mg of a mixture of 26 and 28 (1:19).

Rf (50% EtOAc/X4): 0.48

= +10.2o (c 1.7, chloroform) 1H NMR: (400 MHz, CDCl3) δ: 7.59 (d, 1H, J = 15.9 Hz, H7), 7.45-7.20 (m, 15H, ArH),

7.04 (d, 1H, J = 1.9 Hz, H3), 7.01 (dd, 1H, J = 8.3 and 1.9 Hz, H5), 6.88 (d, 1H, J = 8.3 Hz,

H6), 6.22 (d, 1H, J = 15.9 Hz, H8), 5.64 (d, 1H, J = 2.2 Hz, H2’), 5.26-5.16 (m, 6H,

CH2Ph), 4.86 (br. s, 1H, H3’), 3.93 (s, 3H, OCH3), 3.21 (d, 1H, J = 7.6 Hz, OH). 13C NMR: (400 MHz, CDCl3) δ: 170.7 (C4’), 166.7 (C1’), 165.6 (C9), 150.7 (C1), 149.8

(C2), 146.6 (C7), 136.5, 135.1, 135.0 (3 x Ar), 128.7-128.1 (Ar), 127.3 (C4), 123.0 (C5),

113.9 (C8), 113.3 (C6), 110.3 (C3), 73.0 (C2’), 70.9 (CH2Ph), 70.8 (C3’), 68.2, 67.7 (2 x

CH2Ph), 56.1 (OCH3).

IR (neat) ν: 3478, 2957, 1718, 1133, 731, 695.

LRP (+EI) m/z (%): 596 (M+, 1), 505 (4), 355 (4), 284 (6), 267 (7), 194 (1), 149 (4), 107

(4), 91 (100), 65 (6).



145

Di(1-O-benzyl feruloyl)dibenzyl L-tartaric acid (28)

Dibenzyl (2R,3R)-2,3-bis({(2E)-3-[4-(benzyloxy)-3-methoxyphenyl]prop-2-

enoyl}oxy)butanedioate

Isolated from the synthesis and purification of 26.

Rf (50% EtOAc/X4): 0.58 1H NMR: (400 MHz, CDCl3) δ: 7.61 (d, 1H, J = 15.9 Hz, H7), 7.45-7.15 (m, 10H, ArH),

7.05 (d, 1H, J = 1.8 Hz, H3), 7.01 (dd, 1H, J = 8.4 and 1.8 Hz, H5), 6.89 (d, 1H, J = 8.4 Hz,

H6), 6.26 (d, 1H, J = 15.9 Hz, H8), 5.92 (s, 1H, H2’), 5.26-5.13 (m, 4H, CH2Ph), 3.93 (s,

3H, OCH3). Identification and characterisation based on reported data for the caffeoyl

derivative.154

Attempted syntheses of feruloyl tartrate (8)

Attempt 1:

1-O-Benzyl feruloyl dibenzyl L-tartaric acid (26) (135.1 mg, 0.23 mmol) and 5%

palladium on activated carbon (21.3 mg) were stirred in ethyl acetate (15 mL) followed by

the addition of 1,4-cyclohexadiene (0.22 mL, 0.23 mmol) and the mixture was stirred at

ambient temperature. After 3 hours TLC showed no formation of the desired product and

further 1,4-cyclohexadiene (0.22 mL, 0.23 mmol) was added. After 24 hours analysis by

TLC showed no change and the reaction mixture was filtered through celite, concentrated

in vacuo and the crude mixture analysed by NMR, which suggested formation of the

reduced analogue of the starting material (29).

Rf of reaction mixture (50% EtOAc/X4): 0.48


146

1H NMR: (400 MHz, CDCl3) δ: 5.54 (d, 1H, J = 2.2 Hz, H2’), 4.82 (d, 1H, J = 2.2 Hz, H3’),

2.84 (app. t, 2H, J = 7.9 Hz, H7), 2.62 (m, 1H, H8a), 2.48 (m, 1H, H8b). Unobstructed 1H

signals in addition to those of the remaining starting material (26).

Attempt 2:

Palladium acetate (86.0 mg, 0.38 mmol) and triethylamine (0.05 mL, 0.36 mmol) were

dissolved in dry dichloromethane (5 mL) and after 5 minutes of stirring at ambient

temperature 1-O-benzyl feruloyl dibenzyl L-tartaric acid (26) (80.8 mg, 0.16 mmol) in dry

dichloromethane (5 mL) was added dropwise. After a further 5 minutes triethylsilane (0.20

mL, 1.252 mmol) was added slowly and the mixture stirred at ambient temperature for 24

hours before being diluted with methanol (5 mL) filtered through celite and concentrated in

vacuo. The crude oil was taken up in ethyl acetate/water (6:1, 20 mL), the aqueous layer

separated and washed with ethyl acetate (2 x 10 mL). The combined organics were washed

successively with phosphoric acid solution (1 M, 10 mL) and brine solution (10 mL) until a

pH of 7 was achieved, then dried (MgSO4) and concentrated in vacuo. The clear oil was

dissolved in ethyl acetate and X4 added until a precipitate formed, which was filtered,

washed with X4 and analysed by 1H NMR which suggested formation of the reduced

analogue of the desired product (30).

Rf of crude mixture (50% EtOAc/X4): 0.00 1H NMR: (400 MHz, CD3OD) δ: 5.46 (d, 1H, J = 2.3 Hz, H2’), 4.72 (d, 1H, J = 2.3 Hz,

H3’), 2.94 (app. t, 2H, J = 7.8 Hz, H7), 2.71 (app. t, 2H, J = 7.8 Hz, H8). Unobstructed 1H

signals in the crude mixture.

Attempt 3:

Debenzylation was attempted as described above, employing half the amount of

triethylsilane (0.1 mL, 0.626 mmol). Analysis of the product by 1H NMR displayed the

same peaks as above suggesting formation of the reduced product, with minor signals

corresponding with formation of the desired product (8).

Rf of crude mixture (50% EtOAc/X4): 0.00


147

O,O’-Diacetyl L-tartaric anhydride (31)

(3R,4R)-2,5-Dioxotetrahydrofuran-3,4-diyl diacetate

L-Tartaric acid (10.07 g, 67.08 mmol) in acetyl chloride (65 mL, 0.91 mol) was heated

under reflux under a nitrogen atmosphere for 48 hours. The reaction mixture was allowed

to cool to room temperature, concentrated in vacuo and the resulting oil recyrstallised from

EtOAc/X4 to afford 11.82 g (82%) of 31 as a white crystalline solid. m.p. 135-136 oC (lit.

m.p. 133-135 oC).240 1H NMR: (400 MHz, CDCl3) δ: 5.68 (s, 1H, CH), 2.23 (s, 3H, OCH3). 1H NMR: (400 MHz, Acetone-d6) δ: 6.17 (s, 1H, CH), 2.19 (s, 3H, OCH3). Spectral

properties in chloroform were as previously reported.161, 240-241

O,O’-Diacetyl L-tartaric acid (32)

(2R,3R)-2,3-bis(Acetyloxy)butanedioic acid

O,O’-Diacetyl L-tartaric anhydride (31) (1.03 g, 4.75 mmol) was dissolved in acetone (5

mL) followed by the addition of water (0.17 mL, 9.44 mmol) and the mixture was stirred at

room temperature for 12 hours before being concentrated. Trituration with X4 gave 1.10 g

(99%) of 32 as a white solid. m.p. 117-118 oC (lit. m.p. 118 oC).242 1H NMR: (400 MHz, Acetone-d6) δ: 5.72 (s, 1H, CH), 2.11 (s, 3H, OCH3). Physical and

chemical properties were as previously reported.161, 242

O,O’-Diacetyl-di-tert-butyl L-tartrate (33)

Di-tert-butyl (2R,3R)-2,3-bis(acetyloxy)butanedioate


148

Anhydrous magnesium sulphate (3.98 g) and sulphuric acid (0.5 mL) were stirred in dry

dichloromethane (120 mL) for 15 minutes followed by the addition of O,O’-diacetyl L-

tartaric acid (32) (1.03 g, 4.42 mmol) and dry tert-butanol (4.0 mL, 41.82 mmol), and the

reaction vessel was stoppered tightly. After 3 days the mixture was poured into saturated

sodium bicarbonate solution (100 mL) and stirred until the MgSO4 dissolved, then

extracted in dichloromethane (3 x 50 mL), washed with brine (2 x 50 mL), dried (MgSO4)

and concentrated in vacuo. Column chromatography (20% EtOAc/X4) gave 0.76 g (49%)

of 33 as a clear oil and 0.32 g of a mixture of 33 and the mono-acetate (80:20 as

determined by proton integration).

Rf (50% EtOAc/X4): 0.66 1H NMR: (400 MHz, CDCl3) δ: 5.62 (s, 1H, CH), 2.16 (s, 3H, OCH3), 1.44 (s, 9H, t-Bu). 1H NMR: (400 MHz, DMSO-d6) δ: 5.51 (s, 1H, CH), 2.11 (s, 3H, OCH3), 1.38 (s, 9H, t-

Bu). Spectral properties in chloroform were as previously reported.159

Mono-acetyl-di-tert-butyl L-tartrate (Tentative characterisation)

Di-tert-butyl (2R,3R)-2-(acetyloxy)-3-hydroxybutanedioate

Obtained during the purification of 33.

Rf (50% EtOAc/X4): 0.60 1H NMR: (400 MHz, CDCl3) δ: 5.34 (d, 1H, J = 2.3 Hz, CH), 4.60 (dd, 1H, J = 7.1 and 2.3

Hz, CH), 3.09 (d, 1H, J = 7.1 Hz, OH), 2.12 (s, 3H, OCH3), 1.49 (s, 9H, t-Bu), 1.46 (s, 9H,

t-Bu). Data was extracted from the spectrum of the mixture.

Di-tert-butyl L-tartrate (34)

Di-tert-butyl (2R,3R)-2,3-dihydroxybutanedioate

Diacetyl-di-tert-butyl L-tartrate (33) (3.50 g, 10.10 mmol) was dissolved in methanol (25

mL) followed by the addition of powdered potassium hydroxide (115.7 mg, 2.06 mmol)

and the mixture stirred at room temperature. After 45 minutes the mixture was


149

concentrated and purified by column chromatography (20% EtOAc/X4) yielding 0.79g

(29%) of 34 as a white solid, m.p. 86.1-89.8 oC (lit. m.p. 91 oC),159 as well as tert-butyl

methyl L-tartrate as a clear oil.

Rf (40% EtOAc/X4): 0.49

= +11.0o (c 1.0, acetone) 1H NMR: (400 MHz, CDCl3) δ: 4.36 (d, 1H, J = 6.9 Hz, CH), 3.09 (d, 1H, J = 6.9 Hz,

OH), 1.52 (s, 9H, t-Bu). All physical and chemical properties were as previously

reported.159

tert-Butyl methyl L-tartrate

tert-Butyl methyl (2R,3R)-2,3-dihydroxybutanedioate

Isolated during purification of 34.

Rf (40% EtOAc/X4): 0.27 1H NMR: (400 MHz, CDCl3) δ: 4.51 (dd, 1H, J = 7.6 and 1.8 Hz, CH), 4.41 (dd, 1H, J =

6.4 and 1.8 Hz, CH), 3.86 (s, 3H, OCH3), 3.18 (d, 1H, J = 6.4 Hz, OH), 3.05 (d, 1H, J =

7.6 Hz, OH), 1.52 (s, 9H, t-Bu). Spectral properties were as previously reported.159

General procedure for esterification with di-tert-butyl L-tartrate

The hydroxycinnamate (19-22) (0.50 mmol) was heated under reflux in dry benzene (10

mL) containing thionyl chloride (6.89 mmol). After 5 hours the mixture was allowed to

cool to room temperature and then concentrated in vacuo. The crude residue was taken up

in dry benzene (5 mL) and added dropwise to a solution of di-tert-butyl L-tartrate (34)

(0.65 mmol) in dry pyridine (5 mL), then stirred at ambient temperature overnight. The

mixture was concentrated and pyridine azeotropically removed with toluene. Purification

with column chromatography (20% EtOAc/X4) gave the desired product (35-38).


150

1-O-Acetyl p-coumaroyl di-tert-butyl L-tartrate (35)

Di-tert-butyl (2R,3R)-2-({(2E)-3-[4-(acetyloxy)phenyl]prop-2-enoyl}oxy)-3-

hydroxybutanedioate

From 19 (101.4 mg, 0.49 mmol), and recrystallisation from 30% EtOAc/X4 gave 68.1 mg

(31%) of white crystals. m.p. 143.6-144.2 oC.

Rf (50% EtOAc/X4): 0.57

= +2.04o (c 0.5, acetone) 1H NMR: (400 MHz, CDCl3) δ: 7.73 (d, 1H, J = 16.0 Hz, H7), 7.54 (app. d, 2H, J = 8.7 Hz,


Hz, H2’), 4.67 (dd, 1H, J = 6.9 and 2.3 Hz, H3’), 3.20 (d, 1H, J = 6.9 Hz, OH), 2.31 (s, 3H,

OCOCH3), 1.51 (s, 9H, t-Bu4), 1.44 (s, 9H, t-Bu1). 13C NMR: (600 MHz, CDCl3) δ: 170.2 (C4’), 169.3 (OCOCH3), 165.8 (C9), 165.5 (C1’),

152.4 (C1), 145.5 (C7), 131.9 (C4), 129.6 (C3,5), 122.3 (C2,6), 116.7 (C8), 84.0 (C1(CH3)3),

83.4 (C4(CH3)3), 73.5 (C2’), 71.0 (C3’), 28.1 (C4(CH3)3), 28.0 (C1(CH3)3), 21.3 (OCOCH3).

IR (neat) ν: 2982, 1713, 1128, 1055, 1033, 1015.

LRP (+EI) m/z (%): 450 (M+, <1), 408 (2), 352 (10), 338 (5), 321 (12), 296 (63), 278 (6),

251 (6), 206 (7), 189 (46), 164 (79), 147 (100), 119 (14), 57 (37), 41 (13).


Details of crystal structure determination of 35

Crystal data for C23H30O9: M = 450.47, T = 100(2) K, orthorhombic, P212121, a =

5.7183(2), b = 8.7309(3), c = 46.9988(19) Å, V = 2346.46(15) Å3, Z = 4, Dx = 1.275,

F(000) = 960, µ = 0.822 mm-1, no. of unique data (Agilent Technologies SuperNova Dual

diffractometer with Atlas detector using Cu Kα radiation so that θmax = 74.6°) = 4603, no.

of parameters = 300, R (3842 data with I ≥ 2σ(I)) = 0.053, wR (all data) = 0.131. The

structure was solved by direct-methods (SHELXS-97) and refined (anisotropic

displacement parameters, C-bound H atoms in the riding model approximation, full

refinement of the hydroxyl-H atom, and a weighting scheme w = 1/[σ2(Fo2) + (0.067P)2]

where P = (Fo2 + 2Fc

2)/3) with SHELXL-97 on F2. The value of the Flack parameters =

0.0(2).


151

1-O-Acetyl feruloyl tert-butyl L-tartrate (36)

Di-tert-butyl (2R,3R)-2-({(2E)-3-[4-(acetyloxy)-3-methoxyphenyl]prop-2-enoyl}oxy)-3-

hydroxybutanedioate

From 20 (159.1 mg, 0.67 mmol), and recrystallisation from 30% EtOAc/X4 gave 154.0 mg

(48%) of white crystals. m.p. 140.5-142.0 oC.

Rf (30% EtOAc/X4): 0.34

= -4.51o (c 1.3, acetone) 1H NMR: (600 MHz, CDCl3) δ: 7.70 (d, 1H, J = 16.0 Hz, H7), 7.12-7.11 (m, 2H, H3,5),

7.05 (d, 1H, J = 8.6 Hz, H6), 6.45 (d, 1H, J = 16.0 Hz, H8), 5.50 (d, 1H, J = 2.3 Hz, H2’),

4.68 (dd, 1H, J = 6.8 and 2.3 Hz, H3’), 3.87 (s, 3H, OCH3), 3.21 (d, 1H, J = 6.8 Hz, OH),

2.33 (s, 3H, OCOCH3), 1.51 (s, 9H, t-Bu4), 1.44 (s, 9H, t-Bu1). 13C NMR: (600 MHz, CDCl3) δ: 170.2 (C4’), 168.9 (OCOCH3), 165.9 (C1’), 165.5 (C9),

151.5 (C7), 145.9 (C2), 141.8 (C1), 133.2 (C4), 123.4 (C6), 121.8 (C5), 116.8 (C8), 111.3

(C3), 84.1 (C1(CH3)3), 83.5 (C4(CH3)3), 73.5 (C2’), 71.0 (C3’), 56.1 (OMe), 28.1

(C4(CH3)3), 28.0 (C1(CH3)3), 20.8 (OCOCH3).

IR (neat) ν: 1755, 1710, 1260, 1218, 1195, 1148, 1120, 1074, 1030, 981.

Calc. C 59.99, H 6.71, O 33.30. Anal. C 59.79, H 6.73, O 33.48.

Details of crystal structure determination of 36

Crystal data for C24H32O10: M = 450.47, T = 100(2) K, monoclinic, P21, a = 5.9894(1), b =

10.6483(1), c = 19.6676(2) Å, β = 96.324(1)º, V = 1246.71(3) Å3, Z = 2, Dx = 1.280,

F(000) = 512, µ = 0.837 mm-1, no. of unique data (Agilent Technologies SuperNova Dual

diffractometer with Atlas detector using Cu Kα radiation so that θmax = 74.5°) = 4800, no.

of parameters = 326, R (4762 data with I ≥ 2σ(I)) = 0.057, wR (all data) = 0.159. The

structure was solved by direct-methods (SHELXS-97) and refined (anisotropic

displacement parameters, all H atoms in the riding model approximation, and a weighting

scheme w = 1/[σ2(Fo2) + (0.098P)2+ 0.941P] where P = (Fo

2 + 2Fc2)/3) with SHELXL-97

on F2. Two orientations, of equal weight, were discerned for a significant portion of the

molecule. The aromatic rings were refined as hexagons (C–C = 1.39 Å), equivalent pairs of


152

atoms were constrained to have identical anisotropic displacement parameters and these

were constrained to be nearly isotropic. The value of the Flack parameters = 0.0(2).

Attempted synthesis of p-coumaroyl L-tartrate (7)

1-O-Acetyl p-coumaroyl tert-butyl L-tartrate (35) (68.1 mg, 0.15 mmol) was dissolved in

dry dichloromethane (5 mL) followed by the addition of trifluoroacetic acid (0.30 mL, 3.92

mmol) and the mixture was stirred at ambient temperature. After disappearance of the

starting material (as shown by TLC) the mixture was concentrated in vacuo and taken up in

acetone/3 M hydrochloric acid (3:1, 6 mL) and heated under reflux for 3 hours before

being allowed to cool to ambient temperature. The mixture was diluted with ethyl actetate

(20 mL), washed with brine (3 x 20 mL), dried (MgSO4) and concentrated in vacuo. 1H

NMR of the crude product indicated formation of the desired product with major

impurities, and attempted purification using reverse-phase chromatography (linear gradient

from 0.1% formic acid/water to 0.1% formic acid/acetonitrile) failed to separate the by-

products. To protonate any potentially occuring salts, the crude product was taken up in

methanol (5 mL), the pH adjusted to 1 with 2M HCl and the mixture stirred at ambient

temperature for 16 hours before being concentrated in vacuo and analysed by 1H NMR

which showed no change in the ratio of product and impurities.

Rf of mixture (50% EtOAc/X4): 0.00

1-O-Chloroacetyl p-coumaroyl tert-butyl L-tartrate (37)

Di-tert-butyl (2R,3R)-2-{[(2E)-3-{4-[(chloroacetyl)oxy]phenyl}prop-2-enoyl]oxy}-3-

hydroxybutanedioate

From 21 (91.4 mg, 0.38 mmol), gave 54.7 mg (30%) of 37 (m.p 126.5-127.2 oC), as well as

a mixture of 37 and di-ester in an 84:16 ratio as determined by proton integration.


153

Rf (50% EtOAc/X4): 0.65

= -3.22o (c 1.6, acetone) 1H NMR: (400 MHz, CDCl3) δ: 7.72 (d, 1H, J = 16.0 Hz, H7), 7.55 (app. d, 2H, J = 8.7 Hz,


Hz, H2’), 4.67 (d, 1H, J = 2.1 Hz, H3’), 4.31 (s, 2H, OCH2Cl), 3.25 (br. s, 1H, OH), 1.50 (s,

9H, t-Bu4), 1.43 (s, 9H, t-Bu1). 13C NMR: (600 MHz, CDCl3) δ: 170.2 (C4’), 165.8 (OCOCH3), 165.7 (C9), 165.4 (C1’),

152.0 (C1), 145.1 (C7), 132.5 (C4), 129.8 (C3,5), 121.8 (C2,6), 117.1 (C8), 84.0 (C1(CH3)3),

83.5 (C4(CH3)3), 73.8 (C2’), 70.7 (C3’), 40.9 (OCOCH2Cl), 28.1 (C4(CH3)3), 27.6

(C1(CH3)3).

IR (neat) ν: 3460, 2975, 2929, 1718, 1127, 839.

LRP (+EI) m/z (%): 484 (M+, <1), 428 (3), 372 (8), 355 (12), 327 (31), 296 (18), 278 (8),

240 (19), 223 (100), 206 (2), 164 (94), 147 (94), 119 (15), 57 (74), 41 (17).

HRMS calculated for C23H29ClO9 [M] + 484.1500, found 484.1494.

1-O-Chloroacetyl feruloyl tert-butyl L-tartrate (38)

Di-tert-butyl (2R,3R)-2-{[(2E)-3-{4-[(chloroacetyl)oxy]-3-methoxyphenyl}prop-2-

enoyl]oxy}-3-hydroxybutanedioate

From 22 (50.1 mg, 0.19 mmol), gave 32.0 mg (34%) of 38 as a white solid (m.p 106.0-

106.8 oC), as well as 19.1 mg (24%) of the dechloroacetylated product (40).

Rf (50% EtOAc/X4): 0.60

= -6.69o (c 1.5, acetone) 1H NMR: (600 MHz, CDCl3) δ: 7.70 (d, 1H, J = 16.0 Hz, H7), 7.12-7.08 (m, 3H, H3,5,6),

6.45 (d, 1H, J = 16.0 Hz, H8), 5.50 (d, 1H, J = 2.3 Hz, H2’), 4.68 (br. s, 1H, H3’), 4.34 (s,

2H, CH2Cl), 3.86 (s, 3H, OCH3), 3.22 (br. s, 1H, OH), 1.51 (s, 9H, t-Bu4), 1.44 (s, 9H, t-

Bu1). 13C NMR: (600 MHz, CDCl3) δ: 170.2 (C4’), 165.8 (OCOCH3), 165.4 (C1’), 165.3 (C9),

151.3 (C7), 145.5 (C2), 141.2 (C1), 133.7 (C4), 123.2 (C6), 121.7 (C5), 116.9 (C8), 111.6


154

(C3), 84.0 (C1(CH3)3), 83.5 (C4(CH3)3), 73.8 (C2’), 70.8 (C3’), 56.5 (OMe), 40.7

(OCOCH2Cl), 28.3 (C4(CH3)3), 27.6 (C1(CH3)3).

IR (neat) ν: 3479, 2925, 1717, 1259, 1130, 845, 815, 764.

LRP (+EI) m/z (%): 514 (M+, 1), 458 (1), 402 (2), 382 (2), 357 (6), 326 (14), 308 (4), 270

(10), 253 (16), 236 (13), 194 (100), 177 (26), 145 (9), 133 (6), 117 (4), 89 (4), 77 (6), 57

(8), 41 (6).

HRMS calculated for C24H31ClO10 [M] + 514.1606, found 514.1601.

p-Coumaroyl tert-butyl L-tartrate (39)

Di-tert-butyl (2R,3R)-2-hydroxy-3-{[(2E)-3-(4-hydroxyphenyl)prop-2-

enoyl]oxy}butanedioate

1-O-Chloroacetyl p-coumaric acid (21) (0.21 g, 0.86 mmol) was heated under reflux in dry

benzene (10 mL) containing thionyl chloride (0.50 mL, 6.89 mmol). After 5 hours the

mixture was allowed to cool to ambient temperature and concentrated in vacuo. The crude

residue was taken up in dry benzene (5 mL) and added dropwise to a solution of di-tert-

butyl L-tartrate (34) (0.33 g, 1.24 mmol) in dry pyridine (5 mL), then stirred at room

temperature for 45 hours. The mixture was concentrated and pyridine azeotropically

removed with toluene. Purification by column chromatography (20% EtOAc/X4) gave

68.6 mg (19 %) of 39 as a white solid (m.p 162.3-163.7 oC), as well as 47.1 mg (11%) of

37.

Rf (50% EtOAc/X4): 0.38

= -5.57o (c 1.0, acetone) 1H NMR: (400 MHz, CDCl3) δ: 7.60 (d, 1H, J = 15.9 Hz, H7), 7.29 (app. d, 2H, J = 8.6 Hz,


Hz, H2’), 4.68 (d, 1H, J = 2.3 Hz, H3’), 3.34 (br. s, 1H, OH), 1.52 (s, 9H, t-Bu4), 1.44 (s,

9H, t-Bu1). 13C NMR: (400 MHz, CDCl3) δ: 170.2 (C4’), 166.8 (C1’), 166.2 (C9), 159.0 (C1), 146.8

(C7), 130.4 (C3,5), 126.4 (C4), 116.2 (C2,6), 113.2 (C8), 84.2 (C1(CH3)3), 84.0 (C4(CH3)3),

73.4 (C2’), 71.1 (C3’), 28.2 (C4(CH3)3), 28.0 (C1(CH3)3).


155

IR (neat) ν: 3389, 2977, 1712, 1145, 1129, 988, 831.

LRP (+EI) m/z (%): 408 (M+, 1), 352 (6), 296 (64), 279 (16), 251 (17), 206 (1), 164 (61),

147 (100), 119 (18), 91 (11), 57 (20), 41 (12).


Feruloyl tert-butyl L-tartrate (40)

Di-tert-butyl (2R,3R)-2-hydroxy-3-{[(2E)-3-(4-hydroxy-3-methoxyphenyl)prop-2-

enoyl]oxy}butanedioate

1-O-Chloroacetyl ferulic acid (22) (0.21 g, 0.76 mmol) was submitted to the same reaction

conditions as described for 39 (above). This afforded 92.7 mg (28%) of 40 as a white solid

(m.p 172.6-174.0 oC), as well as 24.8 mg (6%) of 38.

Rf (50 % EtOAc/X4): 0.38

= -8.87o (c 1.0, acetone) 1H NMR: (600 MHz, CDCl3) δ: 7.68 (d, 1H, J = 15.9 Hz, H7), 7.07 (dd, 1H, J = 8.1 and


Hz, H8), 5.50 (d, 1H, J = 2.3 Hz, H2’), 4.67 (dd, 1H, J = 6.9 and 2.3 Hz, H3’), 3.93 (s, 3H,

OCH3), 3.20 (d, 1H, J = 6.9 Hz, OH), 1.51 (s, 9H, t-Bu4), 1.44 (s, 9H, t-Bu1). 13C NMR: (400 MHz, CDCl3) δ: 170.3 (C4’), 166.0 (C1’), 165.9 (C9), 148.4 (C1), 146.9

(C2), 146.7 (C7), 126.8 (C4), 123.7 (C5), 114.8 (C6), 113.8 (C8), 109.3 (C3), 84.0

(C1(CH3)3), 83.3 (C4(CH3)3), 73.3 (C2’), 71.0 (C3’), 56.1 (OCH3), 28.1 (C4(CH3)3), 28.0

(C1(CH3)3).

IR (neat) ν: 3465, 2929, 1726, 1146, 1120, 981, 844.

HRMS calculated for C22H30O9 [M + Na]+ 461.1788, found 461.1772.


156

p-Coumaroyl L-tartrate (7)

(2R,3R)-2-Hydroxy-3-{[(2E)-3-(4-hydroxyphenyl)prop-2-enoyl]oxy}butanedioic acid

p-Coumaroyl tert-butyl L-tartrate (39) (46.2 mg, 0.11 mmol) was dissolved in dry

dichloromethane (5 mL) followed by the addition of trifluoroacetic acid (0.18 mL, 2.29

mmol) and the mixture stirred at room temperature under a nitrogen atmosphere for 24

hours before being concentrated. Purification by reversed-phase chromatography (C18,

eluted with acetonitrile/H2O/formic acid, 30:69:1) gave 7 as an amorphous solid, 27.5 mg

(82%).

Rf (20% MeOH/DCM): 0.00 1H NMR: (400 MHz, CD3OD) δ: 7.74 (d, 1H, J = 15.9 Hz, H7), 7.48 (app. d, 2H, J = 8.7

Hz, H3,5), 6.81 (app. d, 2H, J = 8.7 Hz, H2,6), 6.38 (d, 1H, J = 15.9 Hz, H8), 5.55 (d, 1H, J =

2.3 Hz, H2’), 4.77 (d, 1H, J = 2.3 Hz, H3’). 13C NMR: (400 MHz, CDCl3) δ: 174.0 (C4’), 170.8 (C1’), 168.0 (C9), 161.5 (C1), 147.8

(C7), 131.4 (C3,5), 127.1 (C4), 116.8 (C2,6), 114.1 (C8), 74.9 (C2’), 71.7 (C3’).

All physical and chemical properties were as previously reported.243

Feruloyl L-tartrate (8)

(2R,3R)-2-Hydroxy-3-{[(2E)-3-(4-hydroxy-3-methoxyphenyl)prop-2-

enoyl]oxy}butanedioic acid

Feruloyl tert-butyl L-tartrate (40) (35.9 mg, 0.082 mmol) was dissolved in dry

dichloromethane (2 mL) followed by the addition of trifluoroacetic acid (0.13 mL, 1.63

mmol) and the mixture stirred at room temperature under a nitrogen atmosphere for 24

hours before being concentrated. Purification by reversed-phase chromatography (C18,

eluted with acetonitrile/H2O/formic acid, 30:69:1) gave 8 as an off-white amorphous solid,

11.0 mg (41%).

Rf (20% MeOH/DCM): 0.00


157

1H NMR: (400 MHz, CD3OD) δ: 7.73 (d, 1H, J = 16.0 Hz, H7), 7.20 (d, 1H, J = 1.9 Hz,


16.0 Hz, H8), 5.57 (d, 1H, J = 2.4 Hz, H2’), 4.78 (d, 1H, J = 2.4 Hz, H3’), 3.89 (s, 3H,

OCH3). 13C NMR: (400 MHz, CDCl3) δ: 173.8 (C4’), 170.6 (C1’), 167.9 (C9), 150.8 (C1), 149.4

(C2), 148.1 (C7), 127.6 (C4), 124.3 (C5), 116.4 (C6), 114.3 (C8), 111.7 (C3), 74.9 (C2’), 71.7

(C3’), 56.5 (OCH3).

All physical and chemical properties were as previously reported.243-244

6.2.3 Synthesis of Hydroxycinnamoyl Glucose Esters

Chloroacetyl chloride (46)

Chloroacetic acid (100.25 g, 1.06 mol) and thionyl chloride (75 mL, 1.03 mol) were heated

under reflux for 2 hours under a nitrogen atmosphere. Distillation under nitrogen gave

chloroacetyl chloride (45-55% yield over a number of attempts) as a clear liquid, b.p. 102-

106 oC (lit. b.p. 106 oC),245 and left chloroacetic anhydride as a clear solid, m.p. 49-50 oC

(lit. m.p 48-60 oC)246 which could be separated from residual chloroacetic acid by

kugelrohr distillation under reduced pressure (105 oC at 20 mmHg). (Chloroacetic acid lit.

b.p. 189 oC, 120-123 oC at 20 mmHg).246

1,2,3,4,6-Penta-O-chloroacetyl-D-glucopyranoside (41)

1,2,3,4,6-Pentakis-O-(chloroacetyl)-D-glucopyranose

D-Glucose (4.02 g, 22.4 mmol) was dissolved in dry dichloromethane (90 mL) and dry

pyridine (10 mL). Chloroacetyl chloride (46) (17.6 mL, 221.0 mmol) was added dropwise

at 0 oC, following which the reaction mixture was heated under reflux for 24 hours. The

reaction mixture was poured onto ice water (100 mL), the organics were separated then

washed with 2 M HCl solution (3 x 100 mL), saturated sodium bicarbonate solution (3 x


158

100 mL), saturated brine solution (2 x 100 mL), dried (MgSO4) and concentrated. The

crude mixture was purified by column chromatography (DCM) to give 11.42 g (91 %) of

41 as a pale yellow gum.

Rf (50% EtOAc/X4): 0.56 1H NMR: (600 MHz, CDCl3) δ: 6.43 (d, 0.55H, J = 3.7 Hz, H1α), 5.82 (d, 0.45H, J = 8.2

Hz, H1β), 5.59 (dd, 0.55H, J = 9.9 Hz, H3α), 5.40 (dd, 0.45H, J = 9.5 Hz, H3β), 5.27-5.22

(m, 2H, H2,4), 4.40-4.36 (m, 1H, H6a), 4.33-4.30 (m, 1.45H, H6b,5β), 4.25 (ddd, 0.55H, J =

10.3, 3.9 and 2.3 Hz, H5α), 4.19-4.00 (m, 10H, OCH2Cl). Spectral properties were as

previously reported.142, 182

αααα-Anomer

Isolated as the remaining starting material in the synthesis of 42 (below). 1H NMR: (400 MHz, CDCl3) δ: 6.43 (d, 1H, J = 3.7 Hz, H1), 5.58 (dd, 1H, J = 10.0 and

9.6 Hz, H3), 5.28-5.21 (m, 2H, H2,4), 4.38 (dd, 1H, J = 12.6 and 4.0 Hz, H6a), 4.31 (dd, 1H,

J = 12.6 and 2.3 Hz, H6b), 4.25 (ddd, 1H, J = 10.2, 4.0 and 2.3 Hz, H5), 4.18, 4.12 (2 x

app.s, 2H, OCH2Cl), 4.04, 4.03 (2 x app. s, 1H, OCH2Cl), 4.01, 4.00 (2 x app. s, 2H,

OCH2Cl).

2,3,4,6-Tetra-O-chloroacetyl-D-glucopyranoside (42)

2,3,4,6-Tetrakis-O-(chloroacetyl)-D-glucopyranose

1,2,3,4,6-Penta-O-chloroacetyl-D-glucopyranoside (41) (11.15 g, 19.82 mmol) was

dissolved in THF (150 mL) followed by the addition of hydrazine acetate (1.83 g, 19.88

mmol) and the reaction mixture stirred at ambient temperature. After 5 hours the reaction

mixture was concentrated in vacuo and the crude product purified by column

chromatography (DCM - 5% Et2O/DCM) to afford 5.32 g (55 %) of 42 as an amorphous

solid.

Rf (50% EtOAc/X4): 0.36 1H NMR: (400 MHz, CDCl3) δ: 5.66 (dd, 0.7H, J = 10.0 and 9.7 Hz, H3α), 5.54 (d, 0.7H, J

= 3.6 Hz, H1α), 5.38 (dd, 0.3H, J = 9.7 and 9.6 Hz, H3β), 5.19 (dd, 0.3H, J = 9.7 and 9.7

Hz, H4β), 5.18 (dd, 0.7H, J = 9.7 and 9.6 Hz, H4α), 5.02-4.98 (m, 1H, J = 3.6 and 10.0 Hz,


159

H2), 4.85 (d, 0.3H, J = 8.0 Hz, H1β), 4.40-4.32 (m, 3H, H5,6a,6b), 4.14-4.00 (m, 8H,

OCH2Cl). Spectral properties were as previously reported.142, 182

2,3,4,6-Tetra-O-chloroacetyl-D-glucopyranosyltrichloroacetimidate (43)

2,3,4,6-Tetrakis-O-(chloroacetyl)-1-O-(2,2,2-trichloroethanimidoyl)-D-glucopyranose

2,3,4,6-Tetra-O-chloroacetyl-D-glucopyranoside (42) (5.04 g, 10.36 mmol) was dissolved

in dry dichloromethane (100 mL) followed by the addition of trichloroacetonitrile (10.38

mL, 103.52 mmol), DBU (0.31 mL, 2.07 mmol) and the reaction mixture stirred at ambient

temperature. After 4 hours the reaction mixture was concentrated and the crude product

purified by column chromatography (DCM - 2% Et2O/DCM) to give 4.92 g (75%) of 43 as

a pale yellow amorphous solid.

Rf (30 % EtOAc/X4): 0.27 1H NMR: (600 MHz, CDCl3) δ: 8.79 (s, 0.33H, NHβ), 8.77 (s, 0.67H, NHα), 6.60 (d,

0.67H, J = 3.7 Hz, H1α), 5.94 (d, 0.33H, J = 7.6 Hz, H1β), 5.68 (dd, 0.67H, J = 9.8 and 9.8

Hz, H3α), 5.42 (dd, 0.33H, J = 8.4 and 8.4 Hz, H3β), 5.30-5.24 (m, 2H, H2,4), 4.39-4.31 (m,

3H, H5,6a,6b), 4.13-3.99 (m, 8H, OCH2Cl). Spectral properties were as previously

reported.142, 144

General procedure for hydroxycinnamate glycosylation

2,3,4,6-Tetra-O-chloroacetyl-D-glucopyranosyltrichloroacetimidate (43) (0.55 mmol) was

dissolved in dry dichloromethane (10 mL) containing 4Å molecular sieves, followed by the

addition of the hydroxycinnamate (3, 4, 15, 16, 19-22, 47) (0.45 mmol). After 20 minutes

of stirring at ambient temperature trimethylsilyl triflate (0.33 mmol) was added slowly.

The reaction mixture was stirred at room temperature for a further 4 hours, after which it

was quenched with saturated sodium bicarbonate solution (15 mL), washed with water (3 x

15 mL), then brine (2 x 15 mL), dried (MgSO4) and concentrated in vacuo. Purification

with column chromatography (20-30% EtOAc/X4) yielded the product (44, 45, 48-50, 54,


160

55, 57, 58). Analogues that were achieved as a gum could be solidified via trituration with

methanol, and the melting points quoted are the point at which the solid reverted back to a

gum.

2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl cinnamate (48)

2,3,4,6-Tetrakis-O-(chloroacetyl)-1-O-[(2E)-3-phenylprop-2-enoyl]-ββββ-D-glucopyranose

From 47 (122.3 mg, 0.83 mmol), gave 0.27 g (54 %) of 48 as a white solid.

Rf (50 % EtOAc/X4): 0.53 1H NMR: (400 MHz, CDCl3) δ: 7.77 (d, 1H, J = 16.0 Hz, H7), 7.56-7.53 (m, 2H, H3,5),

7.42-7.41 (m, 3H, H1,2,6), 6.41 (d, 1H, J = 16.0 Hz, H8), 5.91 (d, 1H, J = 8.2 Hz, H1’), 5.44

(dd, 1H, J = 9.5 and 9.4 Hz, H3’), 5.34 (dd, 1H, J = 9.5 and 8.2 Hz, H2’), 5.27 (dd, 1H, J =

10.0 and 9.4 Hz, H4’), 4.42 (dd, 1H, J = 12.6 and 4.2 Hz, H6a’), 4.33 (dd, 1H, J = 12.6 and

2.3 Hz, H6b’), 4.12 (s 2H, OCH2Cl), 4.05 (d, 1H, J = 14.5 Hz, OCHaHbCl), 4.03-3.99 (m,

5H, 2x OCH2Cl and H5’), 4.02 (d, 1H, J = 14.5 Hz, OCHaHbCl). Spectral properties were

as previously reported.144

2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl 1-O-benzyl p-coumarate (44)

1-O-{(2E)-3-[4-(Benzyloxy)phenyl]prop-2-enoyl}-2,3,4,6-tetrakis-O-(chloroacetyl)-ββββ-D-

glucopyranose

From 15 (106.9 mg, 0.42 mmol), yielded 158.4 mg (52%) of 44 as a white solid. m.p

126.2-127.7 oC.

Rf (50% EtOAc/X4): 0.74

= -16.99o (c 1.5, chloroform)


161

1H NMR: (400 MHz, CDCl3) δ: 7.71 (d, 1H, J = 16.0 Hz, H7), 7.49 (app. d, 2H J = 8.8 Hz,


H8), 5.91 (d, 1H, J = 8.2 Hz, H1’), 5.44 (dd, 1H, J = 9.5 and 9.5 Hz, H3’), 5.33 (dd, 1H, J =

9.5 and 8.2 Hz, H2’), 5.26 (dd, 1H, J = 9.7 and 9.5 Hz, H4’), 5.10 (s, 2H, CH2Bn), 4.41 (dd,

1H, J = 12.5 and 4.3 Hz, H6a’), 4.32 (dd, 1H, J = 12.5 and 2.3 Hz, H6b’), 4.12 (app. s, 2H,

OCH2Cl), 4.04-3.99 (m, 7H, 3x OCH2Cl and H5’). 13C NMR: (400 MHz, CDCl3) δ: 167.1, 167.0, 166.4, 166.3 (4 x OCOCH2Cl), 164.9 (C9),

161.4 (C1), 148.1 (C7), 136.4 (Ar), 130.6 (C3,5), 129.0 (2 x Ar), 128.5 (Ar), 127.7 (2 x Ar),

126.8 (C4), 115.5 (C2,6), 113.0 (C8), 91.5 (C1’), 73.9 (C3’), 72.0 (C5’), 71.4 (C2’), 70.3

(CH2Ph), 69.2 (C4’), 62.9 (C6’), 40.9, 40.7, 40.4, 40.3 (4 x OCH2Cl).

IR (neat) ν: 2960, 1754, 1149, 1101, 1064, 826, 744.

HRMS calculated for C30H28Cl4O12 [M + Na]+ 745.0203, found 745.0163.

2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl 1-O-benzyl ferulate (45)

1-O-{(2E)-3-[4-(Benzyloxy)-3-methoxyphenyl]prop-2-enoyl}-2,3,4,6-tetrakis-O-

(chloroacetyl)-ββββ-D-glucopyranose

From 16 (127.0 mg, 0.45 mmol), gave 183.1 mg (54%) of 45 as a white residue. Addition

of methanol followed by evaporation under reduced pressure gave a white solid. m.p 55.7-

59.9 oC.

Rf (50% EtOAc/X4): 0.66

= -8.85o (c 1.1, chloroform) 1H NMR: (400 MHz, CDCl3) δ: 7.68 (d, 1H, J = 15.9 Hz, H7), 7.44-7.42 (m, 2H, ArH),

7.39-7.36 (m, 2H, ArH), 7.33-7.30 (m, 1H, ArH), 7.09-7.04 (m, 2H, H3,5), 6.88 (d, 1H, J =

8.2 Hz, H6), 6.26 (d, 1H, J = 15.9 Hz, H8), 5.90 (d, 1H, J = 8.2 Hz, H1’), 5.43 (dd, 1H, J =

9.5 and 9.5 Hz, H3’), 5.33 (dd, 1H, J = 9.5 and 8.2 Hz, H2’), 5.26 (dd, 1H, J = 9.6 and 9.5

Hz, H4’), 5.20 (s, 2H, CH2Ph), 4.42 (dd, 1H, J = 12.5 and 4.4 Hz, H6a’), 4.32 (dd, 1H, J =

12.5 and 2.3 Hz, H6b’), 4.12 (app. s, 2H, OCH2Cl), 4.04 and 4.03 (2 x app. s, 2 x 1H, O

CH2Cl), 4.01 and 4.01 (m, 5H, 2 x OCH2Cl and H5’), 3.93 (s, 3H, OCH3).


162

13C NMR: (400 MHz, CDCl3) δ: 167.3, 167.1, 166.6, 166.5 (4 x OCOCH2Cl), 165.0 (C9),

151.3 (C1), 150.1 (C2), 148.4 (C7), 136.6 (Ar), 129.0 (2 x Ar), 128.4 (Ar), 127.5 (2 x Ar),

127.2 (C4), 123.6 (C5), 113.5 (C6), 113.4 (C8), 110.5 (C3), 91.6 (C1’), 73.5 (C3’), 72.2 (C5’),

71.1 (C2’), 70.7 (CH2Bn), 69.3 (C4’), 62.8 (C6’), 56.2 (OCH3), 40.8, 40.5, 40.5, 40.5 (4 x

OCH2Cl).

IR (neat) ν: 2958, 1754, 1135, 1068, 1000, 792, 697.


2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl p-coumarate (49)

2,3,4,6-Tetrakis-O-(chloroacetyl)-1-O-[(2E)-3-(4-hydroxyphenyl)prop-2-enoyl]-ββββ-D-

glucopyranose

From 3 (53.5 mg, 0.33 mmol), afforded 60.0 mg (29 %) of a mixture of 49, and a minor

unidentified impurity.

Rf (50% EtOAc/X4): 0.53 1H NMR: (400 MHz, CDCl3) δ: 7.70 (d, 1H, 15.8 Hz, H7), 7.45 (app. d, 2H, J = 8.7 Hz,

H3,5), 6.86 (app. d, 2H, J = 8.7 Hz, H2,6), 6.52 (d, 1H, J = 8.2 Hz, H1’), 4.43-4.30 (m, 2H,

H6a’,6b’), 4.12-3.99 (4 x OCH2Cl). Assignment and identification was based on

unobstructed proton shifts, and the known spectrum of the feruloyl analogue (50).

2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl ferulate (50)

2,3,4,6-Tetrakis-O-(chloroacetyl)-1-O-[(2E)-3-(4-hydroxy-3-methoxyphenyl)prop-2-

enoyl]-ββββ-D-glucopyranose

From 4 (157.4 mg, 0.81 mmol), trituration of the columned product with methanol gave

207.6 mg (39%) of 50 as a white solid. m.p. 157.8-158.7 oC.


163

Rf (50 % EtOAc/X4): 0.48 1H NMR: (400 MHz, CDCl3) δ: 7.69 (d, 1H, J = 15.9 Hz, H7), 7.10 (dd, 1H, J = 8.2 and


Hz, H8), 5.90 (d, 1H, J = 8.2 Hz, H1’), 5.43 (dd, 1H, J = 9.6 and 9.5 Hz, H3’), 5.33 (dd, 1H,

J = 9.5 and 8.2 Hz, H2’), 5.26 (dd, 1H, J = 9.7 and 9.6 Hz, H4’), 4.42 (dd, 1H, J = 12.6 and

4.2 Hz, H6a’), 4.32 (dd, 1H, J = 12.6 and 2.4 Hz, H6b’), 4.12 (app. s, 2H, OCH2Cl), 4.04-

3.99 (m, 7H, 3 x OCH2Cl and H5’), 3.95 (s, 3H, OCH3). Spectral properties were as

previously reported.183

2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl 1-O-acetyl p-coumarate (54)

1-O-{(2E)-3-[4-(Acetyloxy)phenyl]prop-2-enoyl}-2,3,4,6-tetrakis-O-(chloroacetyl)-ββββ-D-

glucopyranose

From 19 (133.7 mg, 0.65 mmol) , gave 146.3 mg (40 %) of 54 as a white residue. Addition

of methanol followed by evaporation under reduced pressure gave a white solid. m.p 55.6-

58.7 oC.

Rf (50% EtOAc/X4): 0.60

= -11.58o (c 1.6, chloroform) 1H NMR: (400 MHz, CDCl3) δ: 7.74 (d, 1H, J = 16.0 Hz, H7), 7.56 (app. d, 2H, J = 8.7 Hz,


Hz, H1’), 5.44 (dd, 1H, J = 9.5 and 9.5 Hz, H3’), 5.33 (dd, 1H, J = 9.5 and 8.3 Hz, H2’), 5.26

(dd, 1H, J = 9.6 and 9.5 Hz, H4’), 4.42 (dd, 1H, J = 12.5 and 4.2 Hz, H6a’), 4.33 (dd, 1H, J

= 12.5 and 2.2 Hz, H6b’), 4.12 (app. s, 2H, OCH2Cl), 4.04-4.00 (m, 7H, 3 x OCH2Cl and

H5’), 2.32 (s, 3H, OCOCH3). 13C NMR: (400 MHz, CDCl3) δ: 169.2 (OCOCH3), 167.1, 167.0, 166.4, 166.3 (4 x

OCOCH2Cl), 164.5 (C9), 152.8 (C1), 147.0 (C7), 131.5 (C4), 129.9 (C3,5), 122.5 (C2,6),

115.9 (C8), 91.6 (C1’), 72.3 (C3’), 72.1 (C5’), 71.3 (C2’), 69.2 (C4’), 62.8 (C6’), 40.7, 40.4,

40.4, 40.3 (4 x OCH2Cl), 20.8 (OCOCH3).

IR (neat) ν: 2959, 1757, 1147, 1069, 1005, 912, 791.



164

2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl 1-O-acetyl ferulate (55)

1-O-{(2E)-3-[4-(Acetyloxy)-3-methoxyphenyl]prop-2-enoyl}-2,3,4,6-tetrakis-O-

(chloroacetyl)-ββββ-D-glucopyranose

From 20 (132.4 mg, 0.56 mmol), afforded 168.7 mg (44%) of 55. Addition of methanol

followed by evaporation under reduced pressure gave a white solid. m.p 59.8-62.3 oC.

Rf (50 % EtOAc/X4): 0.50

= -4.63o (c 1.1, chloroform) 1H NMR: (400 MHz, CDCl3) δ: 7.72 (d, 1H, J = 16.0 Hz, H7), 7.14 (dd, 1H, J = 8.1 and




4.2 Hz, H6a’), 4.33 (dd, 1H, J = 12.5 and 2.2 Hz, H6b’), 4.13 (app. s, 2H, OCH2Cl), 4.04-

4.00 (m, 7H, 3 x OCH2Cl and H5’), 3.88 (s, 3H, OCH3), 2.33 (s, 3H, OCOCH3). 13C NMR: (400 MHz, CDCl3) δ: 168.9 (OCOCH3), 167.1, 167.0, 166.4, 166.3 (4 x

OCOCH2Cl), 164.4 (C9), 151.7 (C1), 147.5 (C7), 142.3 (C2), 132.7 (C4), 123.6 (C6), 122.0

(C5), 115.9 (C8), 111.5 (C3), 91.5 (C1’), 73.9 (C3’), 72.3 (C5’), 71.2 (C2’), 69.2 (C4’), 62.8

(C6’), 56.1 (OCH3), 40.7, 40.4, 40.3, 40.3 (4 x OCH2Cl), 20.8 (OCOCH3).

IR (neat) ν: 2959, 1756, 1147, 1070, 1004, 791.


2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl 1-O-chloroacetyl p-coumarate (57)

2,3,4,6-Tetrakis-O-(chloroacetyl)-1-O-[(2E)-3-{4-[(chloroacetyl)oxy]phenyl}prop-2-

enoyl]-ββββ-D-glucopyranose

From 21 (210.6 mg, 0.88 mmol), gave 0.30 g (48%) of 57 as a white honeycomb. m.p

48.0-51.0 oC.


165

Rf (50% EtOAc/X4): 0.59

= -4.60o (c 0.9, chloroform) 1H NMR: (400 MHz, CDCl3) δ: 7.73 (d, 1H, J = 16.0 Hz, H7), 7.58 (app. d, 2H, J = 8.7 Hz,


Hz, H1’), 5.44 (dd, 1H, J = 9.6 and 9.5 Hz, H3’), 5.33 (dd, 1H, J = 9.5 and 8.2 Hz, H2’), 5.26

(dd, 1H, J = 9.7 and 9.6 Hz, H4’), 4.42 (dd, 1H, J = 12.5 and 4.3 Hz, H6a’), 4.33 (m, 3H,

ArOCOCH2Cl and H6b’), 4.12 (app. s, 2H, OCH2Cl), 4.04-4.01 (m, 7H, 3 x OCH2Cl and

H5’). 13C NMR: (400 MHz, CDCl3) δ: 167.1, 166.9, 166.4, 166.2 (4 x Glc-OCOCH2Cl), 165.6

(ArOCOCH2Cl), 164.3 (C9), 152.3 (C1), 146.7 (C7), 132.0 (C4), 129.9 (C3,5), 122.0 (C2,6),

116.4 (C8), 91.5 (C1’), 73.9 (C3’), 72.2 (C5’), 71.4 (C2’), 69.2 (C4’), 62.8 (C6’), 40.9

(ArOCOCH2Cl), 40.6, 40.3, 40.3, 40.3 (4 x Glc-OCOCH2Cl).

IR (neat) ν: 2958, 1752, 1138, 1069, 1002, 926, 792, 754.


2,3,4,6-Tetra-O-chloroacetyl-ββββ-D-glucopyranosyl 1-O-chloroacetyl ferulate (58)

2,3,4,6-Tetrakis-O-(chloroacetyl)-1-O-[(2E)-3-{4-[(chloroacetyl)oxy]-3-

methoxyphenyl}prop-2-enoyl]-ββββ-D-glucopyranose

From 22 (224.0 mg, 0.83 mmol), afforded 0.38 g (64%) of 58 as a pale yellow honeycomb.

m.p 62.0-64.7 oC.

Rf (40% EtOAc/X4): 0.46 1H NMR: (400 MHz, CDCl3) δ: 7.71 (d, 1H, J = 16.0 Hz, H7), 7.15 (dd, 1H, J = 8.1 and




4.3 Hz, H6a’), 4.33 (m, 3H, ArOCH2Cl and H6b’), 4.12 (app, s. 2H, OCH2Cl), 4.04-4.01 (m,

7H, 3 x OCH2Cl and H5’), 3.88 (s, 3H, OCH3). 13C NMR: (400 MHz, CDCl3) δ: 167.1, 166.9, 166.4, 166.3 (4 x Glc-OCOCH2Cl), 165.2

(ArOCOCH2Cl), 164.3 (C9), 151.4 (C1), 147.2 (C7), 141.6 (C2), 133.2 (C4), 123.2 (C6),


166

121.9 (C5), 116.4 (C8), 111.7 (C3), 91.5 (C1’), 73.9 (C3’), 72.3 (C5’), 71.4 (C2’), 69.2 (C4’),

62.8 (C6’), 56.2 (OCH3), 40.6-40.3 (5 x OCH2Cl).


Physical and chemical properties were as previously reported.183

General procedure for de-chloroacetylation (ambient light)

2,3,4,6-Tetra-O-chloroacetyl-β-D-glucopyranosyl hydroxycinnamate (48-50) (100.0 mg)

was dissolved in pyridine/water (1:1, 10 mL) and stirred at room temperature for 4 hours.

The reaction mixture was concentrated and the crude mixture purified using XAD-8 resin

(eluted with 60% MeOH/H2O) to give a mixture of cis- and trans-β-D-glucopyranosyl

hydroxycinnamate as a colourless residue (53, 9, 10).

1-O-ββββ-D-Glucopyranosyl cinnamate (53)

1-O-(3-Phenylacryloyl)-ββββ-D-glucopyranose

From 48 (95.6 mg, 0.16 mmol), gave 10 mg (21%) of 53 as a mixture of cis/trans-isomers.

Rf (20% MeOH/DCM): 0.40

1-O-ββββ-D-Glucopyranosyl trans-cinnamate

1-O-[(2E)-3-Phenylprop-2-enoyl]-ββββ-D-glucopyranose 1H NMR: (400 MHz, CD3OD) δ: 7.81 (d, 1H, J = 16.0 Hz, H7), 7.64-7.62 (m, 2H, ArH),

7.44-7.40 (m, 3H, ArH), 6.58 (d, 1H, J = 16.0 Hz, H8), 5.60 (d, 1H, J = 7.7 Hz, H1’), 3.86

(dd, 1H, J = 12.1 and 2.0 Hz, H6a’), 3.70 (dd, 1H, J = 12.1 and 4.8 Hz, H6b’), 3.50-3.35 (m,

4H, H2’,3’,4’,5’).

1-O-ββββ-D-Glucopyranosyl cis-cinnamate

1-O-[(2Z)-3-Phenylprop-2-enoyl]-ββββ-D-glucopyranose 1H NMR: (400 MHz, CD3OD) δ: 7.69-7.66 (m, 2H, ArH), 7.35-7.33 (m, 3H, ArH), 7.10

(d, 1H, J = 12.7 Hz, H7), 6.02 (d, 1H, J = 12.7 Hz, H8), 5.54 (d, 1H, J = 8.1 Hz, H1’), 3.87-

3.83 (m, 1H, H6a’), 3.71-3.66 (m, 1H, H6b’), 3.49-3.34 (m, 4H, H2’,3’,4’,5’).


167

The spectrum of each isomer was extracted from the mixture. Spectral properties for the

trans-isomer were as previously reported.247-248 Assignment and identification of the cis-

isomer was performed using the known trans-isomer (trans-53) and the data for the cis-

aglycone (cis-47).235

1-O-ββββ-D-Glucopyranosyl p-coumarate (9)

1-O-[3-(4-Hydroxyphenyl)acryloyl]-ββββ-D-glucopyranose

From 49 (60.0 mg, 0.10 mmol), yielded 10.3 mg (33%) of 9 as a mixture of cis/trans-

isomer, as well as 9.6 mg (25%) of 51.

Rf (20% MeOH/DCM): 0.29

1-O-ββββ-D-Glucopyranosyl trans-p-coumarate

1-O-[(2E)-3-(4-Hydroxyphenyl)prop-2-enoyl]-ββββ-D-glucopyranose 1H NMR: (400 MHz, CD3OD) δ: 7.73 (m, 1H, H7), 7.48 (app. d, 2H, J = 8.5 Hz, H3,5), 6.82

(app. d, 2H, J = 8.5 Hz, H2,6), 6.37 (d, 1H, J = 15.9 Hz, H8), 5.58 (d, 1H, J = 8.0 Hz, H1’),

3.85 (m, 1H, H6a’), 3.68 (m, 1H, H6b’), 3.47-3.32 (m, 4H, H2’,3’,4’,5’).

1-O-ββββ-D-Glucopyranosyl cis-p-coumarate

1-O-[(2Z)-3-(4-Hydroxyphenyl)prop-2-enoyl]-ββββ-D-glucopyranose 1H NMR: (400 MHz, CD3OD) δ: 7.73 (m, 2H, H3,5), 6.94 (d, 1H, J = 12.9 Hz, H7), 6.82

(app. d, 2H, J = 8.8 Hz, H2,6), 5.82 (d, 1H, J = 12.9 Hz, H8), 5.55 (d, 1H, J = 8.0 Hz, H1’),

3.85 (m, 1H, H6a’), 3.68 (m, 1H, H6b’), 3.47-3.32 (m, 4H, H2’, 3’, 4’, 5’).


trans-isomer were as previously reported.108, 179 Assignment and identification of the cis-

isomer was performed using the known trans-isomer (trans-9) and the data for the cis-

aglycone (cis-3).214, 235


168

6-O-Chloroacety-ββββ-D-glucopyranosyl trans-p-coumarate (51)

6-O-(Chloroacetyl)-1-O-[(2E)-3-(4-hydroxyphenyl)prop-2-enoyl]-ββββ-D-glucopyranose

Isolated during the synthesis of 9.

Rf (20% MeOH/DCM): 0.58 1H NMR: (400 MHz, CD3OD) δ: 7.73 (d, 1H, J = 15.9, H7), 7.49 (app. d, 2H, J = 8.6 Hz,


Hz, H1’), 4.49 (dd, 1H, J = 12.0 and 2.2 Hz, H6a’), 4.33 (dd, 1H, J = 12.0 and 5.6 Hz, H6b’),

4.23 (s, 2H, OCH2Cl), 3.49-3.36 (m, 4H, H2’,3’,4’,5’).

1-O-ββββ-D-Glucopyranosyl ferulate (10)

1-O-[3-(4-Hydroxy-3-methoxyphenyl)acryloyl]-ββββ-D-glucopyranose

From 50 (219.3 mg, 0.33 mmol), gave 24.1 mg (20%) of 10 as a mixture of cis/trans-

isomer.

Rf (20 % MeOH/DCM): 0.32

ββββ-D-Glucopyranosyl trans-ferulate

1-O-[(2E)-3-(4-Hydroxy-3-methoxyphenyl)prop-2-enoyl]-ββββ-D-glucopyranose 1H NMR: (400 MHz, CD3OD) δ: 7.73 (d, 1H, J = 15.9 Hz, H7), 7.21 (d, 1H, J = 1.9 Hz,


15.9 Hz, H8), 5.58 (d, 1H, J = 7.5 Hz, H1’), 3.90 (s, 3H, OCH3), 3.86 (m, 1H, H6a’), 3.70

(m, 1H, H6b’), 3.49-3.35 (m, 4H, H2’, 3’,4’,5’).

ββββ-D-Glucopyranosyl cis-ferulate

1-O-[(2Z)-3-(4-Hydroxy-3-methoxyphenyl)prop-2-enoyl]-ββββ-D-glucopyranose 1H NMR: (400 MHz, CD3OD) δ: 7.87 (d, 1H, J = 1.9 Hz, H3), 7.17 (dd, 1H, J = 8.3 and

1.9 Hz, H5), 6.94 (d, 1H, J = 13.0 Hz, H7), 6.77 (d, 1H, J = 8.3 Hz, H6), 5.83 (d, 1H, J =


169

13.0 Hz, H8), 5.56 (d, 1H, J = 7.8 Hz, H1’), 3.88 (s, 3H, OCH3), 3.88-3.84 (m, 1H, H6a’),

3.72-3.66 (m, 1H, H6b’), 3.49-3.35 (m, 4H, H2’,3’,4’,5’).


trans-isomer were as previously reported.108, 179, 183 Assignment and identification of the

cis-isomer was performed using the known trans-isomer (trans-10) and the data for the cis-

aglycone (cis-4).184, 235

General procedure for de-chloroacetylation (red light)

2,3,4,6-Tetra-O-chloroacetyl-β-D-glucopyranosyl hydroxycinnamate (54, 57, 58) (300.0

mg) was dissolved in pyridine/water (1:1, 20 mL) and stirred at room temperature in the

dark for 4 hours (54) or for 6 hours (57 and 58). Only being exposed to red light, the

reaction mixture was concentrated and the crude mixture purified with XAD-8 resin

(eluted with 60% MeOH/H2O) to give a trans-β-D-glucopyranosyl hydroxycinnamate as a

colourless residue.

1-O-ββββ-D-Glucopyranosyl 1-O-acetyl p-coumarate (56)

1-O-{(2E)-3-[4-(Acetyloxy)phenyl]prop-2-enoyl}-ββββ-D-glucopyranose

From 54 (113.8 mg, 0.17 mmol), gave 12.7 mg (20%) of 56 as a white residue containing

minor impurities of 9, as well as 2.3 mg (4%) of 9.

Rf (20 % MeOH/DCM): 0.39 1H NMR: (400 MHz, CD3OD) δ: 7.80 (d, 1H, J = 16.0 Hz, H7), 7.67 (app. d, 2H J = 8.5


8.0 Hz, H1’), 3.86 (dd, 1H, J = 12.1 and 1.9 Hz, H6a’), 3.70 (dd, 1H, J = 12.1 and 4.8 Hz,

H6b’), 3.47-3.38 (m, 4H, H2’,3’,4’,5’), 2.29 (s, 3H, OCOCH3).


170

1-O-ββββ-D-Glucopyranosyl trans-p-coumarate (9)

1-O-[(2E)-3-(4-Hydroxyphenyl)prop-2-enoyl]-ββββ-D-glucopyranose

From 57 (261.7 mg, 0.37 mmol), gave 51.5 mg (43%) of 9 as a white residue, determined

to have undergone acyl migration, largely consisting of the 1-O-β-ester (approx. 80%). The

migrated mixture was found to revert back to the 1-O-β-ester after standing in pH 3.5

model wine media.

Rf (20 % MeOH/DCM): 0.29 1H NMR: (400 MHz, CD3OD) δ: 7.73 (d, 1H, J = 15.9 Hz, H7), 7.48 (app. d, 2H, J = 8.5


7.9 Hz, H1’), 3.85 (dd, 1H, J = 12.1 and 1.8 Hz, H6a’), 3.69 (dd, 1H, J = 12.1 and 4.6 Hz,

H6b’), 3.45-3.38 (m, 4H, H2’,3’,4’,5’).

MS (-EI) m/z (%): 325.7 (M-, 100), 265.5 (7), 187.7 (8), 163.4 (21), 145.2 (44).

Physical and chemical properties for the 1-O-β-ester were as previously reported.108, 131, 179

1-O-ββββ-D-Glucopyranosyl trans-ferulate (10)

1-O-[(2E)-3-(4-Hydroxy-3-methoxyphenyl)prop-2-enoyl]-ββββ-D-glucopyranose

From 58 (501.8 mg, 0.68 mmol), 47.5 mg (20%) of 10 as an off-white residue, determined

to have undergone acyl migration, largely consisting of the 1-O-β-ester (approx. 90%). The

migrated mixture was found to revert back to the 1-O-β-ester after standing in pH 3.5

model wine media.

Rf (20% MeOH/DCM): 0.32 1H NMR: (400 MHz, CD3OD) δ: 7.73 (d, 1H, J = 15.9 Hz, H7), 7.21 (d, 1H, J = 1.9 Hz,


15.9 Hz, H8), 5.58 (d, 1H, J = 7.5 Hz, H1’), 3.90 (s, 3H, OCH3), 3.86 (dd, 1H, J = 12.1 and

2.1 Hz, H6a’), 3.70 (dd, 1H, J = 12.1 and 4.5 Hz, H6b’), 3.49-3.35 (m, 4H, H2’,3’,4’,5’).


171

MS (-EI) m/z (%): 355.3 (M-, 100), 295.5 (8), 217.2 (20), 193.6 (25), 175.4 (32).

Physical and chemical properties for the 1-O-β-ester were as previously reported.108, 179, 183,

200


172


Theoretical studies into the thermodynamics of glucose ester migration

The ten possible glucose esters (1/2/3/4/6-O-α/β-) were drawn for both p-coumaroyl

glucose (9) and feruloyl glucose (10) using the equilibrium geometry optimised in four

different solvents (water, dichloromethane, ethanol and toluene), and the energies given in

Hartrees (a.u.) were converted to kJ/mol using a factor of 2625.5. The final ester energies

were calculated relative to the 1-O-β-ester in each case. The raw data is displayed in

Appendix 1.

Theoretical studies into the kinetics of glucose ester migration

The four intermediates for each migration were drawn and the equilibrium geometry

optimised for three different conditions (vacuum, water and dichloromethane). The

energies calculated were converted to kJ/mol and expressed relative to intermediate 1. The

raw data is displayed in Appendix 2.

Wine samples for analysis

One white wine (Stanley Classic Dry White) and one red wine (Yalumba 1997 Shiraz)

were extracted with and without spikes of p-coumaroyl and feruloyl glucose (5 mg/L).

Liquid-liquid extraction was performed with 50 mL of wine, extracting alternatively with

diethyl ether (3 x 50 mL) and ethyl acetate (3 x 50 mL) before being concentrated under

reduced pressure at 30 oC and taken up in methanol (2 mL). Solid-phase extraction was

performed using XAD-8 resin, where 50 mL of wine was loaded, washed with water,

eluted with 25%, 50%, 75% methanol in water and then 100% methanol, fractions were

individually concentrated and taken up in methanol (2 mL). Concentrated wine samples

were prepared from 50 mL of wine at 30 oC under reduced pressure until the volume had

reduced to 5 mL. Neat wine samples were passed through a 45 µm syringe filter and

analysed directly. Standards of p-coumaroyl glucose (9) and feruloyl glucose (10) were

prepared to determine retention times and response factors (in methanol 10 mg/L, 100

mg/L), extraction efficiencies were estimated using the prepared 3 point calibration curve,

along with the 2 point curves produced by analysing spiked and unspiked wine samples.


173

HPLC analysis of wine samples

Analyses were performed on an Agilent 1100 instrument (Agilent, Forest Hill, Vic,

Australia) equipped with a quaternary pump and diode array detector (DAD). The column

was a 250 x 4.6 mm, 3 µm, 100 Å Luna C18, operated at 25 ºC and protected by a

C18 guard column (4 x 2 mm) (Phenomenex, Lane Cove, NSW, Australia). The eluents

were formic acid/water (0.5:99.5 v/v, Eluent A), formic acid/acetonitrile/water

(0.5:25.0:74.5 v/v, Eluent B) and methanol (Eluent C) with a flow rate of 1 mL/min. A

gradient was applied as follows: 20% to 30% B linear from 0 to 20 minutes; 30% to 50% B

linear from 20 to 50 minutes; 50% B to 100% C linear from 50 to 60 minutes; 100% C to

20% B from 60 to 65 minutes. The column was equilibrated with 20% B for 10 minutes

prior to an injection. A 20 µL injection volume was used for each sample and DAD

signals were recorded at all available wavelengths for compound identification, and

quantified using 280 and 320 nm. Compounds in each sample were identified by

comparison of their retention times and UV/Vis spectra with those of authentic standards.

LCMS analysis of wine samples

HPLC-MS or MS/MS analysis was carried out using a 4000 Q TRAP hybrid tandem mass

spectrometer interfaced with a Turbo V ion source for elecrospray ionization (AB Sciex

AB Sciex, Foster City, CA), combined with an Agilent 1200 HPLC system equipped with

a binary pump, degasser, autosampler, column oven, and photodiode array (PDA) detector.

HPLC conditions:

A 10 µL aliquot of the samples was injected and chromatographed using the same column

and elution profile as described for HPLC, above. The column temperature was maintained

at 25˚C during the HPLC run. The eluent from the HPLC was split by use of a splitter (a

tee) and delivered at a follow rate of 0.45 mL/min to the mass spectrometer and at 0.55

mL/min to the PDA detector with monitoring wavelengths at 290, 320 and 370 nm with a

slit width of 4 and a bandwidth of 16 nm.

Electrospray and mass spectrometric conditions:

All mass spectrometric data were obtained in negative ion mode. Nitrogen gas was used

for the curtain, nebulizer, turbo and collision gases. The Turbo V ion source parameter

were set at -3500 V for the ion spray potential, -60 V for the declustering potential, -10 V


174

for the entrance potential, 50 psi for gas 1 (nebulizer) and gas 2 (turbo), 15 psi for the

curtain gas, and 500 °C for the turbo gas (gas 2) temperature.

For tandem mass spectrometry, the collision potential was set in an appropriate range from

-15 to -25 V and the collision gas pressure was set at high. Product ion spectra of m/z 325

for p-coumaroyl glucose and m/z 355 for feruloyl glucose were recorded in a mass range

from m/z 50 to 400 with a scan time of 1 s and a step mass of 0.1. For selected reaction

monitoring, the following mass transitions were monitored with a dwell time of 50 ms; m/z

325 119, 145, 163 and 187 for p-coumaroyl glucose, and m/z 355 119, 175, 193 and 217

for feruloyl glucose.


175


cis-Ethyl coumarate (cis-11)

Ethyl (2Z)-3-(4-hydroxyphenyl)prop-2-enoate

cis-Ethyl coumarate (cis-11) was achieved as a minor product from the aforementioned

Wittig reaction (Synthesis of 11, Chapter 6.2.1), or by irradiation under ultra-violet light

(365 nm), whereby isolation using column chromatography (10% EtOAc/X4) yielded a

white solid. m.p. 72.3-74.4 oC (lit. m.p.73-74 oC).185

Rf (50% EtOAc/X4): 0.51 1H NMR: (400 MHz, CDCl3) δ: 7.63 (app. d, 2H, J = 8.6 Hz, H3,5), 6.85 (d, 1H, J = 12.7

Hz, H7), 6.80 (app. d, 2H, J = 8.6 Hz, H2,6), 5.83 (d, 1H, J = 12.7 Hz, H8), 4.21 (q, 2H, J =

7.1 Hz, OCH2CH3), 1.29 (t, 3H, J = 7.1 Hz, OCH2CH3).


cis-Ethyl ferulate (cis-12)

Ethyl (2Z)-3-(4-hydroxy-3-methoxyphenyl)prop-2-enoate

cis-Ethyl ferulate (cis-12) was achieved as a minor product from the aforementioned Wittig

reaction (Synthesis of 12, Chapter 6.2.1), or by irradiation under ultra-violet light (365

nm), whereby isolation using column chromatography (10% EtOAc/X4) yielded a

colourless oil.

Rf (50% EtOAc/X4): 0.55 1H NMR: (400 MHz, CDCl3) δ: 7.77 (d, 1H, J = 1.9 Hz, H3), 7.11 (dd, 1H, J = 8.5 and 1.9

Hz, H5), 6.88 (d, 1H, J = 8.5 Hz, H6), 6.79 (d, 1H, J = 12.9 Hz, H7), 5.81 (d, 1H, J = 12.9


OCH2CH3).



176

cis-p-Coumaric acid (cis-3)

(2Z)-3-(4-Hydroxyphenyl)prop-2-enoic acid

cis-Ethyl coumarate (cis-11) (0.11 g, 0.59 mmol) was dissolved in 1:1 aqueous ethanol

(v/v, 10 mL) followed by the addition of potassium hydroxide (0.10 g, 1.78 mmol), then

the reaction mixture was stirred at room temperature for 3 days. The mixture was then

diluted with water (5 mL), unwanted organics extracted with diethyl ether (2 x 10 mL), the

aqueous layer acidified to pH 3 with 2 M hydrochloric acid solution and extracted with

ethyl acetate (2 x 10 mL). Concentration at reduced pressure gave 97 mg (99%) of a 80:20

mixture of trans- and cis-p-coumaric acid as an off-white solid. 1H NMR: (400 MHz, DMSO-d6) δ: 9.55 (br. s, 0.80H, trans-COOH), 9.84 (br. s, 0.20H,

cis-COOH), 7.63 (app. d, 0.40H, J = 8.6 Hz, cis-H3,5), 7.51 (app. d, 1.60H, J = 8.6 Hz,

trans-H3,5), 7.49 (d, 0.80H, J = 15.9 Hz, trans-H7), 6.79 (app. d, 1.60H, J = 8.6 Hz, trans-

H2,6), 6.78-6.73 (m, 0.60H, cis-H2,6,7), 6.28 (d, 0.80H, J = 15.9 Hz, trans-H8), 5.72 (d,

0.20H, J = 12.8 Hz, cis-H8).

Spectral properties were as previously reported.214, 235

cis-p-Coumaric acid by UV irradiation

trans-p-Coumaric acid (3) was dissolved in acetone and exposed to UV light at 365 nm for

4 days. The resulting mixture of isomers was concentrated and analysed by NMR to show

a trans:cis-ratio of approximately 61:39.

cis-Ferulic acid (cis-4)

(2Z)-3-(4-Hydroxy-3-methoxyphenyl)prop-2-enoic acid

Using the same procedure as described above (for cis-3), reaction of cis-ethyl ferulate (cis-

12) (0.31 g, 1.39 mmol) gave 268 mg (99%) of a 65:35 mixture of trans- and cis-ferulic

acid as a yellow solid.


177

1H NMR: (400 MHz, DMSO-d6) δ: 9.55 (br. s, 0.65H, trans-COOH), 9.46 (br. s, 0.35H,

cis-COOH), 7.65 (d, 0.35H, J = 2.0 Hz, cis-H3), 7.48 (d, 0.65H, J = 15.9 Hz, trans-H7),

7.27 (d, 0.65H, J = 1.9 Hz, trans-H3), 7.15 (dd, 0.35H, J = 8.3 and 2.0 Hz, cis-H5), 7.08

(dd, 0.65H, J = 8.2 and 1.9 Hz, trans-H5), 6.79 (d, 0.65H, J = 8.2 Hz, trans-H6), 6.75 (d,

0.35H, J = 12.9 Hz, cis-H7), 6.75 (d, 0.35H, J = 8.2 Hz, cis-H6), 6.36 (d, 0.65H, J = 15.9

Hz, trans-H8), 5.73 (d, 0.35H, J = 12.9 Hz, cis-H8), 3.81 (s, 1.95H, trans-OCH3), 3.75 (s,

1.05H, cis-OCH3).

Spectral properties were as previously reported.184, 235

cis-Ferulic acid by UV irradiation

trans-Ferulic acid (4) was dissolved in acetone and exposed to UV light at 365 nm for 4

days. The resulting mixture of isomers was concentrated and analysed by NMR to show a

trans:cis-ratio of approximately 50:50.

Stability of cis/trans-p-coumaric acid mixture

Portions of cis/trans-p-coumaric acid (3) (61:39 cis:trans ratio, 5 mg) either as a solid, or

dissolved in acetone (5 mL) were stored under four different conditions. Both a solid and

liquid sample was stored at ambient temperature in the dark, at ambient temperature under

ambient light conditions, at 4 oC or at -20 oC. Samples stored at 4 or -20 oC were not

exposed to light. After two weeks of storage, liquid samples were concentrated in vacuo

and all samples were analysed by 1H NMR and the isomeric ratio determined by

integration of the signals. The ratios determined are shown in Table 4.2.

Effect of group 10 metals on ethyl coumarate isomerisation

Experiment 1:

Into four separate flasks were placed:

1) trans-Ethyl coumarate (11) (60.3 mg, 0.31 mmol) and dichloromethane (3 mL).

2) trans-Ethyl coumarate (11) (56.0 mg, 0.29 mmol), dichloromethane (3 mL) and 10%

palladium on activated carbon (15.7 mg, 0.015 mmol of Pd, 5% by moles).


palladium on activated carbon (311.9 mg, 0.29 mmol of Pd, 132% by moles).


178


platinum on activated carbon (4.9 mg, 0.002 mmol of Pt, 1.6% by moles).

The four mixtures were irradiated under 365 nm light for 18 hours before being filtered

through celite and concentrated in vacuo. Analysis of the crude mixtures by 1H NMR

allowed for determination of the isomeric ratios by integration, which existed as follows:

1) 68:32, trans:cis-ethyl coumarate.


3) 100 % trans-ethyl coumarate.


Experiment 2:

Into two separate flasks were placed:

1) cis-Ethyl coumarate (cis-11) (12.8 mg, 0.07 mmol) and dichloromethane (4 mL).

2) cis-Ethyl coumarate (cis-11) (10.3 mg, 0.05 mmol), dichloromethane (4 mL) and

palladium acetate (7.8 mg, 0.03 mmol of Pd).

Both mixtures were stirred at ambient temperature under ambient light conditions for 48

hours before being concentrated in vacuo and subjected to analysis by 1H NMR. Analysis

of mixture 1 indicated minor conversion from cis- to trans-ethyl coumarate while mixture

2 contained only trans-ethyl coumarate.

Theoretical studies into the photoisomerisation of hydroxycinnamates

Energy profiles were initially calculated in a vacuum at the S0 and T1 state using either a

dynamic dihedral constraint between 180 and 0o with the geometry optimised every 10o, or

by manually constraining the dihedral in 10o increments with the geometry optimised for

the resulting 19 structures. Energies were calculated in Hartrees and converted to kJ/mol,

with the energy displayed relative to the S0 180o dihedral geometry. Solvated energy

profiles were produced in an analogous fashion optimising the geometries in water. The

energy profile geometries were initially calculated from the MMFF geometry, then from

the MMFF conformer. The raw data is displayed in Appendix 3.


179

Vertical excitation energies were calculated by optimising the unconstrained trans-

configuration, followed by single-point energy calculation at the T1 level. HOMO-LUMO

gaps were calculated by comparing the orbitals in the S0 state. The raw data is displayed in

Appendix 4.


180


6.5.1 General Procedures for Chapter 5

Media for yeast growth

YPD – Yeast extract (1% w/v), peptone (2% w/v), D-glucose (2% w/v) in Milli-Q water

was autoclaved and stored at room temperature.

YNB – US Biologicals Yeast Nitrogen Base in Milli-Q water (6.76 g/L) was supplemented

with glucose (20 g/L), the pH adjusted to 3.5 with 10% HCl solution, sterile filtered

(StericapTM PLUS 0.22 µm) and stored at room temperature.

Starter cultures

Yeasts were obtained from the AWRI culture collection on MYPG plates, transferred to

YPD broth and stored at 28 oC with constant shaking (150 r.p.m.) until the cell count

surpassed 1 x 108 cells/mL as determined by haemocytometry.

Fermentation experiments

Fermentations were performed in triplicate in 250 mL fermentation flasks equipped with a

gas-lock. YNB media (200 mL) was spiked with the specified compound and inoculated

with 1 x 106 cells/mL of yeast from a starter culture. Where specified, control flasks

containing media and the spike were established in the same manner, without inoculation,

and all experiments were conducted at 28 oC. The experiments were concluded after the

yeast stationary phase, as determined by the optical density.

Ferment sampling

An aliquot of 5 mL was taken from each ferment and from this the yeast growth was

determined. The remaining sample was centrifuged (4000 r.p.m. for 5 minutes at 25 oC),

the supernatant decanted from the yeast pellet, and stored at -20 oC until required for

analysis. Sampling was initially performed every two days, with additional samples taken

near the completion of fermentation (as the results specify).


181

Yeast growth

Was determined from fermentation samples by optical density as measured with a

Beckman Coulter DU 530 Life Sciences UV/Vis Spectrophotometer. Aliquots (100 µL)

were diluted with water (10 x) and the yeast growth determined by measuring the

absorbance at 600 nm using water as a blank.

Model wine

A saturated solution of potassium hydrogen tartrate in demineralised water was acidified to

pH 3.5 with tartaric acid, stored at 4 oC overnight, then decanted from the precipitate.

4-Ethylphenol / 4-ethylguaiacol analysis

Analysis of 4-ethylphenol and 4-ethylguaiacol was performed as described by Pollnitz et

al.63 The concentration of 4-ethylphenol and 4-ethylguaiacol in the fermentation samples

were measured in µg/L and the percentage conversion calculated using a full molar

conversion from substrate to ethylphenol. The average percentage conversion across the

three replicates is displayed in Appendix 5, which was used to produce figures as shown in

Chapter 5.

Results for analysis of uninoculated controls are not shown, but all were found to contain

no traces of ethylphenols.


182

cis / trans-Ferulic acid analysis

An unpublished HPLC method developed by the AWRI was utilised, with the solvent

parameters altered to achieve maximum resolution of isomers.

Analyses were performed on an Agilent 1100 instrument (Agilent, Forest Hill, VIC,

Australia) equipped with a quaternary pump and diode array detector (DAD). The column

consisted of a mixed-mode RP/weak anion exchange (WAX) stationary phase based on N-

(10-undecenoyl)-3-aminoquinuclidine, bonded to thiol-functionalised silica (150 x 2 mm, 5

µm, 100 Å VDS Optilab, Berlin, Germany), operated at 25 ºC and protected by an NH2

guard column (4 x 2 mm) (Phenomenex, Lane Cove, NSW, Australia). The eluents were

formic acid/water (0.1:99.9 v/v, Eluent A), and formic acid/acetonitrile (0.1:99.9 v/v,

Eluent B) with a flow rate of 0.400 mL/min. Isocratic elution was performed using 70% A

and 30% B with run time of 7 minutes. A 50% aqueous acetonitrile solution wash was used

as the column wash eluent. A 20 µL injection volume was used for each sample and DAD

signals were recorded at 280, 320, 353, 370 and 520 nm. Compounds in each sample were

identified by comparison of their retention times and UV/Vis spectra with those of

standards, with quantifications calculated using absorbance at 280 and 320 nm only.

trans-Ferulic acid was dissolved in ethanol and serial dilutions were prepared (75 mg/L,

62.5 mg/L, 50 mg/L, 25 mg/L, 10 mg/L, 5 mg/L and 1 mg/L). The produced trans-ferulic

acid calibration curve was used to quantify trans-ferulic acid in prepared isomeric mixtures

of cis/trans-ferulic acid (50:50 cis/trans-ferulic acid, concentrations used as above,

equating to 37.5 mg/L, 31.25 mg/L, 25 mg/L, 12.5 mg/L, 5 mg/L, 2.5 mg/L and 0.5 mg/L).

The calibration was checked by running cis/trans-mixtures of known concentration and

ratio (25:75 cis/trans-ferulic acid, 10 mg/L, 20 mg/L; 12.5:87.5 cis/trans-ferulic acid, 10

mg/L, 20 mg/L).

cis / trans-Ethyl hydroxycinnamate analysis

Analysis of trans-ethyl ferulate and trans-ethyl coumarate was performed using the

method described by Sleep.141, 249 Quantification of cis-ethyl coumarate and cis-ethyl

ferulate was achieved by determining the differences in extraction efficiency and mass

spectral responses from the trans-isomers through extracting and analysing mixtures of

known concentration and ratio (1:1, 5 mg/L and 10 mg/L).


183

Rapid GCMS quantifications were performed using the same method, quantified using a 4

point calibration curve of the trans-internal standard and cis/trans-ethyl esters (0 mg/L, 5

mg/L, 10 mg/L and 15 mg/L).

HPLC quantifications of the trans-isomers were performed using the same

chromatographic conditions described in Chapter 6.3 for analysis of wine samples,

employing external standards of trans-ethyl ferulate and trans-ethyl coumarate prepared at

5, 10, 15 and 20 mg/L. The produced calibration curve was used to quantify trans-ethyl

ferulate and trans-ethyl coumarate in the fermentation samples, which were syringe filtered

(0.45 µm) and analysed without dilution.


184

6.5.2 Fermentation of trans-Hydroxycinnamate Esters

Fermentation of ethyl hydroxycinnamates

Stock solutions of ethyl coumarate (1 mg/mL) and ethyl ferulate (1 mg/mL) were made up

in ethanol, and the fermentations spiked at 10 mg/L. These were inoculated with AWRI

1499, and along with uninoculated controls, were stored and mixed manually at each

sampling point. Substrate selectivity was tested in a similar manner, inoculating with either

AWRI 1499, AWRI 1608 or AWRI 1613 and stored with occasional shaking.

Fermentation of hydroxycinnamoyl tartrate esters

Stock solutions of p-coumaroyl tartrate (1 mg/mL) and feruloyl tartrate (1 mg/mL) were

made up in ethanol, and the fermentations spiked at 10 mg/L. These were inoculated with

AWRI 1499, and along with uninoculated controls, were stored with constant shaking.

Hydroxycinnamoyl tartrate ester strain dependence was performed in a similar manner,

with fermentations inoculated with either AWRI 1499, AWRI 1608 or AWRI 1613 and

stored with constant shaking.

Fermentation of hydroxycinnamoyl glucose esters

Stock solutions of p-coumaroyl glucose (1.03 mg/mL) and feruloyl glucose (0.95 mg/mL)

were made up in ethanol under red light and stored in the dark. The fermentations were

spiked at 10 mg/L and inoculated with AWRI 1499. These, along with uninoculated

controls, were covered in foil to maintain isomeric purity, stored and mixed manually at

each sampling point with.

6.5.3 Stereoselectivity of D. bruxellensis Enzyme Activities

Decarboxylase stereoselectivity

Stock solutions of trans-ferulic acid (5 mg/mL), cis/trans-ferulic acid (50:50 ratio, 5

mg/mL), trans-p-coumaric acid (5 mg/mL) and cis/trans-p-coumaric acid (39:61 ratio, 5

mg/mL) were made up in ethanol. Ferulate and p-coumarate fermentations were performed

separately, being spiked with either trans-acid (50 mg/L) or cis/trans-acid (50 mg/L),

inoculated with AWRI 1499, stored and mixed manually at each sampling point.


185

NMR spectra of uninoculated flasks containing cis/trans-acid (50 mg/L) were used to

monitor the isomeric ratio over the fermentation period. NMR samples were produced by

extraction of the entire 200 mL ferment with ethyl acetate (3 x 50 mL), then removal of the

solvent by rotary evaporation. The ratio was determined by the integration of either the H8

or the H3 signals.

Ethyl esterase stereoselectivity

Stock solutions cis-ethyl coumarate (1 mg/mL) and cis-ethyl ferulate (1 mg/mL) were

made up in ethanol under red light and stored in the dark. The fermentations were spiked

with both esters at 10 mg/L, inoculated with AWRI 1499, wrapped in alfoil to maintain

isomeric purity and along with uninocluated controls, were stored and mixed manually at

each sampling point. Samples for analysis were taken under red light.

Appendix 1: Data for Migration Thermodynamics

186

Appendix 1: Data for Migration Thermodynamics.

Table A1.1: Energies of p-coumaroyl and feruloyl glucose esters in water, using DFT

B3LYP 6-31G*.

Hartrees (a.u.) kJ/mol Relative to 1-O -ββββ Hartrees (a.u.) kJ/mol Relative to 1-O -ββββ1-O -ββββ -1184.20519 -3109130.3 0.0 -1298.72015 -3409789.3 0.01-O -αααα -1184.18255 -3109070.8 59.4 -1298.70484 -3409749.1 40.22-O -ββββ -1184.19626 -3109106.8 23.4 -1298.71136 -3409766.2 23.12-O -αααα -1184.20248 -3109123.2 7.1 -1298.71770 -3409782.8 6.43-O -ββββ -1184.19781 -3109110.9 19.4 -1298.71293 -3409770.3 19.03-O -αααα -1184.19715 -3109109.2 21.1 -1298.71205 -3409768.0 21.34-O -ββββ -1184.19503 -3109103.6 26.7 -1298.71037 -3409763.6 25.74-O -αααα -1184.19211 -3109095.9 34.3 -1298.71263 -3409769.5 19.76-O -ββββ -1184.19020 -3109090.9 39.4 -1298.70643 -3409753.2 36.06-O -αααα -1184.20156 -3109120.7 9.5 -1298.71677 -3409780.4 8.9

p -Coumaroyl glucose Feruloyl glucose

Table A1.2: Energies of p-coumaroyl and feruloyl glucose esters in dichloromethane, using

DFT B3LYP 6-31G*.

Hartrees (a.u.) kJ/mol Relative to 1-O -ββββ Hartrees (a.u.) kJ/mol Relative to 1-O -ββββ1-O -ββββ -1184.20123 -3109119.9 0.0 -1298.71642 -3409779.5 0.01-O -αααα -1184.18389 -3109074.4 45.5 -1298.70860 -3409758.9 20.52-O -ββββ -1184.19515 -3109103.9 16.0 -1298.71055 -3409764.1 15.42-O -αααα -1184.20189 -3109121.6 -1.7 -1298.71734 -3409781.9 -2.43-O -ββββ -1184.19669 -3109108.0 11.9 -1298.71201 -3409767.9 11.63-O -αααα -1184.19673 -3109108.1 11.8 -1298.71190 -3409767.6 11.94-O -ββββ -1184.19374 -3109100.2 19.7 -1298.70910 -3409760.2 19.24-O -αααα -1184.19117 -3109093.5 26.4 -1298.71229 -3409768.6 10.86-O -ββββ -1184.19349 -3109099.6 20.3 -1298.70875 -3409759.3 20.16-O -αααα -1184.20361 -3109126.1 -6.2 -1298.71927 -3409786.9 -7.5


Appendix 1: Data for Migration Thermodynamics

187

Table A1.3: Energies of p-coumaroyl and feruloyl glucose esters in ethanol, using DFT

B3LYP 6-31G*.

Hartrees (a.u.) kJ/mol Relative to 1-O -ββββ Hartrees (a.u.) kJ/mol Relative to 1-O -ββββ1-O -ββββ -1184.21194 -3109148.0 0.0 -1298.72783 -3409809.4 0.0

1-O -αααα -1184.19874 -3109113.3 34.7 -1298.72248 -3409795.4 14.02-O -ββββ -1184.20851 -3109139.0 9.0 -1298.72467 -3409801.1 8.32-O -αααα -1184.21514 -3109156.4 -8.4 -1298.73133 -3409818.6 -9.23-O -ββββ -1184.21175 -3109147.5 0.5 -1298.72781 -3409809.4 0.13-O -αααα -1184.21154 -3109146.9 1.1 -1298.72747 -3409808.5 0.94-O -ββββ -1184.20785 -3109137.3 10.7 -1298.72393 -3409799.2 10.24-O -αααα -1184.20617 -3109132.8 15.1 -1298.72677 -3409806.6 2.86-O -ββββ -1184.20733 -3109135.9 12.1 -1298.72339 -3409797.8 11.76-O -αααα -1184.21677 -3109160.7 -12.7 -1298.73306 -3409823.2 -13.7


Table A1.4: Energies of p-coumaroyl and feruloyl glucose esters in toluene, using DFT

B3LYP 6-31G*.

Hartrees (a.u.) kJ/mol Relative to 1-O -ββββ Hartrees (a.u.) kJ/mol Relative to 1-O -ββββ1-O -ββββ -1184.20094 -3109119.1 0.0 -1298.71557 -3409777.2 0.0

1-O -αααα -1184.18122 -3109067.3 51.8 -1298.70599 -3409752.1 25.22-O -ββββ -1184.19381 -3109100.4 18.7 -1298.70870 -3409759.2 18.02-O -αααα -1184.20068 -3109118.4 0.7 -1298.71557 -3409777.2 0.03-O -ββββ -1184.19488 -3109103.2 15.9 -1298.70959 -3409761.5 15.73-O -αααα -1184.19458 -3109102.4 16.7 -1298.70918 -3409760.5 16.84-O -ββββ -1184.19246 -3109096.9 22.3 -1298.70733 -3409755.6 21.64-O -αααα -1184.18910 -3109088.0 31.1 -1298.71038 -3409763.6 13.66-O -ββββ -1184.19104 -3109093.1 26.0 -1298.70580 -3409751.6 25.76-O -αααα -1184.20182 -3109121.4 -2.3 -1298.71679 -3409780.4 -3.2


Table A1.5: Energies of p-coumaroyl glucose esters in different solvents, relative to the 1-

O-α-ester.

Water Dichloromethane Ethanol Toluene St. Dev.1-O -ββββ -59.44 -45.53 -34.66 -51.77 10.482-O -ββββ -36.00 -29.56 -25.65 -33.06 4.472-O -αααα -52.33 -47.26 -43.06 -51.09 4.183-O -ββββ -40.07 -33.61 -34.16 -35.86 2.923-O -αααα -38.33 -33.71 -33.61 -35.08 2.204-O -ββββ -32.77 -25.86 -23.92 -29.51 3.934-O -αααα -25.10 -19.11 -19.51 -20.69 2.756-O -ββββ -20.09 -25.20 -22.55 -25.78 2.626-O -αααα -49.91 -51.77 -47.34 -54.09 2.86

Energy (kJ/mol) relatve to the 1-O -αααα -ester

Appendix 2: Data for Migration Kinetics

188

Appendix 2: Data for Migration Kinetics.

Appendix 2: Data for Migration Kinetics

189

Appendix 3: Data for Energy Profiles

190

Appendix 3: Data for Energy Profiles.

Table A3.1: Data for p-coumaric acid energy profile in a vacuum.

Dihedral Angle Hartrees (a.u.) kJ/mol Relative to S0 180oHartrees (a.u.) kJ/mol Relative to S0 180o

180 -573.429091 -1505537.9 0.0 -573.341720 -1505308.5 229.4

170 -573.428809 -1505537.1 0.7 -573.342894 -1505311.6 226.3160 -573.427103 -1505532.6 5.2 -573.344340 -1505315.3 222.5150 -573.424026 -1505524.6 13.3 -573.345986 -1505319.7 218.2140 -573.419539 -1505512.8 25.1 -573.347703 -1505324.2 213.7130 -573.413554 -1505497.1 40.8 -573.349252 -1505328.2 209.6120 -573.406317 -1505478.1 59.8 -573.350502 -1505331.5 206.3110 -573.397875 -1505455.9 82.0 -573.351420 -1505333.9 203.9100 -573.388198 -1505430.5 107.4 -573.351946 -1505335.3 202.590 -573.372181 -1505388.4 149.4 -573.352043 -1505335.6 202.380 -573.383155 -1505417.3 120.6 -573.352070 -1505335.6 202.270 -573.393105 -1505443.4 94.5 -573.351687 -1505334.6 203.260 -573.401892 -1505466.4 71.4 -573.350894 -1505332.6 205.350 -573.409397 -1505486.2 51.7 -573.349653 -1505329.3 208.640 -573.415538 -1505502.3 35.6 -573.347991 -1505324.9 212.930 -573.420179 -1505514.5 23.4 -573.345932 -1505319.5 218.320 -573.423281 -1505522.6 15.3 -573.343416 -1505312.9 224.910 -573.424826 -1505526.7 11.2 -573.340433 -1505305.1 232.80 -573.424840 -1505526.7 11.2 -573.337018 -1505296.1 241.7

S0 T1

Table A3.2: Data for ethyl coumarate energy profile in a vacuum.

Dihedral Angle kJ/mol Relative to S0 180okJ/mol Relative to S0 180o

kJ/mol Relative to S0 180okJ/mol Relative to S0 180o

180 -1712001.0 0.0 -1711999.3 0.0 -1711770.6 230.4170 -1711998.9 2.1 -1711997.5 1.8 -1711770.5 230.4160 -1711993.2 7.8 -1711992.0 7.3 -1711771.0 229.9150 -1711983.7 17.3 -1711982.9 16.3 -1711772.1 228.9140 -1711970.2 30.8 -1711970.2 29.1 -1711774.1 226.9130 -1711952.9 48.0 -1711953.8 45.5 -1711776.5 224.5120 -1711932.4 68.6 -1711933.9 65.4 -1711778.6 222.4110 -1711908.4 92.6 -1711910.7 88.6 -1711780.5 220.5100 -1711881.6 119.4 -1711884.3 115.0 -1711781.6 219.490 -1711852.3 148.6 -1711855.1 144.2 -1711835.2 165.8 -1711782.1 218.980 -1711820.9 180.1 -1711822.6 176.7 -1711862.8 138.2 -1711781.5 219.570 -1711787.8 213.2 -1711787.6 211.6 -1711887.6 113.3 -1711780.1 220.960 -1711750.4 248.9 -1711908.2 92.8 -1711777.9 223.050 -1711924.3 76.7 -1711774.6 226.440 -1711958.0 41.3 -1711936.3 64.7 -1711770.6 230.430 -1711959.8 41.2 -1711965.2 34.1 -1711944.8 56.2 -1711765.6 235.320 -1711968.1 32.9 -1711969.7 29.6 -1711950.5 50.5 -1711759.4 241.610 -1711972.6 28.4 -1711971.7 27.6 -1711952.4 48.6 -1711752.4 248.60 -1711973.1 27.9 -1711971.4 27.9 -1711953.9 47.1 -1711744.5 256.5

S0 Dynamic Forwards T1 S0 Manual S0 Dynamic Backwards

Appendix 3: Data for Energy Profiles

191

Table A3.3: Data for p-coumaroyl glucose energy profile in a vacuum.

Dihedral Angle kJ/mol Relative to S0 180o kJ/mol Relative to S0 180o kJ/mol Relative to S0 180o

180 -3109035.5 0.0 -3108809.4 226.2

170 -3109033.9 1.7 -3108809.4 226.1160 -3109028.3 7.2 -3108809.5 226.1150 -3109019.1 16.4 -3108809.7 225.8140 -3109006.3 29.3 -3108810.1 225.4130 -3108989.9 45.6 -3108810.3 225.2120 -3108969.9 65.6 -3108813.0 222.6110 -3108946.5 89.0 -3108814.2 221.3100 -3108919.9 115.7 -3108815.1 220.490 -3108890.4 145.2 -3108909.1 126.5 -3108815.1 220.480 -3108934.5 101.0 -3108814.3 221.270 -3108955.9 79.7 -3108812.9 222.660 -3108972.9 62.7 -3108810.9 224.650 -3108985.7 49.8 -3108807.8 227.840 -3108995.9 39.7 -3108803.3 232.330 -3109003.1 32.4 -3108797.5 238.020 -3109004.3 31.2 -3108790.7 244.810 -3109006.6 28.9 -3108783.4 252.10 -3109006.6 28.9 -3108790.3 245.3

S0 Dynamic Forwards T1 S0 Dynamic Backwards

Table A3.4: Data for T1 energy profiles in water.

Dihedral Angle kJ/mol Relative to S0 180okJ/mol Relative to S0 180o

S0 180 -1505539.2 -3109133.9

180 -1505309.7 229.4 -3108912.2 221.7170 -1505307.3 231.8 -3108911.8 222.1160 -1505316.6 222.5 -3108910.5 223.4150 -1505320.9 218.2 -3108909.2 224.7140 -1505325.4 213.8 -3108908.4 225.6130 -1505329.5 209.6 -3108907.5 226.4120 -1505332.9 206.3 -3108908.8 225.2110 -1505335.3 203.9 -3108909.4 224.5100 -1505336.6 202.5 -3108910.2 223.790 -1505336.8 202.3 -3108909.9 224.080 -1505336.3 202.9 -3108909.0 224.970 -1505334.5 204.660 -1505331.5 207.7 -3108905.1 228.850 -1505327.1 212.1 -3108902.3 231.740 -1505321.5 217.6 -3108897.7 236.230 -1505315.0 224.2 -3108892.4 241.520 -1505307.8 231.3 -3108886.8 247.110 -1505300.3 238.9 -3108882.1 251.80 -1505293.1 246.1 -3108883.2 250.7

T1 p -Coumaric acid T1 p -Coumaroyl glucose

Appendix 4: Data for Vertical Excitations and HOMO-LUMO Gaps

192

Appendix 4: Data for Vertical Excitations and HOMO-LUMO

Gaps.


193


194


195

Appendix 5: Data from Ethylphenol Analyses

196

Appendix 5: Data from Ethylphenol Analyses.

Table A5.1: Percentage conversion from trans-ethyl esters (11 and 12) to ethylphenols in

fermentations with AWRI 1499.

2 4 6 8 9

4-EP 0.1 ± 0.1 8.8 ± 0.4 28.8 ± 1.2 45.8 ± 1.5 51.4 ± 2.44-EG 0.0 ± 0.0 0.5 ± 0.0 1.9 ± 0.1 3.3 ± 0.1 4.0 ± 0.2


Table A5.2: Percentage of ethyl esters (11 and 12) remaining in fermentations.

2 4 6 8 9

Ethyl coumarate 96.0 ± 5.4 85.7 ± 3.7 65.3 ± 6.9 44.3 ± 3.3 33.3 ± 1.3Ethyl ferulate 93.7 ± 5.9 95.0 ± 4.6 89.7 ± 10.8 90.7 ± 9.0 72.3 ± 3.3


Table A5.3: Percentage conversion from trans-ethyl esters (11 and 12) to ethylphenols in

fermentations with different strains of D. bruxellensis.

AWRI 1499 AWRI 1608 AWRI 1613

4-EP 67.0 ± 0.1 55.6 ± 0.8 N.D

4-EG 7.7 ± 0.2 3.0 ± 0.1 N.D

Table A5.4: Percentage conversion from hydroxycinnamoyl glucose esters (9 and 10) to

ethylphenols in fermentations with AWRI 1499.

2 4 6 84-EP 6.4 ± 2.7 12.6 ± 1.0 15.8 ± 1.1 20.6 ± 1.74-EG 10.3 ± 3.0 14.6 ± 0.6 18.1 ± 1.4 24.2 ± 2.9

10 12 14 164-EP 33.3 ± 2.9 34.4 ± 1.8 34.3 ± 0.6 36.0 ± 0.74-EG 33.5 ± 2.4 38.6 ± 0.8 33.7 ± 1.7 36.1 ± 1.4


Appendix 5: Data from Ethylphenol Analyses

197

Table A5.5: Percentage conversion from trans- and cis/trans-ferulic acid (4) to 4-

ethylguaiacol in D. bruxellensis fermentations.

2 3 4 5 6 7 9trans-ferments 3.0 ± 1.2 17.2 ± 1.9 46.8 ± 4.5 67.1 ± 4.0 78.2 ± 5.2 75.3 ± 2.5 69.8 ± 4.5

cis/trans-ferments 1.7 ± 0.5 8.2 ± 0.6 22.6 ± 1.6 30.8 ± 1.7 37.3 ± 2.4 38.4 ± 0.7 39.6 ± 0.3


Table A5.6: Percentage conversion from trans- and cis/trans-p-coumaric acid (3) to 4-

ethylphenol in D. bruxellensis fermentations.

2 4 6 8 10 12trans-ferments 7.9 ± 0.6 44.8 ± 1.1 56.4 ± 1.6 56.7 ±2.6 66.8 ± 1.6 61.7 ± 3.6

cis/trans-ferments 5.2 ± 0.4 29.9 ± 1.8 37.3 ± 1.2 37.3 ± 2.7 42.4 ± 0.5 40.6 ± 2.4


Table A5.7: Percentage conversion from cis-ethyl ferulate (cis-12) and cis-ethyl coumarate

(cis-11) to ethylphenols in D. bruxellensis fermentations.

2 4 6 8 10

4-EP 0.00 ± 0.00 0.07 ± 0.12 0.51 ± 0.04 0.74 ± 0.04 0.86 ± 0.024-EG 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.13 ± 0.11 0.25 ± 0.03


References

198

References

1. Hornsey, I., The Chemistry and Biology of Winemaking. RSC Publishing:

Cambridge, UK, 2007; p 457.

2. Gent, C., Mixed Dozen: The Story of Australian Winemaking Since 1788. Duffy &

Snellgrove: Sydney, 2003; p 344.

3. Beeston, J., A Concise History of Australian Wine. Allen & Unwin: New South

Wales, 1994; p 338.

4. Farkas, J., Technology and Biochemistry of Wine. Gordon and Breach Scientific

Publishers: Montreux, Switzerland, 1988; Vol. 1, p 387.

5. Wine Australia. Australian Wine Sector at a Glance, 2010.

http://www.wineaustralia.com (21/01/2012),

6. Wine Australia. Australian Wine Sales at a Glance - 2010.


7. Wine Australia. Australian Wine Sector, State by State - 2010.


8. Moreno-Arribas, M. V.; Polo, M. C., Winemaking biochemistry and microbiology:

Current knowlegde and future trends. Critical Reviews in Food Science and

Nutrition 2005, 45, (4), 265-286.

9. Pretorius, I. S., Tailoring wine yeast for the new millennium: novel approaches to

the ancient art of winemaking. Yeast 2000, 16, (8), 675-729.

10. Loureiro, V.; Malfeito-Ferreira, M., Spoilage yeasts in the wine industry. Int. J.

Food Microbiol. 2003, 86, (1-2), 23-50.

11. Fugelsang, K. C.; Osborn, M. M.; Muller, C. J., Brettanomyces and Dekkera

implications in wine-making. In Beer and Wine Production: Analysis,

Characterization, and Technological Advances, Gump, B. H.; Pruett, D. J., Eds.

American Chemical Society: 1993; Vol. 536, pp 110-129.

12. Heresztyn, T., Formation of substituted tetrahyropyridines by species of

Brettanomyces and Lactobacillus isolated from mousy wines. Am. J. Enol. Vitic.

1986, 37, (2), 127-132.

13. Boulton, R. B.; Singleton, V. L.; Bisson, L. F.; Kunkee, R. E., Principles and

Practices of Winemaking. Thompson Hall: New York, USA, 1995; p 604.

14. Fugelsang, K. C.; Edwards, C. G., Wine Microbiology. 2nd ed.; Springer: New

York, 2007.

References

199

15. Licker, J. L.; Acree, T. E.; Henick-Kling, T., What is "Brett" (Brettanomyces)

flavour?: A preliminary investigation. In Chemistry of Wine Flavour, Waterhouse,

A. L.; Ebeler, S. E., Eds. American Chemical Society: 1999; Vol. 714, pp 96-115.

16. Barnett, J. A., A history of research on yeasts 8: Taxonomy. Yeast 2004, 21, (14),

1141-1193.

17. Barnett, J.; Payne, R.; Yarrow, D., Yeasts: Characteristics and Identification. Third

ed.; Cambridge University Press: 2000; p 1139.

18. Smith, M. T.; Yamazaki, M.; Poot, G. A., Dekkera, Brettanomyces and Eeniella -

Electrophoretic comparison of enzymes and DNA-DNA homology. Yeast 1990, 6,

(4), 299-310.

19. Boekhout, T.; Kurtzman, C. P.; Odonnell, K.; Smith, M. T., Phylogeny of the yeast

genera Hanseniaspora (anamorph Kloeckera), Dekkera (anamorph Brettanomyces),

and Eeniella as inferred from partial 26S ribosomal DNA nucleotide-sequences. Int.

J. Syst. Bacteriol. 1994, 44, (4), 781-786.

20. Loureiro, V.; Malfeito-Ferreira, M., Dekkera/Bruxellensis spp. In Food Spoilage

Microorganisms, de W. Blackburn, C., Ed. Woodhead Publishing Ltd.: Cambridge,

UK, 2006; pp 354-400.

21. Curtin, C. D. Personal communication, 2012.

22. Curtin, C. D.; Bellon, J. R.; Henschke, P. A.; Godden, P. W.; de Barros Lopes, M.

A., Genetic diversity of Dekkera bruxellensis yeasts isolated from Australian

wineries. Federation of European Microbiological Societies Yeast Research 2007,

7, 471-481.

23. Conterno, L.; Joseph, C. M. L.; Arvik, T. J.; Henick-Kling, T.; Bisson, L. F.,

Genetic and physiological characterization of Brettanomyces bruxellensis strains

isolated from wines. Am. J. Enol. Vitic. 2006, 57, (2), 139-147.

24. Barbin, P.; Cheval, J. L.; Gilis, J. F.; Strehaiano, P.; Thillandier, P., Diversity in

spoilage yeast Dekkera/Brettanomyces bruxellensis isolated from French red wine.

Assessment during fermentation of synthetic wine medium. J. Inst. Brew. 2008,

114, (1), 69-75.

25. Egli, C. M.; Henick-Kling, T., Identification of Brettanomyces/Dekkera species

based on polymorphism in the rRNA internal transcribed spacer region. Am. J.

Enol. Vitic. 2001, 52, (3), 241-247.

26. Cocolin, L.; Rantsiou, K.; Iacumin, L.; Zironi, R.; Comi, G., Molecular detection

and identification of Brettanomyces/Dekkera bruxellensis and

References

200

Brettanomyces/Dekkera anomalus in spoiled wines. Appl. Environ. Microbiol.

2004, 70, (3), 1347-1355.

27. Stender, H.; Kurtzman, C.; Hyldig-Nielsen, J. J.; Sorensen, D.; Broomer, A.;

Oliveira, K.; Perry-O'Keefe, H.; Sage, A.; Young, B.; Coull, J., Identification of

Dekkera bruxellensis (Brettanomyces) from wine by fluorescence in situ

hybridization using peptide nucleic acid probes. Appl. Environ. Microbiol. 2001,

67, (2), 938-941.

28. Silva, L. R.; Andrade, P. B.; Valentao, P.; Seabra, R. M.; Trujillo, M. E.;

Velazquez, E., Analysis of non-coloured phenolics in red wine: Effect of Dekkera

bruxellensis yeast. Food Chem. 2005, 89, (2), 185-189.

29. Grbin, P. R.; Henschke, P. A., Mousy off-flavour production in grape juice and

wine by Dekkera and Brettanomyces yeasts. Australian Journal of Grape and Wine

Research 2000, 6, (3), 255-262.

30. Romano, A.; Perello, M. C.; de Revel, G.; Lonvaud-Funel, A., Growth and volatile

compound production by Brettanomyces/Dekkera bruxellensis in red wine. J. Appl.

Microbiol. 2008, 104, 1577-1585.

31. Vigentini, I.; Romano, A.; Compagno, C.; Merico, A.; Molinari, F.; Tirelli, A.;

Foschino, R.; Volonterio, G., Physiological and oenological traits of different

Dekkera/Brettanomyces bruxellensis strains under wine-model conditions. FEMS

Yeast Research 2008, 8, (7), 1087-1096.

32. Harris, V.; Ford, C. M.; Jiranek, V.; Grbin, P. R., Survey of enzyme activity

responsible for phenolic off-flavour production by Dekkera and Brettanomyces

yeast. Appl. Microbiol. Biotechnol. 2009, 81, (6), 1117-1127.

33. Agnolucci, M.; Vigentini, I.; Capurso, G.; Merico, A.; Tirelli, A.; Compagno, C.;

Foschino, R.; Nuti, M., Genetic diversity and physiological traits of Brettanomyces

bruxellensis strains isolated from Tuscan Sangiovese wines. Int. J. Food Microbiol.

2009, 130, (3), 238-244.

34. Oelofse, A.; Lonvaud-Funel, A.; du Toit, M., Molecular identification of

Brettanomyces bruxellensis strains isolated from red wines and volatile phenol

production. Food Microbiol. 2009, 26, (4), 377-385.

35. Conterno, L.; Lasik, G.; Tomasino, E.; Schneider, K.; Hesford, F.; Henick-Kling,

T., Influence of sugar and nitrogen sources on growth and phenolic off-flavor

production by Brettanomyces bruxellensis isolated from wine. Am. J. Enol. Vitic.

2007, 58, (3), 411A-411A.

References

201

36. Fugelsang, K. C.; Zoecklein, B. W., Population dynamics and effects of

Brettanomyces bruxellensis strains on Pinot noir (Vitis vinifera L.) wines. Am. J.

Enol. Vitic. 2003, 54, (4), 294-300.

37. Renouf, V.; Lonvaud-Funel, A., Development of an enrichment medium to detect

Dekkera/Brettanomyces bruxellensis, a spoilage yeast, on the surface of grape

berries. Microbiological Research 2007, 162, (2), 154-167.

38. Oelofse, A.; Pretorius, I. S.; du Toit, M., Significance of Brettanomyces and

Dekkera during winemaking: A synoptic review. South African Journal of Enology

and Viticulture 2008, 29, (2), 128-144.

39. Uscanga, M. G. A.; Delia, M. L.; Strehaiano, P., Nutritional requirements of

Brettanomyces bruxellensis: Growth and physiology in batch and chemostat

cultures. Canadian Journal of Microbiology 2000, 46, (11), 1046-1050.

40. Barata, A.; Caldeira, J.; Botelheiro, R.; Pagliara, D.; Malfeito-Ferreira, M.;

Loureiro, V., Survival patterns of Dekkera bruxellensis in wines and inhibitory

effect of sulphur dioxide. Int. J. Food Microbiol. 2008, 121, (2), 201-207.

41. Chatonnet, P.; Dubourdieu, D.; Boidron, J. N.; Pons, M., The origin of ethylphenols

in wines. J. Sci. Food Agric. 1992, 60, (2), 165-178.

42. Dias, L.; Pereira-da-Silva, S.; Tavares, M.; Malfeito-Ferreira, M.; Loureiro, V.,

Factors affecting the production of 4-ethylphenol by the yeast Dekkera bruxellensis

in enological conditions. Food Microbiol. 2003, 20, (4), 377-384.

43. Coulter, A.; Robinson, E.; Cowey, G.; Francis, I. L.; Lattey, K.; Capone, D.;

Gishen, M.; Godden, P. In Dekkera/Brettanomyces yeast - an overview of recent

AWRI investigations and some recommendations for its control, ASVO

Proceedings. Grapegrowing at the edge, managing the wine business, impacts on

wine flavour, Barossa Convention Centre, Tanunda, SA, 10-11 July 2003, 2003;

Bell, S. M.; de Garis, K. A.; Dundon, C. G.; Hamilton, R. P.; Partridge, S. J.; Wall,

G. S., Eds. Barossa Convention Centre, Tanunda, SA, 2003; pp 41-50.

44. Ciani, M.; Ferraro, L., Role of oxygen on acetic acid production by

Brettanomyces/Dekkera in winemaking. J. Sci. Food Agric. 1997, 75, (4), 489-495.

45. Aguilar Uscanga, M. G.; Delia, M. L.; Strehaiano, P., Brettanomyces bruxellensis:

Effect of oxygen on growth and acetic acid production. Applied Microbiology &

Biotechnology 2003, 61, (2), 157-162.

References

202

46. Ciani, M.; Maccarelli, F.; Fatichenti, F., Growth and fermentation behaviour of

Brettanomyces/Dekkera yeasts under different conditions of aerobiosis. World

Journal of Microbiology & Biotechnology 2003, 19, (4), 419-422.

47. du Toit, M.; Pretorius, I. S.; Lonvaud-Funel, A., The effects of sulphur dioxide and

oxygen on the viability and culturability of a strain of Acetobacter pasteurianus and

a strain of Brettanomyces bruxellensis isolated from wine. J. Appl. Microbiol. 2005,

98, 862-871.

48. Millet, V.; Lonvaud-Funel, A., The viable but non-culturable state of wine micro-

organisms during storage. Lett. Appl. Microbiol. 2000, 30, (2), 136-141.

49. Umiker, N. L., Impact of SO2 on culturability and viability of Brettanomyces in

wine. Am. J. Enol. Vitic. 2007, 58, (3), 417A-417A.

50. Agnolucci, M.; Rea, F.; Sbrana, C.; Cristani, C.; Fracassetti, D.; Tirelli, A.; Nuti,

M., Sulphur dioxide affects culturability and volatile phenol production by

Brettanomyces/Dekkera bruxellensis. Int. J. Food Microbiol. 2010, 143, (1-2), 76-

80.

51. Renouf, V.; Perello, M. C.; de Revel, G.; Lonvaud-Funell, A., Survival of wine

microorganisms in the bottle during storage. Am. J. Enol. Vitic. 2007, 58, (3), 379-

386.

52. Renouf, V.; Strehaiano, P.; Lonvaud-Funel, A., Effectiveness of

dimethlydicarbonate to prevent Brettanomyces bruxellensis growth in wine. Food

Control 2008, 19, (2), 208-216.

53. Delfini, C.; Gaia, P.; Schellino, R.; Strano, M.; Pagliara, A.; Ambro, S.,

Fermentability of grape must after inhibition with dimethyl dicarbonate (DMDC).

J. Agric. Food Chem. 2002, 50, (20), 5605-5611.

54. Benito, S.; Palomero, F.; Morata, A.; Calderon, F.; Suarez-Lepe, J. A., Factors

affecting the hydroxycinnamate decarboxylase/vinylphenol reductase activity of

Dekkera/Brettanomyces: Application for Dekkera/Brettanomyces control in red

wine making. J. Food Sci. 2009, 74, (1), M15-M22.

55. Barata, A.; Pagliara, D.; Piccininno, T.; Tarantino, F.; Ciardulli, W.; Malfeito-

Ferreira, M.; Loureiro, V., The effect of sugar concentration and temperature on

growth and volatile phenol production by Dekkera bruxellensis in wine. FEMS

Yeast Research 2008, 8, (7), 1097-1102.

References

203

56. Couto, J. A.; Neves, F.; Campos, F.; Hogg, T., Thermal inactivation of the wine

spoilage yeasts Dekkera/Brettanomyces. Int. J. Food Microbiol. 2005, 104, (3),

337-344.

57. Lustrato, G.; Vigentini, I.; Leonardis, A. D.; Alfano, G.; Tirelli, A.; Foschino, R.;

Ranalli, G., Inactivation of wine spoilage yeasts Dekkera bruxellensis using low

electric current treatment (LEC). J. Appl. Microbiol. 2010, 109, (2), 594-604.

58. Yap, A.; Jiranek, V.; Grbin, P.; Barnes, M.; Bates, D., Studies on the application of

high-powered ultrasonics for barrel and plank cleaning and disinfection. Australian

Wine Industry Journal 2007, 22, (3), 96-104.

59. Couto, J. A.; Barbosa, A.; Hogg, T., A simple cultural method for the presumptive

detection of the yeasts Brettanomyces/Dekkera in wines. Lett. Appl. Microbiol.

2005, 41, (6), 505-510.

60. Freer, S. N., Acetic acid production by Dekkera/Brettanomyces yeasts. World

Journal of Microbiology & Biotechnology 2002, 18, (3), 271-275.

61. Chatonnet, P.; Dubourdieu, D.; Boidron, J. N., The influence of

Brettanomyces/Dekkera sp. yeasts and lactic acid bacteria on the ethylphenol

content of red wines. Am. J. Enol. Vitic. 1995, 46, (4), 463-468.

62. Chatonnet, P.; Viala, C.; Dubourdieu, D., Influence of polyphenolic components of

red wines on the microbial synthesis of volatile phenols. Am. J. Enol. Vitic. 1997,

48, (4), 443-448.

63. Pollnitz, A. P.; Pardon, K. H.; Sefton, M. A., Quantitative analysis of 4-ethylphenol

and 4-ethylguaiacol in red wine. J. Chromatogr. A 2000, 874, (1), 101-109.

64. Gerbaux, V.; Vincent, B.; Bertrand, A., Influence of maceration temperature and

enzymes on the content of volatile phenols in Pinot noir wines. Am. J. Enol. Vitic.

2002, 53, (2), 131-137.

65. Edlin, D. A. N.; Narbad, A.; Dickinson, J. R.; Lloyd, D., The biotransformation of

simple phenolic compounds by Brettanomyces anomalus. FEMS Microbiol. Lett.

1995, 125, (2-3), 311-315.

66. Heresztyn, T., Metabolism of volatile phenolic compounds from hydroxycinnamic

acids by Brettanomyces yeast. Archives of Microbiology 1986, 146, 96-98.

67. Silva, P.; Cardoso, H.; Geros, H., Studies on the wine spoilage capacity of

Brettanomyces/Dekkera spp. Am. J. Enol. Vitic. 2004, 55, (1), 65-72.

References

204

68. Godoy, L.; Martinez, C.; Carrasco, N.; Ganga, M. A., Purification and

characterization of a p-coumarate decarboxylase and a vinylphenol reductase from

Brettanomyces bruxellensis. Int. J. Food Microbiol. 2008, 127, (1-2), 6-11.

69. Tchobanov, I.; Gal, L.; Guilloux-Benatier, M.; Remize, F.; Nardi, T.; Guzzo, J.;

Serpaggi, V.; Alexandre, H., Partial vinylphenol reductase purification and

characterization from Brettanomyces bruxellensis. FEMS Microbiol. Lett. 2008,

284, (2), 213-217.

70. Edlin, D. A. N.; Narbad, A.; Gasson, M. J.; Dickinson, J. R.; Lloyd, D., Purification

and characterization of hydroxycinnamate decarboxylase from Brettanomyces

anomalus. Enzyme and Microbial Technology 1998, 22, (4), 232-239.

71. Larcher, R.; Nicolini, G.; Bertoldi, D.; Nardin, T., Determination of 4-ethylcatechol

in wine by high-performance liquid chromatography-coulometric electrochemical

array detection. Anal. Chim. Acta 2008, 609, (2), 235-240.

72. Gomez-Alonso, S.; Garcia-Romero, E.; Hermosin-Gutierrez, I., HPLC analysis of

diverse grape and wine phenolics using direct injection and multidetection by DAD

and fluorescence. J. Food Compos. Anal. 2007, 20, (7), 618-626.

73. Betes-Saura, C.; Andres-Lacueva, C.; Lamuela-Raventos, R. M., Phenolics in white

free run juices and wines from Penedes by high-performance liquid

chromatography: Changes during vinification. J. Agric. Food Chem. 1996, 44, (10),

3040-3046.

74. Hesford, F.; Schneider, K.; Porret, N.; Gafner, J., Identification and analysis of 4-

ethylcatechol in wines tainted by Brettanomyces off-flavour. Abstract. Am. J. Enol.

Vitic. 2004, 55, (3), 304A.

75. Buron, N.; Coton, M.; Legendre, P.; Ledauphin, J.; Kientz-Bouchart, V.; Guichard,

H.; Barillier, D.; Coton, E., Implications of Lactobacillus collinoides and

Brettanomyces/Dekkera anomala in phenolic off-flavour defects of ciders. Int. J.

Food Microbiol. 2012, 153, (1-2), 159-165.

76. Chatonnet, P.; Dubourdieu, D.; Boidron, J. N.; Lavigne, V., Synthesis of volatile

phenols by Saccharomyces cerevisiae in wines. J. Sci. Food Agric. 1993, 62, (2),

191-202.

77. Vanbeneden, N.; Gils, F.; Delvaux, F.; Delvaux, F. R., Formation of 4-vinyl and 4-

ethyl derivatives from hydroxycinnamic acids: Occurrence of volatile phenolic

flavour compounds in beer and distribution of Pad1-activity among brewing yeasts.

Food Chem. 2008, 107, (1), 221-230.

References

205

78. Hernandez-Orte, P.; Cersosimo, M.; Loscos, N.; Cacho, J.; Garcia-Moruno, E.;

Ferreira, V., The development of varietal aroma from non-floral grapes by yeasts of

different genera. Food Chem. 2008, 107, (3), 1064-1077.

79. Couto, J. A.; Campos, F. M.; Figueiredo, A. R.; Hogg, T. A., Ability of lactic acid

bacteria to produce volatile phenols. Am. J. Enol. Vitic. 2006, 57, (2), 166-171.

80. Cavin, J. F.; Andioc, V.; Etievant, P. X.; Divies, C., Ability of wine lactic-acid

bacteria to metabolize phenol carboxylic-acids. Am. J. Enol. Vitic. 1993, 44, (1),

76-80.

81. Hashidoko, Y.; Tahara, S., Stereochemically specific proton transfer in

decarboxylation of 4-hydroxycinnamic acids by 4-hydroxycinnamate decarboxylase

from Klebsiella oxytoca. Archives of Biochemistry and Biophysics 1998, 359, (2),

225-230.

82. Hashidoko, Y.; Urashima, M.; Yoshida, T.; Mizutani, J., Decarboxylative

conversion of hydroxycinnamic acids by Klebsiella-oxytoca and Erwinia-

uredovora, epiphytic bacteria of Polymnia-sonchifolia leaf, possibly associated

with formation of microflora on the damaged leaves. Bioscience Biotechnology and

Biochemistry 1993, 57, (2), 215-219.

83. Parry, R. J., Stereochemistry of decarboxylation of trans-4-hydroxycinnamic acid

by Aerobacter. Proc. Natl. Acad. Sci. U. S. A. 1975, 75, (5), 1681-1683.

84. Harada, T.; Mino, Y., Some properties of p-coumarate decarboxylase from

Cladosporium phlei. Canadian Journal of Microbiology 1976, 22, (9), 1258-1262.

85. Bayne, H. G.; Finkle, B. J.; Lundin, R. E., Decarboxylative conversion of

hydroxycinnamic acids to hydroxystyrenes by Polyporus circinata. Journal of

General Microbiology 1976, 95, (1), 188-190.

86. Cavin, J. F.; Dartois, V.; Divies, C., Gene cloning, transcriptional analysis,

purification, and characterization of phenolic acid decarboxylase from Bacillus

subtilis. Appl. Environ. Microbiol. 1998, 64, (4), 1466-1471.

87. Dias, L.; Dias, S.; Sancho, T.; Stender, H.; Querol, A.; Malfeito-Ferreira, M.;

Loureiro, V., Identification of yeasts isolated from wine-related environments and

capable of producing 4-ethylphenol. Food Microbiol. 2003, 20, (5), 567-574.

88. Barata, A.; Nobre, A.; Correia, P.; Malfeito-Ferreria, M.; Loureiro, V., Growth and

4-ethylphenol production by the yeast Pichia guilliermondii in grape juices. Am. J.

Enol. Vitic. 2006, 57, (2), 133-138.

References

206

89. Romano, A.; Perello, M. C.; Lonvaud-Funel, A.; Sicard, G.; de Revel, G., Sensory

and analytical re-evaluation of "Brett character". Food Chem. 2009, 114, (1), 15-19.

90. Curtin, C. D.; Bellon, J. R.; Coulter, A. D.; Cowey, G. D.; Robinson, E. M. C.; de

Barros Lopes, M. A.; Godden, P. W.; Henschke, P. A.; Pretorius, I. S., The six

tribes of 'Brett' in Australia - Distribution of genetically divergent Dekkera

bruxellensis strains across Australian winemaking regions. Australian Wine

Industry Journal 2005, 20, (6), 28-36.

91. Fulcrand, H.; dosSantos, P. J. C.; SarniManchado, P.; Cheynier, V.; FavreBonvin,

J., Structure of new anthocyanin-derived wine pigments. J. Chem. Soc.-Perkin

Trans. 1 1996, (7), 735-739.

92. Ugarte, P.; Agosin, E.; Bordeu, E.; Villalobos, J. I., Reduction of 4-ethylphenol and

4-ethylguaiacol concentration in red wines using reverse osmosis and adsorption.

Am. J. Enol. Vitic. 2005, 56, (1), 30-36.

93. Palomero, F.; Ntanos, K.; Morata, A.; Benito, S.; Suarez-Lepe, J. A., Reduction of

wine 4-ethylphenol concentration using lyophilised yeast as a bioadsorbent:

influence on anthocyanin content and chromatic variables. Eur. Food Res. Technol.

2011, 232, (6), 971-977.

94. Fernández de Simón, B.; Hernández, T.; Estrella, I.; Gómez-Cordovés, C.,

Variation in phenol content in grapes during ripening: low-molecular-weight

phenols. Zeitschrift für Lebensmitteluntersuchung und -Forschung A 1992, 194, (4),

351-354.

95. Fernandez de Simon, B.; Hernandez, T.; Estrella, I., Phenolic composition of white

grapes (Var. Airen). Changes during ripening. Food Chem. 1993, 47, (1), 47-52.

96. Lee, C. Y.; Jaworski, A., Major phenolic-compounds in ripening white grapes. Am.

J. Enol. Vitic. 1989, 40, (1), 43-46.

97. Mayen, M.; Merida, J.; Medina, M., Flavonoid and nonflavonoid compounds

during fermentation and postfermentation standing of musts from Cabernet-

Sauvignon and Tempranillo grapes. Am. J. Enol. Vitic. 1995, 46, (2), 255-261.

98. Boido, E.; García-Marino, M.; Dellacassa, E.; Carrau, F.; Rivas-Gonzalo, J. C.;

Escribano-Bailón, M. T., Characterisation and evolution of grape polyphenol

profiles of Vitis vinifera L. cv. Tannat during ripening and vinification. Australian

Journal of Grape and Wine Research 2011, 17, (3), 383-393.

References

207

99. Somers, T. C.; Verette, E.; Pocock, K. F., Hydroxycinnamate esters of Vitis vinifera

- Changes during white vinification, and effects of exogenous enzymatic-

hydrolysis. J. Sci. Food Agric. 1987, 40, (1), 67-78.

100. Monagas, M.; Bartolomé, B.; Gómez-Cordovés, C., Evolution of polyphenols in

red wines from Vitis vinifera L. during aging in the bottle. Eur. Food Res. Technol.

2005, 220, (3), 331-340.

101. Hernandez, T.; Estrella, I.; Carlavilla, D.; Martin-Alvarez, P. J.; Moreno-Arribas,

M. V., Phenolic compounds in red wine subjected to industrial malolactic

fermentation and ageing on lees. Anal. Chim. Acta 2006, 563, (1-2), 116-125.

102. Gil-Munoz, R.; Gomez-Plaza, E.; Martinez, A.; Lopez-Roca, J. M., Evolution of

phenolic compounds during wine fermentation and post-fermentation: Influence of

grape temperature. J. Food Compos. Anal. 1999, 12, (4), 259-272.

103. Hernandez, T.; Estrella, I.; Perez-Gordo, M.; Alegria, E. G.; Tenorio, C.; Ruiz-

Larrrea, F.; Moreno-Arribas, M. V., Contribution of malolactic fermentation by

Oenococcus oeni and Lactobacillus plantarum to the changes in the

nonanthocyanin polyphenolic composition of red wine. J. Agric. Food Chem. 2007,

55, (13), 5260-5266.

104. Vrhovsek, U.; Wendelin, S., The effect of fermentation, storage and fining on the

content of hydroxycinnamoyltartaric acids and on browning of Pinot blanc wines.

Viticultural and Enological Sciences 1998, 53, (2), 87-94.

105. Perez-Magarino, S.; Jose, M. L. G., Effect of ripening stage of grapes on the low

molecular weight phenolic compounds of red wines. Eur. Food Res. Technol. 2005,

220, (5-6), 597-606.

106. Goldberg, D. M.; Tsang, E.; Karumanchiri, A.; Soleas, G. J., Quercetin and p-

coumaric acid concentrations in commercial wines. Am. J. Enol. Vitic. 1998, 49,

(2), 142-151.

107. Monagas, M.; Bartolome, B.; Gomez-Cordoves, C., Updated knowledge about the

presence of phenolic compounds in wine. Critical Reviews in Food Science and

Nutrition 2005, 45, (2), 85-118.

108. Baderschneider, B.; Winterhalter, P., Isolation and characterization of novel

benzoates, cinnamates, flavonoids, and lignans from Riesling wine and screening

for antioxidant activity. J. Agric. Food Chem. 2001, 49, (6), 2788-2798.

109. Stead, D., The effect of hydroxycinnamic acids and potassium sorbate on the

growth of 11 strains of spoilage yeasts. J. Appl. Bacteriol. 1995, 78, (1), 82-87.

References

208

110. Harris, V.; Ford, C. M.; Jiranek, V.; Grbin, P. R., Dekkera and Brettanomyces

growth and utilisation of hydroxycinnamic acids in synthetic media. Appl.

Microbiol. Biotechnol. 2008, 78, (6), 997-1006.

111. Chen, J. Y.; Wen, P. F.; Kong, W. F.; Pan, Q. H.; Wan, S. B.; Huang, W. D.,

Changes and subcellular localizations of the enzymes involved in phenylpropanoid

metabolism during grape berry development. Journal of Plant Physiology 2006,

163, (2), 115-127.

112. Dugelay, I.; Gunata, Z.; Sapis, J. C.; Baumes, R.; Bayonove, C., Role of cinnamoyl

esterase-activities from enzyme preparations on the formation of volatile phenols

during winemaking. J. Agric. Food Chem. 1993, 41, (11), 2092-2096.

113. Spaepen, M.; Verachtert, H., Esterase-activity in the genus Brettanomyces. J. Inst.

Brew. 1982, 88, (1), 11-17.

114. Daenen, L.; Saison, D.; Sterckx, F.; Delvaux, F. R.; Verachtert, H.; Derdelinckx,

G., Screening and evaluation of the glucoside hydrolase activity in Saccharomyces

and Brettanomyces brewing yeasts. J. Appl. Microbiol. 2008, 104, (2), 478-488.

115. Ribereau-Gayon, P., Identification of cinnamic acid esters of tartaric acid in the

limbs and berries of Vitis vinifera. Comptes Rendus de l'Academie des Sciences de

Paris 1965, 260, (1), 341-343.

116. Dumazert, G.; Margulis, H.; Montreau, F.-R., Evolution des composes phenoliques

au cours de la maturation d'un Vitis vinifera blanc: Le Mauzac. Annales de

Technologie Agricole 1973, 22, (2), 137-151.

117. Kramling, T. E.; Singleton, V. L., An estimate of the non-flavonoid phenols in

wines. Am. J. Enol. Vitic. 1969, 20, (2), 86-92.

118. Singleton, V. L., The phenolic cinnamates of white grapes and wine. J. Sci. Food

Agric. 1978, 29, 403-410.

119. Okamura, S.; Watanabe, M., Determination of phenolic cinnamates in white wine

and their effect on wine quality. Agricultural and Biological Chemistry 1981, 45,

(9), 2063-2070.

120. Ong, B. Y.; Nagel, C. W., High-pressure liquid chromatographic analysis of

hydroxycinnamic acid-tartaric acid esters and their glucose esters in Vitis vinifera.

Journal of Chromatography 1978, 157, 345-355.

121. Romeyer, F. M.; Macheix, J. J.; Goiffon, J. P.; Reminiac, C. C.; Sapis, J. C., The

browning capacity of grapes. 3. Changes and importance of hydroxycinnamic acid-

References

209

tartaric acid esters during development and maturation of the fruit J. Agric. Food

Chem. 1983, 31, (2), 346-349.

122. Ong, B. Y.; Nagel, C. W., Hydroxycinnamic acid-tartaric acid ester content in

mature grapes and during the maturation of white riesling grapes. Am. J. Enol.

Vitic. 1978, 29, (4), 277-281.

123. Vrhovsek, U., Extraction of hydroxycinnamoyltartaric acids from berries of

different grape varieties. J. Agric. Food Chem. 1998, 46, (10), 4203-4208.

124. Lee, C. Y.; Jaworski, A., Phenolic-compounds in white grapes grown in New York.

Am. J. Enol. Vitic. 1987, 38, (4), 277-281.

125. Jaworski, A. W.; Lee, C. Y., Fractionation and HPLC determination of grape

phenolics. J. Agric. Food Chem. 1987, 35, (2), 257-259.

126. Montealegre, R. R.; Peces, R. R.; Vozmediano, J. L. C.; Gascuena, J. M.; Romero,

E. G., Phenolic compounds in skins and seeds of ten grape Vitis vinifera varieties

grown in a warm climate. J. Food Compos. Anal. 2006, 19, (6-7), 687-693.

127. Singleton, V. L.; Zaya, J.; Trousdale, K., Caftaric and coutaric acids in fruit of

Vitis. Phytochemistry 1986, 25, (9), 2127-2133.

128. Nagel, C. W.; Baranowski, J. D.; Wulf, L. W.; Powers, J. R., The hydroxycinnamic

acid tartaric acid ester content of musts and grape varieties grown in the pacific

northwest. Am. J. Enol. Vitic. 1979, 30, (3), 198-201.

129. Nagel, C. W.; Wulf, L. W., Changes in the anthocyanins, flavanoids and

hydroxycinnamic acid esters during fermentation and aging of Merlot and Cabernet

Savignon. Am. J. Enol. Vitic. 1979, 30, (2), 111-116.

130. Bautista-Ortin, A. B.; Fernandez-Fernandez, J. I.; Lopez-Roca, J. M.; Gomez-Plaza,

E., The effects of enological practices in anthocyanins, phenolic compounds and

wine colour and their dependence on grape characteristics. J. Food Compos. Anal.

2007, 20, (7), 546-552.

131. Monagas, M.; Suarez, R.; Gomez-Cordoves, C.; Bartolome, B., Simultaneous

determination of nonanthocyanin phenolic compounds in red wines by HPLC-

DAD/ESI-MS. Am. J. Enol. Vitic. 2005, 56, (2), 139-147.

132. Monagas, M.; Gomez-Cordoves, C.; Bartolome, B., Evaluation of different

Saccharomyces cerevisiae strains for red winemaking. Influence on the

anthocyanin, pyranoanthocyanin and non-anthocyanin phenolic content and colour

characteristics of wines. Food Chem. 2007, 104, (2), 814-823.

References

210

133. Reschke, A.; Herrmann, K., Occurrence of 1-O-hydroxycinnamyl-β-D-glucoses in

fruits. 15. Phenolics of fruits. Zeitschrift für Lebensmitteluntersuchung und -

Forschung A 1981, 173, (6), 458-463.

134. Winterhalter, P., Application of countercurrent chromatography for wine research

and wine analysis. Am. J. Enol. Vitic. 2009, 60, (2), 123-129.

135. Mathew, S.; Abraham, T. E., Ferulic acid: An antioxidant found naturally in plant

cell walls and feruloyl esterases involved in its release and their applications.

Critical Reviews in Biotechnology 2004, 24, (2-3), 59-83.

136. Donaghy, J. A.; Kelly, P. F.; McKay, A., Conversion of ferulic acid to 4-vinyl

guaiacol by yeasts isolated from unpasteurised apple juice. J. Sci. Food Agric.

1999, 79, (3), 453-456.

137. Somers, T. C.; Verette, E.; Pocock, K. F.; Strauss, C. R., Spectral characteristics of

hydroxycinnamate esters of Vitis vinifera. Bulletin de Liaison - Groupe

Polyphenols 1986, 13, 502-505.

138. Hufnagel, J. C.; Hofmann, T., Orosensory-directed identification of astringent

mouthfeel and bitter-tasting compounds in red wine. J. Agric. Food Chem. 2008,

56, (4), 1376-1386.

139. Guntert, M.; Rapp, A.; Takeoka, G. R.; Jennings, W., HRGC and HRGC-MS

applied to wine constituents of lower volatility. Z. Lebensm.-Unters.-Forsch. 1986,

182, (3), 200-204.

140. Nickenig, R.; Pfeilsticker, K., Investigating oxidative browning of white wine .1.

HPLC separation of preparatively obtained wine phenols. Dtsch. Lebensm.-

Rundsch. 1980, 76, (4), 115-119.

141. Sleep, N. The developement of stable isotope dilution assays (SIDAs) for the

quantification of important aroma precursors in wine. Honours Thesis, Flinders

University of South Australia, 2003.

142. Galland, S.; Mora, N.; Abert-Vian, M.; Rakotomanomana, N.; Dangles, O.,

Chemical synthesis of hydroxycinnamic acid glucosides and evaluation of their

ability to stabilize natural Colors via anthocyanin copigmentation. J. Agric. Food

Chem. 2007, 55, (18), 7573-7579.

143. Zhao, H.; Burke, T. R., Facile syntheses of (2R,3R)-(-)- and (2S,3S)-(+)-chicoric

acids. Synth. Commun. 1998, 28, (4), 737-740.

References

211

144. Ziegler, T.; Pantkowski, G., Preparation of 1-O-acyl-D-glycopyranoses via

chloroacetylated glycopyranosyl donors. J. Carbohydr. Chem. 1993, 12, (3), 357-

370.

145. Zhang, S. Q.; Li, Z. J.; Wang, A. B.; Cai, M. S.; Feng, R., Total synthesis of the

phenylpropanoid glycoside, grayanoside A. Carbohydr. Res. 1997, 299, (4), 281-

285.

146. Synoradzki, L.; Ruskowski, P.; Bernas, U., Tartaric acid and its O-acyl derivatives.

Part 1. Synthesis of tartaric acid and O-acyl tartaric acids and anhydrides. Org.

Prep. Proced. Int. 2005, 37, (1), 37-63.

147. Bernas, U.; Hajmowicz, H.; Madura, I. D.; Majcher, M.; Synoradzki, L.; Zawada,

K., Tartaric acid and its acyl derivatives. Part 5. Direct synthesis of

monoacyltartaric acids and novel mono(benzoyl)tartaric anhydride: unusual

findings in tartaric acid acylation. ARKIVOC 2010, 11, 1-12.

148. Hu, Y.; Yamada, K. A.; Chalmers, D. K.; Annavajjula, D. P.; Covey, D. F.,

Enantioselective synthesis of cyclothiazide analogues: Novel probes of the

stereospecific actions of benzothiadiazine at AMPA-type glutamate receptors. J.

Am. Chem. Soc. 1996, 118, (19), 4550-4559.

149. Ishihara, K.; Gao, Q. Z.; Yamamoto, H., Enantioselective diels-alder reaction of α-

bromo α,β-enals with dienes under catalysis by CAB. J. Org. Chem. 1993, 58, (24),

6917-6919.

150. Sato, M.; Sunami, S.; Sugita, Y.; Kaneko, C., Use of 1,3-dioxin-4-ones and related-

compounds in synthesis .44. Asymmetric aldol reaction of 4-trimethylsiloxy-6-

methylene-1,3-dioxines - use of tartaric acid-derived (acyloxy)borane complex as

the catalyst. Chem. Pharm. Bull. 1994, 42, (4), 839-845.

151. Buschhaus, B.; Bauer, W.; Hirsch, A., Synthesis and chiroptical properties of a new

type of chiral depsipeptide dendrons. Tetrahedron 2003, 59, (22), 3899-3915.

152. Furuta, K.; Gao, Q. Z.; Yamamoto, H., Chiral (acyloxy)borane complex-catalyzed

asymmetric diels-alder reaction - (1R)-1,3,4-Trimethyl-3-cyclohexene-1-

carboxaldehyde - (3-Cyclohexene-1-carboxaldehyde, 1,3,4-trimethyl-, (-)-). In

Organic Syntheses, Vol 72, John Wiley & Sons Inc: New York, 1995; Vol. 72, pp

86-94.

153. Scarpati, M. L.; Oriente, G., Chicoric acid (dicaffeyltartic acid): Its isolation from

chicory (Chicorium intybus) and synthesis. Tetrahedron 1958, 4, (1-2), 43-48.

References

212

154. Lamidey, A. M.; Fernon, L.; Pouysegu, L.; Delattre, C.; Quideau, S.; Pardon, P., A

convenient synthesis of the Echinacea-derived immunostimulator and HIV-1

integrase inhibitor (-)-(2R,3R)-chicoric acid. Helv. Chim. Acta 2002, 85, (8), 2328-

2334.

155. Clough, J. M.; Jones, R. V. H.; McCann, H.; Morris, D. J.; Wills, M., Synthesis and

hydrolysis studies of a peptide containing the reactive triad of serine proteases with

an associated linker to a dye on a solid phase support. Org. Biomol. Chem. 2003, 1,

(9), 1486-1497.

156. Meier, C.; Ruppel, M. F. H.; Vukadinovic, D.; Balzarini, J., Second generation of

cycloSal-pronucleotides with esterase-cleavable sites: The "lock-in"-concept.

Nucleosides Nucleotides Nucleic Acids 2004, 23, (1-2), 89-115.

157. Patel, C. K.; Owen, C. P.; Aidoo-Gyamfi, K.; Ahmed, S., Sturcture-activity

relationship determination study of a series of novel compounds as potential

inhibitors of the enzyme estrone sulfatase (ES). Letters in Drug Design and

Discovery 2004, 1, (1), 35-44.

158. Percec, V.; Peterca, M.; Sienkowska, M. J.; Ilies, M. A.; Aqad, E.; Smidrkal, J.;

Heiney, P. A., Synthesis and retrostructural analysis of libraries of AB(3) and

constitutional isomeric AB(2) phenylpropyl ether-based supramolecular

dendrimers. J. Am. Chem. Soc. 2006, 128, (10), 3324-3334.

159. Uray, G.; Lindner, W., tert-Butyl esters and ethers of (R,R)-tartaric acid.

Tetrahedron 1988, 44, (14), 4357-4362.

160. Wright, S. W.; Hageman, D. L.; Wright, A. S.; McClure, L. D., Convenient

preparations of t-butyl esters and ethers from t-butanol. Tetrahedron Lett. 1997, 38,

(42), 7345-7348.

161. MacFarland, D. K.; Landis, C. R., Synthesis and characterization of novel ligands

designed for secondary interactions. Organometallics 1996, 15, (2), 483-485.

162. Barros, J. C.; da Silva, J. M.; Calazans, A. R.; Tanuri, A.; Brindeiro, R. D. M.;

Williamson, J. S.; Antunes, O. A. C., Synthesis of pseudopeptides derived from

(R,R)-tartaric acid as potential inhibitors of HIV-protease. Lett. Org. Chem. 2006, 3,

(12), 882-886.

163. Schmidt, M.; Amstutz, R.; Crass, G.; Seebach, D., Preparation of some chiral

aminodiols from tartaric acid - Chiral lithium aluminum-hydride derivatives for

asymmetric ketone reductions. Chem. Ber.-Recl. 1980, 113, (5), 1691-1707.

References

213

164. Whitfield, D. M.; Douglas, S. P., Glycosylation reactions - Present status future

directions. Glycoconjugate J. 1996, 13, (1), 5-17.

165. Zhu, X. M.; Schmidt, R. R., New principles for glycoside-bond formation. Angew.

Chem.-Int. Edit. 2009, 48, (11), 1900-1934.

166. Garegg, P. J., Synthesis and reactions of glycosides. In Advances in Carbohydrate

Chemistry and Biochemistry, Academic Press: 2004; Vol. Volume 59, pp 69-134.

167. Handbook of Chemical Glycosylation: Advances in Stereoselectivity and

Therapeutic Relevance. Wiley-VCH Verlag GmbH & Co.: Weinheim, Germany,

2008; p 524.

168. Modern Methods in Carbohydrate Synthesis. Harwood Academic: 1996.

169. Preparative Carbohydrate Chemistry. Marcel Dekker: New York, USA, 1997.

170. Overend, W., Glycosides. In The Carbohydrates: Chemistry and Biochemistry,

Second ed.; Pigman, W.; Horton, D., Eds. Academic Press: New York, USA, 1970;

Vol. 1A, pp 243-283.

171. Juaristi, E.; Cuevas, G., Recent studies on the anomeric effect. Tetrahedron 1992,

48, (24), 5019-5087.

172. Koenigs, W.; Knorr, E., Some derivatives of grape sugars and gallactose. Berichte

der Deutschen Chemischen Gesellschaft 1901, 34, 957-981.

173. Igarashi, K., The Koenigs-Knorr reaction. In Advances in Carbohydrate Chemistry

and Biochemistry, Tipson, R. S.; Derek, H., Eds. Academic Press: 1977; Vol.

Volume 34, pp 243-283.

174. Skouroumounis, G. K. β-Damascenone Precursors in Grapes and Wine. Ph.D.

Thesis, The University of Adelaide, 1991.

175. Wilkinson, K. L. Oak derived flavour compounds and their contribution to wine

and spirits. Ph.D. Thesis, Flinders University of South Australia, 2004.

176. Daniel, M. A.; Puglisi, C. J.; Capone, D. L.; Elsey, G. M.; Sefton, M. A.,

Rationalizing the formation of damascenone: Synthesis and hydrolysis of

damascenone precursors and their analogues, in both aglycone and glycoconjugate

forms. J. Agric. Food Chem. 2008, 56, (19), 9183-9189.

177. Wilkinson, K. L.; Elsey, G. M.; Prager, R. H.; Tanaka, T.; Sefton, M. A.,

Precursors to oak lactone. Part 2: Synthesis, separation and cleavage of several β-D-

glucopyranosides of 3-methyl-4-hydroxyoctanoic acid. Tetrahedron 2004, 60, (29),

6091-6100.

References

214

178. Hixson, J. L. Glucose esters as precursors to volatile phenols in wine. Honours

Thesis, Flinders University of South Australia, 2007.

179. Birkofer, L.; Kaiser, C.; Kosmol, H.; Romussi, G.; Donike, M.; Michaelis, G., D-

Glucose- und L-Rhamnoseester der p-Cumar- und Ferulasaure. Justus Liebig's

Annalen der Chemie 1966, 699, 223-231.

180. Bertolini, M.; Glaudemans, C. P. J., The chloroacetyl group in synthetic

carbohydrate chemistry. Carbohydr. Res. 1970, 15, (2), 263-270.

181. Schmidt, R. R., Recent developments in the synthesis of glycoconjugates. Pure

Appl. Chem. 1989, 61, (7), 1257-1270.

182. Lubineau, A.; Meyer, E.; Place, P., Synthesis of aryl D-gluco-pyranosides and D-

galacto-pyranosides and 1-O-acyl-D-gluco-pyranoses and 1-O-acyl-D-galacto-

pyranoses exploiting the mitsunobu reaction - Influence of the pKa of the acid on

the stereoselectivity of the reaction. Carbohydr. Res. 1992, 228, (1), 191-203.

183. Zhu, Y. M.; Ralph, J., Stereoselective synthesis of 1-O-β-feruloyl and 1-O-β-

sinapoyl glucopyranoses. Tetrahedron Lett. 2011, 52, (29), 3729-3731.

184. Machida, K.; Kikuchi, M., Norisoprenoids from Viburnum dilatatum.

Phytochemistry 1996, 41, (5), 1333-1336.

185. Nonnenmacher, A.; Mayer, R.; Plieninger, H., High pressure experiments. XII.

Application of high pressure in Wittig reactions with resonance stabilised ylides.

Liebigs Annalen Der Chemie 1983, 12, 2135-40.

186. Kahnt, G., trans-cis-Equilibrium of hydroxycinnamic acids during irradiation of

aqueous solutions at different pH. Phytochemistry 1967, 6, (5), 755-&.

187. Kahnt, G., Uber das gleichgewicht zwischen den stereoisomeren einiger zimt-

saurederivate in abhangigkeit von der molaren konzentration und ihre quantitative

spektrophotometrische messung bei pflanzenanalysen. Biologisches Zentralblatt

1966, 85, (5), 545-554.

188. Das, B.; Banerjee, J.; Ramu, R.; Pal, R.; Ravindranath, N.; Ramesh, C., Efficient,

selective deprotection of aromatic acetates catalyzed by Amberlyst-15 or iodine.

Tetrahedron Lett. 2003, 44, (29), 5465-5468.

189. Brecker, L.; Mahut, M.; Schwarz, A.; Nidetzky, B., In situ proton NMR study of

acetyl and formyl group migration in mono-O-acyl D-glucose. Magnetic Resonance

in Chemistry 2009, 47, (4), 328-332.

190. Whistler, R. L.; Anisuzza.Ak; Kim, J. C., Silica gel catalyzed migration of acetyl

groups from a sulfur to an oxygen atom. Carbohydr. Res. 1973, 31, (2), 237-243.

References

215

191. Iddon, L.; Richards, S. E.; Johnson, C. H.; Harding, J. R.; Wilson, I. D.; Nicholson,

J. K.; Lindon, J. C.; Stachulski, A. V., Synthesis of a series of phenylacetic acid 1-

β-O-acyl glucosides and comparison of their acyl migration and hydrolysis kinetics

with the corresponding acyl glucuronides. Org. Biomol. Chem. 2011, 9, (3), 926-

934.

192. Horrobin, T.; Tran, C. H.; Crout, D., Esterase-catalysed regioselective 6-

deacylation of hexopyranose per-acetates, acid-catalysed rearrangement to the 4-

deprotected products and conversions of these into hexose 4- and 6-sulfates. J.

Chem. Soc.-Perkin Trans. 1 1998, (6), 1069-1080.

193. Yoshimoto, K.; Tsuda, Y., On the possibility of direct O-1-β- to -6 acyl migration

in 1-O-acyl-b-D-glucose derivatives. Chem. Pharm. Bull. 1983, 31, (12), 4335-

4340.

194. Yoshimoto, K.; Tsuda, Y., General path of O-acyl migration in D-glucose

derivatives - Acyl migration of methyl mono-O-myristoyl-α and β-D-

glucopyranosides and mono-O-myristoyl-D-glucopyranoses. Chem. Pharm. Bull.

1983, 31, (12), 4324-4334.

195. Thevenet, S.; Wernicke, A.; Belniak, S.; Descotes, G.; Bouchu, A.; Queneau, Y.,

Esterification of unprotected sucrose with acid chlorides in aqueous medium:

kinetic reactivity versus acyl- or alkyloxycarbonyl-group migrations. Carbohydr.

Res. 1999, 318, (1-4), 52-66.

196. Molinier, V.; Wisnienwski, K.; Bouchu, A.; Fitremann, J.; Queneau, Y.,

Transesterification of sucrose in organic medium: Study of acyl group migrations.

J. Carbohydr. Chem. 2003, 22, (7-8), 657-669.

197. Rangelov, M. A.; Vayssilov, G. N.; Petkov, D. D., Quantum chemical model study

of the acyl migration in 2 '(3 ')-formylnucleosides. Int. J. Quantum Chem. 2006,

106, (6), 1346-1356.

198. Roslund, M. U.; Aitio, O.; Warna, J.; Maaheimo, H.; Murzin, D. Y.; Leino, R.,

Acyl group migration and cleavage in selectively protected β-D-galactopyranosides

as studied by NMR spectroscopy and kinetic calculations. J. Am. Chem. Soc. 2008,

130, (27), 8769-8772.

199. Perez-Magarino, S.; Ortega-Heras, M.; Cano-Mozo, E., Optimization of a solid-

phase extraction method using copolymer sorbents for isolation of phenolic

compounds in red wines and quantification by HPLC. J. Agric. Food Chem. 2008,

56, (24), 11560-11570.

References

216

200. Lunkenbein, S.; Bellido, M.; Aharoni, A.; Salentijn, E. M. J.; Kaldenhoff, R.;

Coiner, H. A.; Munoz-Blanco, J.; Schwab, W., Cinnamate metabolism in ripening

fruit. Characterization of a UDP-glucose: Cinnamate glucosyltransferase from

strawberry. Plant Physiol. 2006, 140, (3), 1047-1058.

201. Hartley, R. D.; Jones, E. C., Effect of ultraviolet-light on substituted cinnamic acids

and estimation of their cis and trans isomers by gas-chromatography. Journal of

Chromatography 1975, 107, (1), 213-218.

202. Fenton, T. W.; Mueller, M. M.; Clandinin, D. R., Isomerization of some cinnamic

acid-derivatives. Journal of Chromatography 1978, 152, (2), 517-522.

203. Kleinhofs, A.; Haskins, F. A.; Gorz, H. J., Ultraviolet-induced isomerization of β-

D-glucosyl O-hydroxycinnamic acid on filter paper. Journal of Chromatography

1966, 22, (1), 184-186.

204. Conkerton, E. J.; Chapital, D. C., High-performance liquid-chromatography

separation of the cis-trans isomers of cinnamic acid-derivatives - Ultraviolet and

electrochemical detection. Journal of Chromatography 1983, 281, (DEC), 326-329.

205. Katase, T., Stereoisomerization of para-coumaric and ferulic acids during their

incubation in peat soil extract solution by exposure to fluorescent light. Soil Sci.

Plant Nutr. 1981, 27, (4), 421-427.

206. Singleton, V. L.; Timberlake, C. F.; Whiting, G. C., Chromatography of natural

phenolic cinnamate derivatives on sephadex LH-20 and G-25. Journal of

Chromatography 1977, 140, (1), 120-124.

207. Challice, J. S.; Williams, A. H., Paper chromatographic separation and behaviour of

cis-and trans-isomers of cinnamic acid derivatives. Journal of Chromatography

1966, 21, (2), 357-&.

208. Li, Q. S.; Fang, W. H., Ab initio study on the structures and properties of trans-p-

coumaric acid in low-lying electronic states. Chem. Phys. 2005, 313, (1-3), 71-75.

209. Yoda, M.; Houjou, H.; Inoue, Y.; Sakurai, M., Spectral tuning of photoactive

yellow protein. Theoretical and experimental analysis of medium effects on the

absorption spectrum of the chromophore. J. Phys. Chem. B 2001, 105, (40), 9887-

9895.

210. Molina, V.; Merchan, M., On the absorbance changes in the photocycle of the

photoactive yellow protein: A quantum-chemical analysis. Proc. Natl. Acad. Sci. U.

S. A. 2001, 98, (8), 4299-4304.

References

217

211. Sergi, A.; Gruning, M.; Ferrario, M.; Buda, F., Density Functional study of the

photoactive yellow protein's chromophore. J. Phys. Chem. B 2001, 105, (19), 4386-

4391.

212. Ko, C.; Levine, B.; Toniolo, A.; Manohar, L.; Olsen, S.; Werner, H. J.; Martinez, T.

J., Ab initio excited-state dynamics of the photoactive yellow protein chromophore.

J. Am. Chem. Soc. 2003, 125, (42), 12710-12711.

213. Yamada, A.; Yamamoto, S.; Yamato, T.; Kakitani, T., Ab initio MO study on

potential energy surfaces for twisting around C7=C8 and C4-C7 bonds of coumaric

acid. Theochem-J. Mol. Struct. 2001, 536, (2-3), 195-201.

214. Kort, R.; Vonk, H.; Xu, X.; Hoff, W. D.; Crielaard, W.; Hellingwerf, K. J.,

Evidence for trans-cis isomerization of the p-coumaric acid chromophore as the

photochemical basis of the photocycle of photoactive yellow protein. FEBS Lett.

1996, 382, (1-2), 73-78.

215. Gomez-Plaza, E.; Gil-Munoz, R.; Lopez-Roca, J. M.; Martinez, A., Color and

phenolic compounds of a young red wine as discriminanting variables of its ageing

status. Food Res. Int. 1999, 32, (7), 503-507.

216. Romeyer, F. M.; Sapis, J. C.; Macheix, J. J., Hydroxycinnamic esters and browning

potential in mature berries of some grape varieties. J. Sci. Food Agric. 1985, 36,

728-732.

217. Gramatica, P.; Ranzi, B. M.; Manitto, P., Decarboxylation of cinnamic acids by

Saccharomyces cerevisiae. Bioorganic Chemistry 1981, 10, (1), 14-21.

218. Clausen, M.; Lamb, C. J.; Megnet, R.; Doerner, P. W., PAD1 encodes

phenylacrylic acid decarboxylase with confers resistance to cinnamic acid in

Saccharomyces cerevisiae. Gene 1994, 142, (1), 107-112.

219. Goodey, A. R.; Tubb, R. S., Genetic and biochemical analysis of the ability of

Saccharomyces cerevisiae to decarboxylate cinnamic acids. Journal of General

Microbiology 1982, 128, (11), 2615-2620.

220. Canovese, L.; Santo, C.; Visentin, F., Palladium(0)-catalyzed cis-trans alkene

isomerizations. Organometallics 2008, 27, (14), 3577-3581.

221. Lowry, T. H.; Richardson, K. S., Mechanism and Theory in Organic Chemistry. 3rd

ed.; Harper and Row: New York, 1987; p 1090.

222. Beltran, J. L.; Sanli, N.; Fonrodona, G.; Barron, D.; Ozkan, G.; Barbosa, J.,

Spectrophotometric, potentiometric and chromatographic pK(a) values of

References

218

polyphenolic acids in water and acetonitrile-water media. Anal. Chim. Acta 2003,

484, (2), 253-264.

223. Salameh, D.; Brandam, C.; Medawar, W.; Lteif, R.; Strehaiano, P., Highlight on the

problem generated by p-coumaric acid analysis in wine fermentations. Food Chem.

2008, 107, 1661-1667.

224. Barberousse, H.; Roiseux, O.; Robert, C.; Paquot, M.; Deroanne, C.; Blecker, C.,

Analytical methodologies for quantification of ferulic acid and its oligomers. J. Sci.

Food Agric. 2008, 88, 1494-1511.

225. Singleton, V. L.; Zaya, J.; Trousdale, E.; Salgues, M., Caftaric acid in grapes and

conversion to a reaction-product during processing. Vitis 1984, 23, (2), 113-120.

226. Gottlieb, H. E.; Kotlyar, V.; Nudelman, A., NMR chemical shifts of common

laboratory solvents as trace impurities. The Journal of Organic Chemistry 1997, 62,

(21), 7512-7515.

227. Lang, R. W.; Hansen, H. J., Simple synthesis of alkyl allenecarboxylates (allenic

esters) by the Wittig-reaction. Helv. Chim. Acta 1980, 63, (2), 438-455.

228. Wu, J.; Zhang, D.; Wei, S., Wittig reactions of stabilized phosphorus ylides with

aldehydes in water. Synth. Commun. 2005, 35, 1213-1222.

229. Li, N.-G.; Shi, Z.-H.; Tang, Y.-P.; Li, B.-Q.; Duan, J.-A., Highly efficient

esterification of ferulic acid under microwave irradiation. Molecules 2009, 14,

2118-2126.

230. Reddy, S. H. K.; Lee, S.; Datta, A.; Georg, G. I., Efficient synthesis of the 3 '-

phenolic metabolite of paclitaxel. J. Org. Chem. 2001, 66, (24), 8211-8214.

231. Pearl, I. A.; Beyer, D. L., Reactions of vanillin and its derived compounds .11.

Cinnamic acids derived from vanillin and its related compounds. J. Org. Chem.

1951, 16, (2), 216-220.

232. Xia, Y.; Wang, W., Asymmetric synthesis of machilin C and its analogue.

Chemical Papers 2010, 64, (5), 630-636.

233. Lebedev, A. V.; Lebedeva, A. B.; Sheludyakov, V. D.; Kovaleva, E. A.; Ustinova,

O. L.; Kozhevnikov, I. B., Competitive formation of β-amino acids, propenoic, and

ylidenemalonic acids by the rodionov reaction from malonic acid, aldehydes, and

ammonium acetate in alcoholic medium. Russ. J. Gen. Chem. 2005, 75, (7), 1113-

1124.

234. Leschot, A.; Tapia, R. A.; Eyzaguirre, J., Efficient synthesis of 4-methyl-

umbelliferyl dihydroferulate. Synth. Commun. 2002, 32, (20), 3219.

References

219

235. Salum, M. a. L.; Robles, C. J.; Erra-Balsells, R., Photoisomerization of ionic liquid

ammonium cinnamates: One-pot synthesis−isolation of Z-cinnamic acids. Org. Lett.

2010, 12, (21), 4808-4811.

236. Prachayasittikul, S.; Suphapong, S.; Worachartcheewan, A.; Lawung, R.;

Ruchirawat, S.; Prachayasittikul, V., Bioactive metabolites from Spilanthes acmella

Murr. Molecules 2009, 14, (2), 850-867.

237. Shimizu, T.; Kojima, M., Partial purification and characterization of UDPG: t-

Cinnamate glucosyltransferase in the root of sweet potato, Ipomoea batatas Lam.

Journal of Biochemistry 1984, 95, (1), 205-212.

238. Kiviranta, P. H.; Leppanen, J.; Rinne, V. M.; Suuronen, T.; Kyrylenko, O.;

Kyrylenko, S.; Kuusisto, E.; Tervo, A. J.; Jarvinen, T.; Salminen, A.; Poso, A.;

Wallen, E. A. A., N-(3-(4-hydroxyphenyl)-propenoyl)-amino acid tryptamides as

SIRT2 inhibitors. Bioorg. Med. Chem. Lett. 2007, 17, (9), 2448-2451.

239. Hosoda, A.; Nomura, E.; Mizuno, K.; Taniguchi, H., Preparation of a (+/-)-1,6-di-

O-feruloyl-myo-inositol derivative: An efficient method for introduction of ferulic

acid to 1,6-vicinal hydroxyl groups of myo-inositol. J. Org. Chem. 2001, 66, (21),

7199-7201.

240. Dobashi, Y.; Hara, S., A chiral stationary phase derived from (R,R)-tartramide with

broadened scope of application to the liquid-chromatographic resolution of

enantiomers

J. Org. Chem. 1987, 52, (12), 2490-2496.

241. Lucas, R. L.; Benjamin, M.; Reineke, T. M., Comparison of a tartaric acid derived

polymeric MRI contrast agent to a small molecule model chelate. Bioconjugate

Chem. 2008, 19, (1), 24-27.

242. Austin, P. C.; Park, J. R., The rotatory dispersion of derivatives of tartaric acid. Part

II. Acetyl derivatives. Journal of the Chemical Society 1925, 127, 1926-1934.

243. Baderschneider, B. Isolierung und Strukturaufklarung antioxidativ wirksamer

Verbindungen aus Weisswein. Ph.D. Thesis, Universitat Braunschweig, Germany,

2000.

244. Opitz, S.; Schnitzler, J. P.; Hause, B.; Schneider, B., Histochemical analysis of

phenylphenalenone-related compounds in Xiphidium caeruleum (Haemodoraceae).

Planta 2003, 216, (5), 881-889.

245. Yella, R., Chloroacetylchloride: A versatile reagent in heterocyclic synthesis.

Synlett 2010, 2010, (EFirst), 835,836.

References

220

246. Sigma-Aldrich http://www.sigmaaldrich.com/australia.html (19/01/2012),

247. Schroeder, C.; Lutterbach, R.; Stockigt, J., Preparative biosynthesis of natural

glucosides and fluorogenic substrates for β-glucosidases followed by in vivo 13C

NMR with high density plant cell cultures. Tetrahedron 1996, 52, (3), 925-934.

248. Plusquellec, D.; Roulleau, F.; Bertho, F.; Lefeuvre, M.; Brown, E., Sugar chemistry

without protective groupings .1. Regioselective esterification of anomeric hydroxyl

of lactose, maltose and glucose. Tetrahedron 1986, 42, (9), 2457-2467.

249. Hixson, J. L.; Sleep, N.; Capone, D. L.; Elsey, G. M.; Curtin, C.; Sefton, M. A.;

Taylor, D. K., Hydroxycinnamic Acid Ethyl Esters as Precursors to Ethylphenols in

Wine. J. Agric. Food Chem. 2012.

chemistry of hydroxycinnamate esters and their role as ... · bruxellensis . two classes of berry...

Documents