mass spectrometry based investigation of chlorogenic acid
TRANSCRIPT
Mass Spectrometry Based Investigation of
Chlorogenic Acid Reactivity and Profile in Model
Systems and Coffee Processing
by
Sagar Deshpande
A Thesis submitted in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
in Chemistry
Approved Dissertation Committee
Prof. Dr. Nikolai Kuhnert (Chair)
Prof. Dr. Gerd-Volker Röschenthaler (Reviewer)
Prof. Dr. Michael N. Clifford (External Reviewer)
Date of Defense: 24th January 2014
School of Engineering and Science
i
Abstract
Beneficial health and biological effects of coffee as well as its sensory properties are largely
associated with chlorogenic acids (CGAs) since; coffee is the richest dietary source of CGAs
and their derivatives. From green coffee beans to the beverage, chemical components of the
green coffee undergo enormous transformations, which have been studied in great details in
the past. Roasted coffee melanoidines are extensively contributed by the products formed by
the most relevant secondary metabolite- chlorogenic acids. For every 1% of the dry matter of
the total CGA content in the green coffee beans, 8-10% of the original CGAs are transformed
or decomposed into respective derivatives of cinnamic acid and quinic acid. The non-volatile
fraction of the roasted coffee remains relatively unravelled in the aspects of its chemistry and
structural information. Coffee roasting, along with the other processes brings about
considerable changes in the chlorogenic acid profile of green coffee through number of
chemical processes. In roasting, chlorogenic acids evidently undergo various processes such
as, acyl group migration, transesterification, thermal trans-cis isomerization, dehydration and
epimerization. To understand the chemistry behind roasted coffee melanoidines, it is of
utmost importance to study the changes occurring in CGAs and their derivatives through food
processing.
The isomeric transformations of the chlorogenic acids resulting due to the migration of
hydroxycinnamoyl group from any of the four hydroxyl groups of quinic acid to another have
been thoroughly investigated in this work by LC-MSn. In this thesis, the acyl migration
phenomenon under the treatment of tetramethylammonium hydroxide (TMAH) hydrolysis,
model roasting experiments and by brewing at pH 5 (water reflux, 5 h) of the seven
commercially available mono- and di-caffeoylqunic acids was studied in detail.
Intermolecular acyl migration (transesterification) was also studied by tetramethylammonium
hydroxide (TMAH) hydrolysis and model roasting experiments in between 5-CQA and p-
coumaric acid as well as in 5-CQA and ferulic acid.
In this thesis, four diastereoisomers of quinic acid have been synthesized selectively, namely,
epi-quinic acid, muco-quinic acid, cis-quinic acid and scyllo-quinic acid by applying
appropriate hydroxyl group protection and deprotection strategies in order to study their
behavior in LC-MSn along with commercially available (-)-quinic acid. We report for the first
time that these diastereoisomers are distinguishable on the basis of their fragmentation
behavior as well as their chromatographic elution order. In this study, we also observed that
ii
muco-quinic acid, scyllo-quinic acid and epi-quinic acid are present in hydrolyzed Guatemala
roasted coffee sample as possible products of roasting. The synthetic work accomplished in
present work will provide for the generation of the reference standards to identify remaining
epimers of CGAs in roasted coffee.
Considering the fact that relatively large amount of the degradation products of CGAs such
as, quinic acid, caffeic acid along with the most prominent member of the CGAs profile in
roasted coffee, 5-caffeoylquinic acid itself are present in the roasted coffee along with free
small, non-volatile organic acids, we examined in details the further esterification
phenomenon among themselves. To investigate transesterification in roasted coffee in details
we designed a thorough analytical plan involving four experiments. A selection of small
organic acids were heated in the presence of 5-CQA to check if simulated roasting conditions
facilitate the formation of the transesterification, caffeic acid and quinic acid with the mixture
of all the organic acids separately to check, which of the organic acid show greater affinity
towards the formation of the condensed esters. With the experimental results in hand, we then
identified transesterification products in different roasted coffee samples by LC-MSn, LC-
TOF-MS and FT-ICR-MS.
Data generated by different analytical techniques such as, NMR-, CD-, IR spectroscopy and
LC-MS was used to differentiate the Arabica and Robusta green coffee extracts by principal
component analysis (PCA) to determine, which spectroscopic technique allows the best
discrimination of coffee varieties. A total of 38 green bean extracts were characterized using
NMR-, CD-, and IR spectroscopy along with LC-MS and the data was further analyzed by
PCA using different PCA processing parameters by unsupervised non-targeted approach.
Distinction between different groups of samples, in particular, Arabica versus Robusta green
coffee beans successfully achieved using IR- spectroscopy and LC-MS. Surprisingly, both
CD- and NMR spectroscopy fail to achieve in this case, an adequate level of distinction. This
is the first study that directly compares the value of various spectroscopic techniques if
multivariant statistical techniques are employed to them.
iii
Acknowledgements
I, hereby acknowledge that most of the times, this specific segment of the thesis write-up is
supposed to be formal. But, I want to take up the liberty of using the language in which, I am
comfortable in to thank people who actually thought they would make a possible ‘scientist’
out of something like ‘me’ risking a doubt which may arise in my examiners mind of being
too non-scientific in my thesis. I am taking this risk because, to my personal belief, I think the
apology and the gratefulness should come by heart when you are down two beers.
Firstly, I would like to thank Mr. Yadav who awaken the interest in me to peruse chemistry as
a career as opposed to being a play-writer or a Himalayan monk in my adolescent age. Big
thanks with a little judgment goes to the committee, which accepted my application for
masters in Nanomolecular science in Jacobs University after having a degree in organic
chemistry with moderate grades. I must thank Mr. Sunil Joshi for giving me an opportunity to
work in a reputed research institute such as, National Chemical Laboratory (NCL), Pune,
which made my application to Jacobs look considerable. I would like to thank Prof. Dr.
Nugent who pumped a great deal of professionalism into me starting from how I should reply
to the formal emails. Prof. Nugent pulled me out of some tough times and taught me to be
serious about the opportunities I have in present and may have in my future. I am grateful to
Ms. Shalaka Shah, Mr. Ketan Kulkarni and my lecturer first and a very close friend later, Mr.
Suparna Tambe for making my masters days in India and in Germany so easy and joyful. My
lab-mates, Rakesh, Tina, Aga, Hande, Marius, Maria, Rohan, Boris, Mohammed, Abhinandan
(names not in order of importance!) and their partners have been the family away from my
home. All of whom, suffered me, my inappropriate jokes and occasional emotional outbursts
with so much of tolerance, that I am unable express my gratitude towards them even if they
might not understand the necessity in doing so.
I cannot even begin to thank Prof. Dr. Nikolai Kuhnert for trusting me to perform this work as
a PhD candidate. My intelligence level matches to a Chimp in front of him; still it’s his skills
that he managed to pull a significant contribution out of me to the coffee chemistry. I want to
mention that if I manage to incorporate 10% of his knowledge and 1% of his modesty and
humbleness in myself when I am of his age, I would consider myself to be very successful in
my life. I want to thank Prof. Dr. Clifford for accepting to be an external examiner for my
dissertation although; it is of considerable inconvenience for him to travel from U.K. and
Prof. Dr. Gerd-Volker Röschenthaler for taking time to revise my thesis. I am thankful to Ms.
iv
Anja Mueller for her technical support for five years. I appreciate my collaborating professor,
Prof. Dr. Materny and his group, in particular Dr. Rasha El-Abasssy for very fruitful
collaborative work. I would like to thank Dr. Bassem Bassil from the group of Prof. U. Kortz
for the measurement and solution of the single crystal X-ray structures.
I have been in Germany for more than five years now and most of the times my wife Neha
and me, had to live apart from each other. She has been and will be a definition of
unconditional love for me. She stretched me to the extremes of the feeling of happiness and
(rarely) suicidal throughout the entire time before and after marriage. She trusted me with the
fatherhood of our child, Raghav. Although, she would doubt it, there is no one close to my
heart than her. We are a proud evidence for a working long distance happy relationship.
There are actually very few words in which I could be expressing my gratefulness towards my
father, mother, sister and uncle completely. I am very aware of the fact that it is due to their
financial and emotional sacrifices I am what I am today.
I am thankful to Jacobs community for teaching me how to accept and respect diverse
nationalities and religions. It is the healthy international environment of Jacobs University,
which gave me conversational confidence and a feeling of not being special or unique than
other communities all over the world. I am thankful to my Indian friends in India and in
Jacobs as well, for support and long lasting memories throughout the entire stay in Germany.
My former flat mate turned very good friend Naveen, thanks to you too.
Last but not the least at all, I am very thankful to Jacobs University and Kraft Food for
financial support.
v
Contents
Abstract………………………………………………………………………………………..i
Acknowledgements…………………………………………………………………………..iii
List of Figures………………………………………………………………………………...ix
List of Tables………………………………………………………………………………..xiii
Abbreviations………………………………………………………………………………..xv
CHAPTER 1: Introduction ..................................................................................................... 1
1.1 Introduction ...................................................................................................................... 1
1.2 Coffee: Commercial aspects and biological relevance ..................................................... 1
1.3 Classes of compounds present in coffee ........................................................................... 4
1.3.1 Carbohydrates....... ......................................................................................................4
1.3.2 Lipids........ .................................................................................................................5
1.3.3 Amino acids and protein... ..........................................................................................6
1.3.4 Acids and phenolic compounds in coffee ..................................................................7
1.4 Chlorogenic acids: Definition, history, occurrence and biological relevance .................. 9
1.5 Fate of CGAs in food processing and Aim of the project .............................................. 17
1.5.1 Acyl migration...... ....................................................................................................18
1.5.2 Epimerization........ ....................................................................................................19
1.5.3 Transesterification ....................................................................................................20
1.5.4 Suitability of an analytical technique to differentiate large set of green coffee
extracts from various origins by PCA... ............................................................................20
References...............................................................................................................22
CHAPTER 2: LC-MSn identification of CGAs in green and roasted coffee ………….31
2.1 Introduction .................................................................................................................... 31
2.2 LC-MSn identification of CGAs in green coffee ............................................................ 32
2.3 LC-MSn identification of CGAs in roasted coffee ......................................................... 45
References……………………………………………………………………………..50
vi
CHAPTER 3: Acyl migration in mono- and di-caffeoylquinic acids under basic and
aqueous acidic conditions and dry roasting conditions ...................................................... 52
3.1 Introduction .................................................................................................................... 52
3.2 Materials and methods .................................................................................................... 53
3.3 Results and discussion .................................................................................................... 55
3.3.1 Intramolecular acyl migration: hydrolysis by TMAH of 2-5 and 8-10.................... 55
3.3.2 Intermolecular acyl migration (Transesterification): hydrolysis by TMAH (Cross-
over experiment) ............................................................................................................... 63
3.3.3 Intramolecular acyl migration: model roasting of 2-5 and 8-10 .............................. 72
3.3.4 Transesterification: model roasting (Cross-over experiment) ................................. 77
3.3.5 Intramolecular acyl migration: Brewing of CGAs ................................................... 80
3.4 Conclusions……………………………………………………………………………….83
References……………………………………………………………………………..86
CHAPTER 4: Synthesis, structure and tandem MS investigation of diastereomers of
quinic acid ............................................................................................................................... 89
4.1 Introduction .................................................................................................................... 89
4.2 Experimental ................................................................................................................... 91
4.2.1 Synthesis of the mixture of the epimers of (-)-quinic acid ...................................... 91
4.2.2 Synthesis of the epi-quinic acid (2).......................................................................... 92
4.2.3 Synthesis of the muco-quinic acid (3) ...................................................................... 95
4.2.4 Synthesis of the cis-quinic acid (4) .......................................................................... 95
4.2.5 Synthesis of the scyllo-quinic acid (5) ..................................................................... 99
4.2.6 Hydrolysis of the CGAs in roasted coffee ............................................................. 101
4.2.7 Synthesis of the methyl esters of epi-, muco-, cis-, scyllo-quinic acids and (-)-quinic
acid .................................................................................................................................. 101
4.2.8 X-ray crystallography ............................................................................................ 102
4.3 Results and discussions ................................................................................................ 102
4.4 Discussion of the X-ray structures ................................................................................ 120
vii
4.5 Conclusions .................................................................................................................. 122
References.....................................................................................................................123
CHAPTER 5: Transesterification of chlorogenic acids with small organic acids present
in the coffee bean .................................................................................................................. 125
5.1 Introduction .................................................................................................................. 125
5.2 Materials and methods .................................................................................................. 127
5.2.1 Chemicals and materials ........................................................................................ 127
5.2.2 Model roasting ....................................................................................................... 128
5.2.3 Aqueous extract of roasted coffee .......................................................................... 129
5.2.4 Roasted coffee samples for ESI-FT-ICR-MS analysis .......................................... 129
5.3 Results and discussion .................................................................................................. 135
5.3.1 Transesterification of 5-CQA(2) in model roasting and in roasted coffee samples
......................................................................................................................................... 136
5.3.2 Transesterification of quinic acid (1) in model roasting and in roasted coffee
samples ............................................................................................................................ 151
5.3.3 Transesterification of caffeic acid (3) in model roasting and in roasted coffee
samples ............................................................................................................................ 158
5.4 Conclusions....................................................................................................................160
References......................................................................................................................161
CHAPTER 6: Which spectroscopic technique allows best differentiation of coffee
varieties: Comparing principal component analysis using data derived from CD-, NMR-
, IR- spectroscopy and LC-MS in the analysis of the chlorogenic acid fraction in green
coffee beans ........................................................................................................................... 163
6.1 Introduction .................................................................................................................. 163
6.2 Materials and methods .................................................................................................. 166
6.2.1 Statistical analysis .................................................................................................. 166
6.3 Experimental ................................................................................................................. 167
6.4 Results and discussion .................................................................................................. 168
viii
6.4.2 Circular Dichroism spectroscopy ........................................................................... 170
6.4.3 Infrared spectroscopy ............................................................................................. 177
6.4.4 1H NMR spectroscopy ........................................................................................... 179
6.5 Conclusions...................................................................................................................181
References....................................................................................................................182
Conclusions............................................................................................................................184
List of Publications................................................................................................................186
Curriculum Vitae..................................................................................................................188
ix
List of figures
Figure 1.1 Structures of caffeine and adenosine 3
Figure 1.2 Basic chemical structures of aliphatic carboxylic acids which are present in coffee
in free form and in mixed esters of CGAs 8
Figure 1.3 Basic structures of quinic acid, derivatives of cinnamic acid and typical
chlorogenic acids 10
Figure 1.4 Basic structures of feruloyl- and p-coumaroylquinic acid and mono- and di-
caffeoylquinic acids 14
Figure 1.5 Chlorogenic acids and their derivatives 15
Figure 1.6 Representative scheme showing possible chemical transformations in CGA 18
Figure 1.7 Stereoisomers of quinic acid 19
Figure 2.1 TIC of green Robusta coffee extract in negative ion mode 32
Figure 2.2 Structures of the fragments generated by quinic and cinnamic acid derivatives 36
Figure 2.3 MS2 and MS3 spectra of 3-acyl chlorogenic acids in negative ion mode 37
Figure 2.4 MS2 and MS3 spectra of 4-acyl chlorogenic acids in negative ion mode 38
Figure 2.5 MS2 and MS3 spectra of 5-acyl chlorogenic acids in negative ion mode 39
Figure 2.6 MS2, MS3, and MS4 spectra of 3,4-diCQA, 3,5-diCQA, and 4,5-diCQA in negative
ion mode (m/z 515) 41
Figure 2.7 MS2, MS3, and MS4 spectra of 3,4-diFQA 16 and 3D-4FQA 34 in negative ion
mode (m/z 543 and m/z 557, respectively) 42
Figure 3.1 Structure of mono and di caffeoylquinic, p-coumaroylquinic and feruloylquinic
acids 57
Figure 3.2 UV Chromatograms (318-322 nm) at 2, 5, 10 and 30 minutes of base hydrolysis of
5-CQA (3) 58
Figure 3.3 Amount of the transformation products after base hydrolysis for different time
intervals of 5-CQA (3), 4-CQA (4) and 3-CQA (2) 59
x
Figure 3.4 Mechanism of the acyl migration through an ortho-ester intermediate formation 60
Figure 3.5 Amount of the transformation products after base hydrolysis for different time
intervals of di-acylated reference standards 63
Figure 3.6 UV Chromatograms (318-322 nm) at 2, 5, 10, 30 and 60 minutes of base
hydrolysis of 3, 5-diCQA (9) 66
Figure 3.7 Compounds identified during acyl migration studies 67
Figure 3.8 Comparison between the peak areas of compounds formed during TMAH
treatment of 5-CQA (3) with p-coumaric acid (pCoA) 69
Figure 3.9 EIC and fragmentation patterns for 1-cis-caffeoylquinic acid (49) at m/z 353 and
caffeoyl-feruloylquinic acid (51) at m/z 529 in transesterification induced by TMAH 70
Figure 3.10 Comparison between the peak areas of compounds formed during TMAH
treatment of 5-CQA (3) with ferulic acid (FA) 72
Figure 3.11 EIC and fragmentation patterns for m/z 671 observed during model roasting 76
Figure 3.12 EIC and fragmentation patterns for m/z 497 observed during model roasting 77
Figure 3.13 MS3 and MS4 of 4-pCoQA (14) and 5-FQA (23) respectively observed during
cross-over experiment by model roasting 79
Figure 3.14 Structures identified after brewing of the reference standards 84
Figure 4.1 Stereoisomers of quinic acid 88
Figure 4.2 Reaction scheme for obtaining scyllo-quinic acid (5) and epi-quinic acid (2) 91
Figure 4.3 X-ray structure of methyl 3,5-Di-O-(tert-butyldimethylsilyl)quinate (7)
Conformer A 92
Figure 4.4 X-ray structure of methyl 3,5-Di-O-(tert-butyldimethylsilyl)quinate (7)
Conformer B 93
Figure 4.5 Reaction scheme for obtaining cis-quinic acid (4) and Methyl-cis-quinate (17) 95
Figure 4.6 X-ray structure 3,4-O-Cyclohexylidene-1,5-quinide (13) 96
xi
Figure 4.7 X-ray structure of cis-quinic acid (4) 98
Figure 4.8 X-ray structures of 3-O-cinnamoyl-1,4-scyllo-quinide (9) and 3-O-cinnamoyl-1,5-
quinide (10) 99
Figure 4.9 X-ray structure of methyl cis-quinate (17) 101
Figure 4.10 EIC at m/z 191 in negative mode and MSn of quinic acid (1), epi-quinic acid (2),
muco-quinic acid (3), cis-quinic acid (4) and scyllo-quinic acid (5) 104
Figure 4.11 MSn of the acidic fraction of non-selectively isomerized quinic acid 109
Figure 4.12 Diastereomers identified in the EIC of the hydrolyzed roasted coffee in negative
ion mode 110
Figure 4.13 MSn of the diastereomers of the quinic acid and reference compounds obtained
from the direct infusion experiments 113
Figure 4.14 Proposed mechanisms for the fragmentation of 1, 2, 3 and 4 118
Figure 5.1 Structures of all the reactants involved in the model roasting experiments 127
Figure 5.2 Tentative structures of transesterification products 130
Figure 5.3 UV chromatogram at 254 nm of the model roasting experiment sample generated
by heating 5-CQA (2) with succinic acid (6) 144
Figure 5.4 Fragmentation scheme for compound 19 (glutaric acid+ caffeoylquinic acid) 145
Figure 5.5 Fragmentation patterns for m/z 481 (21) and m/z 463 (22) 148
Figure 5.6 Total ion chromatogram in the negative mode of the model roasting experiment
sample generated by heating QA (1) with glutaric acid (7) 151
Figure 5.7 EIC at m/z 341 159
Figure 6.1 Representative chromatogram of green coffee extract of sample No. 33 (Tanzania
Robusta), a) TIC in negative ion mode; b) UV-VIS chromatogram monitored at 320 nm 171
Figure 6.2 CD spectra of 3-CQA and 3,5-diCQA 172
Figure 6.3 The PCA score and loading plots of the obtained CD spectral data 177
Figure 6.4 ATR-IR spectrum of Panama Boguete Arabica extract 179
Figure 6.5 The PCA score and loading plots of the obtained IR spectral data 179
xii
Figure 6.6 1H-NMR spectra of Tanzania Robusta in DMSO-d6 181
Figure 6.7 The PCA score plot of the obtained NMR spectral data 181
xiii
List of tables
Table 1.1 Typical quantities of CGAs found in vegetables and fruits 11
Table 1.2 Biological activities associated with CGAs 13
Table 2.1 Chlorogenic acids identified in green coffee beans 32
Table 2.2 MS2 and MS3 data of monoacyl CGAs in negative ion mode 43
Table 2.3 MS2, MS3, and MS4 data of diacyl CGAs in negative ion mode 44
Table 2.4 Chlorogenic acids identified in roasted coffee 46
Table 2.5 Negative ion mode MS2, MS3 and MS4 fragmentation data for the
cinnamoylshikimate esters and chlorogenic acid lactones 48
Table 3.1 Compounds identified after base treatment of CGA with p-coumaric acid and CGA
with ferulic acid for various time intervals 64
Table 3.2 Compounds identified after heating (model roasting) reference standards 73
Table 3.3 Compounds identified after heating (Model roasting) of 5-CQA (3) and p-coumaric
acid 78
Table 3.4 Compounds identified after heating (Model roasting) 5-CQA (3) and ferulic acid 80
Table 3.5 Compounds identified after hydrolysis (Brewing) of reference standards 81
Table 4.1 MS2 data of quinic acid diastereomers in negative ion mode at 75% collision
energy 103
Table 4.2 Crystal data and structure refinement for compounds 17, 4, 9, 10, 15 and 7 120
Table 5.1 Transesterification products of 4-11 with 5-CQA (2) identified with LC-TOF-MS
in the samples generated by heating each acid separately with 5-CQA 137
Table 5.2 Transesterification products of 4-11 with 5-CQA (2) identified with LC-MSn in the
samples generated by heating each acid separately with 5-CQA 141
Table 5.3 Compounds transesterified with 5-CQA identified in FT-ICR-MS data of roasted
coffee samples 146
Table 5.4 Transesterification products of 4-11 with 5-CQA (2) identified with LC-TOF-MS
in the roasted coffee samples 143
xiv
Table 5.5 Transesterification products of 4-11 with 5-CQA (2) identified with targeted LC-
MSn in the samples generated by heating all of the acids collectively with 5-CQA 149
Table 5.6 Transesterification products of 4-11 with 5-CQA (2) identified with LC-TOF-MS
in the samples generated by heating all of the acids collectively with 5-CQA 149
Table 5.7 Transesterification products of 4-11 with quinic acid (1) identified with LC-TOF-
MS in the samples generated by heating each acid separately with quinic acid (QA) 153
Table 5.8 Compounds transesterified with quinic acid (QA) identified in FT-ICR-MS data of
roasted coffee samples 156
Table 5.9 Transesterification products of 4-11 with quinic acid identified with LC-TOF-MS
in the samples generated by heating all of the acids collectively with quinic acid 157
Table 5.10 Compounds transesterified with caffeic acid (CA) identified in FT-ICR-MS data
of roasted coffee samples 158
Table 5.11 Transesterification products of 4-11 with CA (3) identified with LC-TOF-MS in
the roasted coffee samples 155
Table 6.1 Origins, nature and grouping of green bean coffee samples anayzed and included in
PCA analysis 173
Table 6.2 Numbering, nomenclature and high resolution MS data of selected secondary
metabolites identified in green bean coffee samples 174
xv
Abbreviations
Ac Acetyl
APCI Atmospheric Pressure Chemical Ionisation
Bn Benzyl
CD Circular Dichorism
CFQA Caffeoyl-feruloylquinic Acid
CGA Chlorogenic Acid
CGAs Chlorogenic Acids
CQ Caffeoylquinate
CQA Caffeoylquinic Acid
CQL Caffeoylquinic Acid Lactone/ Caffeoylquinide
CSA Caffeoylshikimic Acid / Caffeoylshikimate
CSiQA Caffeoyl-Sinapoylquinic Acid
DAD Diode Array Detector
DCC N,N’-Dicyclohexylcarbodiimide
DCE Dichloroethane
DCM Dichloromethane
DEAD Diethyl Azodicarboxylate
DIAD Diisopropyl Azodicarboxylate
DiCQA Dicaffeoylquinic Acid
DMAP N,N’-Dimethylaminopyridine
DMP 2,2-Dimethoxypropane
DMSO Dimethylsulfoxide
DMF N,N’-Dimethylformamide
ESI Electrospray Ionisation
FQA Feruloylquinic Acid
FQL Feruloylquinic acid lactone/ Feruloylquinide
FSA Feruloylshikimic Acid/ Feruloyl Shikimate
FSiQA Feruloyl-Sinapoylquinic Acid
xvi
FT-ICR Fourier transform ion cyclotron resonance
HPLC High Performance Liquid Chromatography
HRMS High Resolution Mass Spectometry
ISCID In Source Collision Induced Dissociation
LC Liquid Chromatography
LC-MSn Liquid Chromatography Tandem Mass Spectrometry
MeOH Methanol
MRM Multi Reaction Monitoring
MS Mass Spectrometry
PCA Principal Component Analysis
p-CoQA p-Coumaroylquinic Acid
RP Reverse Phase
SIM Selected Ion Monitoring
SiQA Sinapoylquinic Acid
THF Tetrahydrofuran
TLC Thin Layer Chromatography
TMB 2,2,3,3-Tetramethoxybutane
TBDMS-Cl tert-butyldimethylsilyl chloride
TOF Time of Flight
TriCQA Tricaffeoylquinic Acid
Troc 2,2,2,-Trichloroethylformyl
Troc-Cl 2,2,2-Trichloroethylformyl Chloride
1
CHAPTER 1: INTRODUCTION
1.1 Introduction
The profound health and biological effects of coffee have received a lot of scientific attention
in last decade. Beneficial health effects of coffee are largely associated with chlorogenic acids
(CGAs) and their derivatives present in coffee, which have been extensively investigated in
animals, in vitro models and in humans also through numerous epidemiological studies. 1-7
CGA content of coffee also contributes considerably towards the uniqueness of coffee’s
sensory and organoleptic properties. 8
In this thesis, the work carried out and accomplished results are presented in six chapters: first
chapter covers the introduction of coffee underlying its commercial aspects, classes of
compounds present in coffee, introduction to CGAs; definition, occurrence in coffee and
biological relevance, food processing and CGAs, and the statement of the aim of the project.
The second chapter covers profiling, structural elucidation and identification of CGAs by
tandem mass spectrometry (MS) in green and roasted coffee. The following three chapters
describe the outcomes from three different projects undertaken to understand in details, the
fate of CGAs in food processing. The last chapter deals with the challenges in verification of
the possible adulterations in coffee varieties and focuses on, which spectrometric technique
allows the best differentiation of coffee varieties by comparing the principle component
analysis data generated from CD-, NMR-, IR- spectroscopy and LC-MS in the analysis of the
chlorogenic acid fraction in green coffee beans.
1.2 Coffee: Commercial aspects and biological relevance
Coffee is one of the most valued agricultural commodities in terms of the economic aspects of
the exports from the developing coffee producing countries, accounting to ca. 8 million metric
tonnes per year. About 70 to 80% of the total human population consumes coffee as a
beverage regularly. More than 70 countries in the world cultivate coffee, which grows in the
form of ‘cherries’ on a coffee plant. Approximately, 2.3 billion cups of coffee are consumed
worldwide per day. 9 After water and black tea, coffee is the third most consumed beverage
on this planet with a market value in excess of 5 Billion US $ of the raw material alone. In
countries like USA and Germany coffee is second most consumed beverage. Approximately,
450 million cups of coffee are consumed every day in United States only. Green coffee beans
2
are produced in two varieties, Coffea Arabica (known as Arabica coffee) and Coffea
canephora (known as Robusta coffee) holding 70% and 30%, respectively of the total coffee
market in the world.7 The conventional coffee whether instant, filter or freshly ground is made
from roasting the green beans of coffee obtained from cherry fruits of either Arabica or
Robusta varieties of the coffee plant.
Similar to any other plant food material, coffee contains a complex mixture of over 1000
chemical components. In recent times, it has been acknowledged on a wider platform that
consumption of coffee under specific conditions provide consumers with certain physiological
benefits sourced from certain chemical components of coffee beyond basic nutritional
functions. Assessment of the impact of the coffee consumption henceforth on the human
health must consider potential beneficial as well as adverse health effects arising from coffee
and its constituents. Specially, if we consider the fact that although coffee consumption goes
back to over 1000 years in history, until recently most of the studies on health effects of
coffee were based on potential adverse and toxic effects. More than 100 diseases have been
linked to the coffee consumption in humans in literature.10 Among other diseases,
hypertension, cardiovascular disease, cancer, fatalities to the fetus and child birth such as,
early abortions and low birth weight and osteoporosis were often linked directly to coffee
consumption. Roasted coffee contains series of compounds that have been shown to be
carcinogens in animal models including acrylamide and furan derivatives. Regardless of the
huge research attention put into the establishment of the direct link between these diseases
and coffee consumption, the evidence to support the links has been inconsistent and limited.
For example, a large epidemiological study involving 43,000 people was conducted in
Norway which confirmed that there is no association between overall risk of cancer and
coffee consumption 11 this phenomenon has been described as the coffee paradox. Not to
mention that Norway is the country where per capita coffee consumption is among the largest
in the world.
The beneficial health effects of coffee are attributed to the polyphenolic content present in the
coffee. It varies based on the concentration of the polyphenols present in the green bean and
roasted bean because, different roasting conditions gives rise to the different derivatives of the
phenolic compounds in different concentrations in roasted coffee, which can affect the
antioxidative properties of coffee variety directly. Commonly, protective effects of coffee
consumption are generally accompanied with the antioxidant activity resulting in protection of
the cells from the oxidative damage in body. However, it is now clear that these antioxidative
3
properties are caused by the phenolic compounds and their metabolites. Apart from CGAs,
caffeoyl-tryptophan and caffeine present in green coffee beans have also been shown to
contribute to the antioxidative activity of coffee 12, 13 along with roasted coffee melanoidines
and phenylindans. The mechanism of the antioxidative activity of coffee is complex and
believed to be a collective contribution towards different processes such as, radical
scavenging, transition metal chelation and active oxygen trapping. The importance of anti-
oxidant activity for the observed health effects of coffee and other phenol rich dietary
materials is under intense discussion.
O
N
N
NN
O
Caffeine
OH
OH
HO
O
NN
H2N
N N
Adenosine
Figure 1.1 Structures of caffeine and adenosine
Neuroactive behavioral modification induced by coffee consumption is also counted amongst
positive health effects of coffee. Caffeine plays a key role as a coffee component in these
effects. Caffeine blocks adenosine receptors in the central nervous system and at high doses
exceeding 500 mg it interferes with gamma amino butyric acid (GABA) transmission in the
brain (Figure 1.1). A typical serving of coffee contains 80-200 mg of caffeine. LD50 has
been established for humans to be around 150-200 mg/ kg (192 mg/ kg in rats) with dose
above 500 mg leading to initial intoxication symptoms. Caffeine has been reported to be
responsible for increased alertness, performance improvement and fatigue reduction in the
literature.14 Behavioral modification through coffee- induced neuroactivity has also been
documented to possess properties towards prevention of suicidal tendencies and caffeine has
directly been related as active anti-depression factor. Epidemiological studies report three to
five fold decreases in the risk of suicide in both men and women compared to the placebo
group, directly related to the coffee intake.15 Griffiths et al. established a direct relationship
between the experimental administrations of caffeine to the increase in subjective feeling of
wellness, self-confidence and motivation and decrease in the social anxiety. 16 Caffeine has
also been reported as a responsible factor for improved mood 17 and decreased irritability 18 in
psychiatric studies.
4
As stated earlier, coffee consumption has been attempted to be linked as one of the causes of
various negative health effects by the scientific community in the past. But, recent
epidemiological evidences and results from different in vivo and in vitro studies on the effect
of coffee constituents on human health suggests that coffee consumption may prove
protective against some type of cancers such as, colon cancer. 19 It is postulated that the anti-
oxidant activity induced by the constituents of coffee stimulates the chemo-detoxification
processes in body based on the experimental data, resulting in the chemoprotective properties
of coffee. As reported by Klatsky et al. and Corrao et al., risk of acquiring cirrhosis through
excessive alcohol consumption can be reduced by coffee consumption.20-22 It was also
observed that consumption of coffee adversely affect the advancement of hepatitis B and C
infections on development of cirrhosis thus indicating positive effect of coffee in case of non-
alcoholic cirrhosis. Stelzer et al. published an epidemiological study in which he found that
coffee consumption in patients undergoing radiotherapy as a part of cervical cancer treatment
significantly decreases the chances of severe late radiation injuries. 23
Caffeine and certain chlorogenic acids have been reported to have glycemic effects in
humans. 102, 105,106 Johnston and Clifford reported that coffee consumption modulated plasma
glucose levels, gastrointestinal hormone and insulin secretion in humans. Their study
confirmed the role of caffeine in glucose uptake, gastrointestinal hormone and insulin
secretion also, 5-CQA was observed to impart an antagonistic effect on glucose transport.24
1.3 Classes of compounds present in coffee
Carbohydrates, proteins and amino acids, lipids and organic acids and phenolics are the most
prominent classes of compounds in the non-volatile fraction of coffee, which have been
investigated in details.
1.3.1 Carbohydrates
Carbohydrates content in coffee is very important, as it contributes approximately half of the
dry weight of the green coffee bean. 25 The roasting process brings out extensive changes in
the chemical constituents of the green coffee in which, both low and high molecular weight
carbohydrates present in green coffee bean play a major role. Monosaccharides are found in
negligible amounts in coffee with sucrose being present as major component of the low
molecular weight sugar content. Arabica coffee variety contains about twice amount of
sucrose than Robusta variety. Clifford summarized the reported amount in literature to be in
the range of 2- 5% for Robusta and 5- 8.5% for Arabica coffee green beans 26 also, Robustas
were found to contain more reducing sugars than Arabicas. There was no evidence found for
5
the existence of other simple oligomeric sugars such as, raffinose or stachyose except for
sucrose and mannose in very minute quantity (0.1%). 27 However, upon roasting the sucrose
content is degraded almost completely and roasted coffee contains only negligible amounts of
sucrose totaling to 0.24- 0.33%. 28 The hydrolysis products of sucrose, which are reducing
sugars, were expected to be identified in the roasted coffee samples in quantitative amount
but, evidently glucose and fructose undergo thermal degradation even more rapidly than
sucrose only to leave their trace amounts in light roasted coffee. As reported by Noyes and
Chu, 21 samples from roasted Arabica or Robusta coffees of Brazilian origin showed the
presence of only 0.1% sugars in total with 0.8% of sucrose itself. 29 In the roasting conditions
involving higher temperature since, water is absent for the glycosidic bond cleavage, cellulose
is expected to form through pyrolysis of oligomers or polymers of monosaccharides. 30
However, anhydro-sugars were also detected in the range of 0.1%.
On the other hand, high molecular weight sugars or polysaccharides were found to be
relatively stable in the roasting process since, they are the principle cell wall component of
coffee bean, which through its thickness, provide the characteristic hardness to the bean.
Polysaccharide content in green coffee bean is dominated by arabinogalactan, mannan and/or
galacto-mannan and cellulose. 31 Cellulose was found to be the most stable polysaccharide
during the roasting procedure while, arabinogalactan being the most labile. Mannan remains
more stable to roasting than galacto-mannan polysaccharides. 31 Rich crema or foam
generated in espresso coffee signifies the quality of the brew. Nunes et al. established a direct
relationship between the stability of the foam and concentration of the polysaccharide content
present in the coffee variety. Brazilian Arabicas and Ugandan Robustas found out to contain
highest polysaccharide content hence, producing richest ‘crema’ of espresso brew. 32
Polysaccharides present in coffee are not hydrolyzed by mammalian enzymes therefore, are
considered as dietary fibers. Accordingly, Rao et al. reported potential properties of coffee
fiber as anti-colon cancer agent. 33
1.3.2 Lipids
Most amount of the coffee-oil or the lipids are found to be present in the endosperm of the
coffee bean. 34 The lipid content varies in green Arabica and Robusta coffee. About 15% of
the dry weight of green Arabica coffee bean is comprised of lipids while, green Robusta
coffee bean possesses 10% of the lipids on a dry weight basis. 35 Out of the total green
coffee-oil, about 75% is comprised of triglycerides, which mean most of the lipid content of
green coffee is unsaponifiable. Free and esterified diterpene alcohols contribute about 19% to
6
the rest 25% of the coffee oil whereas, esterifed sterols contribute 5%. Very small quantities
of other substances such as, tocopherols are also present in the composition of the green
coffee oil. 36 Upon roasting, fatty acid content does not undergo major changes as far as the
amounts are concerned. Trans fatty acid levels are found to be increasing after roasting green
Arabica and Robusta coffee beans. Free fatty acids (FFAs) were found to exist at around 1 g/
100 g of total lipids whereas, green Robustas FFA content was found at comparatively higher
amount; about 2 g/ 100 g of lipids. Roasting does not bring about noticeable changes in the
amount of distribution of FFA content with only exception of linoleic acid, which decreases
slightly during roasting at higher temperatures. 37
In this work, we selected linoleic acid and palmitic acid as representatives of FFA content of
lipids in green coffee for the purpose of studying the possibility of transesterification between
FFA and quinic acid (1) during model roasting experiments since, both of them are observed
in significant amounts in their free form in roasted coffee. Detailed results from this
experiment are presented in Chapter 4.
1.3.3 Amino acids and protein
Both amino acids and proteins contribute towards the color, flavor and aroma of the brewed
coffee. Quantity and variety of the amino acids and proteins present in coffee play a vital role
to the aromatic qualities of coffee variety. 38 Free amino acid (FAA) content of the green
coffee bean is largely transformed by the roasting process resulting in trace amounts in
roasted coffee.
The post-harvesting processes such as, drying, fermentation and storage affect the content of
free amino acids (FAAs) considerably. For example, if adhering pulp was not removed from
the coffee bean before drying, amount of glutamic acid increases after drying the freshly
harvested coffee beans to 500 mg/ kg dry basis. Whereas, in most samples, aspartic acid was
found to be decreasing by 110-780 mg/ kg dry basis. 39 Decaffeination by steam treatment
decreases free amino acids content significantly. It was observed that decrease in the levels of
FAA by steaming is grater in Arabica coffee than Robusta coffee. The amount of protein
bound amino acids was also found to be decreasing with the increase in the duration of steam
treatment. The industrially steamed coffee bean contains 10% less of protein bound amino
acids and 50% less of FAAs compared to their original amounts. On the other hand, protein
bound amino acids were found to be more stable in the roasting environment than free amino
acids, this observation is supported by the recent studies confirming the role of proteins in the
formation of aroma and metal-chelating compounds in coffee brew. 38
7
Derivatives of amino acids with caffeic acid and other hydroxycinnamic acids such as, ferulic
acid and p-coumaric acid have been detected in green coffee bean. 40-42 However, in roasted
coffee, no information about these derivatives found reported in the literature.
1.3.4 Acids and phenolic compounds in coffee
Acids present in coffee are the major contributors to the perceived taste characteristic to
coffee. Acidity in coffee is considered as one of the important parameters for the quality of
the coffee variety. 43 11 % of the total weight of the green coffee bean is contributed by the
acid content, which decreases upon roasting to 6 %. 44 This acid content is contributed by
various volatile and non-volatile acids. As Clifford reported earlier, in brewed coffee citric
acid (9), phosphoric acid (3), phytic acid (2), quinic acid (1), chlorogenic acids and malic acid
(10) are the most important acids contributing to the perceived acidity. 43 Other free organic
acids present in roasted coffee such as oxalic acid (4), malonic acid (5), glutaric acid (7),
adipic acid (8), tartaric acid (11) and succinic acid (6) 45-54 need to be considered as factors
affecting the acidic taste as well. These acids do not contribute to the titrable acidity of the
coffee as established by Engelhard and Maier 54, 55 but they might exist as anions in a coffee
brew or in a coffee extract providing protons.
Volatile acids, which can be isolated from coffee sample by distillation processes and
detected by GC directly are basically low molecular weight aliphatic compounds bound to a
carboxylic functionality. Rancid smelling volatile acids such as 2- and/or 3-methylbutyric
acid (13, 12) contribute to the aromatic properties of coffee. 56 Apart from formic and acetic
acid, propanoic, butanoic, isomers of methyl propanoic and methyl butanoic acids are also
found in Arabica and Robusta roasted coffee. Small organic acid content was found to be
increasing in proportion with the higher degree of roasting; however, higher saturated fatty
acids did not show significant increase with higher degree roasts. Straight chain fatty acids
from C5 to C10 were also observed in roasted coffee as hydrolysis products of high molecular
weight fatty acids. 57
Approximately 10% of the total composition of the processed seeds of green Coffea
canephora (robusta coffee) and around 6-10% of green Arabica coffee variety comes from
chlorogenic acids on dry basis, out of the total content, 5-O-caffeoylquinic acid (32)
contributes about half. 7 3-O-caffeoylquinic acid (33) and 4-O-caffeoylquinic acid (34)
contribute significantly after 5-O-caffeoylquinic acid to the total composition of the coffee
chemistry accompanied by feruloylquinic acids (35), dicaffeoylquinic acids such as, 3,4-O-,
4,5-O- and 3,5-O-dicaffeoylquinic acids (40-42). Mono- and di-acyl chlorogenic acids
8
involving p-coumaric acid (24) and 3,4-dimethoxycinnamic acid (26) have also been observed
to contribute in minor proportions to the chlorogenic acid profile of green coffee. 7
O
PHO
OHO
O
POH
HO
OOP
OH
HO
OO
P
HO
HOO
OP
OHHO
OO P
OH
OH
O
phytic acid
OH
OHO
O
HO OH
OO
OH
O
O
HOOH
OO
HOOH
O
HO
O
OH
OO
HO
O OH
OH
HO
O OH
OH
O
OH
HO
O OH
OH
O
OH
HO
HO
OH
O
OH
1
2
3 4
5 6 7 8
9 10 11
Oxalic acid
Malonic acid Succinic acid Glutaric acid Adipic acid
Citric acid Malic acid Tartaric acid
(-)-qunic acid
P
O
HO
OH
OH
Phosphoric acid
O
HO
3-methylbutyric acid
12
O
HO
2-methylbutyric acid
13
O OHO
OH
citraconic acid
14
O
HO
O
OH
itaconic acid
15
O
HO
O
OH
mesaconic acid
16
O
OH
O
HO
fumaric acid
17
O
HO
HO
maleic acid
18
O
O
O
OHO
methoxyoxalic acid
19
HO
O
OH
OOH
CH3
3-Hydroxy-3-methylglutaric acid
20
O
Figure 1.2 Basic chemical structures of aliphatic carboxylic acids which are present in coffee
in free form and in mixed esters of CGAs
50% of the chlorogenic acids are lost through decomposition during roasting thus, producing
half of the decomposition products such as, quinic acid and hydroxycinnamic acids (Figure
1.3) through hydrolysis. 7, 9, 58 Quinic acid (1) ranges from 3 to 6 g/kg in green robusta and
9
Arabica coffee beans from various origins in the free form as reported earlier. 53, 59 After
steaming the green coffee beans as a part of the decaffeination process, original quinic acid
content rises by up to 15% as shown by Hucke and Maier. 60 Similarly, roasting also helps to
elevate the quinic acid content. 44, 60, 61 Although chlorogenic acid lactones and quinic acid
lactones are among the degradation products from chlorogenic acid, free quinic acid in
roasted coffee is still found out in roasted coffee to maintain its existence between the ranges
of 6.63 to 9.47 g/kg. 62 Among the other non-volatile organic acids, citric acid (9) and malic
acid (10) are present in green coffee in the ranges of 5 to 15 g/kg in arabica and 3 to 10 g/kg
in robusta respectively. 59, 63 12 to 18 % of these acids are degraded during roasting process.
Citric acid mainly yields citraconic acid (14), glutaric acid (7), itaconic acid (15), mesaconic
acid (16) and succinic acid (6) as decomposition products during roasting whereas, malic acid
(10) generate fumaric acid (17) and maleic acid (18). Among these degradation products of
citric and malic acids, succinic acid, glutaric acid and fumaric acid are included in this work
to study their significance towards the formation of transesterification products produced in
coffee roasting (Chapter 4). 62
1.4 Chlorogenic acids: Definition, history, occurrence and biological relevance
Chlorogenic acids (CGAs) are generally defined as a family of the esters between quinic acids
and certain trans-cinnamic acids, most commonly caffeic acid (21), p-coumaric acid (24) and
ferulic acid (22) 64, 65 (Figure 1.3). This report will use the nomenclature defined by the
IUPAC system for (-)-quinic acid as, 1L-1(OH),3,4/5-tetrahydroxycyclohexane carboxylic
acid. 66
The most common chlorogenic acid (CGA) is 5-O-caffeoylquinic acid (32) (5-caffeoylquinic
acid or 5-CQA), which formerly referred to as 3-CQA (pre-IUPAC). In this work, the current
nomenclature is used as per IUPAC guidelines and the numbering in some of the references is
changed as per requirement of consistency. Shorthand, which is used for abbreviation of the
CGAs in this work is as follows: A-XQA or AX,BY-QA where A and B signify the position
of acyl substituent and X and Y define the chemical nature of the substituent e.g. C =
caffeoyl, pCo = p-coumaroyl, F = feruloyl, Si = sinapoyl, D = dimethoxycinnamoyl. Quinic
acid will be referred as, QA, shikimic acid as, SA and quinic acid lactone as, QL (L= lactone).
Subsequently, 3-caffeoylquinide will be written as, 3-CQL (44) and 5-caffeoylshikimic acid
as, 5-CSA. Similarly, 4Si,5-CQA will stand for 4-sinapoyl-5-caffoylquinic acid (43).
The term, ‘chlorogenic acid’ arises from the chemical reaction of CGAs to generate green
pigment when reacted with ferric chloride falsely indicating the presence of chlorine, which
10
was first introduced by Payen. 67 Fischer and Dangschat firstly identified CGAs to be the
esters of caffeic and quinic acid. 68
OH
HOOCOH
OH
OH
1
3
45
OR
HOOCOR
OR
OR
1
3
4
5
(-) Quinic acid Chlorogenic acid OR
HOOCO
OR
OR
1
3
4
5
O
OH
OH
R= H, 5-Caffeoylquinic acid
Common name: Chlorogenic acid
OR= Cinnamoyl or Alkoyl
HO
HO
OH
O
O
HO
OH
O
O
O
OH
O
O
HO
OH
O
O
HO
OH
O
OH
O
O
O
OH
O
O
Caffeic acid Ferulic acid
Dimethoxycinnamic acid
Sinapic acid p-Coumaric acid
m-Coumaric acid Trimethoxycinnamic acid
HO
HO
OH
O
OH
Trihydroxycinnamic acid
HO
HO
O
OH
O
Isoferulic acid
21 22 23 24
25 26 27 28
29 1
32
Figure 1.3 Basic structures of quinic acid, derivatives of cinnamic acid and typical
chlorogenic acids
CGAs may be subdivided by the identity, number, and positions of the individual acyl
residues. The following subgroups can be identified:
1. Mono esters of hydroxycinnamic acids (caffeic acid (21), ferulic acid (22), sinapic
acid (23), etc.). 69-71
11
2. Diesters, triesters, and a single tetraester of a single hydroxycinnamate moiety (e.g.
diferuloyl, tricaffeoyl, or tetracaffeoyl quinic acid). 69, 71-74 This class of compounds
can be referred to as homo-di or homo-tri esters.
3. Mixed diesters, triesters of caffeic acid (21) and ferulic acid (22) or any other
hydroxycinnamate moieties (referred to as hetero-diesters or hetero-triesters). 71, 73, 74
4. Mixed esters involving various permutations of a hydroxycinnamate and other
aromatic or aliphatic ester substituent (e.g. oxalic (4), methoxyoxalic (19), fumaric
(17), succinic (6), malic (10), glutaric acid (7) characteristic for many plants of the
Asteraceae family and 3-hydroxy-3-methylglutaric acid found in Gardeniae Fructus,
Rubiaceae family) (Figure 1.2). 75
5. Other derivatives including cis-hydroxycinnamate esters or esters of diastereoisomers
of quinic acid (1). 73
The dietary burden of CGAs with their occurrence in plants as a secondary metabolite was
reviewed comprehensively by Clifford. 64, 65 Additionally, an excellent and comprehensive
review summarizing the chemistry of CGAs has also been published by Clifford. 76 Table 1.1
summarizes the typical quantities of CGAs found in vegetables and fruits.
Table 1.1 Typical quantities of CGAs found in vegetables and fruits
Food source Source Amount (mgkg-1) Ref
Coffee Roast coffee 20-675 mg(200ml)-1 77
Tea Black tea 10-50 gkg-1 78, 79
Maté Maté 107-133 mg(200ml)-1 80
Pome fruits Apple 62-385 mgkg-1 81, 82
Pear 60-280 mgkg-1 83, 84
Stone fruits Cherries, apricot 150-600 mgkg-1 85
Berry fruits Blueberries 0.5-2 gkg-1 86
Blackcurrants 140 mgkg-1 87
Blackberries 70 mgkg-1 88
Raspberries 20-30 mgkg-1 88
Strawberries 20-30 mgkg-1 88
12
Redcurrant 20-30 mgkg-1 88
Gooseberries 20-30 mgkg-1 88
Citrus fruits Oranges 170-250 mgkg-1 89, 90
Grapefruit 27-62 mgkg-1 89, 91
Lemon 55-67 mgkg-1 89, 91
Grapes and wines Grape juice 10-430 mgl-1 81
American wine 9-116 mgl-1 92
Other fruits Pine apple 3 mgl-1 93
Kiwi 11 mgl-1 94
Brassica vegetables Kale 6-120 mgkg-1 89, 95
Cabbage 104 mgkg-1 89, 95
Brussels sprouts 37 mgkg-1 89, 95
Broccoli 60 mgl-1 96
Cauliflower 20 mgkg-1 96
Radish 240-500 mgkg-1 97
Chenopodiaceae Spinach 200 mgkg-1 98, 99
Asteraceae Lettuce 50-120 mgkg-1 98
Endive 200-500 mgkg-1 98
Chicory 20 mgkg-1 98
Solanaceae Potato 500-1200 mgkg-1 100
Aubergines 600 mgkg-1 100
Tomatoes 10-80 mgkg-1 101
Apiaceae Carrot 20-120 mgkg-1 89
Cereals Barley bran 50 mgkg-1 94
Rice 12 gkg-1 102
13
Table 1.2 Biological activities associated with CGAs
Source Chlorogenic Acid/
Derivative
Biological
Activity
Ref
Food materials (Diets), 32 Antioxidant 103-105
Synthetic 45 Increase hepatic
glucose utilization
106
Synthetic 47 Antidiabetic 102
Synthetic 46-70 Antidiabetic 102
- 32 Antidiabetic 107
Coffee 32 Glycemic effects,
Anticancer
108, 109
Baccharis genistelloides 38, 40, 41, 42 Anti-HIV 110
Achyrocline satureioides 71 Anti-HIV 110
- 32, 38, 40, 41, 42,
71
Anti-HIV 111
Aster scaber 73 Anti-HIV 112
Evolvulus alsinoides 37 Antistress 113
Ipomoea pes-caprae 73, 74 Collagenase
inhibitor
114
Lonicera japonica 40, 41, 42 Anti-HIV,
inhibition of DNA
polymerase-α
115
Lonicera japonica 74, 76 Anti-HIV,
inhibition of DNA
polymerase-α
115
Lonicera bournei 32, 75, 77, 78, 79,
80, 81, 82, 83, 84,
85, 86
Hepatocyte
protective activity
116
Hedera helix 32, 41, 42, 87 Antispasmodic 116
Roasted coffee 88, 89, 90 Antidiabetic,
Inhibition of
adenosine
transporter
117, 118
Approximately, a normal human being ingests 1 to 2.5 g of chlorogenic acids per day. The
figures for average daily intake vary considerably depending on the diet types and food habits
of the consumer as well as the insufficiency of the accurate data on these compounds. Exact
quantities of CGAs present in different food materials are rarely known accurately.
Inadequate knowledge on statistical variances between the content of CGAs in plant materials
of different geographical origins and species adds up to the difficulties in describing CGA
profile in particular food material. In many cases, the analytical data is obtained through the
techniques based on derivatization followed by colorimetry, thus quantifying only one
14
particular class of CGAs e.g., caffeoyl or feruloyl esters, and therefore resulting in a possible
underestimation of actual CGA content present in the sample. 119 Old data in many cases
proves to be obsolete to be considered as relevant currently, because of the change in
agricultural practices.
O
HOOCOH
OH
O
1
3
45
O
HOOCO
OH
OH
1
3
4
5
OO O
OH
HO
OH
HO
O
OH
OH
O
HOOCOH
O
OH
1
3
4
5
O
OH
OH
O
OHHO
OH
HOOCOH
O
O
1
3
45
O
OHHO
O
OH
OH
OH
HOOCO
OH
O
1
3
45
O
OH
OH
O
OH
HO
OH
HOOCO
O
OH
1
3
4
5
O
OH
OH
O
OH
OH
1,3-diCQA (Cynarin)1,5-diCQA
1,4-diCQA
3,4-diCQA
3,5-diCQA
4,5-diCQA
OH
HOOCOH
OH
O
1
3
4
5
O
OH
HO
3-CQA
OH
HOOCOH
O
OH
1
3
4
5
O
OHHO
4-CQA
OH
HOOCOH
OH
O O
OH
O
OH
HOOCOH
OH
O O
OH
p-CoQA
33
34
35 36
37
38
39
40
41
42
FQAOH
OH
OH
HOOCO
O
OH
1
3
4
5
O
OH
OH
O
OH
O
4Si,5-CQA
43
O
OH
O
OH
O
1
3
4
5
O
OH
HO
O
44
3-CQL
Figure 1.4 Basic structures of feruloyl- and p-coumaroylquinic acid and mono- and di-
caffeoylquinic acids
15
O OH
O
OHHOOC
O
OH
H
H
Cl
S-3483
OH OH
O
OHHOOC
O
OH
OH
OH OH
O
OHHOOC
O
OH
OH
OH
OH OH
O
OHHOOC
O
N
OH OH
O
OHHOOC
O
S
46 47 48
49OH
O
OHHOOC
O
OH
OH OH
O
OHH3COOC
O
OH
OH
OH
NH
OH
O
OH
HOOC
50 51 52
45
O
OHHOOC
O
OH
OH
O
OH
O
OH
HOOC OH OH
O
OHHOOC
O
OH
OHOH OH
O
OHHOOC
O
OH
OH
OH
53 54 55 56
O OH
O
OHHOOC
O
OH
O OH
O
OHHOOC
O
OH
O OH
O
OHHOOC
O
OH
O OH
O
OHHOOC
O
OH
57 58 59 60
O OH
O
OHHOOC
O
OH
S
O OH
O
OHHOOC
O
OH
O61 62
O OH
O
OHHOOC
O
OH
O OH
O
OHHOOC
O
OH
63 64
O OH
O
OHHOOC
O
OH
Cl
O OH
O
OHHOOC
O
OH
Cl
O OH
O
OHHOOC
O
O
Cl
O OH
O
OHHOOC
O
Cl65 66 67 68
16
O OH
O
OHHOOC
O
Cl
O OH
O
OHHOOC
O
Cl
S
NN
69 70
O O
O
OHHOOC
O
OH
OH
O
OH
OH
O
O
O
71
OH
HOOCO
OH1
3
4
5
O
OH
OH
O
O
HOOH
OH
HOOCO
O
O
1
3
4
5
O
OH
OH
O
OH
HO
O
OH O
HOOCO
O
O
1
3
4
5
O
OH
OH
O
OH
O
OH
OH
O
OH72 73 74
OH
ROOCO
OH
O
1
3
4
5
O
OH
OH
O
OH
HO
OH
ROOCOH
O
O
1
3
4
5
O
OH
HO
O
OH
OH
R=Methyl 77
R=Ethyl 78
R=Methyl 75
OH
ROOCO
O
OH
1
3
4
5
O
OH
OH
O
OH
OH
R=Ethyl 80
R=Methyl 79
R=Ethyl 76
R=Ethyl 82
OH
ROOCO
OH
OH
1
3
45
O
OH
OH
R=Methyl 81
OH
H3COOCOH
OH
O
1
3
4
5
O
OH
HO
OH
ROOCOH
O
OH
1
3
4
5
O
OH
OH
R=Methyl 84
R=Ethyl 85
83
OH
HOOCO
OH
OH
1
3
4
5
O
OH
OHOH
HOOCOH
OH
O
1
3
4
5
O
OH
OH
O
O
HO
O
O
HOO
O
15
3 4
O
OH
O
86 8788
O
O
HO
O
O
HOOH
O
15
3 4
O
OH
OH
O
O
HO
O
O
HO
O
15
3 4
O
OH
89 90
Figure 1.5 Chlorogenic acids and their derivatives
17
Coffee is the richest source of chlorogenic acids and most of the beneficial health effects from
coffee consumption including anti-diabetic, antioxidant, anticancer, and protective effects on
Parkinson’s disease and Alzheimer’s disease 104, 120-123 have been associated with the CGA
content present in coffee. CGAs are ubiquitous plant material and have been reported
previously to possess important positive biological effects such as, antioxidant capacity,
radical scavenging activity, antimutagenic/anticarcinogenic effect, anti-diabetic, anti-HIV,
anti-bacterial, anti-HBV, inflammation inhibiting, endothelial protective properties and so on.
102, 103, 110, 111 Table 1.2 gives a brief summary of the biological effects associated with
specific CGAs present in the respective food source.
1.5 Fate of CGAs in food processing and Aim of the project
The volatile compounds formed in coffee roasting have been extensively researched in the
past. More than 800 different volatile compounds have been identified in roasted coffee,
which includes heterocycles such as, pyrazines, furanethiols, disulfides and aldehydes. 9
However, attempts to unravel the chemistry and structures of the non-volatile fraction of the
roasted coffee have been reported on very few occasions. The non-volatile fraction; so called,
‘melanoidines’ is a very complex mixture composed from several fractions of diverse
molecular weight, which can be separated using dialysis as shown by Schols and co-workers.
124, 125 Lack of analytical techniques and strategies to undertake complex task to obtain useful
structural information of the majority of the compounds in melanoidines has been an
impossible venture until now. Main constituents of the green coffee are chlorogenic acids as
major secondary metabolites in green coffee, proteins and carbohydrates. Considering the fact
that the CGA fraction accounts for an estimated value of 10 w% of the green coffee bean, in
this work we have undertaken the task of unravelling the fate of CGA fraction during
processes like roasting and brewing. During food processing, CGAs undergo chemical
processes such as, acyl migration, Transesterification, thermal trans-cis isomerization,
dehydration and epimerization (Figure 1.6).
18
HO
OH
O
O
OH
OH
HO
OH
O
O
O
OH
O OH
OHO
OH
O OH
HO
OH
O
O
OH
OH
O
HO
OMe
O
HO
OMe
O
HO
OMe
O
HO
OMe
O
HO
OMe
HO
OH
HO
O
OH
O
O
OH
OMe
HO
O
O
O
OH
OH
O
HO
OMe
O OH
OH
H
HO
OH
O
O
OH
OH
O
OH
OMe
ab
c
d
e
f
gAcyl migration
Oxidation
Epimerisation
Hydration andtrans-cis isomerisation
Dehydration
Lactonisation
Figure 1.6 Representative scheme showing possible chemical transformations in CGA
1.5.1 Acyl migration
Roasted coffee melanoidines are extensively contributed by the products formed by the most
relevant secondary metabolite- chlorogenic acids. For every 1% of the dry matter of the total
CGA content in the green coffee beans, 8-10% of the original CGAs are transformed or
decomposed into respective cinnamic acid derivatives and quinic acid. 126 Hydroxycinnamoyl
group in chlorogenic acids undergoes positional exchanges during coffee roasting. The
isomeric transformations of the chlorogenic acids resulting due to the migration of a
hydroxycinnamoyl group from any of the four hydroxyl groups of quinic acid to another have
been thoroughly investigated in this work by LC-MSn. Previously, Clifford described the acyl
group migration in selected chlorogenic acids in aqueous basic solutions and Dawidowicz
reported on acyl migration in 5-caffeoylquinic acid (32) in aqueous acidic solutions.
However, mechanistic study comparing inter- and intra-molecular acyl migration involving a
number of chlorogenic acids in basic conditions and in dry roasting conditions as well as in
19
brewing process is not yet reported. In this contribution, we studied the acyl migration
phenomenon under the treatment of tetramethylammonium hydroxide (TMAH) hydrolysis,
model roasting experiments and by brewing at pH 5 (water reflux, 5 h) of the seven
commercially available mono- and di-caffeoylquinic acids. Intermolecular acyl migration
(transesterification) was also studied by tetramethylammonium hydroxide (TMAH)
hydrolysis and model roasting experiments in between 5-CQA (32) and p-coumaric acid (24)
as well as in 5-CQA and ferulic acid (22).
1.5.2 Epimerization
CGA derivatives undergo isomerization to form CGA derivatives based on one or more of the
six possible diastereoisomers of quinic acid. To verify this hypothesis the synthesis of CGA
derivatives (mono and diacyl derivatives) based on the diastereoisomers of (-)-quinic acid
(muco-, epi-, scyllo-, cis-, neo-, iso-quinic acid) should be accomplished.
OHOH
OHO
HO
OH
OH
OHO
HO
OH
OH
(-)-quinic acid (1) (-)-epi-quinic acid (91)
OH
OHO
HO
OHOH
muco-quinic acid (92)
OHOH
OHO
HO OH
cis-quinic acid (93)
OH
OHO
HO OHOH
scyllo-quinic acid (94)
OHO
HO
OHOH
OH
neo-quinic acid (95)
C4 inverted C3 inverted C5 inverted
C4 and C5 inverted C3 and C4 inverted
(+)-quinic acid
OHOH
HO
O
OHOH
OH
OHOH
HO
O
OH
(+)-epi-quinic acid
Figure 1.7 Stereoisomers of quinic acid
In this contribution, we have selectively synthesized four isomers namely, epi-quinic acid
(91), muco-quinic acid (92), cis-quinic acid (93) and scyllo-quinic acid (94) in order to study
their behavior in LC-MSn along with commercially available (-)-quinic acid (1) (Figure 1.7).
We report for the first time that these isomers are distinguishable on the basis of their
fragmentation behavior as well as their chromatographic elution order. In this study, we also
observed that muco-quinic acid (92), scyllo-quinic acid (94) and epi-quinic acid (91) are
present in hydrolyzed Guatemala roasted coffee sample as possible products of roasting. Non
selective isomerization of (-)-quinic acid using acetic acid/conc. H2SO4 was performed from
20
which, we could identify epi-quinic acid , scyllo-quinic acid and (-)-quinic acid using newly
assigned fragmentation schemes and retention times characteristic to the specific compound.
1.5.3 Transesterification
The perceived taste of the coffee is largely contributed by the acids present in the coffee. In
fact, acidity in coffee is considered as one of the important parameters for the quality of the
coffee variety. 11 % of the total weight of the green coffee bean is contributed by the acid
content, which decreases upon roasting to 6 %. 63 This acid content is contributed by various
volatile and non-volatile acids. Considering the fact that relatively large amount of the
degradation products of CGAs such as, quinic acid (1), caffeic acid (21) along with the most
prominent member of the CGAs profile in roasted coffee, 5-caffeoyl quinic (32) acid itself are
present in the roasted coffee along with free small, non-volatile organic acids. In this work,
we report the further condensation of the CGAs and their decomposition products with the
non-volatile fraction of the total acid content. We have taken a set of small organic acids and
heated each of them individually with 5-caffeoyl quinic acid to check if simulated roasting
conditions facilitate the formation of the transesterification products. Same experimental
conditions were used incorporating caffeic acid and quinic acid as well. Also, we heated 5-
caffeoyl quinic acid, caffeic acid and quinic acid with the mixture of all the organic acids
separately to check, which of the organic acid show greater affinity towards the formation of
the condensed esters. The set of eight organic acids contained oxalic acid (4), malonic acid
(5), succinic acid (6), glutaric acid (7), adipic acid (8), citric acid (9), malic acid (10) and
dextrotartaric acid (11) (Figure 1.2). In each of the experiment, the roasting conditions were
simulated by keeping the temperature at 200 oC for the duration of 12 minutes. All the
samples acquired from these experiments were analyzed by high resolution ESI-TOF-MS.
Four green coffee samples were also roasted in the conditions described earlier and then
analyzed by ESI-FT-ICR-MS to identify the transesterification products in roasted coffee
samples.
1.5.4 Suitability of an analytical technique to differentiate large set of green coffee
extracts from various origins by PCA
In order to investigate parameters like geographic origin, varieties, adulterations, processing
conditions, sensory properties, beneficial health effects, shelf-life or any other desirable or
undesirable property of a food, a detailed knowledge of its composition and chemistry is
required and therefore becomes foremost a problem of analytical chemistry.
Once the chemical constituents of food have been elucidated comparison of the chemical
21
profile of different samples allows differentiation between samples and identification of
variations that are of interest to both producer and consumer. Most popular statistical method
aimed at data reduction in current times is Principle component analysis (PCA). The question
however arises when choosing the most suitable analytical technique which can provide
optimum differentiation between given set of food samples. In this work, we tried to answer
this question in terms of differentiation of green coffee extracts.
Within this contribution we have analysed aqueous methanolic extracts of a total of 38 green
bean coffee samples, which vary in terms of coffee variety and processing conditions. We
have characterized these extracts using NMR-, IR- and CD spectroscopy along with LC-MS.
All spectroscopic data have been analysed by principal component analysis (PCA) using
different PCA processing parameters using an unsupervised non-targeted approach. We could
show, that distinction between different groups of samples, in particular, Arabica versus
Robusta green coffee beans can be successfully carried out using IR- spectroscopy and LC-
MS. Surprisingly both CD- and NMR spectroscopy fail to achieve in this case, an adequate
level of distinction. This is to our knowledge the first study that directly compares the value
of various spectroscopic techniques if multivariant statistical techniques are employed to
them.
22
References
1. Czok, G. Coffee and health. Z. Ernaehrungswiss. 1977, 16, 248-255.
2. Higdon, J.V.; Frei, B. Coffee and health: a review of recent human research.
Crit. Rev. Food Sci. Nutr. 2006, 46, 101-123.
3. Dorea, J.G.; da Costa, T.H.M. Is coffee a functional food?. Br. J. Nutr. 2005, 93, 773-782.
4. Hamer, M. Coffee and health: explaining conflicting results in hypertension. J. Hum.
Hypertens. 2006, 20, 909-912.
5. van Dam, R.M. Coffee consumption and risk of type 2 diabetes, cardiovascular diseases,
and cancer. Appl Physiol Nutr Metab 2008, 33, 1269-1283.
6. Tavani, A.; La, V.C. Coffee and cancer: a review of epidemiological studies, 1990-1999.
Eur J Cancer Prev 2000, 9, 241-256.
7. Crozier, A.; Jaganath, I.B.; Clifford, M.N. Dietary phenolics: chemistry, bioavailability and
effects on health. Nat. Prod. Rep. 2009, 26, 1001-1043.
8. Frank, O.; Blumberg, S.; Kunert, C.; Zehentbauer, G.; Hofmann, T. Structure
Determination and Sensory Analysis of Bitter-Tasting 4-Vinylcatechol Oligomers and Their
Identification in Roasted Coffee by Means of LC-MS/MS. J. Agric. Food Chem. 2007, 55,
1945-1954.
9. Jaiswal, R.; Matei, M.F.; Golon, A.; Witt, M.; Kuhnert, N. Understanding the fate of
chlorogenic acids in coffee roasting using mass spectrometry based targeted and non-targeted
analytical strategies. Food Funct. 2012, 3, 976-984.
10. Leviton, A.; Pagano, M.; Allred, E.N.; el Lozy, M. Why those who drink the most coffee
appear to be at increased risk of disease. A modest proposal. Ecol. Food Nutr. 1994, 31, 285-
293.
11. Stensvold, M.I.; Jacobsen, B.K. Coffee and cancer: a prospective study of 43,000
Norwegian men and women. Cancer Causes & Control 1994, 5, 401-408.
12. Ohnishi, M.; Morishita, H.; Toda, S.; Yase, Y.; Kido, R. Inhibition in vitro linoleic acid
peroxidation and haemolysis by caffeoyltryptophan. Phytochemistry 1998, 47, 1215-1218.
13. Devasagayam, T.; Kamat, J.; Mohan, H.; Kesavan, P. Caffeine as an antioxidant:
inhibition of lipid peroxidation induced by reactive oxygen species. Biochimica et Biophysica
Acta (BBA)-Biomembranes 1996, 1282, 63-70.
14. Smith, A. Effects of caffeine on human behavior. Food and chemical toxicology 2002, 40,
1243-1255.
15. Kawachi, I.; Willett, W.C.; Colditz, G.A.; Stampfer, M.J.; Speizer, F.E. A prospective
study of coffee drinking and suicide in women. Arch. Intern. Med. 1996, 156, 521.
23
16. Griffiths, R.R.; Evans, S.M.; Heishman, S.J.; Preston, K.L.; Sannerud, C.; Wolf, B.;
Woodson, P. Low-dose caffeine discrimination in humans. J. Pharmacol. Exp. Ther. 1990,
252, 970-978.
17. Furlong, F. Possible psychiatric significance of excessive coffee consumption. The
Canadian Psychiatric Association Journal/La Revue de l'Association des psychiatres du
Canada 1975
18. Stephenson, P.E. Physiologic and psychotropic effects of caffeine on man. A review. J.
Am. Diet. Assoc. 1977, 71, 240-247.
19. Giovannucci, E. Meta-analysis of coffee consumption and risk of colorectal cancer. Am. J.
Epidemiol. 1998, 147, 1043-1052.
20. Klatsky, A.L.; Armstrong, M.A.; Friedman, G.D. Coffee, tea, and mortality. Ann.
Epidemiol. 1993, 3, 375-381.
21. Armstrong, M.A. Alcohol, Smoking, Cofee, and Cirrhosis. Am. J. Epidemiol. 1992, 136,
1248-1257.
22. Corrao, G.; Lepore, A.; Torchio, P.; Valenti, M.; Galatola, G.; D'Amicis, A.; Arico, S.; Di
Orio, F. The effect of drinking coffee and smoking cigarettes on the risk of cirrhosis
associated with alcohol consumption. Eur. J. Epidemiol. 1994, 10, 657-664.e
23. Stelzer, K.J.; Koh, W.; Kurtz, H.; Greer, B.E.; Griffin, T.W. Caffeine consumption is
associated with decreased severe late toxicity after radiation to the pelvis. International
Journal of Radiation Oncology* Biology* Physics 1994, 30, 411-417.
24. Johnston, K.L.; Clifford, M.N.; Morgan, L.M. Coffee acutely modifies gastrointestinal
hormone secretion and glucose tolerance in humans: Glycemic effects of chlorogenic acid and
caffeine. Am. J. Clin. Nutr. 2003, 78, 728-733.
25. Trugo, L.C.; Macrae, R. The use of the mass detector for sugar analysis of coffee
products. Colloq. Sci. Int. Cafe, [C. R. ] 1985, 11th, 245-251.
26. Clifford, M. Chemical and physical aspects of green coffee and coffee products, In Coffee,
Anonymous ; Springer: 1985; pp. 305-374.
27. Silwar, R.; Luellmann, C. The determination of mono- and disaccharides in green Arabica
and Robusta coffees using high-performance liquid chromatography. Cafe, Cacao, The 1988,
32, 319-322.
28. Hughes, W.; Thorpe, T. Determination of Organic Acids and Sucrose in Roasted Coffee
by Capiillary Gas Chromatography. J. Food Sci. 1987, 52, 1078-1083.
29. Noyes, R.; Chu, C. Material balance on free sugars in the production of instant coffee.
1993
30. Maga, J.A. Thermal decomposition of carbohydrates. An overview. ACS Symp. Ser.
1989, 409, 32-39.
24
31. Bradbury, A.G.W.; Halliday, D.J. Chemical structures of green coffee bean
polysaccharides. J. Agric. Food Chem. 1990, 38, 389-392.
32. Nunes, F.M.; Coimbra, M.A.; Duarte, A.C.; Delgadillo, I. Foamability, Foam Stability,
and Chemical Composition of Espresso Coffee As Affected by the Degree of Roast.
J. Agric. Food Chem. 1997, 45, 3238-3243.
33. Rao, C.V.; Chou, D.; Simi, B.; Ku, H.; Reddy, B.S. Prevention of colonic aberrant crypt
foci and modulation of large bowel microbial activity by dietary coffee fiber, inulin and
pectin. Carcinogenesis 1998, 19, 1815-1819.
34. Wilson, A.; Petracco, M.; Illy, E. Some preliminary investigations of oil biosynthesis in
the coffee fruit and its subsequent re-distribution within green and roasted beans. 1997
35. Folstar, P. The composition of wax and oil in green coffee beans.
Versl. Landbouwkd. Onderz. 1976, 854, 65.
36. Maier, H.G. Kaffee. Parey: 1981; Vol. 18
37. Speer, K.; Sehat, N.; Montag, A. Fatty acids in coffee. 1993
38. Macrae, R. Nitrogenous components, In Coffee, Anonymous ; Springer: 1985; pp. 115-
152.
39. Steinhart, H.; Luger, A. Amino acid pattern of steam treated coffee.
Colloq. Sci. Int. Cafe, [C. R. ] 1995, 16th, 278-285.
40. Morishita, H.; Takai, Y.; Yamada, H.; Fukuda, F.; Sawada, M.; Iwahashi, H.; Kido, R.
Caffeoyltryptophan from green robusta coffee beans. Phytochemistry 1987, 26, 1195-1196.
41. Murata, M.; Okada, H.; Homma, S. Hydroxycinnamic acid derivatives and p-coumaroyl-
(L)-tryprophan, a novel hydroxycinnamic acid derivative, from coffee beans. Bioscience
Biotechnology and Biochemistry 1995, 59, 1887-1895.
42. Balyaya, K.; Clifford, M. Individual chlorogenic acids and caffeine contents in
commercial grades of wet and dry processed Indian green robusta coffee. JOURNAL OF
FOOD SCIENCE AND TECHNOLOGY-MYSORE- 1995, 32, 104-108.
43. Clifford, M. What factors determine the intensity of coffee’s sensory attributes. Tea
Coffee Trade J 1987, 159, 35-39.
44. Maier, H. The acids of coffee. Proceedings 12th Asic College 1987, 229-237.
45. Mabrouk, A.F.; Deatherage, F.E. Organic acids in brewed coffee. Food
Technol. (Chicago, IL, U. S. ) 1956, 10, 194-197.
46. Lentner, C.; Deatherage, F.E. Organic acids in coffee in relation to the degree of roast.
Food Res. 1959, 24, 483-492.
47. Nakabayashi, T. Chemical studies on the quality of coffee. VI. Changes in organic acids
and pH of roasted coffee. Nippon Shokuhin Kogyo Gakkaishi 1978, 25, 142-146.
25
48. Engelhardt, U.H.; Maier, H.G. Acids in coffee. XI. The proportion of individual acids in
the total titratable acid. Z Lebensm Unters Forsch 1985, 181, 20-23.
49. Woodman, J.S.; Giddey, A.; Egli, R.H. Carboxylic acids of brewed coffee.
Int. Colloq. Chem. Coffee, 3rd 1968, 137-143.
50. Hughes, W.J.; Thorpe, T.M. Determination of organic acids and sucrose in roasted coffee
by capillary gas chromatography. J. Food Sci. 1987, 52, 1078-1083.
51. Engelhardt, U.H. Nichtflüchtige Säuren im Kaffee. Technische Universität Carolo-
Wilhelmina zu Braunschweig: 1984
52. Clifford, M. Chemical and physical aspects of green coffee and coffee products, In Coffee,
Anonymous ; Springer: 1985; pp. 305-374.
53. Scholze, A. Quantitative Bestimmung von Säuren im Kaffee durch Kapillar-
Isotachophorese. Technische Universität Carolo-Wilhelmina zu Braunschweig: 1983;
54. Engelhardt, U.H.; Maier, H.G. Säuren des Kaffees. Zeitschrift für Lebensmittel-
Untersuchung und Forschung 1985, 181, 206-209.
55. Balzer, H.H. Chemistry I: Non-Volatile Compounds, In Anonymous ; Blackwell Science
Ltd: 2008; pp. 18-32.
56. Holscher, W.; Vitzthum, O.; Steinhart, H. Identification and sensorial evaluation of
aroma-impact-compounds in roasted Colombian coffee. Café, cacao, thé 1990, 34, 205-212.
57. Wöhrmann, R.; Hojabr-Kalali, B.; Maier, H. Volatile minor acids in coffee. I. Contents of
green and roasted coffee. Dtsch. Lebensm. -Rundsch. 1997, 93, 191-194.
58. Maeaettae, K.R.; Kamal-Eldin, A.; Toerroenen, A.R. High-Performance Liquid
Chromatography (HPLC) Analysis of Phenolic Compounds in Berries with Diode Array and
Electrospray Ionization Mass Spectrometric (MS) Detection: Ribes Species. J. Agric. Food
Chem. 2003, 51, 6736-6744.
59. Kampmann, B.; Maier, H.G. Säuren des Kaffees. Zeitschrift für Lebensmittel-
Untersuchung und Forschung 1982, 175, 333-336.
60. Hucke, J.; Maier, H.G. Quinic acid lactone in coffee. Z Lebensm Unters Forsch 1985,
180, 479-484.Journal Article
61. Galli, V.; Barbas, C. Capillary electrophoresis for the analysis of short-chain organic acids
in coffee. Journal of Chromatography A 2004, 1032, 299-304.Journal Article
62. Ginz, M.; Balzer, H.H.; Bradbury, A.G.; Maier, H.G. Formation of aliphatic acids by
carbohydrate degradation during roasting of coffee. European Food Research and
Technology 2000, 211, 404-410.
63. Balzer, H.H. Chemistry I: Non-volatile compounds. Acids in coffee. Coffee 2001, 18-32.
26
64. Clifford, M.N. Chlorogenic acids and other cinnamates - nature, occurrence and dietary
burden. J. Sci. Food Agric. 1999, 79, 362-372.
65. Clifford, M.N. Chlorogenic acids and other cinnamates - nature, occurrence, dietary
burden, absorption and metabolism. J. Sci. Food Agric. 2000, 80, 1033-1043.
66. Anonymous IUPAC Commission on the Nomenclature of Organic Chemistry (CNOC)
and IUPAC-IUB Commission on Biochemical Nomenclature (CBN). Nomenclature of
cyclitols. Recommendations, 1973. Biochem J 1976, 153, 23-31.
67. Payen, S. Untersuchung des Kaffees. Annalen 1846, 60, 286-294.
68. Fischer, H.O.L.; Dangschat, G. Konstitution der Chlorogensäure (3. Mitteil. über
Chinasäure Derivate). Berichte der deutschen chemischen Gesellschaft (A and B Series) 1932,
65, 1037-1040.
69. Jaiswal, R.; Patras, M.A.; Eravuchira, P.J.; Kuhnert, N. Profile and Characterization of the
Chlorogenic Acids in Green Robusta Coffee Beans by LC-MSn: Identification of Seven New
Classes of Compounds. J. Agric. Food Chem. 2010, 58, 8722-8737.
70. Clifford, M.N.; Johnston, K.L.; Knight, S.; Kuhnert, N. Hierarchical Scheme for LC-MSn
Identification of Chlorogenic Acids. J. Agric. Food Chem. 2003, 51, 2900-2911.
71. Clifford, M.N.; Marks, S.; Knight, S.; Kuhnert, N. Characterization by LC-MSn of four
new classes of p-coumaric acid-containing diacyl chlorogenic acids in green coffee beans. J.
Agric. Food Chem. 2006, 54, 4095-4101.
72. Jaiswal, R.; Kuhnert, N. Hierarchical scheme for liquid chromatography/multi-stage
spectrometric identification of 3,4,5-triacyl chlorogenic acids in green Robusta coffee beans.
Rapid Communications in Mass Spectrometry 2010, 24, 2283-2294.
73. Jaiswal, R.; Sovdat, T.; Vivan, F.; Kuhnert, N. Profiling and Characterization by LC-MSn
of the Chlorogenic Acids and Hydroxycinnamoylshikimate Esters in Mate (Ilex
paraguariensis). J. Agric. Food Chem. 2010, 58, 5471-5484.
74. Clifford, M.N.; Knight, S.; Surucu, B.; Kuhnert, N. Characterization by LC-MSn of four
new classes of chlorogenic acids in green coffee beans: Dimethoxycinnamoylquinic acids,
diferuloylquinic acids, caffeoyl-dimethoxycinnamoylquinic acids, and feruloyl-
dimethoxycinnamoylquinic acids. J. Agric. Food Chem. 2006, 54, 1957-1969.
75. Clifford, M.N.; Wu, W.; Kirkpatrick, J.; Jaiswal, R.; Kuhnert, N. Profiling and
characterisation by liquid chromatography/multi-stage mass spectrometry of the chlorogenic
acids in Gardeniae Fructus. Rapid Communications in Mass Spectrometry 2010, 24, 3109-
3120.
76. Clifford, M. Chlorogenic acids, In Coffee, Anonymous ; Springer: 1985; pp. 153-202.
77. Clifford, M.; Walker, R. Chlorogenic acids—confounders of coffee-serum cholesterol
relationships. Food Chem. 1987, 24, 77-80.
27
78. Cartwright, R.; Roberts, E. Theogallin, a polyphenol occurring in tea. J. Sci. Food Agric.
1954, 5, 593-597.
79. Hara, Y.; Luo, S.; Wickremasinghe, R.; Yamanishi, T. Special issue on tea. Food Rev. Int.
1995, 11, 371-542.
80. Clifford, M.N.; Ramirez-Martinez, J.R. Chlorogenic acids and purine alkaloids contents of
Maté (Ilex paraguariensis) leaf and beverage. Food Chem. 1990, 35, 13-21.
81. Spanos, G.A.; Wrolstad, R.E. Phenolics of apple, pear, and white grape juices and their
changes with processing and storage. A review. J. Agric. Food Chem. 1992, 40, 1478-1487.
82. Mosel, H.; Herrmann, K. The phenolics of fruits. Zeitschrift für Lebensmittel-
Untersuchung und Forschung 1974, 154, 6-11.
83. Igile, G.O.; Oleszek, W.; Jurzysta, M.; Burda, S.; Fafunso, M.; Fasanmade, A.A.
Flavonoids from Vernonia amygdalina and their antioxidant activities. J. Agric. Food Chem.
1994, 42, 2445-2448.
84. Wald, B.; Wray, V.; Galensa, R.; Herrmann, K. Malonated flavonol glycosides and 3, 5-
dicaffeoylquinic acid from pears. Phytochemistry 1989, 28, 663-664.
85. Risch, B.; Herrmann, K. Contents of hydroxycinnamic acid derivatives and catechins in
pome and stone fruit. Z. Lebensm. Unters. 1988, 186, 225-230.
86. Schuster, B.; Herrmann, K. Hydroxybenzoic and hydroxycinnamic acid derivatives in soft
fruits. Phytochemistry 1985, 24, 2761-2764.
87. Gao, L.; Mazza, G. Quantitation and distribution of simple and acylated anthocyanins and
other phenolics in blueberries. J. Food Sci. 1994, 59, 1057-1059.
88. Koeppen, B.H.; Herrmann, K. Flavonoid glycosides and hydroxycinnamic acid esters of
blackcurrants (Ribes nigrum). Zeitschrift für Lebensmittel-Untersuchung und Forschung
1977, 164, 263-268.
89. Risch, B.; Herrmann, K. Hydroxyzimtsäure-Verbindungen in Citrus-Früchten. Zeitschrift
für Lebensmittel-Untersuchung und Forschung 1988, 187, 530-534.
90. Winter, M.; Brandl, W.; Herrmann, K. Determination of hydroxycinnamic acid
derivatives in vegetables. Z. Lebensm. -Unters. Forsch. 1987, 184, 11-16.
91. Risch, B.; Herrmann, K.; Wray, V.; Grotjahn, L. 2'-(E)-O-p-coumaroylgalactaric acid and
2'-(E)-O-feruloylgalactaric acid in citrus. Phytochemistry 1987, 26, 509-510.
92. Okamura, S.; Watanabe, M. Determination of phenolic cinnamates in white wine and their
effect on wine quality. Agric. Biol. Chem. 1981, 45, 2063-2070.
93. Fernandez de Simon, B.; Perez-Ilzarbe, J.; Hernandez, T.; Gomez-Cordoves, C.; Estrella,
I. Importance of phenolic compounds for the characterization of fruit juices. J. Agric. Food
Chem. 1992, 40, 1531-1535.
28
94. Hernandez, T.; Auśn, N.; Bartolomé, B.; Bengoechea, L.; Estrella, I.; Gómez-Cordovés,
C. Variations in the phenolic composition of fruit juices with different treatments. Zeitschrift
für Lebensmitteluntersuchung und-Forschung A 1997, 204, 151-155.
95. Brandl, W.; Herrmann, K. Hydroxycinnamic acid esters in cabbage-like vegetables and
garden cress. Z. Lebensm. -Unters. Forsch. 1983, 176, 444-447.
96. Plumb, G.W.; Price, K.R.; Rhodes, M.J.C.; Williamson, G. Antioxidant properties of the
major polyphenolic compounds in broccoli. Free Radical Res. 1997, 27, 429-435.
97. Brandl, W.; Herrmann, K.; Grotjahn, L. Hydroxycinnamoyl esters of malic acid in small
radish (Raphanus sativus L. var. sativus). Z. Naturforsch. , C: Biosci. 1984, 39C, 515-520.
98. Tadera, K.; Mitsuda, H. Isolation and chemical structure of a new fluorescent compound
in spinach leaves. Agr. Biol. Chem. 1971, 35, 1431-1435.
99. Winter, M.; Herrmann, K. Esters and glucosides of hydroxycinnamic acids in vegetables.
J. Agric. Food Chem. 1986, 34, 616-620.
100. Malmberg, A.; Theander, O. Analysis of chlorogenic acid, coumarins and
feruloylputrescine in different parts of potato tubers infected with Phoma.
Swed. J. Agric. Res. 1984, 14, 63-70.
101. Brandl, W.; Herrmann, K. Occurrence of chlorogenic acids in potatoes. Z. Lebensm. -
Unters. Forsch. 1984, 178, 192-194.
102. Shibuya, N. Phenolic acids and their carbohydrate esters in rice endosperm cell walls.
Phytochemistry 1984, 23, 2233-2237.
103. Hemmerle, H.; Burger, H.; Below, P.; Schubert, G.; Rippel, R.; Schindler, P.W.; Paulus,
E.; Herling, A.W. Chlorogenic Acid and Synthetic Chlorogenic Acid Derivatives: Novel
Inhibitors of Hepatic Glucose-6-phosphate Translocase. J. Med. Chem. 1997, 40, 137-145.
104. del Castillo, M.D.; Ames, J.M.; Gordon, M.H. Effect of Roasting on the Antioxidant
Activity of Coffee Brews. J. Agric. Food Chem. 2002, 50, 3698-3703.
105. Luzia, M.R.; Da Paixao, C.C.; Marcilio, R.; Trugo, L.C.; Quinteiro, L.M.C.; De Maria,
C.A.B. Effect of 5-caffeoylquinic acid on soybean oil oxidative stability. Int. J. Food
Sci. Technol. 1997, 32, 15-19.
106. Herling, A.W.; Burger, H.; Schwab, D.; Hemmerle, H.; Below, P.; Schubert, G.
Pharmacodynamic profile of a novel inhibitor of the hepatic glucose-6-phosphatase system.
Am. J. Physiol. 1998, 274, G1087-G1093.
107. Arion, W.J.; Canfield, W.K.; Ramos, F.C.; Schindler, P.W.; Burger, H.; Hemmerle, H.;
Schubert, G.; Below, P.; Herling, A.W. Chlorogenic acid and hydroxynitrobenzaldehyde: new
inhibitors of hepatic glucose 6-phosphatase. Arch. Biochem. Biophys. 1997, 339, 315-322.
108. Stich, H.F.; Rosin, M.P.; Bryson, L. Inhibition of mutagenicity of a model nitrosation
reaction by naturally occurring phenolics, coffee and tea. Mutat. Res. 1982, 95, 119-
128.Journal Article
29
109. Johnston, K.L.; Clifford, M.N.; Morgan, L.M. Coffee acutely modifies gastrointestinal
hormone secretion and glucose tolerance in humans: Glycemic effects of chlorogenic acid and
caffeine. Am. J. Clin. Nutr. 2003, 78, 728-733.
110. Robinson, W.E., Jr.; Cordeiro, M.; Abdel-Malek, S.; Jia, Q.; Cho, S.A.; Reinecke, M.G.;
Mitchell, W.M. Dicaffeoylquinic acid inhibitors of human immunodeficiency virus integrase:
inhibition of the core catalytic domain of human immunodeficiency virus integrase.
Mol. Pharmacol. 1996, 50, 846-855.
111. Mcdougall, B.; King, P.J.; Wu, B.W.; Hostomsky, Z.; Reinecke, M.G.; Robinson,
W.E.,Jr Dicaffeoylquinic and dicaffeoyltartaric acids are selective inhibitors of human
immunodeficiency virus type 1 integrase. Antimicrob. Agents Chemother. 1998, 42, 140-146.
112. Kwon, H.C.; Jung, C.M.; Shin, C.G.; Lee, J.K.; Choi, S.U.; Kim, S.Y.; Lee, K.R. A new
caffeoyl quinic acid from Aster scaber and its inhibitory activity against human
immunodeficiency virus-1 (HIV-1) integrase. Chem. Pharm. Bull. 2000, 48, 1796-1798.
113. Gupta, P.; Akanksha; Siripurapu, K.B.; Ahmad, A.; Palit, G.; Arora, A.; Maurya, R.
Anti-stress constituents of Evolvulus alsinoides: an ayurvedic crude drug.
Chem. Pharm. Bull. 2007, 55, 771-775.
114. Teramachi, F.; Koyano, T.; Kowithayakorn, T.; Hayashi, M.; Komiyama, K.; Ishibashi,
M. Collagenase inhibitory quinic acid esters from Ipomoea pes-caprae. J. Nat. Prod. 2005,
68, 794-796.
115. Chang, C.; Lin, M.; Lee, S.; Liu, K.C.S.C.; Hsu, F.; Lin, J. Differential inhibition of
reverse transcriptase and cellular DNA polymerase-α activities by lignans isolated from
Chinese herbs, Phyllanthus myrtifolius Moon, and tannins from Lonicera japonica Thunb and
Castanopsis hystrix. Antiviral Res. 1995, 27, 367-374.
116. Xiang, T.; Xiong, Q.; Ketut, A.I.; Tezuka, Y.; Nagaoka, T.; Wu, L.; Kadota, S. Studies
on the hepatocyte protective activity and the structure-activity relationships of quinic acid and
caffeic acid derivatives from the flower buds of Lonicera bournei. Planta Med. 2001, 67, 322-
325.
117. de Paulis, T.; Schmidt, D.E.; Bruchey, A.K.; Kirby, M.T.; McDonald, M.P.; Commers,
P.; Lovinger, D.M.; Martin, P.R. Dicinnamoylquinides in roasted coffee inhibit the human
adenosine transporter. Eur. J. Pharmacol. 2002, 442, 215-223.
118. Shearer, J.; Farah, A.; de Paulis, T.; Bracy, D.P.; Pencek, R.R.; Graham, T.E.;
Wasserman, D.H. Quinides of roasted coffee enhance insulin action in conscious rats.
J. Nutr. 2003, 133, 3529-3532.
119. Jaiswal, R. Synthesis and Analysis of the Dietary Relevant Isomers of Chlorogenic
Acids, Their Derivatives and Hydroxycinnamates. 2012. PhD Thesis.
120. Panagiotakos, D.B.; Lionis, C.; Zeimbekis, A.; Makri, K.; Bountziouka, V.; Economou,
M.; Vlachou, I.; Micheli, M.; Tsakountakis, N.; Metallinos, G.; Polychronopoulos, E. Long-
term, moderate coffee consumption is associated with lower prevalence of diabetes mellitus
among elderly non-tea drinkers from the Mediterranean Islands (MEDIS Study). Rev Diabet
Stud 2007, 4, 105-111.
30
121. Daglia, M.; Racchi, M.; Papetti, A.; Lanni, C.; Govoni, S.; Gazzani, G. In Vitro and ex
Vivo Antihydroxyl Radical Activity of Green and Roasted Coffee. J. Agric. Food Chem.
2004, 52, 1700-1704.
122. Tan, E.-.; Tan, C.; Fook-Chong, S.M.C.; Lum, S.Y.; Chai, A.; Chung, H.; Shen, H.;
Zhao, Y.; Teoh, M.L.; Yih, Y.; Pavanni, R.; Chandran, V.R.; Wong, M.C. Dose-dependent
protective effect of coffee, tea, and smoking in Parkinson's disease: a study in ethnic Chinese.
J Neurol Sci 2003, 216, 163-167.
123. Lindsay, J.; Laurin, D.; Verreault, R.; Hebert, R.; Helliwell, B.; Hill, G.B.; McDowell, I.
Risk factors for Alzheimer's disease: a prospective analysis from the Canadian Study of
Health and Aging. Am J Epidemiol 2002, 156, 445-453.
124. Bekedam, E.K.; Roos, E.; Schols, H.A.; Van Boekel, M.A.J.S.; Smit, G. Low molecular
weight melanoidins in coffee brew. J. Agric. Food Chem. 2008, 56, 4060-4067.
125. Bekedam, E.K.; Loots, M.J.; Schols, H.A.; Van Boekel, M.A.J.S.; Smit, G. Roasting
effects on formation mechanisms of coffee brew melanoidins. J. Agric. Food Chem. 2008,
56, 7138-7145.
126. Crozier, A.; Jaganath, I.B.; Clifford, M.N. Dietary phenolics: chemistry, bioavailability
and effects on health. Nat. Prod. Rep. 2009, 26, 1001-1043.
31
CHAPTER 2: LC-MSn identification of CGAs in green and roasted coffee
2.1 Introduction
Mono-caffeoylquinic acids are generally identified in their tandem mass spectra in negative
ion mode, a pseudomolecular ion [M-H] at m/z 353, which as precursor ions in tandem MS
yields fragment ions at m/z 191 or m/z 173, which is characteristic to the quinic acid and
dehydrated quinic acid moiety respectively. If these two ions are observed in MS2, it signifies
the presence of mono-acyl CGAs and if observed in MS3, they are originated by di-acyl
CGAs. Similarly, tri-acyl CGAs generate m/z 191 or m/z 173 in MS4. Retention times and
resulting elution order on a reverse phase column help in the preliminary assessment of the
regio-isomers of CGAs. For example, mono-acyl CGAs normally elute in the following order:
1-CQA (76) > 5-CQA (3)> 3-CQA (1)> 4-CQA (2). Also, the m/z value of the parent ion will
in most cases reveal the chemical nature of the cinnamic acid moiety e.g. m/z 353-
caffeoylquinic acid, m/z 367- feruloylquinic acid, m/z 515-dicaffeoylquinic acid etc.
Furthermore, targeted MSn experiments in the negative ion mode can be performed on the
intact acyl moiety if observed. Quinic acid moiety stabilizes the negative charge hence,
observed in negative ion mode very often. On the other hand, the side chain of the intact ion
of the acyl moiety particularly stabilizes the positive charge in order to be observed in positive
ion mode. In this manner, targeted MSn experiments in positive ion mode also help to
elucidate the structure of the acyl substituent.
Additionally, confirmation of the molecular formulas of the CGAs is achieved by high
resolution mass measurement typically through LC-TOF-MS or FT-ICR-MS techniques. For
publication standards, mass error lower than five ppm is considered to be acceptable.
For the assignment of the regio-chemistry of the CGAs through tandem MS spectra, Clifford
and Kuhnert’s hierarchical schemes can be exclusively employed. Direct comparison of the
experimentally obtained fragment spectra with those already published allows the
unambiguous assignment.
In this work, we have achieved to synthesize four diastereomers of (-)-quinic acid in order to
distinguish quinic acid stereoisomers by tandem MS. This will enable us in future to
discriminate between the hydroxycinnamic esters, which are bound to the diastereomers of
quinic acid and do show their existence in roasted coffee. The detailed results are discussed in
Chapter 4.
32
2.2 LC-MSn identification of CGAs in green coffee
Presently, around 48 different chlorogenic acids in green coffee have been identified, which
are characterized to their regio-isomeric level based on their fragmentation behavior in
tandem MS and retention times in LC (Table 2.1). 1-5 Unlike in Artichoke, no mono- or di-
acylated CGA shows the presence of acyl group on C1 of quinic acid in the case of green
coffee beans. A representative total ion chromatogram in negative mode of the methanolic
extract of the green coffee bean is shown in the Figure 2.1. With the peaks assigned to the
CGAs identified.
Figure 2.1 TIC of green Robusta coffee extract in negative ion mode 6 (peak numbers
corresponds to CGAs listed in Table 2.1)
Table 2.1 Chlorogenic acids identified in green coffee beans
No. Compound Abbreviation R3 R4 R5
1 3-O-caffeoylquinic acid 3-CQA C H H
2 4-O-caffeoylquinic acid 4-CQA H C H
3 5-O-caffeoylquinic acid 5-CQA H H C
4 3-O-feruloylquinic acid 3-FQA F H
5 4-O-feruloylquinic acid 4-FQA H F H
6 5-O-feruloylquinic acid 4-FQA H H F
7 3-O-p-coumaroylquinic acid 3-pCoQA pCo H H
TIC -All MS
0
100
[%]
10 20 30 40 50 Time [min]
1
4
3
2
6
9
5
41
14
13
15
26
25
22
18
36
33
8 4-O-p-coumaroylquinic acid 4-pCoQA H pCo H
9 5-O-p-coumaroylquinic acid 5-pCoQA H H pCo
10 3-O-dimethoxycinnamoylquinic acid 3-DQA D H H
11 4-O-dimethoxycinnamoylquinic acid 4-DQA H D H
12 5-O-dimethoxycinnamoylquinic acid 5-DQA H H D
13 3,4-di-O-caffeoylquinic acid 3,4-diCQA C C H
14 3,5-di-O-caffeoylquinic acid 3,5-diCQA C H C
15 4,5-di-O-caffeoylquinic acid 4,5-diCQA H C C
16 3,4-di-O-feruloylquinic acid 3,4-diFQA F F H
17 3,5-di-O-feruloylquinic acid 3,5-diFQA F H F
18 4,5-di-O-feruloylquinic acid 4,5-diFQA H F F
19 3,4-di-O-p-coumaroylquinic acid 3,4-dipCoQA pCo pCo H
20 3,5-di-O-p-coumaroylquinic acid 3,5-dipCoQA pCo H pCo
21 4,5-di-O-p-coumaroylquinic acid 4,5-dipCoQA H pCo pCo
22 3-O-feruloyl-4-O-caffeoylquinic acid 3F-4CQA F C H
23 3-O-caffeoyl-4-O-feruloylquinic acid 3C-4FQA C F H
24 3-O-feruloyl-5-O-caffeoylquinic acid 3F-5CQA F H C
25 3-O-caffeoyl-5-O-feruloylquinic acid 3C-5FQA C H F
26 4-O-feruloyl-5-O-caffeoylquinic acid 4F-5CQA H F C
27 4-O-caffeoyl-5-O-feruloylquinic acid 4C-5FQA H C F
28 3-O-dimethoxycinnamoyl-4-O-caffeoylquinic acid 3D-4CQA D C H
29 3-O-dimethoxycinnamoyl-5-O-caffeoylquinic acid 3D-5CQA D H C
30 4-O-dimethoxycinnamoyl-5-O-caffeoylquinic acid 4D-5CQA H D C
31 3-O-caffeoyl-4-O-dimethoxycinnamoylquinic acid 3C-4DQA C D H
32 3-O-caffeoyl-5-O-dimethoxycinnamoylquinic acid 3C-5DQA C H D
33 4-O-caffeoyl-5-O-dimethoxycinnamoylquinic acid 4C-5DQA H C D
34 3-O-dimethoxycinnamoyl-4-O-feruloylquinic acid 3D-4FQA D F H
35 3-O-dimethoxycinnamoyl-5-O-feruloylquinic acid 3D-5FQA D F H
36 4-O-dimethoxycinnamoyl-5-O-feruloylquinic acid 4D-5FQA H D F
37 3-O-p-coumaroyl-4-O-caffeoylquinic acid 3pCo-4CQA pCo C H
34
38 3-O-caffeoyl-4-O-p-coumaroylquinic acid 3C-4pCoQA C pCo H
39 3-O-p-coumaroyl-5-O-caffeoylquinic acid 3pCo-5CQA pCo H C
40 3-O-caffeoyl-5-O-p-coumaroylquinic acid 3C-5pCoQA C H pCo
41 4-O-caffeoyl-5-O-p-coumaroylquinic acid 4C-5pCoQA H C pCo
42 4-O-p-coumaroyl-5-O-caffeoylquinic acid 4pCo-5CQA H pCo C
43 3-O-p-coumaroyl-4-O-feruloylquinic acid 3pCo-4FQA pCo F H
44 3-O-p-coumaroyl-5-O-feruloylquinic acid 3pCo-5FQA pCo H F
45 4-O-p-coumaroyl-5-O-feruloylquinic acid 4pCo-5FQA H pCo F
46 4-O-dimethoxycinnamoyl-5-O-p-coumaroylquinic acid 4D-5pCoQA H D pCo
47 3-O-p-coumaroyl-4-O-dimethoxycinnamoylquinic acid 3pCo-4DQA pCo D H
48 3-O-p-coumaroyl-5-O-dimethoxycinnamoylquinic acid 3pCo-5DQA pCo H D
C = caffeoyl; F = feruloyl; D = dimethoxycinnamoyl; H = hydrogen; pCo = p-coumaroyl.
Mono-acyl CGAs being more polar, elute earlier than the di-acylated CGAs in the reverse
phase packings. 1-53, 4, 7, 8 Generally, water, acetic acid, methanol, and acetonitrile have been
used as mobile phases for HPLC with reverse phase stationary phases like C18, C8,
phenylhexyl, and diphenyl etc. 1-4, 7-12 Clifford et al. developed the hierarchical scheme for the
identification of the CGAs considering relative hydrophobicity, fragmentation patterns and
retention times. 1-4, 8 In this work, we have applied the same strategies to identify mono- and
di-acylated CGAs especially, when we studied the acyl migration in CGAs during roasting,
brewing and by base treatment (Chapter 3). Figure 2.2 Shows the postulated fragment
structures in negative ion mode generated by different chlorogenic acids.
The following guidelines could be useful to identify mono-acylated CGAs:
1. 3-acylated mono-acyl CGAs such as, 3-FQA (4), 3-pCoQA (7), 3-DQA (10) and 3-
sinapoylquinic acid (49) (3-SiQA) base peak ions in MS2 and MS3 are generated from
cinnamic acid moiety (A4, A1, A2 and A3 respectively). 1, 6 Whereas, remaining mono-
acyl CGAs such as, 3-CQA generates base peak ions in MS2 and MS3, which are
derivatives of quinic acid moiety. Figure 2.3 shows representative MSn spectra of 3-
acyl CGAs.
2. 4-acylated CGAs can be readily identified by the dehydrated base peak in MS2 at m/z
173, followed by MS3 base peak at m/z 93 (Q6) and Q7 at m/z 111. 1-5 Figure 2.4
shows representative MSn spectra of 4-acyl CGAs.
35
3. 3-CQA (1) and 5-CQA (3) generate same base peak ion in MS2 at m/z 191. However,
intense secondary ion A1 at m/z 179 in the MS2 of 3-CQA allows the distinction
between the two (Figure 2.5).
OH
HOOC
O
OH
OH OH
HOOC
O
OH OH
HOOC
O
O OH
HOOC
O
OC
HC CH
O
OH
R1
R2
O
OH OH
O
OH
O
O
HCCH
CO O
OH
R1R2
HCCH
OH
R1R2
HCCH
OH
OR2
O
Q1 Q2 Q3 Q4
Q5 Q6 Q7 Q8
A B C
Fragment R1 R2 Cinnamic acid Accurate mass
Q1 191.06
Q2 173.04
Q3 172.04
Q4 OH H Caffeic 335.08
OCH3 H Ferulic 349.08
H H p-Coumaric 319.08
OCH3 OCH3 Sinapic 379.10
36
Q5 85.03
Q6 93.03
Q7 111.04
Q8 127.04
A1 OH H Caffeic 179.04
A2 OCH3 H Ferulic 193.04
A3 H H p-Coumaric 163.04
A4 OCH3 OCH3 Sinapic 223.06
B1 OH H Caffeic 135.04
B2 OCH3 H Ferulic 149.04
B3 H H p-Coumaric 119.04
C1 Ferulic 134.04
C2 Sinapic 164.04
Figure 2.2 Structures of the fragments generated by quinic and cinnamic acid derivatives 15
Figure 2.3 (Continued)
118.8 190.7
162.6
118.8
0
100
[%]
0
100
50 100 150 200 250 300 350 m/z
MS3
MS2 3-pCoQA (7)
37
Figure 2.3 MS2 and MS3 spectra of 3-acyl chlorogenic acids in negative ion mode
134.0
192.8
148.9
133.8
0
100
[%]
0
100
50 100 150 200 250 300 350 m/z
3-FQA (4)
MS3
MS2
135.0
190.8
85.3
126.9 172.8
0
100
[%]
0
100
50 100 150 200 250 300 350 m/z
178.9
3-CQA (1)
MS3
MS2
164.0
222.9
148.9 178.9
163.9
0
100
0
100
100 150 200 250 300 350 400 m/z
[%] MS2
MS3
3-SiQA (49)
38
Figure 2.4 MS2 and MS3 spectra of 4-acyl chlorogenic acids in negative ion mode
(Continued)
0
172.9
110.9 154.8
93.1
0
100
[%]
0
100
100 150 200 250 300 350 m/z
191.0
178.9
MS2
MS3
4-CQA (2)
172.7
71.2
110.8
136.6
154.7
93.0
0
100
[%]
0
100
50 100 150 200 250 300 350 m/z
MS3
MS2 4-pCoQA (8)
39
Figure 2.4 MS2 and MS3 spectra of 4-acyl chlorogenic acids in negative ion mode
Figure 2.5 (Continued)
192.8
172.8
71.4
111.0 154.8
93.1
0
100
[%]
0
100
100 150 200 250 300 350 m/z
4-FQA (5)
MS3
MS2
222.9
172.9
71.6 111.1
154.9
93.3
0
100
[%]
0
100
100 150 200 250 300 350 400 450 m/z
MS2
MS3
4-SiQA (50)
190.8
85.2
111.1 172.9
126.9
0
100
[%]
0
100
50 100 150 200 250 300 350 m/z
5-CQA (3)
MS3
MS2
40
Figure 2.5 MS2 and MS3 spectra of 5-acyl chlorogenic acids in negative ion mode
As mentioned earlier, mono-acylated CGAs elute before di-acyl CGAs. Although in green
coffee, 1-acylated CGAs are not identified, they appear first in the chromatogram in both
mono- and di-acylated CGAs. The elution order for di-acylated CGAs is as follows: 1,3
(80)>1,4 (81)>1,5 (82)>3,4 (13)>3,5 (14)>4,5 (15). 6, 13, 14 This elution order was first
established by Clifford and then Jaiswal pointed out that 3,5- elutes first followed by 3,4- and
4,5-di-acylated CGAs. 15
As in the case of mono-acylated CGAs we can lay out some guidelines for identification of
di-acyl CGAs along with their regio-chemistry in negative ion mode. They are as follows:
1. Similar to the mono-acylated CGAs, di-acyl CGAs generate equivalent parent ion in
MS1 as, [di-acyl CGA -H+]-. In MS2, all di-acyl CGAs either produce [di-acyl CGA –
cinnamoyl -H+]- (Figure 2.6) or [diacyl CGA – cinnamoyl – H2O -H+]- (Figure 2.7). 1-
3, 6, 13
2. Vicinal di-caffeoylquinic acids such as, 3,4-diCQA (13) and 4,5-diCQA (15) remain
consistent with the fragmentation behavior of the 4-acylated mono-acyl CGAs
producing Q2 as a base peak in MS3 at m/z 173 followed by strong MS4 ions at m/z 93
(Q6) and Q7 at m/z 111. (Figure 2.6 )
3. The difference in two vicinal di-caffeoylquinic acids arises from the intensity of the
fragment Q4, which is high in the case of the MS2 of 3,4-diCQA(13) and barely
detected in the MS2 of 4,5-diCQA(15). Similarly, the 3,4-isomer produces Q1 in MS3
and Q7 in MS4 with approximately double the intensities if compared to the 4,5-
isomer. 4
222.9
190.9
71.
93.3
109.1
172.
127.0
0
100
[%]
0
100
100 150 200 250 300 350 400 m/z
85.4
MS2
MS3
5-SiQA (51)
41
On the other hand, tandem MS of the 3,5-diCQA (14) remains consistent with 3-CQA (1) and
5-CQA (3), by producing base peak Q1 at m/z 191, supported by strong MS3 ions at m/z 86
(Q5), m/z 127 (Q8), and m/z 172 (Q3). Fragmentation spectra are shown in the Figure 2.6.
Figure 2.6 MS2, MS3, and MS4 spectra of 3,4-diCQA, 3,5-diCQA, and 4,5-diCQA in negative
ion mode (m/z 515) (Continued)
172.9 254.9
353.0
134.9
172.
9
93.1
154.8
0
100
[%]
0
100
0
100
100 150 200 250 300 350 400 450 500 m/z
178.9 190.9
MS4
MS3
MS2 3,4-diCQA (13)
190.9
353.1
135.0
190.8
85.3
170.8
0
100
[%]
0
100
0
100
100 150 200 250 300 350 400 450 500 m/z
111.0
178.9
3,5-diCQA (14)
MS4
MS3
MS2
42
Figure 2.6 MS2, MS3, and MS4 spectra of 3,4-diCQA, 3,5-diCQA, and 4,5-diCQA in negative
ion mode (m/z 515)
Figure 2.7 MS2, MS3, and MS4 spectra of 3,4-diFQA 16 and 3D-4FQA 34 in negative ion
mode (m/z 543 and m/z 557, respectively)
172.9 202.9 255.0 299.0
353.1
135.0
172.9
93.1 154.8
0
100
[%]
0
100
0
100
100 150 200 250 300 350 400 450 500 m/z
191.0 178.9
4,5-diCQA (15)
MS4
MS3
MS2
172.8 472.7
349.1
269.0
192.8
133.9
0
100
[%]
0
100
0
100
100 150 200 250 300 350 400 450 500 550 m/z
148.9
172.8
3,4-diFQA (16)
MS4
MS3
MS2
206.7
261.0
395.0 486.7
574.6
348.9
133.8
172.8 268.8
304.9
0
100
[%]
0
100
100 20
0
300 400 500 600 m/z
3D-4FQA (34)
MS3
MS2
43
Table 2.2 MS2 and MS3 data of monoacyl CGAs in negative ion mode
No. CGA MS1 MS2 MS3
Parent ion Base peak Secondary peak Base peak Secondary peak
m/z m/z int m/z int m/z int m/z m/z int m/z int
1 3-CQA 353.1 190.9 178.5 50 134.9 7 85.3 127.0 71 172.9 67
2 4-CQA 353.1 172.9 178.9 60 190.8 20 135.0 9 93.2 111.0 48
3 5-CQA 353.2 190.0 178.5 5 135.0 15 85.2 126.9 66 172.9 27
4 3-FQA 367.2 192.9 191.5 2 173.2 2 133.9 148.9 23
5 4-FQA 367.2 172.9 192.9 16 93.1 111.5 44
6 5-FQA 367.2 190.9 172.9 2 85.2 126.9 70
7 3-pCoQA 337.1 162.9 190.0 5 118.9
8 4-pCoQA 337.1 172.7 93.0 111.0 61
9 5-pCoQA 337.2 190.9 162.9 5 85.2
44
Table 2.3 MS2, MS3, and MS4 data of diacyl CGAs in negative ion mode (n.d. = not detected)
No. CGA MS1 MS2 MS3
Parent ion Base peak Secondary peak Base peak Secondary peak
m/z m/z int m/z int m/z int m/z m/z int m/z int m/z int
13 3,4-diCQA 515.2 353.1 335.1 4 172.9 20 172.9 178.9 68 191.0 32 135.1 9
15 4,5-diCQA 515.2 353.1 335.1 2 172.9 6 172.9 178.9 76 190.9 9 135.0 19
18 4,5-diFQA 543.2 367.1 349.1 35 172.9 178.9 60 190.8 20 135.0 9
24 3F-5CQA 529.2 367.1 353.1 60 349.0 32 335.0 32 192.7 172.6 36 133.8 36
26 4F-5CQA 529.2 367.1 335.0 4 172.7 21 172.9 192.9 71 133.8 8
27 4C-5FQA 529.1 353.1 367.1 25 172.9 178.9 49 190.8 35 134.7 10
40 3C-5pCoQA 499.0 353.1 337.0 15 190.7
41 4C-5pCoQA 499.3 353.0 172.8 15 172.9 178.7 66 190.6 29 134.8 12
42 4pCo-5CQA 499.1 337.1 335.1 3 172.7 59 172.9 162.6 8
52 1,3-diCQA 515.2 353.1 335.1 2 173.0 4 190.9 179.0 60 135.1 6
53 1,4-diFQA 543.2 349.1 367.1 25 172.9 17 268.8 9 192.9 172.9 31 268.8 9 133.8 22
54 1C-3FQA 529.1 367.1 353.1 15 192.7 172.6 13 178.6 5 133.8 18
45
No. CGA MS1 MS4
Parent ion Base peak Secondary peak
m/z m/z int m/z int
13 3,4-diCQA 515.2 93.2 111.1 30
15 4,5-diCQA 515.2 93.1 111.0 20
18 4,5-diFQA 543.2 93.1 111.1 40
24 3F-5CQA 529.2 133.7 149.0 19
26 4F-5CQA 529.2 93.2
27 4C-5FQA 529.1 93.2 127.0 n.d.
40 3C-5pCoQA 499.0 85.2 93.0 70 126.9 99
41 4C-5pCoQA 499.3 93.2
42 4pCo-5CQA 499.1 93.2 111.1 98
52 1,3-diCQA 515.2 85.1 111.1 86 172.9 60
53 1,4-diFQA 543.2
54 1C-3FQA 529.1 133.7 149.0 16 127.0 6
2.3 LC-MSn identification of CGAs in roasted coffee
As mentioned earlier, roasting decreases the CGAs content as well as it transforms CGAs to
their isomers by epimerization or by acyl group migration and degrades them to
corresponding cinnamic acid derivatives and quinic acid. Additionally, roasting gives rise to
the dehydrated CGA derivatives in the form of lactones or shikimates. 16-21 C1- acylated
mono- or di-acyl CGAs are not observed in green coffee since; they are essentially the
products of roasting. The workgroup of Kuhnert has been able to develop LC-MSn method for
the identification and distinction between cinnamoylshikimates and CGA lactones (CGLs) to
their regio-isomeric level. 14 In present work, ten new chlorogenic acids have been identified
in roasted coffee, in which organic acids were found conjugated with quinic acid. Detailed
results are presented in Chapter 4. Table 2.4 and Table 2.5 summarize CGAs, CGLs and
CSAs identified in the roasted coffee. CGLs and CSAs are characterized by their parent
pseudo molecular ion at m/z 335 followed by characteristic fragment spectra (Table 2.5).14, 22
46
Table 2.4 Chlorogenic acids identified in roasted coffee 15
No. Compound Abbreviation R1 R3 R4 R5
1 3-O-caffeoylquinic acid 3-CQA H C H H
2 4-O-caffeoylquinic acid 4-CQA H H C H
3 5-O-caffeoylquinic acid 5-CQA H H H C
4 3-O-feruloylquinic acid 3-FQA H F H H
5 4-O-feruloylquinic acid 4-FQA H H F H
6 5-O-feruloylquinic acid 5-FQA H H H F
7 3-O-p-coumaroylquinic acid 3-pCoQA H pCo H H
8 4-O-p-coumaroylquinic acid 4-pCoQA H H pCo H
9 5-O-p-coumaroylquinic acid 5-pCoQA H H H pCo
10 3-O-dimethoxycinnamoylquinic acid 3-DQA H D H H
11 4-O-dimethoxycinnamoylquinic acid 4-DQA H H D H
12 5-O-dimethoxycinnamoylquinic acid 5-DQA H H H D
13 3,4-di-O-caffeoylquinic acid 3,4-diCQA H C C H
14 3,5-di-O-caffeoylquinic acid 3,5-diCQA H C H C
15 4,5-di-O-caffeoylquinic acid 4,5-diCQA H H C C
16 3,4-di-O-feruloylquinic acid 3,4-diFQA H F F H
17 3,5-di-O-feruloylquinic acid 3,5-diFQA H F H F
18 4,5-di-O-feruloylquinic acid 4,5-diFQA H H F F
22 3-O-feruloyl-4-O-caffeoylquinic acid 3F-4CQA H F C H
23 3-O-caffeoyl-4-O-feruloylquinic acid 3C-4FQA H C F H
25 3-O-caffeoyl-5-O-feruloylquinic acid 3C-5FQA H C H F
26 4-O-feruloyl-5-O-caffeoylquinic acid 4F-5CQA H H F C
28 3-O-dimethoxycinnamoyl-4-O-caffeoylquinic acid 3D-4CQA H D C H
29 3-O-dimethoxycinnamoyl-5-O-caffeoylquinic acid 3D-5CQA H D H C
30 4-O-dimethoxycinnamoyl-5-O-caffeoylquinic acid 4D-5CQA H H D C
33 4-O-caffeoyl-5-O-dimethoxycinnamoylquinic acid 4C-5DQA H H C D
36 4-O-dimethoxycinnamoyl-5-O-feruloylquinic acid 4D-5FQA H H D F
47
37 3-O-p-coumaroyl-4-O-caffeoylquinic acid 3pCo-4CQA H pCo C H
38 3-O-caffeoyl-4-O-p-coumaroylquinic acid 3C-4pCoQA H C pCo H
39 3-O-p-coumaroyl-5-O-caffeoylquinic acid 3pCo-5CQA H pCo H C
42 4-O-p-coumaroyl-5-O-caffeoylquinic acid 4pCo-5CQA H H pCo C
43 3-O-p-coumaroyl-4-O-feruloylquinic acid 3pCo-4FQA H pCo F H
45 4-O-p-coumaroyl-5-O-feruloylquinic acid 4pCo-FCQA H H pCo F
56 3-O-sinapoyl-4-O-caffeoylquinic acid 3Si-4CQA H Si C H
57 3-O-caffeoyl-4-O-sinapoylquinic acid 3C-4SiQA H C Si H
60 3-O-feruloyl-4-O-sinapoylquinic acid 3F-4SiQA H F Si H
64 3-O-trimethoxycinnamoyl-5-O-feruloylquinic acid 3T-5FQA H T H F
76 1-O-caffeoylquinic acid 1-CQA C H H H
77 1-O-feruloylquinic acid 1-FQA F H H H
78 1-O-p-coumaroylquinic acid 1-pCoQA pCo H H H
79 1-O-dimethoxycinnamoylquinic acid 1-DQA D H H H
80 1,3-di-O-caffeoylquinic acid 1,3-diCQA C C H H
81 1,4-di-O-caffeoylquinic acid 1,4-diCQA C H C H
82 1,5-di-O-caffeoylquinic acid 1,5-diCQA C H H C
83 1-O-caffeoyl-3-O-feruloylquinic acid 1C-3FQA C F H H
84 1-O-caffeoyl-4-O-feruloylquinic acid 1C-4FQA C H F H
85 1-O-caffeoyl-4-O-dimethoxycinnamoylquinic acid 1C-4DQA C H D H
86 4-O-feruloyl-5-O-dimethoxycinnamoylquinic acid 4F-5DQA H H F D
87 1-O-caffeoyl-3-O-sinapoylquinic acid 1C-3SiQA C Si H H
88 1-O-feruloyl-4-O-sinapoylquinic acid 1F-4SiQA F H Si H
89 1-O-feruloyl-3-O-sinapoylquinic acid 1F-3SiQA F Si H H
90 1-O-caffeoyl-3-O-trimethoxycinnamoylquinic acid 1C-3TQA C T H H
C = caffeoyl; D = dimethoxycinnamoyl; F = feruloyl; pCo = p-coumaroyl; H = hydrogen; T = trimethoxycinnamoyl.
48
Table 2.5 Negative ion mode MS2, MS3 and MS4 fragmentation data for the cinnamoylshikimate esters and chlorogenic acid lactones 22
No. Compd. MS1 MS2 MS3
Parent ion Base peak Secondary peak Base peak Secondary peak
m/z m/z int m/z int m/z int m/z m/z int m/z int m/z int
91 3-CSA 335.1 178.9 178.5 50 134.9 7 85.3 127.0 71 172.9 67
92 4-CSA 335.1 178.9 178.9 60 190.8 20 135.0 9 93.2 111.0 48
93 5-CSA 335.1 178.9 178.5 5 135.0 15 85.2 126.9 66 172.9 27
94 1-CQL 335.1 160.8 172.8 67 132.8 14 132.8
95 3-CQL 335.1 160.8 134.8 82 132.8
96 4-CQL 335.1 160.8 134.8 17 132.9
97 3-FSA 349.1 192.9 155.0 10 148.9 177.9 73 134.0 78
98 4-FSA 349.1 192.9 174.9 24 154.9 24 137.0 13 148.9 177.9 63 134.0 75
99 5-FSA 349.0 192.9 155.0 27 148.9 177.9 81 134.0 71
100 1-FQL 349.0 172.7 192.7 56 175.0 88 159.7 19 93.1 159.7 22 110.9 32
101 3-FQL 349.0 174.7 192.7 42 148.7 65 133.8 32 159.7
102 4-FQL 349.1 174.7 192.7 41 148.7 13 159.7 20 159.7
103 3-DSA 363.1 206.8 154.8 70 136.8 45 294.7 35 115.0 130.8 93
104 4-DSA 363.1 154.8 206.8 15 136.7 50 136.8 111.0 12 93.0 15
105 5-DSA 363.1 154.8 206.8 50 136.7 50 110.9 16 136.8 111.0 15 93.0 20
106 1-DQL 363.1 206.8 154.8 17 132.8 191.8 19 162.8 66 148.9 20
49
107 3-DQL 363.1 206.8 148.8 190.8 23 162.8 32 134.8 34
108 4-DQL 363.1 206.8 148.8 191.8 52 130.8 67
50
References
1. Clifford, M.N.; Johnston, K.L.; Knight, S.; Kuhnert, N. Hierarchical Scheme for LC-MSn
Identification of Chlorogenic Acids. J. Agric. Food Chem. 2003, 51, 2900-2911.
2. Clifford, M.N.; Knight, S.; Surucu, B.; Kuhnert, N. Characterization by LC-MSn of four
new classes of chlorogenic acids in green coffee beans: Dimethoxycinnamoylquinic acids,
diferuloylquinic acids, caffeoyl-dimethoxycinnamoylquinic acids, and feruloyl-
dimethoxycinnamoylquinic acids. J. Agric. Food Chem. 2006, 54, 1957-1969.
3. Clifford, M.N.; Marks, S.; Knight, S.; Kuhnert, N. Characterization by LC-MSn of four
new classes of p-coumaric acid-containing diacyl chlorogenic acids in green coffee beans. J.
Agric. Food Chem. 2006, 54, 4095-4101.
4. Clifford, M.N.; Knight, S.; Kuhnert, N. Discriminating between the six isomers of
dicaffeoylquinic acid by LC-MSn. J. Agric. Food Chem. 2005, 53, 3821-3832.
5. Alonso-Salces, R.M.; Guillou, C.; Berrueta, L.A. Liquid chromatography coupled with
ultraviolet absorbance detection, electrospray ionization, collision-induced dissociation and
tandem mass spectrometry on a triple quadrupole for the on-line characterization of
polyphenols and methylxanthines in green coffee beans. Rapid Commun. Mass Spectrom.
2009, 23, 363-383.
6. Jaiswal, R.; Patras, M.A.; Eravuchira, P.J.; Kuhnert, N. Profile and Characterization of the
Chlorogenic Acids in Green Robusta Coffee Beans by LC-MSn: Identification of Seven New
Classes of Compounds. J. Agric. Food Chem. 2010, 58, 8722-8737.
7. Clifford, M.N.; Wang, Z.; Kuhnert, N. Profiling the chlorogenic acids of Aster by HPLC-
MSn. Phytochem. Anal. 2006, 17, 384-393.
8. Clifford, M.N.; Wu, W.; Kuhnert, N. The chlorogenic acids of Hemerocallis. Food Chem.
2005, 95, 574-578.
9. Clifford, M.N.; Kirkpatrick, J.; Kuhnert, N.; Roozendaal, H.; Salgado, P.R. LC-MSn
analysis of the cis isomers of chlorogenic acids. Food Chem. 2008, 106, 379-385.
10. Clifford, M.N.; Wang, Z.; Kuhnert, N. Profiling the chlorogenic acids of Aster by HPLC-
MSn. Phytochem. Anal. 2006, 17, 384-393.
11. Clifford, M.N.; Stoupi, S.; Kuhnert, N. Profiling and Characterization by LC-MSn of the
Galloylquinic Acids of Green Tea, Tara Tannin, and Tannic Acid. J. Agric. Food Chem.
2007, 55, 2797-2807.
12. Clifford, M.N.; Wu, W.; Kirkpatrick, J.; Kuhnert, N. Profiling the Chlorogenic Acids and
Other Caffeic Acid Derivatives of Herbal Chrysanthemum by LC-MSn. J. Agric. Food
Chem. 2007, 55, 929-936.
13. Clifford, M.N.; Wu, W.; Kirkpatrick, J.; Jaiswal, R.; Kuhnert, N. Profiling and
characterisation by liquid chromatography/multi-stage mass spectrometry of the chlorogenic
51
acids in Gardeniae Fructus. Rapid Communications in Mass Spectrometry 2010, 24, 3109-
3120.
14. Jaiswal, R.; Sovdat, T.; Vivan, F.; Kuhnert, N. Profiling and Characterization by LC-MSn
of the Chlorogenic Acids and Hydroxycinnamoylshikimate Esters in Mate (Ilex
paraguariensis). J. Agric. Food Chem. 2010, 58, 5471-5484.
15. Jaiswal, R. Synthesis and Analysis of the Dietary Relevant Isomers of Chlorogenic Acids,
Their Derivatives and Hydroxycinnamates. 2012. PhD Thesis.
16. Bennat, C.; Engelhardt, U.H.; Kiehne, A.; Wirries, F.M.; Maier, H.G. HPLC ANALYSIS
OF CHLOROGENIC ACID LACTONES IN ROASTED COFFEE. Zeitschrift Fur
Lebensmittel-Untersuchung Und-Forschung 1994, 199, 17-21.
17. Bicchi, C.P.; Binello, A.E.; Pellegrino, G.M.; Vanni, A.C. Characterization of Green and
Roasted Coffees through the Chlorogenic Acid Fraction by HPLC-UV and Principal
Component Analysis. J. Agric. Food Chem. 1995, 43, 1549-1555.
18. del Castillo, M.D.; Ames, J.M.; Gordon, M.H. Effect of Roasting on the Antioxidant
Activity of Coffee Brews. J. Agric. Food Chem. 2002, 50, 3698-3703.
19. Farah, A.; De Paulis, T.; Trugo, L.C.; Martin, P.R. Effect of roasting on the formation of
chlorogenic acid lactones in coffee. J. Agric. Food Chem. 2005, 53, 1505-1513.
20. Farah, A.; De Paulis, T.; Moreira, D.P.; Trugo, L.C.; Martin, P.R. Chlorogenic acids and
lactones in regular and water-decaffeinated arabica coffees. J. Agric. Food Chem. 2006, 54,
374-381.
21. Hucke, J.; Maier, H.G. Quinic acid lactone in coffee. Z Lebensm Unters Forsch 1985,
180, 479-484.
22. Jaiswal, R.; Matei, M.F.; Ullrich, F.; Kuhnert, N. How to distinguish between
cinnamoylshikimate esters and chlorogenic acid lactones by liquid chromatography-tandem
mass spectrometry. Journal of Mass Spectrometry 2011, 46, 933-942.
52
CHAPTER 3: Acyl migration in mono- and di-caffeoylquinic acids under
basic and aqueous acidic conditions and dry roasting conditions
3.1 Introduction
Coffee is the most valued agricultural commodity in terms of the economic aspects of the
exports from the developing coffee producing countries, accounting to ca. 8 million metric
tonnes per year. Approximately, 2.3 billion cups of coffee are consumed worldwide per day.1
Coffea Arabica (known as Arabica coffee) and Coffea canephora (known as Robusta coffee)
are the two types of coffee holding 70% and 30%, respectively of the total coffee market in
the world.2 Chlorogenic acids are present in the range of 6-12% of the dry weight of the green
coffee bean.3
Chlorogenic acids (CGAs) are a large group of esters formed between one or more cinnamic
acid derivatives and D-(-)-quinic acid. CGAs are classified on the basis of the number of the
cinnamoyl residues esterified with the quinic acid as well as the functional groups present on
the aromatic moiety of the cinnamoyl residues. Out of the total content of the CGAs in green
coffee, 5-O-caffoylquinic acid (3) comprises about 50%. Other subclasses like caffoylquinic
acids, dicaffoylquinic acids, feruloylquinic acids and p-coumaroylquinic acids contribute to a
large extent to the other 50% of the total CGAs present in coffee. CGAs are very important
plant secondary metabolites due to their pharmacological properties, like antioxidant
property,4 anti-hepatitis B virus activity,5 antispasmodic activity,3 anti-diabetic activity,6
inhibition of the HIV-1 integrase 7,8 and inhibition of the mutagenicity of carcinogenic
compounds.3
The various roasting conditions affect the concentration and composition of the CGA content
in the green coffee. For every 1% of the dry matter of the total CGA content in the green
coffee beans, 8-10% of the original CGAs are transformed or decomposed into respective
cinnamic acid derivatives and quinic acid. 1,9 The lightest drinkable roast (the so-called
‘Cinnamon’ roast) involves roasting of green coffee beans at around 180 0C until the coffee
beans just encounter the ‘first crack’. It was reported by Clifford et al. that during the early
stage of the roasting process, transformations such as isomerization (acyl migration) or
hydrolysis of the ester bond, take place in the CGAs. Later, chemical transformations like
decarboxylation in cinnamoyl moieties to produce a number of phenylindans, epimerization at
the quinic acid and lactonization take place. 1,2 Clifford et al. also reported the base hydrolysis
induced intramolecular isomerization and transesterification in 5-O-caffoylquinic acid (3), 3-
53
O-caffoylquinic acid (2), 4-O-caffoylquinic acid (4), 3,4-di-O-caffoylqunic acid (8), 3,5-di-O-
caffoylqunic acid (9), 4,5-di-O-caffoylqunic acid (10), 5-O-p-coumaroylquinic acid (13) and
5-O-feruloylqunic acid (23), in which the identification of some of the transformed products
was based on the putative conclusions acquired by analytical HPLC. 10,11 Dawidowicz et al.
found nine transformation products of 5-O-caffoylquinic acid (3) after five hours of reflux in
an acid-water solution including two water addition products. This study only incorporated 5-
CQA. 12 No comprehensive mechanistic study has been previously reported, which comments
on the intra- versus inter-molecular acyl migration under different conditions incorporating all
major commercially available regio-isomers of mono- and di-caffeoyl chlorogenic acids.
The complexity in the data interpretation for the structural analyses of the CGAs present in
the roasted coffee melanoidines arises from the regio- and stereoisomeric compounds in the
natural sources. For this reason, model roasting experiments on the commercially available
mono- and dicaffeoylquinic acids were attempted in the present work in order to study the
transformations taking place in CGAs during the early roasting stages. Also, isomerization
(acyl migration) was induced by both base hydrolysis and simple hydrolysis (brewing) in
these reference standards to observe the isomeric transformations on the basis of relative
quantification. Recently, tandem mass spectrometry has allowed accurate structural
assignment and identification of the CGA regioisomers. The advantage of multi-dimensional
speciphicity of LC-MSn enabled isomeric resolution and relative quantification of the early
roasting transformations in the CGAs. 13
3.2 Materials and methods
Chemicals and materials
All the chemicals (analytical grade) were purchased from Sigma-Aldrich (Bremen, Germany).
Commercially available mono- and di-caffeoylquinic acids such as 5-O-caffoylquinic acid (3),
3-O-caffoylquinic acid (neo-chlorogenic acid) (2), 4-O-caffoylquinic acid (crypto-chlorogenic
acid) (4), 1,3-di-O-caffoylqunic acid (cynarin) (5), 3,4-di-O-caffoylqunic acid (8), 3,5-di-O-
caffoylqunic acid (9), 4,5-di-O-caffoylqunic acid (10) were purchased from PhytoLab GmbH
& Co. KG, Germany.
54
Hydrolysis by tetramethylammonium hydroxide (TMAH)
All the seven CGAs reference standards were treated with aqueous TMAH (25 g/litre). The
initial concentrations of standards were as shown in Table S1 in supplementary information.
Each sample was diluted by 5 ml of aqueous TMAH and stirred at room temperature. 1 ml
solution from each sample was taken out at 2, 5, 10, 30 and 60 minutes time intervals. Each
sample was saturated with brine and extracted twice with ethyl acetate. Combined organic
layers were concentrated in vacuo and each sample was prepared in 1 ml of methanol to
analyse by LC-MSn for intramolecular acyl migration.
To study the intermolecular acyl migration (Cross-over experiment), 5-CQA (25 mg, 0.07062
mmol) was added to a round bottom flask containing ferulic acid (13.7 mg, 0.07062 mmol). 5
ml of 10 times diluted (25 g/litre) TMAH was added to the flask and the mixture was stirred
at room temperature. 1 ml samples were taken out from the flask at 2, 5, 10, 15 and 30
minutes time intervals. Each sample was saturated with brine and extracted twice with ethyl
acetate. The combined organic layers were concentrated in vacuo and each sample was
prepared in methanol to be analysed by LC-MSn. The same procedure was repeated with p-
coumaric acid (11.57 mg, 0.07062 mmol) and 5-CQA (25 mg, 0.07062 mmol).
Model roasting
All the seven CGAs reference standards were heated at 180 0C for 12 minutes separately to
study the intramolecular acyl migration. Equimolar quantities of 5-CQA with p-coumaric acid
and 5-CQA with ferulic acid were heated together at 180 0C for 12 minutes to study the
intermolecular acyl migration (Cross-over experiment). All the samples were heated in a
Buechi Glass Oven B-585 and prepared in 1 ml methanol for LC-MSn analysis.
Brewing of CGAs (2-5 and 8-10)
Commercially available chlorogenic acids standards (each sample 500 µg) were infused in 3
mL of hot water each and stirred for 5 h under reflux. pH of each sample was determined to
be 5 with pH meter. The solvent was removed under low pressure and the samples were
dissolved in 1ml MeOH and used for LC-MSn.
LC-MSn
The LC equipment (Agillent 1100 series, Bremen, Germany) comprised a binary pump, an
auto sampler with a 100 µL loop, and a DAD detector with a light-pipe flow cell (recording at
55
254 and 320 nm and scanning from 200 to 600 nm). This was interfaced with an ion-trap mass
spectrometer fitted with an ESI source (Bruker Daltonics HCT Ultra, Bremen, Germany)
operating in full scan, auto MSn mode to obtain fragment ion m/z. Tandem mass spectra were
acquired in Auto-MSn mode (smart fragmentation) using a ramping of the collision energy.
Maximum fragmentation amplitude was set to 1 Volt, starting at 30% and ending at 200%.
The MS operating conditions (negative mode) had been optimized using 5-caffeoylquinic acid
with a capillary temperature of 365 oC, a dry gas flow rate of 10 L/min and a nebulizer
pressure of 10 psi.
HPLC
Separation was achieved on a 150 x 3 mm i.d. column containing diphenyl 5 µm, with a 5 mm
x 3 mm i.d. guard column (Varian, Darmstadt, Germany). Alternatively, separation was also
achieved on a 250 mm x 3 mm i.d. column containing C18-amide 5 µm, with a 5 mm x 3 mm
i.d. guard column of the same material (Varian, Darmstadt, Germany) for the cases of
hydrolysis (brewing) of reference standards experiments. Solvent A was water/formic acid
(1000:0.05 v/v) and solvent B was methanol. Solvents were delivered at a total flow rate of
500 µL/min. The gradient profile was from 10 % B to 70 % B linearly in 60 min followed by
10 min isocratic, and a return to 10 % B at 90 min and 10 min isocratic to re-equilibrate.
Preliminary assessment of data
All the data for the chlorogenic acids presented in this paper use the recommended IUPAC
numbering system; 14 the same numbering system was adopted for chlorogenic acids, their cis-
isomers, their acyl-migration isomers and water addition products (Figure 3.1). The relative
concentrations of the transformed products are expressed here in terms of the peak areas
obtained in their UV chromatograms assuming the relative response factor in UV close to one,
based on identical absorptivity of all mono CGA. 13,15 In tables and figures, the peak area
values are stated accordingly.
3.3 Results and discussion
3.3.1 Intramolecular acyl migration: hydrolysis by TMAH of 2-5 and 8-10
The concentration of the samples was chose to be sufficiently low (1-1.5 mg/ ml) to prevent
intermolecular acyl migration or transesterification, therefore allowing observation of intra
molecular acyl migration exclusively. In the hydrolysis of 3-CQA (2), 5-CQA (3) and 4-CQA
(4), all the other mono-acyl derivatives were identified during the hydrolysis except for 1-
56
CQA (1). Also, we did not observe the formation of any di-acyl derivatives in the hydrolysis
of the mono-acylquinic acids. This observation confirms that the acyl migration we observed
in this study was in fact an intramolecular process. Figure 3.2 represents the UV
chromatograms of the 5-CQA in basic solution at different time intervals. It should be noted
that the UV response in mono- and di-acylquinic acids has been used in the past as a reliable
relative response factor. 9 Hanson et al. used radiolabelled quinic acid to investigate the acyl
migration pathway in cinnamoylquinic acids 16; in accordance to his findings, we assumed that
the mechanism of the acyl migration follows the ortho ester intermediate formation. This
assumption is also supported by the study reported by Hanson and Cen Xie et al. 17 The
transformations of 5-CQA (3), 4-CQA (4) and 3-CQA (2) with time are presented in Figure
3.3. From the results, we can conclude that 5-CQA is much more stable than 4-CQA and 3-
CQA and the order of the stability is 5>4>3 in terms of the hydrolysis of the caffeoyl ester.
This stability was observed to provide the resistance to decomposition thus allowing 5-CQA
to form the acyl migrated products over longer hydrolysis durations. On the other hand, while
3-CQA being the least stable of the three mono-acylquinic acids, 3-CQA decomposes to form
caffeic acid and quinic acid even before acyl migration takes place.
CHO
OR3
OR4
OR5
Q
OR1
HC
CH
COHO
OH
OH
C
HC
CH
COHO
OH
pCo
HC
CH
COHO
OH
OCH3
F
O
Number Name and abbreviation R1 R3 R4 R5
1 1-O-caffoylquinic acid (1-CQA) C H H H
2 3-O-caffoylquinic acid (3-CQA) H C H H
3 5-O-caffoylquinic acid (5-CQA) H H H C
4 4-O-caffoylquinic acid (4-CQA) H H C H
5 1,3-di-O-caffoylqunic acid (1,3-diCQA) C C H H
6 1,4-di-O-caffoylqunic acid (1,4-diCQA) C H C H
57
7 1,5-di-O-caffoylqunic acid (1,5-diCQA) C H H C
8 3,4-di-O-caffoylqunic acid (3,4-diCQA) H C C H
9 3,5-di-O-caffoylqunic acid (3,5-diCQA) H C H C
10 4,5-di-O-caffoylqunic acid (4,5-diCQA) H H C C
11 1-O-p-coumaroylqunic acid (1-pCoQA) pCo H H H
12 3-O-p-coumaroylqunic acid (3-pCoQA) H pCo H H
13 5-O-p-coumaroylqunic acid (5-pCoQA) H H H pCo
14 4-O-p-coumaroylqunic acid (4-pCoQA) H H pCo H
15 1,3-di-O-p-coumaroylqunic acid (1,3-dipCoQA) pCo pCo H H
16 1,4-di-O-p-coumaroylqunic acid (1,4-dipCoQA) pCo H pCo H
17 1,5-di-O-p-coumaroylqunic acid (1,5-dipCoQA) pCo H H pCo
18 3,4-di-O-p-coumaroylqunic acid (3,4-dipCoQA) H pCo pCo H
19 3,5-di-O-p-coumaroylqunic acid (3,5-dipCoQA) H pCo H pCo
20 4,5-di-O-p-coumaroylqunic acid (4,5-dipCoQA) H H pCo pCo
21 1-O-feruloylquinic acid (1-FQA) F H H H
22 3-O-feruloylquinic acid (3-FQA) H F H H
23 5-O-feruloylquinic acid (5-FQA) H H H H
24 4-O-feruloylquinic acid (4-FQA) H H F H
25 1,3-di-O-feruloylqunic acid (1,3-diFQA) F F H H
26 1,4-di-O-feruloylqunic acid (1,4-diFQA) F H F H
27 1,5-di-O-feruloylqunic acid (1,5-diFQA) F H H F
28 3,4-di-O-feruloylqunic acid (3,4-diFQA) H F F H
29 3,5-di-O-feruloylqunic acid (3,5-diFQA) H F H H
30 4,5-di-O-feruloylqunic acid (4,5-diFQA) H H F F
Q- quinic acid, C- caffeic acid, pCo- p-coumaric acid, F- ferulic acid
Figure 3.1 Structure of mono and di caffeoylquinic, p-coumaroylquinic and feruloylquinic
acids
58
Considering the mechanism of the acyl migration proceeding through an ortho-ester
intermediate formation as 5-CQA ⇋ 4-CQA ⇋ 3-CQA, the reverse equilibrium, through
which the migrated acyl group reverts back to its original position appears to be slower. It is
difficult to comment on the thermodynamic equilibrium of the acyl migration because of the
ongoing simultaneous ester hydrolysis competing with the acyl migration process.
Figure 3.2 UV Chromatograms (318-322 nm) at 2, 5, 10 and 30 minutes of base hydrolysis of
5-CQA (3)
0
25
50
75
100
Intens. [mAU]
0
25
50
75
100
0
10
20
0
5.0
7.5
10.0
10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 Time [min]
2 4
2 3 4
2 3
4
2
3
4
2 Min
5 Min
10 Min
30 Min
3 32
32
32
32
59
Figure 3.3 Amount of the transformation products after base hydrolysis for different time
intervals of 5-CQA (3), 4-CQA (4) and 3-CQA (2)
0
500
1000
1500
2000
2500
3000
2 Min. 5 Min. 10 Min. 30 Min.
Pea
k a
rea
Time (Minutes)
5-CQA
3-CQA
4-CQA
5-CQA
0
5
10
15
20
25
30
2 Min. 5 Min. 10 Min. 30 Min.
Pea
k a
rea
Time (Minutes)
3-CQA
4-CQA
5-CQA
4-CQA
0
1
2
3
4
5
6
7
8
2 Min. 5 Min. 10 Min. 30 Min.
Pea
k a
rea
Time (Minutes)
3-CQA
4-CQA
5-CQA
3-CQA
60
The equilibrium between 3-CQA (2) and 4-CQA (4) is readily achieved because the ortho-
ester intermediate is more stable due to the cis geometry (Figure 3.4). Compound 2 shows a
tendency to hydrolyse to generate caffeic acid and quinic acid rather than to undergo acyl
migration presumably due to the 1,3-syn-diaxial arrangement between the C1 hydroxyl group
and the C3 ester. The hydrogen bonding between C1-OH and the carbonyl oxygen on the ester
on C3 results in steric hindrance preventing the nucleophilic attack on the carbonyl carbon by
the C4 hydroxyl group and facilitates the hydrolysis of the ester bond thus dissociating the
caffeoyl moiety in basic conditions. At the same time this hydrogen bonding presumably
activates the ester at C3 for hydrolytic cleavage.
O
HO
OH OH
O
O R
O
HOH
3
HO
O
OH OH
O
O
O
R
HO
O
OH O
O
OH O
R
4 OH
H
HO
O
OH O
O
OH
R
O
HO
O
OH O
OH
OH
2
R
O
-H2O
-H2O
Figure 3.4 Mechanism of the acyl migration through an ortho-ester intermediate formation
Previous studies have shown that 1,5-diCQA (7) was converted into 1,3-diCQA (5) and 5-
CQA (3) rapidly and extensively by TMAH treatment within 1 minute of hydrolysis. 18,19 In
the present study however, base hydrolysis of 1,3-diCQA (5) did not show any presence of
1,5-diCQA (7) or 1,4-diCQA (6). 1,3-diCQA (5) decomposed largely into 1-CQA (1) rather
than 3-CQA (2) in the ratio 2.2:1 after two minutes of base treatment. 1,3-diCQA (5) did not
transform preferably into 3,5-diCQA (9) whereas 3,4-diCQA (8) and 4,5-diCQA (10) were
formed in very small quantities (Figure 3.5). 1-CQA (1) was found to be present entirely as a
decomposition product rather than a migrated derivative as it was not observed in any other
mono- or di-acylated substrates; this fact also supports the ortho-ester propagation of acyl
61
migration process. Moreover, this observation confirms the hydrolytic lability of the esters in
the 3-acylated position.
In the case hydrolysis of 4,5-diCQA (10), we observed that 4,5-diCQA transformed mainly
into 3,4-diCQA (8) after two minutes of base treatment. 5-CQA (3) and 4-CQA (4) showed
approximately the same peak area at two minutes and it was slightly larger than 3-CQA (2).
This trend continued for five minutes during the base treatment, after which all of the
derivatives were completely decomposed (Figure 3.5). 3,4-diCQA (8) was observed to be the
least stable substrate during the base treatment study. At two minutes, 8 was found to be in
equilibrium with 3,5-diCQA (9) however after five minutes of treatment, 3,5-diCQA (9) was
observed to display a slightly larger peak area than 3,4-diCQA (8). 3-CQA (2) was not
observed in any sample throughout the duration of the base treatment of 3,4-diCQA due to its
tendency to decompose rapidly by hydrolysis.
Observations based on the results from the base treatment on 3,5-diCQA (9) were quite
distinctive. After five minutes 3,5-diCQA (9) was observed to possess surprising stability to
the hydrolysis of the ester as we did not detect any of the mono-CQA derivatives even after
60 minutes of base treatment (Figure 3.5). 3,5-diCQA (9) transformed mainly into 3,4-
diCQA (8) followed by 4,5-diCQA (10). The ratio of 9:8:10 remained approximately constant
throughout the 60 minutes of base treatment.
62
Figure 3.5 (Continued)
0
100
200
300
400
500
600
700
800
900
2 Min. 5 Min. 10 Min. 30 Min.
Pea
k a
rea
Duration (Min)
1, 3-diCQA
1, 3-diCQA
3, 4-diCQA
4, 5-diCQA
3-CQA
5-CQA
4-CQA
1-CQA
0
50
100
150
200
250
300
2 Min. 5 Min. 10
Min.
30
Min.
60
Min.
Pea
k a
rea
Duration (Min)
3, 5-diCQA
3, 5-diCQA
3, 4-diCQA
4, 5-diCQA
1514
10
8
23
0 00
2
4
6
8
10
12
14
16
3, 4-diCQA 3, 5-diCQA 4-CQA 5-CQA
Pea
k a
rea
3, 4-diCQA
2 Min.
5 Min.
63
Figure 3.5 Amount of the transformation products after base hydrolysis for different time
intervals of di-acylated reference standards
3.3.2 Intermolecular acyl migration (Transesterification): hydrolysis by TMAH (Cross-
over experiment)
In this contribution, we studied intermolecular acyl migration by carrying out cross-over
experiments, in which 5-CQA (3) was reacted with the free acids like ferulic and p-coumaric
acids at 1:1 stoichiometry in presence of a base. The intermolecular acyl migration was found
to be simultaneously competing with intramolecular acyl migration as well as hydrolysis of
the CGA. The products identified in the reaction of 5-CQA with ferulic acid and p-coumaric
acid at 2, 5, 10, 15 and 30 minutes of base experiment are summarized in Table 3.1. Along
with the transesterification products of 5-CQA and respective free acid, formation of the cis
isomers was also observed in ferulic, caffeic and p-coumaric acids. All products were
identified according to the fragmentation schemes reported by Kuhnert et al., Jaiswal et al.
and Clifford et al. 18-21
91
165
134
47
61 58
28 5 0 0 00
20
40
60
80
100
120
140
160
180
3,5-diCQA 3, 4-diCQA 4, 5-diCQA 3-CQA 4-CQA 5-CQA
Pea
k a
rea
4, 5-diCQA
2 Min.
5 Min.
64
Table 3.1 Compounds identified after base treatment of CGA with p-coumaric acid and CGA
with ferulic acid for various time intervals
5-CQA + p-coumaric acid
5-CQA + ferulic acid
Reaction Product RT (min) m/z (M-H) Reaction Product RT (min) m/z (M-H)
time(min) number time(min) number
2 33 26.70 162.6 2 31 27.3 192.6
32 19.30 178.6 34 29.6 192.6
37 25.70 334.8 22 19.2 366.7
38 29.40 334.9 23 24.1 366.9
2 11.50 352.8 24 26.4 366.9
3 17.10 352.8 49 3.0 352.8
4 20.80 352.8 2 11.2 352.9
50 13.1 352.8
5 37 25.80 335.0 3 16.9 352.8
38 29.80 335.1 4 20.5 352.8
14 27.60 337.0 37 23.6 334.9
1 10.70 353.1 38 25.7 334.9
2 12.10 353.0 32 7.4 178.6
3 17.30 353.1 35 19.2 178.6
4 21.70 353.1
40 30.60 367.1 5 23 26.9 367.1
32 19.50 178.6 40 28.7 367.1
33 26.60 162.6 1 11.10 353.1
37 25.9 335.1
10 2 12.20 353.1 38 29.6 335.1
37 25.90 335.1 2 12.5 353.1
38 29.80 335.0 3 17.7 353.0
1 10.80 353.1 4 22.2 353.0
3 17.50 353.0 51 30.8 529.0
4 21.90 353.1
33 26.40 162.9 10 37 25.9 335.1
40 30.60 367.1
1 10.9 353.0
65
13 24.30 336.8 2 12.2 353.0
14 28.40 337.0 3 17.6 353.0
32 19.70 178.6 4 22.1 353.1
23 26.8 367.1
15 38 29.80 335.1 40 28.9 367.1
2 12.40 353.1 51 31.0 529.0
3 17.70 353.0
32 19.60 178.9 15 37 26.2 335.1
4 22.00 352.8 2 12.4 353.0
3 17.7 353.0
30 3 18.00 353.1 4 22.3 353.1
4 22.60 352.8 23 27.2 367.1
33 26.60 162.9 40 28.9 367.1
51 31.2 529.0
30 40 28.9 367.1
3 18.0 353.0
66
Figure 3.6 UV Chromatograms (318-322 nm) at 2, 5, 10, 30 and 60 minutes of base
hydrolysis of 3, 5-diCQA (9)
0
2
4
6
8
Intens. [mAU]
0
2
4
6
8
10
0
2
4
6
8
10
0
2
4
6
0
1
2
3
10 15 20 25 30 35 40 45 Time [min]
2 Min
5 Min
10 Min
30 Min
60 Min
9
8
10
9
9
9
9
8
8
8
8
10
10
10
10
67
OH
O
OH
O
31
OH
O
OH
OH
32 33
OH
O
OH
OH
O
34
OH
OH
35 36
OH
OH
O
OH
OO
OH
O
O
OH
O
O
OH
HO
HO
37
O
O
HO
HO
HO
O
O
OH
38
OH
HOO
O
O
O
HO
OH
36
O
O
OH OH
OH
O
O
OH
OH
40
O
O
OH OH
OH
O
O
OH
39
O
HO
O
O
OH
OH
OH
OH
41
HO
O
OH OH
OH
O
O
HOOH
42
OH
O
OO
OH
OH
O
OH
OH
O
HO
43
Figure 3.7 Compounds identified during acyl migration studies (Continued)
68
O
O
OH
OH
O
OHHO
O
OO
HO
46
O
O
O
O
OH
OH
O
OH
HO
OH
OH
O
O
HO
45
O
OH
O
O
OH
OH
O
OH
HO
O
HO
44
OH
OOH
HO
O
HO
O
O
O
OH
OH
OHO
O
48
50
OH
O
O
O
OH
OHO
O
O
HO
OH
OH
OH
47
HO
O
O
O
HO
HO
OH
OH
OH
49
HO
O
OH O
O
OH O
O
OH
HO
O
OH
51
OH
OOH
O
O
HO
O
O
O
O
OH
OH
OHO
HO
HO
52
HO
O
OH OH
O
OH
OH
O
OH
OH
(regio-chemistry randomly selected)
(regio-chemistry randomly selected)
(regio-chemistry randomly selected)
O
Figure 3.7 Compounds identified during acyl migration studies
69
Figure 3.8 Comparison between the peak areas of compounds formed during TMAH
treatment of 5-CQA (3) with p-coumaric acid (pCoA)
In the case of the base treatment of an equimolar mixture of p-coumaric acid and 5-CQA (3),
the intramolecular acyl migration within 5-CQA seemed to be dominating the hydrolysis and
the intermolecular acyl migration. According to the peak areas observed, the formation of the
intermolecular acyl migrated species (transesters) found to be least favoured (Figure 3.8). For
example, 3-CQA (2), 5-CQA (3) and 4-CQA (4) formed predominantly over caffeic acid (32)
during two minutes of TMAH treatment of 5-CQA. p-coumaric acid (33) did not esterify with
the quinic acid generated from the hydrolysis of the 5-CQA (3) after two minutes hence, no p-
coumaroylqunic acids were observed. However, after 5 minutes of the base treatment first
transesterification product appeared in the form of 4-pCoQA (14) and after ten minutes both
5-pCoQA (13) and 4-pCoQA (14) were observed. Unfortunately, due to very low
concentration we could not compare the peak areas of compounds 13 and 14 in the UV
chromatogram and hence cannot comment on the kinetics of the acyl migration. Although it
was clear that 4-pCoQA (14) was formed earlier than 5-pCoQA (13), but it was not obvious
whether 13 was an acyl migration product of 14. Formation of intramolecular acyl migration
products takes place according to the conclusions established earlier in this paper. After two
minutes the rate of hydrolysis of 5-CQA was very low and the amounts of 3-CQA (2), 5-CQA
(3) and 4-CQA (4) were the highest. Between five to ten minutes of base hydrolysis,
equilibrium was reached where 5-CQA was found to be predominant. Formation of the cis
derivatives is supposed to follow a water molecule addition-elimination to the double bond in
the cinnamoyl moiety during the base hydrolysis.
0
2000
4000
6000
8000
10000
12000
2 5 10 15 30
Pea
k A
rea
Duration (Min)
5-CQA + pCoA3-CQA
5-CQA
4-CQA
Caffeic acid
70
Figure 3.9 (Continued)
EIC 353.0
0
1
2
3
7 x10
Intens.
5 10 15 20 25 Time [min]
2 50
3 4
49
191 MS2(353)
110.7
126.7 172.5
MS3(353>191)
110.6 126.6 154.6
170.6 MS4(353 >191>172)
0
100
[%]
0
100
0
100
110 120 130 140 150 160 170 180 190 m/z
49
EIC 529.0
0.0
0.2
0.4
0.6
0.8
7 x10 Intens.
22 24 26 28 30 32 34 Time [min]
51
71
Figure 3.9 EIC and fragmentation patterns for 1-cis-caffeoylquinic acid (49) at m/z 353 and
caffeoyl-feruloylquinic acid (51) at m/z 529 in transesterification induced by TMAH
When an equimolar mixture of ferulic acid (31) and 5-CQA (3) was treated with base we
observed three regio-isomers of feruloylquinic acid resulting from intermolecular
transesterification: 3-FQA (22), 5-FQA (23) and 4-FQA (24) (Figure 3.10). 22 The same was
not observed in the case of the p-coumaric acid experiment. After two minutes, 3-FQA (peak
area = 3215) was formed predominantly over 5-FQA and 4-FQA. 5-FQA (23) and 4-FQA
(24) showed negligible peak areas. Only 5-FQA remained stable enough to be detected after
10 and 15 minutes of base hydrolysis. After 30 minutes all the transesters and the substrate 5-
CQA found to be decomposed completely as we could identify caffeic acid only. Formation
of 3-FQA was found to be kinetically favoured. This was found to be consistent with the fact
that during the first few minutes of the base treatment of 5-CQA to study intramolecular acyl
migration, 3-CQA dominates the acyl migration product spectrum. We also identified cis-1-
O-caffeoylquinic acid (49) in the EIC of m/z 353 and the UV chromatogram on the basis of its
early elution and fragmentation (Figure 3.9). 23 We suspect the formation of the compound 49
is the product of the two step procedure involving water molecule addition-elimination at the
double bond of the caffeic acid moiety. Since, we strictly did not observe any 1-acylated
caffeoylquinic acid when 5-CQA was subjected to intramolecular acyl migration through base
treatment as discussed previously in this paper. 4-cis-CQA (50) was also identified as another
isomerised caffeoylquinic acid derivative, which is assumed to be formed by an addition
elimination of the water molecule across the double bond in caffeic acid. 4-cis-CQA (50) was
observed to be in equilibrium with 4-CQA (4) after two minutes but with an increase in the
290.9
334.9 MS2(529.0)
148.9 192.9
290.9 MS3(529>335)
148.9 MS4(529>335>291)
0
100
[%]
0
100
0
100
50 100 150 200 250 300 350 400 450 500 m/z
51
72
reaction time only 4-CQA was observed to survive the base treatment. Unexpectedly, we
identified 3-CQL (37) and 4-CQL (38) in this sample having negligible peak areas since the
dehydrated product of caffeoylquinic acid must be in continuous equilibrium with
caffeoylquinic acid itself in basic aqueous conditions. Hetero-diacyl chlorogenic acid was
identified in the form of caffeoyl-feruloylquinic acid (51). It was observed after 5 minutes to
15 minutes of base hydrolysis (Figure 3.10). Figure 3.9 shows the fragmentation pathway for
compound 51, in which it loses the ferulic acid moiety and undergoes simultaneous
dehydration to give m/z 335 as a base peak in MS2. Furthermore, in MS3 the dehydrated
caffeoylquinic acid entity undergoes decarboxylation to give a base peak at m/z 291 and also
showing the presence of ferulic acid as a secondary peak at m/z 193. This fragmentation
pathway for a caffeoyl-feruloylquinic acid was found to be inconsistent with the
fragmentation pathways of 1-caffeoyl-3-feruloylquinic acid, 3-feruloyl-5-caffeoylquinic acid,
cis-4-feruloyl-5-caffeoylquinic acid, 4-feruloyl-5-caffeoylquinic acid, 4-caffeoyl-5-
feruloylquinic acid and cis-3-feruloyl-5-caffeoylquinic acid previously reported by Jaiswal et
al. 21 Hence, the regio-chemistry of the acyl groups in caffeoyl-feruloylquinic acid (51)
remains unknown.
Figure 3.10 Comparison between the peak areas of compounds formed during TMAH
treatment of 5-CQA (3) with ferulic acid (FA)
3.3.3 Intramolecular acyl migration: model roasting of 2-5 and 8-10
Compounds 3-CQA (2), 5-CQA (3), 4-CQA (4), 1,3-diCQA (5), 3,4-diCQA (8), 3,5-diCQA
(9) and 4,5-diCQA (10) were heated at 180 0C for 12 minutes separately to study the
0
2000
4000
6000
8000
10000
12000
14000
16000
2 5 10 15 30
Pea
k A
rea
Duration (Min)
5-CQA + FA3-CQA
5-CQA
4-CQA
4-cis-CQA
1-CQA
Caffeic acid
73
intramolecular acyl migration in the absence of other under conditions mimicking coffee
roasting.
Table 3.2 Compounds identified after heating (model roasting) reference standards
Starting
material Compound Product RT(min)
Peak
Area(UV)
3 5-CQA 3 23.7 11035
2 3-CQA 2 18.9 06034
4 4-CQA 37 31.8 02418
38 35.1 01543
41 32.9 00396
48 44.9 00210
48 52.3 00197
52 46.3 00080
48 49.5 00521
5 1,3-diCQA 5 30.2 16245
8 3,4-diCQA 37 31.7 00098
38 35.2 00074
46 51.2 07099
46-cis 52.4 00896
9 3,5-diCQA 46 56.2 03034
46-cis 56.9 00119
10 4,5-diCQA 10 46.3 26560
From the data summarized in Table 3.2, we clearly see that only 4-CQA (4), 3,4-diCQA (8)
and 3,5-diCQA (9) undergo transformations to generate various dehydrated products mainly
in the form of caffeoyl lactones. In case of mono-acylated chlorogenic acids reference
standards, only 4-CQA undergoes acyl migration with simultaneous dehydration. 3-CQL (37)
was found to be the predominant transformation product in the heat treatment of 4-CQA. 5-
74
caffeoyshikimic acid was also identified but was found to be the least favoured dehydration
product after 4-CQL. In this experiment, we observed that 3-CQL (37) was forming
predominantly over 4-CQL (38) irrespective of the substrate speculatively because of the
additional stability awarded by the equatorial position 3-CQL obtains in the inverted chair
conformation, therefore we can conclude that the dehydration processes such as lactone and
shikimic acid formation at the quinic acid moiety follow acyl migration. i.e. in the simulated
roasting environment, acyl migration takes place before dehydration at quinic acid moiety.
Additionally, lactonization dominates over the alternative shikimate formation in the model
roasting of all the substrates (Table 3.2). A peculiar fragmentation was observed for a product
yielding a pseudomolecular ion at m/z 671 leading to structures 48 and 52 (regio-chemistry
unknown). Although, several peaks in the EIC at m/z 671 are observed, they are supposedly
isomers of the two compounds 48 and 52 since only two distinct fragmentation patterns were
observed shown in Figure 3.11, which suggests the possibility of the two different structures.
Compound 48 is present in the form of three different isomers, which are represented in EIC
at m/z 671 as 48a, 48b and 48c (Figure 3.11). All the three isomers generate a base peak at
m/z 335 in MS2 suggesting the presence of di-caffeoyl quinide in the original structure. The
other acyl group is believed to be sharing an ether linkage with an extra unit of quinic acid on
one of its phenolic groups, which disappears as a neutral loss in MS2. On the other hand, in
the MS2 of the compound 52, we see a base peak at m/z 509 and a secondary peak at m/z 353
clearly indicating the presence of di-caffeoylquinic acid as a central structure in which, one of
the caffeoyl moieties is fused with a dehydrated quinic acid unit via ether bond on one of its
phenolic groups. 1,3-diCQA (5) did not undergo noticeable transformation by the heat
treatment whereas 3,4-diCQA (8) generated most of the transformation products among all
four di-CQAs. It is likely that the observed products 3-CQL (37) and 4-CQL (38) resulting
from 3,4-diCQA (8) were formed by lactonization and loss of one of the two acyl moieties
following the order of the processes as dissociation first and dehydration at the quinic acid
moiety later; attributed to the fact that both of the caffeoyl lactones possess similar peak areas.
3,4-di-O-caffoyl-1,5-quinide (46) was preferably formed in the heat treatment of 3,4-diCQA.
The presence of the two different peaks eluting at 51.2 and 52.4 minutes having the same
fragmentation pattern as 3,4-diCQL (46) suggested the presence of a cis isomer of compound
46 (Figure 3.12). Both isomers of 3,4-diCQL (46) generate a base peak at m/z 335 with
virtually non-existent secondary peaks confirming the regio-chemistry of 3,4-diCQL.
75
Among the unchanged substrates throughout the heat treatment at 180 0C for 12 minutes such
as, 5-CQA, 3-CQA, 1,3-diCQA and 4,5-diCQA the stability of 5-CQA (3) and 1,3-diCQA (5)
to high temperatures was confirmed by heating them at 200 0C for 10 minutes.
76
Figure 3.11 EIC and fragmentation patterns for m/z 671 observed during model roasting
EIC at m/z 671
0.0
0.5
1.0
1.5
x10 Intens.
30 35 40 45 50 55 Time [min]
37 38 41
48a
52
48b
48c
509.1 671.1
335.7 MS2(671)
135.1 254.9
160.9 MS3(671>336)
MS4(671>336>161)
0
100
[%]
0
100
0
100
100 200 300 400 500 600 m/z
133
48 (a/b/c)
191.4
353.3 509.2
MS2(671.5)
353.0 190.8
MS3(671 >509)
MS4(671>509>191)
0
100
[%]
0
100
0
100
100 200 300 400 500 600 m/z
93
52
77
Figure 3.12 EIC and fragmentation patterns for m/z 497 observed during model roasting
3.3.4 Transesterification: model roasting (Cross-over experiment)
In this contribution, we explored the possibilities of the transesterification or intermolecular
acyl migration by carrying out cross-over experiments, in which 5-CQA was subjected to the
heat treatment with the free acids like ferulic acid and p-coumaric acid at 1:1 stoichiometry.
In the first case where equimolar mixture of pCoA and 5-CQA (3) was heated together, it was
observed that 5-CQA (3) mostly remained unchanged by the heat treatment this experiment
since the peak area under 5-CQA alone is more than 5 times the peak area of all the
transformation products combined (Table 3.3). It is clear from the list that the
caffeoylactones dominated all of the other transformation products. Since it was established
EIC at m/z 497
0
2
4
6 x10
Intens.
42 44 46 48 50 52 54 56 Time [min]
46-cis
46
335.1 MS2(497)
135.3
160.9 MS3(497>335)
MS4(497>335>161)
0
100
[%]
0
100
0
100
100 200 300 400 500 m/z
133
46/ 46-cis
78
earlier by Clifford et al. that acyl migration takes place before dehydration in chlorogenic acid
when there is still some water present in the sample, 2 it can be assumed that the 5-acyl group
migrates to positions C3 and C4 first and then both products undergo dehydration to yield the
corresponding lactones in the form of 3-CQL and 4-CQL. 4-CQL (38) was formed in larger
quantities than 3-CQL (37). Only transesterification product identified was 4-pCoQA (14),
which formed during model roasting conditions shows the peak area of 330 in UV
chromatogram (Table 3.3 and Figure 3.13). It is speculated that the formation of the 1,5-
quinide takes place earlier making C5 on quinic acid unavailable for the possible condensation
with p-coumaric acid hence, we do not observe 5-pCoQA (13) but 4-pCoQA (14). Also, as
mentioned earlier, 1,3-syn-diaxial arrangement between C1 and C3 hydroxyl groups in quinic
acid generated by the hydrolysis of caffeoylquinic acid sterically hinders ester bond formation
at C3, with additional H-bond activation of the C3 ester by the C1-OH. Hence, the overall
absence of C3 transesterification products in model roasting experiment can be explained.
Table 3.3 Compounds identified after heating (Model roasting) of 5-CQA (3) and p-coumaric
acid
Product number RT (min) Peak area
37 25.9 01093
38 29.6 01538
14 28.0 00330
48 36.9 00145
52 39.8 00642
52 41.6 02137
3 17.5 30503
Model roasting experiment between 5-CQA (3) and ferulic acid (31) generated a larger
number of transformation products compared to the number of transformation products
detected from the same experiment with pCoA (33) and 5-CQA (3). Similar trend in terms of
number of the transformed products was observed in the cross-over experiment by TMAH
treatment. The list of the transformation products in this experiment looks very similar to the
list in Table 3.3. Additional two products are found to be the acyl migration products of 5-
CQA. The acyl group in 5-CQA (3) was found to migrate intramolecularly to produce 3-CQA
(2) and 4-CQA (4) but the corresponding caffeoyllactones 3-CQL (37) and 4-CQL (38) were
formed more in quantity. The ratio of 3-CQL (37) to 3-CQA (2) was 11:1 whereas, the ratio
79
of 4-CQL (38) to 4-CQA (4) was found to be 1:1.3 since, 4-CQL formed in more quantity 3-
CQL by comparison of corresponding peak areas (Table 3.4). In this experiment it was
observed that 5-CQA (3) remained unchanged to even greater extent than previous
experiment by the heat treatment since the peak area under 5-CQA alone is more than 16
times the peak area of all the transformation products combined (Table 3.4). 5-FQA (23) was
formed in considerable quantity confirming intermolecular acyl migration between ferulic
acid and 5-CQA (3) (Figure 3.13).
Figure 3.13 MS3 and MS4 of 4-pCoQA (14) and 5-FQA (23) respectively observed during
cross-over experiment by model roasting
337 MS, 28.0min
173 MS2(337), 28.0min
93.1
110.9
154.7
MS3(337>173), 28.0min
0
100
[%]
0
100
0
100
100 150 200 250 300 350 400 450 m/z
4-pCoQA (14)
191.1 MS2(366.7), 26.8min
110.7 173 127
MS3(367 >191), 26.8min
MS4(367>191 >173), 26.9min
0
100
[%]
0
100
0
100
100 125 150 175 200 225 250 275 300 m/z
109
5-FQA (23)
80
Table 3.4 Compounds identified after heating (Model roasting) 5-CQA (3) and ferulic acid
Product number RT (min) Peak area
37 26.0 00484
38 29.4 00535
3 16.8 44224
2 12.2 00045
4 21.9 00421
23 26.8 00484
48 36.8 00176
52 39.9 00138
52 41.9 00500
Formation of the transesterification products in the cross-over experiments by the heat
treatment was highly unexpected since; we did not observe both caffeic acid (32) and quinic
acid in the model roasting of the mono- and di-caffeoylquinic acids (Table 3.2). It seems that
the dissociation products remain stable enough to be esterified with another entity. Problem is,
we can prove the existence of the hydrolysed products only if there is a chance to form
transesterification products. Hence, we cannot supply any evidence of the hydrolysis of the
esters in mono- and di-caffeoylquinic acids when they are heated separately because, even if
free caffeic acid or quinic acid condenses with each other, it will be counted as acyl migration
product in roasting experiments.
3.3.5 Intramolecular acyl migration: Brewing of CGAs
In this contribution we also studied the acyl migration products of mono- and di-acyl CGAs
(2-5 and 8-10), which were formed during the brewing process (Table 3.5). Typically, the pH
of a coffee brew is at 4.7-5.1, in our case the pH of the brew was measured at 5.0. We
observed that the hot water also serves as a reactive reagent other than just a simple solvent in
coffee brewing similar to the previous work on tea fermentation, where water was shown to
be the key reagent in thearubigin formation.24 Apart from the acyl migration products and
trans-cis isomerization (cis-caffeoylquinic acids) products; the resulting chromatograms
showed transformation products referred here as hydroxy-dihydrocaffeoylquinic acids arising
through conjugate addition of water molecule to the olefinic cinnamoyl moiety.25
81
Table 3.5 Compounds identified after hydrolysis of reference standards (Brewing of CGAs)
Starting material Product name RT (Min) Peak area(UV)
5-CQA(3) 5-CQA 20.1 17610
cis-5-CQA 23.0 00347
5-hCQA I 7.3 NA
5-hCQA II 7.9 NA
4-CQA(4) 3-CQA 13.1 17576
4-CQA 20.6 06098
cis-3-CQA 11.9 00017
cis-4-CQA 16.5 00179
4-hCQA I 6.9 NA
4-hCQA II 7.9 NA
3-CQA(2) 3-CQA 13.1 00639
4-CQA 20.6 00954
cis-3-CQA 11.9 00150
3-hCQA I + II 5.6 00017
1, 3-diCQA(5) 1, 3-diCQA 22.9 03422
3, 4-diCQA(8) 3,4-diCQA 35.2 02558
cis-4,5-diCQA I 38.0 00099
4,5-diCQA 36.9 00520
cis-3,4-diCQA I 34.2 00393
cis-3,4-diCQA II 35.9 00345
3-CQA 12.3 00020
4-CQA 19.3 00027
5-CQA 15.6 00002
3-C-4-hCQA 26.4 00057
3, 5-diCQA(9) 3,4-diCQA 36.5 01406
3,5-diCQA 37.3 01057
4,5-diCQA 41.4 01107
82
3-C-5-hCQAI 23.8 00014
3-hC-5-CQA I 27.8 00022
3-hC-cis-5-CQA 28.3 00004
3-hC-5-CQA II 31.7 00004
3-CQA 12.9 00080
5-CQA 20.3 00126
4, 5-diCQA(10) 3,4-diCQA 43.7 01825
3,5-diCQA 42.6 00512
4,5-diCQA 45.7 02148
cis-4,5-diCQA I 47.2 00036
cis-4,5-diCQA II 52.0 00007
3-CQA 16.8 00019
4-CQA 27.9 00014
5-CQA 23.1 00018
4-hC-5-CQA 34.7 00018
3-hC-5-CQA II 35.5 00018
3-C-5-hCQA II 63.1 00023
NA- Insignificant peak area
Similar to the model roasting experiment, 5-CQA (3) did not show any acyl migrated products
in hydrolysis (brewing). The acyl moiety in 4-CQA (4) and 3-CQA (2) did not migrate to C5
of the quinic acid but acyl moiety interchange between C3 and C4 was observed due to the
stability of the ortho-ester intermediate arising from the cis geometry (Figure 3.4). Acyl
migration to C5 from C3 and C4 was found to be highly pH dependent as we only observed it
in case of base hydrolysis.
Cynarin (1,3-diCQA) did not show any transformation products after 5 h. of refluxing with
water. In rest of the di-acylated reference standards 3,4-di-acylated esters were preferably
formed irrespective of the substrate. This observation can possibly be attributed to the fact
that the parallel displaced 𝜋- 𝜋 stacking arrangement of the cinnamoyl benzene rings provide
the added stability to the 3,4-diCQA in the minimum energy chair conformation of quinic acid
moiety. By comparing the peak areas in UV chromatogram it was observed that the
decomposition of the di-CQAs to produce mono-CQAs was taking place in minute quantity as
83
compared to the base hydrolysis experiment to study intermolecular acyl migration. In the
case of mono-acylated chlorogenic acids, up to 1.5-2.0% of the chlorogenic acids were
transformed into their hydroxylated derivatives and the diacylated chlorogenic acids up to 4-
4.5% if the relative peak areas in EIC are considered.25 But their peak areas in UV
chromatograms were found to be negligible. The structures for the water addition compounds
can be found in in our previous publication 25 and in Figure 3.14.
3.4 Conclusions
In this work we observed that that the acyl migration phenomena occur before dehydration
takes place in quinic acid moiety. Acyl migration is facilitated in presence of the liquid media
as compared to the roasting process. Therefore, the lower temperature roasts like the
‘cinnamon roast’ produce more number of acyl migration products than higher temperature
roasts, which generate dehydration products like lactones and shikimates high in numbers.
Esters present on C3 position of the quinic acid are prone to hydrolysis of the ester bond than
undergoing acyl migration in any experimental condition. In both cross-over experiments
(TMAH and model roasting), 5-CQA (3) and ferulic acid (31) when treated together
generated a large number of transesterification products than 5-CQA (3) treated with p-
coumaric acid (33). Dehydration products in cross-over experiment by roasting are observed
10 times more in abundance than acyl migration or transesterification products. The amount
of esters present on C3 and C4 positions of quinic acid moiety in a cup of coffee after roasting
and brewing processes is highly contributed by C5 positioned esters in case of mono-
caffeoylquinic acids content. In contrast to this observation we found that acyl migration to C5
position from C3 and C4 is only possible in base hydrolysis i.e. it is highly pH dependant. 1,3-
diCQA (5) and 5-CQA (3) were observed to be more stable than the rest of the reference
standards in both roasting and brewing conditions.
84
OH OH
OH
O
HOOC
O
OH
OH
HO
OH OH
OH
O
HOOC
O
OH
OH
HO
OH OH
O
OH
HOOCO
HO
HO
OH
OH OH
O
OH
HOOCO
HO
HO
OH
5-hCQA I 5-hCQA II 4-hCQA I 4-hCQA II
OH O
OH
OH
HOOC
3-hCQA I
O
OH
HO
OH
OH O
OH
OH
HOOC
3-hCQA II
O
OH
HO
OH
OH
O
1-hCQL
HO
O
O
O
OH
OH
HO
OH O
OHHOOC
cis-3-CQA
OH
O OH
OH
OH O
OHOOC
cis-3-C-4-CQA
OH
O OH
OH
O
HOOH
OH O
OHHOOC
3-hC-cis-5-CQA
OO
HO
HO
O
OH
HO
OHOH OH
OHOOC
cis-4-C-5-CQA
OO
OH
OH
O
OHHO
OH OH
OHOOC
4-C-cis-5-CQA
OO
HO
HO
O
HOOH
Figure 3.14 Structures identified after brewing of the reference standards (Continued)
85
OH O
O
OH
HOOC
3-hC-4-CQA
O
OH
HO
OH
O
HOOH
OH O
O
OH
HOOC
O
OH
HO
O
HOOH
OH
3-C-4-hCQA
OH O
OH
O
HOOC
3-hC-5-CQA I
O
OH
HO
OH
O
OHHO
OH O
OH
O
HOOC
3-hC-5-CQA II
O
OH
HO
OH
O
OHHO
OH O
OH
O
HOOC
3-hC-cis-5-CQA
O
OH
HO
OH
OHO
HO
OH O
OH
O
HOOC
3-C-5-hCQA I
O
OH
HO
O
OHHO
HO
OH O
OH
O
HOOC
3-C-5-hCQA II
O
OH
HO
O
OHHO
HO
OH OH
OHOOC
4-hC-5-CQA
OO
HOOH
OHO
OHHO
OH OH
OHOOC
4-C-5-hCQA
OO
HOOH
O
OHHO
HO
Figure 3.14 Structures identified after brewing of the reference standards
86
References
1. Clifford, M.N. Chlorogenic acids and other cinnamates - nature, occurrence, dietary
burden, absorption and metabolism. J. Sci. Food Agric. 2000, 80, 1033-1043.
2. Jaiswal, R.; Matei, M.F.; Golon, A.; Witt, M.; Kuhnert, N. Understanding the fate of
chlorogenic acids in coffee roasting using mass spectrometry based targeted and non-targeted
analytical strategies. Food Funct. 2012, 3, 976-984.
3. Crozier, A.; Jaganath, I.B.; Clifford, M.N. Dietary phenolics: chemistry, bioavailability and
effects on health. Nat. Prod. Rep. 2009, 26, 1001-1043.
4. Farah, A.; De Paulis, T.; Trugo, L.C.; Martin, P.R. Effect of roasting on the formation of
chlorogenic acid lactones in coffee. J. Agric. Food Chem. 2005, 53, 1505-1513.
5. Kweon, M.H.; Hwang, H.J.; Sung, H.C. Identification and antioxidant activity of novel
chlorogenic acid derivatives from bamboo (Phyllostachys edulis). J. Agric. Food Chem. 2001,
49, 4646-4655.
6. Wang, G.; Shi, L.; Ren, Y.; Liu, Q.; Liu, H.; Zhang, R.; Li, Z.; Zhu, F.; He, P.; Tang, W.;
Tao, P.; Li, C.; Zhao, W.; Zuo, J. Anti-hepatitis B virus activity of chlorogenic acid, quinic
acid and caffeic acid in vivo and in vitro. Antiviral Res. 2009, 83, 186-190.
7. Hemmerle, H.; Burger, H.; Below, P.; Schubert, G.; Rippel, R.; Schindler, P.W.; Paulus, E.;
Herling, A.W. Chlorogenic Acid and Synthetic Chlorogenic Acid Derivatives: Novel
Inhibitors of Hepatic Glucose-6-phosphate Translocase. J. Med. Chem. 1997, 40, 137-145.
8. Robinson, W.E.; Reinecke, M.G.; AbdelMalek, S.; Jia, Q.; Chow, S.A. Inhibitors of HIV-1
replication that inhibit HPV integrase. Proc. Natl. Acad. Sci. U. S. A. 1996, 93, 6326-6331.
9. Kwon, H.C.; Jung, C.M.; Shin, C.G.; Lee, J.K.; Choi, S.U.; Kim, S.Y.; Lee, K.R. A new
caffeoyl quinic acid from Aster scaber and its inhibitory activity against human
immunodeficiency virus-1 (HIV-1) integrase. Chem. Pharm. Bull. 2000, 48, 1796-1798.
10. Clifford, M.N.; Kellard, B.; Birch, G.G. Characterisation of chlorogenic acids by
simultaneous isomerisation and transesterification with tetramethylammonium hydroxide.
Food Chem. 1989, 33, 115-123.
87
11. Clifford, M.N.; Kellard, B.; Birch, G.G. Characterization of caffeoylferuloylquinic acids
by simultaneous isomerization and transesterification with tetramethylammonium hydroxide.
Food Chem. 1989, 34, 81-88.
12. Dawidowicz, A.L.; Typek, R. Thermal Stability of 5-o-Caffeoylquinic Acid in Aqueous
Solutions at Different Heating Conditions. J. Agric. Food Chem. 2010, 58, 12578-12584.
13. Trugo, L.C.; Macrae, R. A study of the effect of roasting on the chlorogenic acid
composition of coffee using HPLC. Food Chem. 1984, 15, 219-227.
14. Anonymous IUPAC Commission on the Nomenclature of Organic Chemistry (CNOC)
and IUPAC-IUB Commission on Biochemical Nomenclature (CBN). Nomenclature of
cyclitols. Recommendations, 1973. Biochem J 1976, 153, 23-31.
15. Clifford, M.N.; Williams, T.; Bridson, D. Chlorogenic acids and caffeine as possible
taxonomic criteria in Coffea and Psilanthus. Phytochemistry 1989, 28, 829-838.
16. Hanson, K.R. Chlorogenic acid biosynthesis Chemical synthesis and properties of the
mono-O-cinnamoylquinic acids. Biochemistry 1965, 4, 2719-2731.
17. Xie, C.; Yu, K.; Zhong, D.; Yuan, T.; Ye, F.; Jarrell, J.A.; Millar, A.; Chen, X.
Investigation of Isomeric Transformations of Chlorogenic Acid in Buffers and Biological
Matrixes by Ultraperformance Liquid Chromatography Coupled with Hybrid Quadrupole/Ion
Mobility/Orthogonal Acceleration Time-of-Flight Mass Spectrometry. J. Agric. Food Chem.
2011, 59, 11078-11087.
18. Clifford, M.N.; Johnston, K.L.; Knight, S.; Kuhnert, N. Hierarchical Scheme for LC-MSn
Identification of Chlorogenic Acids. J. Agric. Food Chem. 2003, 51, 2900-2911.
19. Clifford, M.N.; Knight, S.; Kuhnert, N. Discriminating between the six isomers of
dicaffeoylquinic acid by LC-MSn. J. Agric. Food Chem. 2005, 53, 3821-3832.
20. Jaiswal, R.; Matei, M.F.; Ullrich, F.; Kuhnert, N. How to distinguish between
cinnamoylshikimate esters and chlorogenic acid lactones by liquid chromatography-tandem
mass spectrometry. Journal of Mass Spectrometry 2011, 46, 933-942.
88
21. Jaiswal, R.; Sovdat, T.; Vivan, F.; Kuhnert, N. Profiling and Characterization by LC-MSn
of the Chlorogenic Acids and Hydroxycinnamoylshikimate Esters in Mate (Ilex
paraguariensis). J. Agric. Food Chem. 2010, 58, 5471-5484.
22. Kuhnert, N.; Jaiswal, R.; Matei, M.F.; Sovdat, T.; Deshpande, S. How to distinguish
between feruloyl quinic acids and isoferuloyl quinic acids by liquid chromatography/tandem
mass spectrometry. Rapid Commun. Mass Spectrom. 2010, 24, 1575-1582.
23. Clifford, M.N.; Kirkpatrick, J.; Kuhnert, N.; Roozendaal, H.; Salgado, P.R. LC-MSn
analysis of the cis isomers of chlorogenic acids. Food Chem. 2008, 106, 379-385.
24. Kuhnert, N.; Drynan, J.W.; Obuchowicz, J.; Clifford, M.N.; Witt, M. Mass spectrometric
characterization of black tea thearubigins leading to an oxidative cascade hypothesis for
thearubigin formation. Rapid Commun. Mass Spectrom. 2010, 24, 3387-3404.
25. Matei, M.F.; Jaiswal, R.; Kuhnert, N. Investigating the Chemical Changes of Chlorogenic
Acids during Coffee Brewing: Conjugate Addition of Water to the Olefinic Moiety of
Chlorogenic Acids and Their Quinides. J. Agric. Food Chem. 2012, 60, 12105-12115.
89
CHAPTER 4: Synthesis, structure and tandem MS investigation of
diastereomers of quinic acid
4.1 Introduction
(-)-Quinic acid (1) is distributed naturally in a variety of plant materials ranging from coffee
to cinchona bark to tobacco leaves and cranberries in its free form or in the form of its
depsides, chlorogenic acids. Quinic acid was first isolated in 1790 and was given an empirical
formula in 1838. 1 In a single coffee bean, up to 4.0mg free quinic acid is found. 2 In
Colombian Arabica green coffee, up to 7.0 g/kg of quinic acid is present, which increases up
to 10.0 g/kg upon roasting.3 Quinic acid provides characteristic astringent taste to the
beverage hence; it is also used as a flavor enhancer in certain beverages. 4 Quinic acid is
considered as a primary metabolite in most living organisms, being an intermediate in the
shikimic acid pathway in the biosynthesis of aromatic compounds. 5,6 This paper will use the
nomenclature defined by the IUPAC system for (-)-quinic acid as, 1L-1(OH),3,4/5-
tetrahydroxycyclohexane carboxylic acid. 7
OHOH
OHO
HO
OH
OH
OHO
HO
OH
OH
(-)-quinic acid (1) (-)-epi-quinic acid (2)
OH
OHO
HO
OHOH
muco-quinic acid (3)
OHOH
OHO
HO OH
cis-quinic acid (4)
OH
OHO
HO OHOH
scyllo-quinic acid (5)
OHO
HO
OHOH
OH
neo-quinic acid (6)
C4 inverted C3 inverted C5 inverted
C4 and C5 inverted C3 and C4 inverted
(+)-quinic acid
OHOH
HO
O
OHOH
OH
OHOH
HO
O
OH
(+)-epi-quinic acid
Figure 4.1 Stereoisomers of quinic acid
Quinic acid has eight possible stereoisomers: four meso forms and two pairs of enantiomers
(Figure 4.1). While the esters of the 3R, 4S, 5R isomers of the quinic acid dominates the
esters with other diastereomers, they has been reported to be found naturally or as products of
the food processing. As reported by Kuhnert and co-workers, 80 different chlorogenic acid
90
derivatives have been identified in green coffee beans. After roasting and brewing, this
number is increased to 120 derivatives identified on the basis of the presence of fragment ions
corresponding to quinic acid and quinic acid lactones in MSn. 8,9 After ingestion of foods
containing quinic acid esters, metabolism in humans may also give rise to the esters of the
diastereomers of the quinic acid. 8,10 This fact supports the assumption that the roasting or
food processing in general facilitates the isomerization at the stereogenic centers in 3R, 4S,
5R esters of the quinic acid. 11,12 Having stated this, we found that a number of esters of the
diastereomers of the quinic acid are already reported in literature as plant secondary
metabolites. For example, in Lactuca indica L., Asimina triloba and Aster scaber 3,5-
dicaffeoyl-muco-quinic acid was identified. In Asimina triloba 3-caffeoyl-muco-quinic acid
was also identified. 13-15 In Chrysanthemum morifolium 3,5-dicaffeoyl-epi-quinic acid and
1,3-dicaffeoyl-epi-quinic acid was identified. 16 3,5-dicaffeoyl-epi-quinic acid esters were also
reported in Ilex kudingcha. 17 These muco, epi and scyllo esters of diastereomers of quinic
acid are reported to show important biological activities like, hepatoprotectivity, antioxidant
activity and anti HIV-1 integrase activity. 13-17
Considering the fact that regiosomers as well as the esters of the diastereomers of quinic acids
are readily distinguishable by their fragmentation pattern in tandem MS experiments, it is
very important to acquire the authentic synthetic standards for the diastereomers of the quinic
acid to obtain selective diastereomers of chlorogenic acids by organic synthesis. Development
of the mass spectrometrical methods for diastereomers of chlorogenic acids will enable us to
identify the novel chlorogenic acid derivatives present in the biological samples even in very
low concentrations. Since, these biological samples emerge biosynthetically, either through
various processes such as, roasting, brewing, cooking etc. or as products of metabolism; their
mass spectrometrical study will help us understand the changes in the chlorogenic acid profile
through biosynthetic processes in great details. Additionally, this study will contribute
immensely to answer if diastereomeric compounds possess biological activities and if they are
easily distinguishable by mass spectrometry. By this, the role of the tandem MS as a powerful
tool for the structural elucidation will be emphasized.
Almost all of the studies involving analysis of the diastereomers of the quinic acid until now
are either based on the non-selective isomerization of the quinic acid approach, which
involves slightly ambiguous assignments of the diastereomers isolated by chromatographic
techniques or by extraction of the quinic acid derivatives from plant material. In this
contribution we have selectively synthesized muco-, epi-, cis- and scyllo-quinic acids using
91
appropriate functional group protection and deprotection strategies confirmed by analytical
techniques like NMR or single crystal XRD. We also report the behavior of the diastereomers
in the reverse phase HPLC in terms of the retention times and elution order and basic features
of their tandem mass spectra using mass spectrometry as a reliable and predictive tool to
assign the stereochemistry to even comparatively smaller molecules like (-)-quinic acid.
4.2 Experimental
Chemicals
All the chemicals (analytical grade) were purchased from Sigma-Aldrich (Bremen, Germany)
and were used without further purification.
LC/MSn
The LC equipment (Agillent 1100 series, Bremen, Germany) comprised a binary pump, an
auto sampler with a 100 µL loop, and a DAD detector with a light-pipe flow cell (recording at
254 and 320 nm and scanning from 200 to 600 nm). This was interfaced with an ion-trap mass
spectrometer fitted with an ESI source (Bruker Daltonics HCT Ultra, Bremen, Germany)
operating in full scan, auto MSn mode to obtain fragment ion m/z. As necessary, MS2, MS3,
and MS4 fragment-targeted experiments were performed to focus only on compounds
producing a parent ion at m/z 191, 173. Tandem mass spectra were acquired in Manual-MSn
mode using fixed collision energy. The fragmentation amplitude was set to 0.75 V. Also,
direct injection experiments targeting the fragments in MS2, MS3, and MS4 were performed
on all diastereomers of quinic acid keeping the fragmentation amplitude constant at 1.0 volts.
MS operating conditions (negative mode) was optimized using (-)-quinic acid with a capillary
temperature of 365oC, a dry gas flow rate of 10 L/min, and a nebulizer pressure of 10 psi.
HPLC
Separation was achieved on a 250 × 4.6 mm i.d. column containing diphenyl 5 µm and 5 ×
4.6 mm i.d. guard column of the same material (Varian, Darmstadt, Germany). Solvent
(water: formic acid 1000:0.05 v/v) was delivered at a total flow rate of 800 µL/min by 30 min
isocratic.
1.2.1 Synthesis of the mixture of the epimers of (-)-quinic acid
The mixture of the epimers of (-)-quinic acid was obtained non-selectively by the process
already described by Maier and co-workers. 6
92
4.2.2 Synthesis of the epi-quinic acid (2)
Methyl quinate was prepared by refluxing quinic acid (1) (5000 mg, 26.02 mmol) with MeOH
(100 ml) and Amberlite IR120 acidic resin (5000 mg) for 12 hours. The reaction mixture was
then filtered and concentrated in vacuo. The product was obtained in more than 95 % yield
and the purity of the product was confirmed by NMR. It was then subjected to selective silyl
protection on C3 and C5 of the methyl quinate by tert-Butyldimethylsilyl chloride
(TBDMSCl). 18
OH
HO
OH
HO
O
OH
O
HO
OH
O
O
O
Si SiO
HO
O
O
O
O
Si Si
SOO
OH
HO
O
O
O
O
+
HO
OH
O
O
O
O
O
HO
O
O
O
O
Si SiO
HO
OH
O
O
O
Si Si
OH
HO
OH
HO
O
OH
OH
HO
OH
HO
O
OH
+
1 7 8
9 102 1
a, b c
df
g
e
HO O
11
g1+5
12
Reagents and conditions: (a) MeOH, reflux, 12 h, Amberlite IR 120, 100%; (b) 2.6 equiv
TBDMSCl, 2.6 equiv Et3N, DMF 0 0C, 2 h and then to RT, 16 h 75%; (c) 1.5 equiv MsCl,
Py, RT, overnight, 95%; (d) 11, CsF, DMF, 90 0C, 24 h 2% 9 and 10; (e) 2 equiv Dess-
Martin periodinane, CH2Cl2 , overnight, RT, 87%; (f) 1.5 equiv NaBH4, Ethanol, -30 0C, 40
min.; (g) 2M HCl, H2O, 1 h.
Figure 4.2 Reaction scheme for obtaining scyllo-quinic acid (5) and epi-quinic acid (2)
Methyl quinate (2650 mg, 12.86 mmol) was stirred with triethylamine (5 ml) in anhydrous
dimethylformamide (26 ml) in an inert atmosphere and to this solution, tert-
Butyldimethylsilyl chloride (5034 mg, 33.436 mmol) was added and the mixture was stirred
for 2 hours at 0 0C and 16 hours at room temperature. EtOAc was added and the residue was
93
filtered. Concentrated filtrate was purified using flash chromatography (gradient eluent: 20%
EtOAc in petroleum ether) to afford white crystalline methyl 3,5-Di-O-(tert-
butyldimethylsilyl)quinate (7) in 75% yield, which was confirmed by crystal XRD (see
Figure 4.3 and 4.4) to have free C4 on quinic acid skeleton. Compound 7 was used as a
precursor for the synthesis of epi-quinic acid (2) and scyllo-quinic acid (5) as shown in the
Figure 4.2.
Figure 4.3 X-ray structure of methyl 3,5-Di-O-(tert-butyldimethylsilyl)quinate (7) Conformer
A
94
Figure 4.4 X-ray structure of methyl 3,5-Di-O-(tert-butyldimethylsilyl)quinate (7) Conformer
B
Dess-Martin periodinane (390 mg, 0.92 mmol) to the solution of compound 7 (200 mg, 0.46
mmol) and dichloromethane (15 ml) at room temperature. Reaction was stirred overnight and
diluted with Et2O. Then to this, 1:1 saturated mixture of Na2S2O3 and NaHCO3 solution was
added and stirred until reaction becomes almost clear. Organic layer was collected and the
aqueous layer was extracted with 20 ml EtOAc thrice. Combined organic layers were
collected, dried and concentrated in vacuo. Crude product was subjected to flash
chromatography (gradient eluent: 24% EtOAc in petroleum ether), which afforded methyl
3,5-Di-O-(tert-butyldimethylsilyl)-4-oxoquinate (12) in the form of sticky colorless solid in
87%. See Figure 4.2. 13C NMR (100 MHz, CDCl3): 206.51, 173.15, 75.98, 75.33, 69.81,
52.94, 46.71, 41.22, 25.79, 25.61, 18.45, 17.88, -4.73, -5.04, -5.25, -5.30
95
Compound 12 was subjected to reduction at C4 by the action of NaBH4 and L-selectride. Out
of these two reduction procedures, reduction with NaBH4 showed higher selectivity towards
the formation of epi- derivative as compared to L-selectride. 177 mg of compound 12 (0.4074
mmol) was dissolved in ethanol and the reaction flask was immersed in an acetone bath.
Liquid nitrogen was slowly added until the bath temperature reaches to -30 0C. NaBH4 (23
mg, 0.6111 mmol) was added and the mixture was stirred for 40 minutes. Solvent was
removed immediately under reduced pressure at 30 0C in a rotary evaporator. The residue was
then extracted with water and EtOAC mixture three times. Organic layers were collected and
dried over Na2SO4 and concentrated. 13C NMR of the crude product confirmed reduction at
C4 indicated by the loss of a peak at 206 ppm. The crude product was directly subjected to
hydrolysis by 2M HCl and water without further purification. Product of the hydrolysis was
diluted with water and extracted with water and EtOAC mixture thrice. Aqueous layers were
collected and concentrated in vacuo. Resulting white product was used in HPLC/MS analysis.
4.2.3 Synthesis of the muco-quinic acid (3)
muco-quinic acid (3) was obtained from the methyl TMB-muco-quinate, which was
synthesized by the procedure already reported by Jaiswal et al.. 8 Methyl TMB-muco-quinate
(500 mg, 1.56 mmol) was subjected to hydrolysis by 70% aqueous trifluoroacetic acid for 1
hour. Resulting methyl muco-quinate was stirred with 2M KOH for 20 minutes followed by
neutralization by Amberlite IR 120 acidic resin for 10 minutes. The mixture was filtered and
concentrated under reduced pressure. The resulting yellowish solid was analyzed by NMR. 1H
NMR (400 MHz, D2O): 1.73 (dd, 1H, H6ax, 3JHH 12.4), 1.75 (dd, 1H, H2ax,
3JHH 12.4), 1.86
(dd, 1H, H6eq, 2JHH 4.1, 3JHH 11.2), 1.89 (dd, 1H, H2eq,
2JHH 4.6, 3JHH 11.4), 3.21 (t, 1H, H4,
3JHH 9.6), 3.62 (ddd, 2H, 2JHH 4.6, H3 and H5, 3JHH 11.9, 9.1). 13C NMR (100 MHz, D2O):
180.9, 79.5, 74.7, 69.8, 69.7, 40.5, 40.4
4.2.4 Synthesis of the cis-quinic acid (4)
The synthetic scheme is shown in Figure 4.5. 3,4-O-Cyclohexylidene-1,5-quinide (13) was
synthesized by adding quantities of 10.00 g (52.04 mmol) of quinic acid and 200 mg (1.05
mmol) of p-toluenesulfonic acid monohydrate (PTSA·H2O) to 100 mL of cyclohexanone to
give a white suspension. The reaction was then refluxed for 24 h to give a yellow solution,
which was cooled to 50 ºC and neutralized with a solution of NaOEt (71.5 mg) in EtOH (5
mL) to give a yellow clear solution. The solvents were removed under reduced pressure and
to the resulting yellow viscous liquid a volume of 100 mL of EtOAc was added. The organic
phase was washed with 50 mL of H2O and the aqueous phase was back-extracted with 30 mL
96
EtOAc. The combined organic layers were washed with a half-saturated NaHCO3 solution,
dried on Na2SO4, filtered and evaporated. The resulting yellow viscous liquid was
recrystallized successively from a 1:1 n-heptane:EtOAc solution to afford white crystals of
compound 13 (9.26 g, 36.43 mmol, 70%); 1H-NMR (CDCl3): δH 4.73 (dd, 1H, J = 5.6, 2.8
Hz), 4.48 (td, 1H, J = 6.8, 2.8 Hz), 4.29 (ddd, 1H, J = 6.4, 2.8, 1.4 Hz), 2.66 (d, 1H, J = 11.9
Hz), 2.38-2.31 (ddd, 1H, J = 14.7, 7.8, 2.3 Hz), 2.31-2.25 (ddt, 1H, J = 11.9, 6.4, 2.3 Hz),
2.17 (dd, 1H, J = 14.7, 3.2 Hz), 1.73-1.68 (m, 2H), 1.67-1.60 (m, 2H), 1.57-1.51 (m, 4H),
1.43-1.36 (m, 2H); 13C-NMR (CDCl3): δC 178.9 (-COOR), 110.7 (C-1'), 76.1 (C-4), 71.8 (C-
1), 71.6 (C-3), 71.2 (C-5), 38.6 (C-6), 37.0 (C-2), 34.5 (C-6'), 33.7 (C-2'), 25.1 (C-5'), 24.0
(C-3'), 23.6 (C-4').
OH
HO
OH
HO
OH
OHO
O
O
O
OHO
O
O
O
OH
OHO
O
O
O
O
O
HO
O
O
O
O
OH
HO
OH
HO
O
OH
OH
HO
OH
HO
O
O
OH
1 13 14 15
16
17
4
a b c
d
e
f
Reagents and conditions: (a) cyclohexanone, PTSA·H2O, reflux, 24 h; (b) 21%
NaOMe/MeOH, MeOH, rt, overnight; (c) Dess-Martin periodinane, DCM, rt, 18 h; (d)
NaBH4, MeOH/THF (1:1), -30 ˚C, 1 h; (e) KOH, THF, rt, 45 min; (f) HCl (trace amounts in
CDCl3).
Figure 4.5 Reaction scheme for obtaining cis-quinic acid (4) and Methyl-cis-quinate (17)
97
3,4-O-Cyclohexylidene-1,5-quinide (13) (8.75 g, 34.41 mmol) was dissolved in 100 mL
MeOH and a 21% solution NaOMe/MeOH was added (187 mg NaOMe). The clear solution
was stirred overnight, the mixture was then quenched with glacial acetic acid (232 µL) and
the volatile components were removed under vacuum. The resulting mixture was dissolved in
EtOAc and washed 3 times (3x40 mL). The organic layer was dried over Na2SO4, filtered and
the solvent was removed under low pressure. The crude product was purified by column
chromatography on silica gel (20-50% EtOAc/petroleum ether) to give methyl 3,4-O-
cyclohexylidene-quinate (14) as a white solid (5.87 g, 20.51 mmol, 60%). 1H-NMR (CDCl3):
δH 4.46 (dt, 1H, J = 5.9, 3.7 Hz), 4.14-4.07 (m, 1H), 3.97 (t, 1H, J = 5.9 Hz), 3.79 (s, 3H),
2.26 (m, 1H), 2.25 (d, 1H, J = 4.1 Hz), 2.08 (ddd, 1H, J = 13.4, 4.1, 3.0 Hz), 1.86 (dd, 1H,
13.7, 11.0 Hz), 1.74-1.68 (m, 2H), 1.68-1.52 (m, 6H), 1.44-1.33 (m, 2H); 13C-NMR (CDCl3):
δC 175.6 (-COOCH3), 110.1 (C-1'), 79.4 (C-4), 74.1 (C-1), 73.1 (C-3), 68.7 (C-5), 53.0 (-
CH3), 39.0 (C-6), 38.0 (C-2), 34.9 (C-6'), 34.8 (C-2'), 25.0 (C-5'), 24.1 (C-3'), 23.7 (C-4').
Figure 4.6 X-ray structure 3,4-O-Cyclohexylidene-1,5-quinide (13)
To a suspension of Dess-Martin periodinane (2963 mg, 6.99 mmol) in anhydrous CH2Cl2 (65
mL) compound 14 (1.82 g, 6.36 mmol) was added. The reaction mixture was stirred at room
temperature for 18 h, was then diluted with Et2O (100 mL) and a 1:1 mixture (v/v) of
saturated aqueous Na2S2O3 and NaHCO3 solution (100 mL). The mixture was stirred until the
98
solids were dissolved (20 min). The aqueous layer was extracted with Et2O and the combined
organic layers were dried over Na2SO4, filtered and concentrated. Product methyl 3,4-O-
cyclohexylidene-5-oxoquinate (15) (1.81 g, 6.36 mmol, 100%) was used for the next step
without further purification. 1H-NMR (CDCl3): δH 4.72 (m, 1H), 4.41 (d, 1H, J = 5.5 Hz),
3.81 (s, 3H), 2.89 (d, 1H, J = 14.2 Hz), 2.80 (dd, 1H, J = 14.7, 2.3 Hz), 2.56 (t, 1H, J = 2.3
Hz), 2.55 (d, 1H, J = 4.1 Hz), 1.74-1.64 (m, 2H), 1.64-1.56 (m, 4H), 1.56-1.47 (m, 2H), 1.43-
1.32 (m, 2H); 13C-NMR (CDCl3): δC 204.4 (C-5), 173.0 (-COOCH3), 111.7 (C-1'), 78.2 (C-4),
76.9 (C-1), 75.9 (C-3), 53.5 (-CH3), 49.1 (C-6), 37.0 (C-2), 35.3 (C-6'), 34.7 (C-2'), 24.9 (C-
5'), 23.9 (C-3'), 23.8 (C-4').
To obtain methyl 3,4-O-cyclohexylidene-epi-quinate (16), compound 15 (1.53 g, 5.33 mmol)
was dissolved in a 1:1 mixture (v/v) MeOH/THF (100 mL) and was cooled to -30 ˚C with an
acetone/liquid nitrogen bath. NaBH4 (222 mg, 5.86 mmol) was added and the mixture was
stirred at -30 ˚C for 1 h. The solvents were removed in vacuum and the residue was extracted
three times with a water/EtOAc mixture. The organic layers were dried over Na2SO4, filtered
and the solvent was removed under reduced pressure. The product was purified by column
chromatography (40% EtOAc/petroleum ether). 1H-NMR (CDCl3): δH 4.52 (dt, 1H, J = 7.6,
5.0 Hz), 4.30 (dd, 1H, J = 7.6, 4.1 Hz), 3.90 (dt, 1H, J = 9.6, 4.1 Hz), 3.78 (s, 3H), 2.18-2.06
(m, 3H), 2.02 (dd, 1H, J = 14.2, 10.1 Hz), 1.75 (m, 2H), 1.69-1.59 (m, 4H), 1.59-1.51 (m,
2H), 1.44-1.35 (m, 2H); 13C-NMR (CDCl3): δC 175.4 (-COOCH3), 110.0 (C-1'), 73.8 (C-4),
73.4 (C-1), 72.5 (C-3), 66.1 (C-5), 53.0 (-CH3), 38.1 (C-6), 36.4 (C-2), 35.6 (C-6'), 34.0 (C-
2'), 25.2 (C-5'), 24.1 (C-3'), 23.7 (C-4').
99
Figure 4.7 X-ray structure of cis-quinic acid (4)
Crystals of cis-quinic acid (4) (Figure 4.7) suitable for single crystal XRD were obtained in
an NMR tube containing demethylated 16 dissolved in CDCl3 by removal of the acid-labile
cyclohexylidene protection promoted by the trace amounts of HCl present in the deuterated
solvent. 1H-NMR (D2O): δH 3.76 (br, 1H), 3.67 (t, 1H, J = 3.2 Hz), 3.64 (t, 1H, J = 3.2 Hz),
1.97 (dd, 2H, J = 12.4, 4.1 Hz), 1.62 (t, 2H, J = 12.4 Hz); 13C-NMR (D2O): δC 177.1 (-
COOCH3), 72.4 (C-4), 71.1 (C-1), 66.9 (C-3, C-5), 35.9 (C-2, C-6).
4.2.5 Synthesis of the scyllo-quinic acid (5)
Methyl 3,5-Di-O-(tert-butyldimethylsilyl)quinate (7) (500 mg, 1.15 mmol) was mixed and
stirred with 12 ml of anhydrous pyridine and cooled to 0 0C. Methanesulfonyl chloride was
added drop wise to the mixture. Reaction was stirred at room temperature for overnight. 10 ml
NaHCO3 added and after stirring for 10 minutes, aqueous phase was washed two times with
Et2O. Organic layers were collected, dried over Na2SO4 and concentrated. Crude product was
subjected to flash chromatography (gradient eluent: 22% EtOAc in petroleum ether), which
afforded methyl 3,5-Di-O-(tert-butyldimethylsilyl)-4-O-methanesulfonylquinate (8) in 95%
yield. 1H NMR (400 MHz, CDCl3): 0.06 (s, 3H, MeSi), 0.08 (s, 3H, MeSi), 0.11 (s, 3H,
MeSi), 0.12 (s, 3H, MeSi), 0.86 (s, 9H, Me3CSi), 0.88 (s, 9H, Me3CSi), 1.94 (dd, 1H, H6ax,
2JHH 13.74, 3JHH 10.07, Hz), 2.00-2.10 (m, 1H, H2ax), 2.12-2.31 (m, 2H, H6eq, H2eq), 3.04 (s,
3H, MeS), 3.77 (s, 3H, -OMe), 4.25 (dd, 1H, H4, 3JHH 8.7, 2JHH 2.3 Hz), 4.34 (m, 1H, H5) and
4.57 (m, 1H, H3).
100
Compound 8 (520 mg, 1.014 mmol) was dissolved in anhydrous dimethylformamide and to
this was added cesium fluoride (790 mg, 5.2 mmol). The mixture was stirred for 30 minutes
and cinnamic acid (770 mg, 5.2 mmol) was added. The reaction was heated to 90 0C and
allowed to stir for 36 hours. DMF was removed under low pressure and the obtained crude
product was subjected to column chromatography (gradient eluent: 45% EtOAc in petroleum
ether). One fraction collected from the column chromatography confirmed the formation of
the desired product in the NMR analysis and formed a crystal in the NMR tube. The crystal
was analyzed by crystal XRD without detailed assignment of the NMR signals. XRD showed
a rare phenomenon of co-crystallization of 3-O-cinnamoyl-1,4-scyllo-quinide (9) and 3-O-
cinnamoyl-1,5-quinide (10) as can be seen in Figure 4.8.
The mixture of compound 9 and compound 10 was subjected to hydrolysis through the same
procedure as described earlier. Product of the hydrolysis was diluted with water and extracted
with water and EtOAC mixture thrice. Aqueous layers were collected and concentrated in
vacuo. Resulting white product was used in HPLC/MS analysis.
Figure 4.8 X-ray structures of 3-O-cinnamoyl-1,4-scyllo-quinide (9) and 3-O-cinnamoyl-1,5-
quinide (10)
101
4.2.6 Hydrolysis of the CGAs in roasted coffee
10 grams of finely grounded roasted coffee from Guatemala was boiled in distilled water for 1
hour. The mixture was then filtered and the water in the filtrate was removed under reduced
pressure. Resulting residue was subjected to base hydrolysis by the treatment of 25 ml 2M
NaOH for 2 hours. The reaction was neutralized by addition of 2M HCl. The mixture was
diluted with 25 ml water and the aqueous layer was extracted with EtOAC thrice. The
aqueous layers were collected and concentrated in vacuo. Resulting yellowish product was
used directly in HPLC/MS analysis.
4.2.7 Synthesis of the methyl esters of epi-, muco-, cis-, scyllo-quinic acids and (-)-quinic
acid
As described earlier, methyl quinates of all the synthesized diastereomers except the methyl
cis-quinate (17) were prepared by refluxing respective diastereomers in MeOH with equal
amount of Amberlite IR120 acidic resin for 12 hours. The reaction mixture was then filtered
and concentrated in vacuo. The resulting product was used directly in HPLC/MS analysis in
case of methyl epi-quinate, methyl scyllo-quinate and methyl muco-quinate. Crystals of
methyl cis-quinate (17) suitable for single crystal XRD (Figure 4.9) were obtained in an
NMR tube containing methyl 3,4-O-cyclohexylidene-epi-quinate (16) dissolved in CDCl3 by
removal of the acid-labile cyclohexylidene protection promoted by the trace amounts of HCl
present in the deuterated solvent. The methyl quinate of (-)-quinic acid was confirmed by the
NMR and then utilized in HPLC/MS.
102
Figure 4.9 X-ray structure of methyl cis-quinate (17)
4.2.8 X-ray crystallography
Crystals were mounted on a Hampton cryoloop in light oil for data collection at 100 K.
Indexing and data collection were performed on a Bruker D8 SMART APEX II CCD
diffracto-meter with κ geometry and Mo Kα radiation (graphite mono- chromator, λ =
0.71073 Å). Data integration was performed using SAINT. Routine Lorentz and polarization
corrections were applied. The SHELX package was used for structure solution and
refinement. Refinements were full-matrix least-squares against F2 using all data. In the final
refinement, all non-hydrogen atoms were refined anisotropically and hydrogen atoms were
either found directly and refined isotropically or placed in calculated positions.
Crystallographic data are summarized in Table 4.2.
4.3 Results and discussions
Several methods for the synthesis of the epi-quinic acid (2) have been reported, 18 out of
which two methods were selected for further modification. Most of the other methods rely on
formation of 1,5-quinide followed by selective protection of C3 by silyl protecting group and
then achieve inversion of configuration at C4 by oxidation-reduction reactions. This approach
was found out to be unreliable and highly problematic, since the protection was observed to
be non-selective. In the present work, once the methodology to obtain compound 7 in pure
103
form was developed, the ambiguity in the selective protection was eliminated in order to leave
C4-OH free for further treatment (Figure 4.2). Methyl 3,5-Di-O-(tert-
butyldimethylsilyl)quinate (7) served as a precursor for both of the diastereomers, epi-quinic
acid (2) and scyllo-quinic acid (5). Out of which, diastereomer 2 was obtained by oxidation-
reduction pathway and compound 5 was synthesized by incorporating SN2 reaction strategy, in
which cinnamic acid served the role of nucleophile attacking C4 from backside to achieve
inversion of the configuration. Acids such as, benzoic acid and 3,4-dimethoxycinnamic acid
were also unsuccessfully used to serve as nucleophiles. However, in fact the desired product
was found out to have rearranged itself in the form of 3-O-cinnamoyl-1,4-scyllo-quinide (9)
possibly due to the traces of the HCl present in CDCl3 in the NMR sample tube. 3-O-
cinnamoyl-1,4-scyllo-quinide(9) and 3-O-cinnamoyl-1,5-quinide (10) were co-crystallized
and were obtained in very low yield. Hence, attempts towards the isolation of compound 9
were avoided and the crystals were directly hydrolyzed to get the mixture of scyllo-quinic
acid (5) and quinic acid.
Implementation of the protection strategies to obtain cis-quinic acid (4) was observed to be
less complicated if compared to epi-quinic acid (2) and scyllo-quinic acid (5). Protection on
C3 and C4 in 1, 2 acetal formation and formation of 1,5-quinide can be achieved
simultaneously by using cyclohexanone as previously described by Fernandez et al.. 19 Once
the C1- C5 lactone was cleaved with sodium methoxide, C5-OH becomes available for
oxidation-reduction reactions to achieve inversion at C5 (Figure 4.7). From the synthetic
products and intermediates, a total of five single crystal x-ray structures could be obtained.
cis-quinic acid (4) was obtained as a single crystal. scyllo-quinic acid (5) was obtained by
hydrolyzing the co-crystal of Compounds 3-O-cinnamoyl-1,4-scyllo-quinide (9) and 3-O-
cinnamoyl-1,5-quinide (10). muco-quinic acid (3) was obtained in quantitative yield from the
hydrolysis of Methyl TMB-muco-quinate. The co-eluting mixture of epi-quinic acid (2) and
quinic acid (1) was purified by column chromatography and was analyzed as a mixture in LC-
MS. We have investigated all of the above mentioned compounds by chromatography as
single compounds and as a mixture by optimization of the separation methods on reversed
phase HPLC packings. We also performed the non-selective synthesis of all the diastereomers
from (-)quinic acid using prolonged heating in sulfuric acid as described previously 6 in an
attempt of the identification and assignment of the remaining diastereomers of quinic acid,
which have not been obtained in this study by synthetic methods.
104
Table 4.1 MS2 data of quinic acid diastereomers in negative ion mode at 75% collision
energy
Compound MS1 MS2
No.
Parent
ion
Base
peak Secondary peak
m/z m/z int % m/z int % m/z int % m/z int %
1 191 127 173 82 85 56 93 51 111 23
2 191 173 127 86 145 30 85 24 111 17
3 191 127 173 64 85 45 109 20 145 19
4 191 93 173 37 111 34 155 7 61 5
5 191 173 127 91 85 50 111 48 93 46
If we observe the retention times of all the diastereomers, we see that they elute very close to
each other (Figure 4.10). The elution order can be given as, 1>5>2>3=4. Where quinic acid
(1) elutes at 3.9 minutes and muco- and/or cis-quinic acid (3/4) elute at 3.4 minutes. Although
diastereomers 3 and 4 elute at the same retention time, their tandem MS shows clear
distinction as can be seen in Figure 4.10 and Figure 4.13. Diastereomers 2 and 5 elute very
close to each other and they fragment almost identically in MS2 however in MS3, scyllo-
quinic acid (5) produces a base peak at m/z 93 whereas epi-quinic acid (2) gives a base peak at
m/z 111 in MS3. In other case, (-)-quinic acid (1) and muco-quinic acid (3) show similar
fragmentation pattern giving base peaks at m/z 127 and m/z 109 in MS2 and MS3 respectively
but, they differ in the retention times with compound 3 eluting half a minute earlier than
compound 1. Also, to find a difference between diastereomers 1 and 3 on the basis of their
tandem MS, we had to fragment them further into MS4, which will be discussed in detail later
in this paper. Next to tandem LC-MSn experiments, direct infusion tandem MS experiments
were carried out wherever necessary to see the further fragmentation of a particular base peak
ion. The MS2 fragmentation all of the diastereomers was confirmed by keeping collision
energy constant in the LC-MS runs at 0.75 volts (see Table 4.1). Optimum value for the
collision energy was standardized by taking (-)-quinic acid (1) as a standard. Roasted coffee
hydrolysis experiment shows us that diastereomers epi-quinic acid (2), muco-quinic acid (3)
and cis-quinic acid (4) are present along with quinic acid (1) in the roasted coffee as products
of the food processing (Figure 4.12). Based on the fragmentation behavior and retention
times learned from the earlier experiments, we could only identify diastereomers 2, 5 and 1
from the EIC of the non-selectively isomerized quinic acid as shown in Figure 4.11.
105
Figure 4.10 EIC at m/z 191 in negative mode and MSn of quinic acid (1), epi-quinic acid (2),
muco-quinic acid (3), cis-quinic acid (4) and scyllo-quinic acid (5) (Continued)
EIC 191.0
0
1
2
3
7 x10
Intens.
2 4 6 8 10 Time [min]
(-)-quinic acid (1)
61.5
85.2
93.0 110.9
154.9
172.7
188.6
126.8 -MS2(190.7), 3.9min
85.5
108.9 -MS3(190.8>126.9), 4.0min
0
100
[%]
0
100
40 60 80 100 120 140 160 180 200 m/z
(1)
106
Figure 4.10 EIC at m/z 191 in negative mode and MSn of quinic acid (1), epi-quinic acid (2),
muco-quinic acid (3), cis-quinic acid (4) and scyllo-quinic acid (5) (Continued)
EIC 191.0
0.0
0.5
1.0
1.5
6 x10
Intens.
1.5 2.0 2.5 3.0 3.5 4.0 4.5 Time [min]
2
1 epi-quinic acid (2)
59.5
85.1
108.8
126.8
144.8
172.7
-MS2(190.7), 3.6min
111.0
154.8
-MS3(190.8->172.5), 3.7min 0
100
[%]
0
100
50 100 150 200 250 m/z
(2)
107
Figure 4.10 EIC at m/z 191 in negative mode and MSn of quinic acid (1), epi-quinic acid (2),
muco-quinic acid (3), cis-quinic acid (4) and scyllo-quinic acid (5) (Continued)
EIC 191.0
0
2
4
6
7 x10 Intens.
2.5 5.0 7.5 10.0 12.5 15.0 17.5 Time [min]
muco –quinic acid (3)
85.1
108.9
144.8
154.7
172.7
126.8 -MS2(190.7), 3.4min
85.1
108.8
-MS3(190.8->126.8), 3.5min 0
100
[%]
0
100
40 60 80 100 120 140 160 180 200 m/z
(3)
108
Figure 4.10 EIC at m/z 191 in negative mode and MSn of quinic acid (1), epi-quinic acid (2),
muco-quinic acid (3), cis-quinic acid (4) and scyllo-quinic acid (5) (Continued)
EIC 191.0
0
2
4
6
8
7 x10
Intens.
2 4 6 8 10 Time [min]
cis-quinic acid (4)
59.1 86.1
110.9
136.8 154.7
172.7
93.1
-MS2(190.7), 3.4min
0
20
40
60
80
100
[%]
40 60 80 100 120 140 160 180 200 m/z
(4)
109
Figure 4.10 EIC at m/z 191 in negative mode and MSn of quinic acid (1), epi-quinic acid (2),
muco-quinic acid (3), cis-quinic acid (4) and scyllo-quinic acid (5)
EIC 191.0
0
1
2
3
4
6 x10
Intens.
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 Time [min]
scyllo-quinic acid (5)
(1)
71.3
85.1
110.9
126.8
154.7 188.7
172.6
-MS2(190.7), 3.7min
110.9 127.8 152.7
93.0 -MS3(190.8->173.3), 3.8min
0
100
[%]
0
100
50 100 150 200 250 300 m/z
(5)
110
Figure 4.11 MSn of the acidic fraction of non-selectively isomerized quinic acid (Continued)
Acid fraction: TIC -All MS
0.5
1.0
1.5
7 x10 Intens.
3.00 3.50 4.00 4.50 Time [min]
2
5 1
85.1 108.9
126.8
144.8
172.7
-MS2(190.7), 3.5min
71.4
93.1
110.8 126.7 154.8
-MS3(190.8->172.8), 3.6min 0
100
[%]
0
100
50 100 150 200 250 m/z
(2)
71.3
85.1 108.9
126.8
144.8
172.7
-MS2(190.7), 3.7min
93.1
111.0
127.0 154.7
-MS3(190.8->172.6), 3.8min 0
100
[%]
0
100
50 100 150 200 250 m/z
(5)
111
Figure 4.11 EIC at m/z 191 in negative mode and MSn of the acidic fraction of non-
selectively isomerized quinic acid
Figure 4.12 Diastereomers identified in the EIC of the hydrolyzed roasted coffee in negative
ion mode (Continued)
71.4
85.1 110.9 144.8
172.7
126.8 -MS2(190.7), 4.0min
73.3
108.9
-MS3(190.9->126.9), 4.0min 0
100
[%]
0
100
50 100 150 200 250 m/z
(1)
EIC 191.0
0.0
0.2
0.4
0.6
0.8
1.0
6 x10 Intens.
1 2 3 4 5 6 Time [min]
Hydrolyzed roasted coffee
3 1
5 2
112
Figure 4.12 Diastereomers identified in the EIC of the hydrolyzed roasted coffee in negative
ion mode (Continued)
73.3 108.8 144.7
172.6
126.9 -MS2(190.7), 3.5min
108.8
-MS3(190.9>126.8), 3.5min 0
100
[%]
0
100
50 100 150 200 250 300 m/z
3
85.2 110.9 144.7
172.7
126.8 -MS2(190.8), 4.0min
-MS3(190.9>127.0), 4.1min 0
100
[%]
0
100
50 100 150 200 250 300 m/z
108.9
1
71.3
110.9
126.8
144.7
172.7
93.1 -MS2(190.7), 3.6min
0
20
40
60
80
100
[%]
50 100 150 200 250 300 m/z
5
113
Figure 4.12 Diastereomers identified in the EIC of the hydrolyzed roasted coffee in negative
ion mode
85.1 110.9
126.8
144.8 188.7
172.7 -MS2(190.7), 3.7min
65.8
93.1
110.9
152.7
-MS3(190.8>172.6), 3.8min 0
100
[%]
0
100
50 100 150 200 250 300 m/z
4
114
Figure 4.13 MSn of the diastereomers of the quinic acid and reference compounds obtained
from the direct infusion experiments (Continued)
85.1 108.9
126.8
144.8
172.8
188.6
-MS2(191.0)
57.8 85.1
108.9
-MS3(191.0>127.0)
-MS4(191.0>127.0>109.0)
0
100
[%]
0
100
0
100
40 60 80 100 120 140 160 180 200 220 m/z
80.8
1
82.4
126.8
144.8
172.7 172.7 172.7 172.7 172.7 172.7 172.7
-MS2(191.0)
-MS3(191.0>173.0) 0
100
[%]
0
100
60 80 100 120 140 160 m/z
111
93 127.8
2
115
Figure 4.13 MSn of the diastereomers of the quinic acid and reference compounds obtained
from the direct infusion experiments (Continued)
99.0
108.9
-MS3(191.0->127.0)
-MS4(191.0->127.0>109.0) 0
100
[%]
0
100
50 60 70 80 90 100 110 120 130 140 m/z
81.2
3
73.3 83.2
93.1
99.0
110.9
142.8
154.7
-MS2(173.0)
83.2 93.1
136.8
110.9 110.9 110.9 110.9 110.9 110.9 110.9 -MS3(173.0>155.0)
-MS4(173.0>155.0>111.0)
0
100
[%]
0
100
0
100
40 60 80 100 120 140 160 180 m/z
93.1 83.2
Shikimic acid
116
Figure 4.13 MSn of the diastereomers of the quinic acid and reference compounds obtained
from the direct infusion experiments (Continued)
107.9 -MS2(109.0)
-MS3(109.0->108.0) 0
100
[%]
0
100
40 60 80 100 120 140 160 180 200 m/z
77
OH
OHHydroquinone
81.3 -MS2(109.0)
0
100
[%]
50 60 70 80 90 100 110 120 130 140 m/z
OH
OH
Catechol
117
Figure 4.13 MSn of the diastereomers of the quinic acid and reference compounds obtained
from the direct infusion experiments
We observed that the tandem MS of each of the diastereomers is different therefore allow
unambiguous identification. In the following mechanisms of fragmentation, rationalizing
different fragment spectra are discussed and proposed. As shown in scheme A in Figure 4.14,
(-)-quinic acid (1) undergoes dehydration and decarboxylation to produce a base peak at m/z
127 to give fragment Q1 in MS2. For dehydration in quinic acid, there are in principle three
alternate mechanistic pathways available. Firstly, formation of lactone, secondly, elimination
of water leading to a cyclohexene and finally, epoxide formation. To confirm the mechanistic
pathways, wherever possible, fragment ions generated from the precursor ions were compared
to reference substances or derivatives to assign the structure of the fragment ion as lactone,
65.5
107.0
-MS2(109.0)
0
100
[%]
40 60 80 100 120 140 160 180 200 220 m/z
OH
OH
Resorcinol
-MS2(97.0)
0
100
[%]
20 40 60 80 100 120 140 m/z
81.2
O
Cyclohexene oxide
118
cyclohexene or epoxide. We have to assume that these two fragmentations take place
simultaneously as we see no secondary peak at m/z 145, which signifies decarboxylated
quinic acid. Carboxyl group on C1 leaves as a formic acid and the dehydration takes place in
between C2 and C3 due to the fact that the proton on C2 and hydroxyl group on C3 share an
anti-periplanar arrangement. Q1 at m/z 127 further fragments in MS3 to give base peak at m/z
109, which suggests the presence of dihydroxybenzene-type structure. But, since C4 and C5 do
not possess the anti-periplanar arrangement between either protons and hydroxyl groups all
three 1,3-, 1,4- and 1,2-dihydroxybenzenes do not show fragmentation similar to the MS4 of
quinic acid (1). In fact, the fragmentation of cyclohexene oxide in MS2 gives only peak at m/z
81 similar to MS4 of (-)-quinic acid (1) as can be seen in Figure 4.13 suggests the presence of
an epoxide between C4 and C5 (Q2).
Scheme B in Figure 4.14 shows the proposed fragmentation pathway for epi-quinic acid (2).
The fragmentation can be explained if we consider that the elimination of water through an E2
route takes place between C1 and C2 to give shikimic acid with a base peak at m/z 173 (Q3). A
comparison of the MS3 spectrum of shikimic acid with the MS3 spectrum of 2 with Q3 at m/z
173 as a precursor ion, confirms this assignment. Q3 shows a fragment spectrum similar to
that of shikimic acid. The direct injection experiment shows that m/z 111 is a primary peak in
the MS3 of epi-quinic acid (2) (Figure 4.13). It can be only explained through the assumption
of the presence of Q4 arising through the dehydration in between C4 and C5, which possess
the anti-periplanar arrangement of a proton and –OH group as can be seen in the Figure 4.14.
Fragmentation of the muco-quinic acid (3) is explained in scheme C. Carboxyl group on C1
leaves as a formic acid and the dehydration takes place in between C2 and C3 due to anti-
periplanar arrangement similar to the fragmentation in (-)-quinic acid (1) to give base peak in
MS3 at m/z 127 (Q5). As shown in the scheme, when we invert the cyclohexane ring we see
the anti-periplanar geometry between C2-H and C3-OH due to which, the dehydration between
C2 and C3 takes place. As can be observed in the direct injection experiment in Figure 4.13,
m/z 109 fragments further to give m/z 81.2 similar to the MS4 of (-)-quinic acid (1). This
means that the fragments Q2 and Q6 are in fact same molecules. Only difference in the
fragmentation of the diastereomers 1 and 3 is that in the MS2 of 3, we observe a secondary
peak at m/z 145, which signifies the decarboxylated molecule (Figure 4.10). In the structural
arrangement of the diastereomer 3, absence of the hydrogen bonding between 1,3-di axially
positioned oxygen anion on C1 and hydrogen on C3-OH does not facilitate the elimination of
C3-OH as observed in case of (-)-quinic acid (1).
119
O
O
OH
OH
OH
OH
-HCO2H
OH
OH
OH
OH
H
OH
O
O
-H2O
O
O
A
m/z 127
Q1 Q2
m/z 109
O
O
OH OH
OHH-H2O
O
O
OH
OH
-HCO2
OH
OH
1
6
23
45
6
OH
OH
HOH
5
43
1
2
-H2O
OH
HO
m/z 111m/z 173
Q3 Q4
B
HO
O
OH
OHH
C
OH OH OH
OHOH
-HCO2H
OH
OH
H
OH
OH
Om/z 127
OHOH
O
H
H
OH
OH
OH
23
-H2O
Q5 Q6
-H2O
-H2O
D O
O
OH
OH
OHOH
H
HO
O
OH
OHOH
OH
OHOH
H
m/z 93
O
Q9
m/z 173
Q7
OH
O
H
m/z 111
Q8
O
O
m/z 109
O
H
OH
OH
O
O
OH
O
(M-H) of (-)-quinic acid (1)
(M-H) of epi-quinic acid (2)
(M-H) of muco-quinic acid (3)
(M-H) of cis-quinic acid (4)
MS2 MS3
MS2 MS3
MS3MS2 MS4
MS2
Figure 4.14 Proposed mechanisms for the fragmentation of 1, 2, 3 and 4
120
MS2 of the cis-quinic acid (4) gives base peak at m/z 93 along with the secondary peaks at m/z
173 and m/z 111. Base peak at m/z 93 suggests that the decarboxylation and dehydration
processes in the molecule had taken place simultaneously to attain aromaticity in the
molecule. Unfortunately, phenol as an external standard and m/z 93 in compound 4 does not
fragment further as observed in direct injection experiments where the collision energy was
kept constant at 1.0 volts. But, we can propose mechanistic evidence as presented in scheme
D in Figure 4.14 to prove that m/z 93 can be in fact, a phenolic ion. In the first step of the
simultaneous fragmentation phenomena, intermediate Q7 is formed due to the anti-periplanar
arrangement similar to the MS2 of epi-quinic acid (2). We assume that the dehydration takes
place between C1 and C2 before decarboxylation due to the fact that the elimination of C1-OH
is assisted by the hydrogen bonding provided by 1,3-diaxially positioned hydroxyl groups on
C3 and C5. Decarboxylation takes place leaving negative charge on C6. Possibility of the
hydrogen bonding between the oxygen on C3-OH and hydrogen on the C5-OH considering the
1,3-diaxial arrangement, C3-OH leaves leaving a double bond between C6 and C5. The anti-
periplanar geometry between C4-H and C3-OH helps the molecule to achieve aromaticity
through dehydration to give Q8, which shows up as a secondary peak at m/z 111 in MS2.
scyllo-quinic acid (5) shows unique fragmentation pattern if compared to the other
diastereomers in present work. In MS2 it shows the pattern similar to the MS2 of the epi-
quinic acid (2) and in MS3, m/z 173 directly fragments into m/z 93 similar to the MS2 of the
cis-quinic acid (4) as can be seen in Figure 4.10. It is unclear at this point of time that how
the fragmentation scheme in case of 5 can be explained mechanistically.
4.4 Discussion of the X-ray structures
In the crystal of cis-quinic acid (4) (Figure 4.7), we observed two molecules in the
asymmetric unit representing the same structure. Compounds 3-O-cinnamoyl-1,4-scyllo-
quinide (9) and 3-O-cinnamoyl-1,5-quinide (10) were co-crystallized along with one molecule
of chloroform in one asymmetric unit. The crystal structure of 3,5-Di-O-(tert-
butyldimethylsilyl)quinate (7) shows conformational disorder in one of the silyl groups (Si1A
and Si1B), both conformations are shown in Figure 4.3 (Conformer A) and Figure 4.4
(Conformer B). The disorder was modeled and showed a value of 50% for each conformation.
Table 4.2 illustrates the crystal data and structure refinement for compounds methyl cis-
quinate (17), cis-quinic acid (4), 3-O-cinnamoyl-1,4-scyllo-quinide(9), 3-O-cinnamoyl-1,5-
quinide (10), 3,4-O-Cyclohexylidene-1,5-quinide (13) and methyl 3,5-Di-O-(tert-
butyldimethylsilyl)quinate (7).
121
Table 4.2 Crystal data and structure refinement for compounds 17, 4, 9, 10, 15 and 7
Compound 17 Compound 4 Compounds
9 and 10
Compound 15 Compound 7
Formula Unit C8H14O6 C14H24O12 C33H33Cl3O12 C13H18O5 C19H27O6Si2
Formula
weight
(g/mol)
206.19 384.33 727.94 254.27 407.59
Crystal
System
Orthorhombic Monoclinic Monoclinic Orthorhombic Orthorhombic
Space group P212121 P21/c C2 P212121 P212121
a (Å) 6.3490(2) 6.2971(2) 26.5045(12) 5.5977(6) 7.3775(5)
b (Å) 9.6841(3) 19.7482(6) 5.9464(3) 10.6426(11) 11.1129(7)
c (Å) 14.8574(5) 12.8471(4) 22.5590(12) 19.6591(19) 31.4879(19)
α (°) 90 90 90.00 90 90
β (°) 90 94.218(2) 112.380(3) 90 90
δ (°) 90 90 90.00 90 90
Volume (Å3) 913.50(5) 1593.29(9) 3287.6(3) 1171.2(2) 2581.5(3)
Z 4 4 4 4 4
Dcalc (g/cm3) 1.499 1.602 1.471 1.442 1.049
Abs. Coeff. µ
(mm-1)
0.129 0.142 0.344 0.110 0.163
Temperature 100(2) K 100(2) K 100(2) K 100(2) K 100(2) K
Total
reflections
35008 95659 36946 45621 66496
Min-max θ
(°)
3.49 – 30.51 3.40 – 30.50 3.50 – 24.71 3.65 – 30.51 3.56 – 27.48
Unique
reflections
2762 4848 5559 3521 5552
Calculated
reflection (I >
2σ)
2532 4111 4229 3432 4739
Final R1* 0.0308 0.0355 0.0565 0.0289 0.0883
wR2* 0.0813 0.1010 0.1500 0.0787 0.2501
Rint 0.0525 0.0557 0.1044 0.0527 0.0733
Goodness of
Fit
1.006 1.001 1.018 1.016 1.019
Parameters 183 331 434 166 249
Restraints 0 0 1 0 0
Largest Peak/
Deepest Hole
0.291/-0.229 0.543 /-0.210 0.467/-0.533 0.365/-0.199 1.262/-0.506
Flack
Parameter
0.0(6) n/a -0.02(9) -0.3(5) 0.0 (3)
* R1 = Σ║F(obs)│- │F(calc) ║/ Σ │F(obs) │; wR2 = { Σ[w(Fo2 – Fc
2)2] / Σ[w(Fo2)2]}1/2
122
4.5 Conclusions
In conclusion, we have selectively synthesized all the diastereomers of the quinic acid except
for the neo-quinic acid (6) to prove that they can be chromatographically resolved and can be
identified by their characteristic fragmentation behavior in tandem MS spectra. This study
also illustrates the importance of the tandem mass spectrometry in the area of identification
and confirmation of the stereochemistry of comparatively smaller molecules like the
diastereomers of the quinic acid. It is also worth mentioning that it is the first time the crystal
structures of 3-O-cinnamoyl-1,4-scyllo-quinide (9) , 3-O-cinnamoyl-1,5-quinide (10), cis-
quinic acid (4), methyl cis-quinate (17) and 3,4-O-Cyclohexylidene-1,5-quinide (13) are
being reported. The fragmentation mechanisms are explained on the basis of the
conformational differences between the diastereomers of the quinic acid. Presence of the
diastereomers such as, muco-quinic acid (3), scyllo-quinic acid (5) and epi-quinic acid (2) in
hydrolyzed roasted coffee sample proves the existence of the diastereomerized chlorogenic
acids in roasted coffee. However, it still remains unclear if the epimerization occurs after the
esterification with cinnamoyl functionality or the cinnamoyl group esterifies with epimerized
quinic acid during the food processing.
123
References
1. E.L. Eliel, M.B. Ramirez. (-)-quinic acid: configurational (stereochemical) descriptors Tetrahedron-
Asymmetry. 1997, 8, 3551.
2. Y. Koshiro, M.C. Jackson, R. Katahira, M. Wang, C. Nagai, H. Ashihara. Biosynthesis of
chlorogenic acids in growing and ripening fruits of Coffea arabica and Coffea canephora plants. Z.
Naturforsch., C: J. Biosci. 2007, 62, 731.
3. M. Weers, H. Balzer, A. Bradbury, O.G. Vitzthum. Analysis of acids in coffee by capillary
electrophoresis. Colloq. Sci. Int. Cafe, [C. R.]. 1995, 16th, 218.
4. O. Frank, G. Zehentbauer, T. Hofmann. Bioresponse-guided decomposition of roast coffee beverage
and identification of key bitter taste compounds. Eur. Food Res. Technol. 2006, 222, 492.
5. A. Crozier, I.B. Jaganath, M.N. Clifford. Dietary phenolics: chemistry, bioavailability and effects
on health. Nat. Prod. Rep. 2009, 26, 1001.
6. B.M. Scholz-Boettcher, L. Ernst, H.G. Maier. New stereoisomers of quinic acid and their lactones.
Liebigs Ann. Chem. 1991, 1029.
7. Anonymous. IUPAC Commission on the Nomenclature of Organic Chemistry (CNOC) and
IUPAC-IUB Commission on Biochemical Nomenclature (CBN). Nomenclature of cyclitols.
Recommendations, 1973 Biochem J. 1976, 153, 23.
8. R. Jaiswal, M.H. Dickman, N. Kuhnert. First diastereoselective synthesis of methyl caffeoyl- and
feruloyl-muco-quinates. Org. Biomol. Chem. 2012, 10, 5266.
9. M.F. Matei, R. Jaiswal, N. Kuhnert. Investigating the Chemical Changes of Chlorogenic Acids
during Coffee Brewing: Conjugate Addition of Water to the Olefinic Moiety of Chlorogenic Acids and
Their Quinides. J. Agric. Food Chem. 2012, 60, 12105.
10. A. Crozier. Recent studies on the bioavailability of dietary (poly)phenolics. Abstracts of Papers,
245th ACS National Meeting & Exposition, New Orleans, LA, United States, April 7-11, 2013. 2013,
AGFD.
11. C. Bennat, U.H. Engelhardt, A. Kiehne, F.M. Wirries, H.G. Maier. HPLC ANALYSIS OF
CHLOROGENIC ACID LACTONES IN ROASTED COFFEE Zeitschrift Fur Lebensmittel-
Untersuchung Und-Forschung. 1994, 199, 17.
124
12. K. Schrader, A. Kiehne, U.H. Engelhardt, H.G. Maier. Determination of chlorogenic acids with
lactones in roasted coffee J.Sci.Food Agric. 1996, 71, 392.
13. K.H. Kim, Y.H. Kim, K.R. Lee. Isolation of quinic acid derivatives and flavonoids from the aerial
parts of Lactuca indica L. and their hepatoprotective activity in vitro Bioorg.Med.Chem.Lett. 2007, 17,
6739.
14. M. Haribal, P. Feeny, C.C. Lester. A caffeoylcyclohexane-1-carboxylic acid derivative from
Asimina triloba Phytochemistry. 1998, 49, 103.
15. H.C. Kwon, C.M. Jung, C.G. Shin, J.K. Lee, S.U. Choi, S.Y. Kim, K.R. Lee. A new caffeoyl
quinic acid from Aster scaber and its inhibitory activity against human immunodeficiency virus-1
(HIV-1) integrase. Chem. Pharm. Bull. 2000, 48, 1796.
16. H.J. Kim, Y.S. Lee. Identification of new dicaffeoylquinic acids from Chrysanthemum morifolium
and their antioxidant activities Planta Med. 2005, 71, 871.
17. P.T. Thuong, N.D. Su, T.M. Ngoc, T.M. Hung, N.H. Dang, N.D. Thuan, K. Bae, W.K. Oh.
Antioxidant activity and principles of Vietnam bitter tea Ilex kudingcha Food Chem. 2009, 113, 139.
18. L. Sanchez-Abella, S. Fernandez, N. Armesto, M. Ferrero, V. Gotor. Novel and Efficient
Syntheses of (-)-Methyl 4-epi-Shikimate and 4,5-Epoxy-Quinic and -Shikimic Acid Derivatives as
Key Precursors to Prepare New Analogues. J. Org. Chem. 2006, 71, 5396.
19. S. Fernandez, M. Diaz, M. Ferrero, V. Gotor. New and efficient enantiospecific synthesis of (-)-
methyl 5-epi-shikimate and methyl 5-epi-quinate from (-)-quinic acid. Tetrahedron Lett. 1997, 38,
5225.
125
CHAPTER 5: Transesterification of chlorogenic acids with small organic
acids present in the coffee bean
5.1 Introduction
Coffee is one of the most traded agricultural commodities in the world. More than 70
countries in the world cultivate coffee, which grows in the form of ‘cherries’ on a coffee
plant. Brewed coffee is the third most consumed beverage after water and black tea.1 In
countries like USA and Germany coffee is second most consumed beverage. Approximately,
450 million cups of coffee are consumed every day in United States only. The conventional
coffee whether instant, filter or freshly ground is made from roasting, grinding and brewing
the green beans of coffee obtained from cherry fruits of either Arabica or Robusta varieties of
the coffee plant.
The perceived taste of the coffee is largely contributed by the acids present in the coffee. In
fact, acidity in coffee is considered as one of the important parameters for the quality of the
coffee variety.2 11 % of the total weight of the green coffee bean is contributed by the acid
content, which decreases upon roasting to 6 %.3 This acid content is contributed by various
volatile and non-volatile acids. In this work we will discuss about the non-volatile fraction of
the total acid content. As Clifford reported earlier, in brewed coffee citric acid, phosphoric
acid, phytic acid, quinic acid (1), chlorogenic acids and malic acid are the most important
acids contributing to the perceived acidity.2 Other free organic acids present in roasted coffee
such as oxalic acid (4), malonic acid (5), glutaric acid (7), adipic acid (8), tartaric (11) and
succinic acid (6) 4-15 need to be considered as factors affecting the acidic taste as well. These
acids do not contribute to the titrable acidity of the coffee as established by Engelhard and
Maier 16 but they might exist as anions in a coffee brew or in a coffee extract providing
protons.
Apart from small organic acids, various authors have also studied the presence of free fatty
acids (FFA) in coffee.17-19 FFA content of various green arabica and robusta coffee ranges
from 0.8 to 3.0 g/100 g lipid.16 In the commercially available roasted coffees in German
market it was found to be in the ranges of 0.8-1.8 g/100 g lipid, whereas 1.2-2.5 g/100 g lipid
in French market.16 So, it is clear that although upon roasting FFA decrease slightly, they are
present in coffee in roasted coffee in considerable amount. Among the reported FFA
constituents in the roasted coffee, linoleic acid and palmitic acid are prominently present.16, 20
126
Approximately 10% of the total composition of the processed seeds of green Coffea
canephora (robusta coffee) comes from chlorogenic acids on a dry weight basis.21 50% of the
chlorogenic acids are lost through decomposition during roasting by producing half of the
decomposition products such as, quinic acid and hydroxycinnamic acids through hydrolysis.1,
21, 22 Quinic acid (1) ranges from 3 to 6 g/kg in green robusta and Arabica coffee beans from
various origins in the free form as reported earlier.15, 23 After steaming the green coffee bean,
original quinic acid content rises by up to 15% as shown by Hucke and Maier.24 Similarly,
roasting also helps to elevate the quinic acid content.3, 25, 26 Although chlorogenic acid
lactones and quinic acid lactones are among the degradation products from chlorogenic acid,
free quinic acid in roasted coffee is still found out in roasted coffee to maintain its existence
between the ranges of 6.63 to 9.47 g/kg.27 Among the other non-volatile organic acids, citric
and malic acid (9 and 10) are present in green coffee in the ranges of 5 to 15 g/kg in arabica
and 3 to 10 g/kg in robusta respectively16 12 to 18 % of these acids are degraded during
roasting process. Among many other decomposition products of citric acid occurring as a
result of the various processes, succinic acid (6) and glutaric acid (7) are included in this work
to study their significance towards the formation of transesterification products produced in
coffee roasting.28
Despite considerably large amount of free non-volatile organic acids, degradation products
and chlorogenic acids themselves exist in the roasted coffee, the understanding about the fate
of these so-called melanoidine products remains hugely obscure. The green or roasted coffee
beans contain a number of compounds such as, caffeine, trigonelline, lipids, phytosterols and
series of small organic acids, which needs to be accounted for in analytical studies in future.
Kuhnert et al. used mass spectrometry to characterize the chemical components of coffee
melanoidines. They observed that the CGA derivatives undergo further condensation during
coffee roasting to produce transesterified products such as, isomers of acetyl-caffeoyl quinic
acid. They also found CGA derivatives condensed or transesterified with quinic acid and
shikimic acid.1 Additionally, their work indicated the presence of the further esterification
products of chlorogenic acids with small organic acids, which are present in the roasted coffee
such as, maloyl caffeoyl quinic acid (at m/z 469.09876, C20H21O13) by studying the data
generated by FT-ICR-MS measurements of the roasted coffee samples.1 This finding stands as
an inspiration for the present work.
Considering the fact that relatively large amount of the degradation products of CGAs such
as, quinic acid (1), caffeic acid along with the most prominent member of the CGAs profile in
127
roasted coffee, 5-caffeoylquinic acid itself are present in the roasted coffee along with free
small, non-volatile organic acids, we must entertain the possibility of the further esterification
phenomenon among themselves. To investigate transesterification in roasted coffee in details
we designed a thorough analytical plan involving four experiments. We have taken a set of
small organic acids and heated each of them individually with 5-CQA (2) to check if
simulated roasting conditions facilitate the formation of the transesterification products. Same
experimental conditions were used incorporating caffeic acid and quinic acid as well. Also,
we heated 5-CQA (2), caffeic acid (3) and quinic acid (1) with the mixture of all the organic
acids separately to check, which of the organic acid show greater affinity towards the
formation of the condensed esters. The set of eight organic acids contained oxalic acid,
malonic acid, succinic acid, glutaric acid, adipic acid, citric acid, malic acid and dextrotartaric
acid (Figure 5.1). They were chosen on the basis of their reported occurrence in green or
roasted coffee and solid state. In each of the experiment, the roasting conditions were
simulated by keeping the temperature at 200 oC for the duration of 12 minutes. All the
samples acquired from these experiments were analyzed by high resolution ESI-TOF-MS.
Four green coffee samples were also roasted in the conditions described earlier and then
analyzed by ESI-FT-ICR-MS to identify the transesterification products in roasted coffee
samples.
In addition to these experiments, to study the possibility of the transesterification between
quinic acid and free fatty acids present in the roasted coffee, we performed similar
experiments involving quinic acid and linoleic and/or palmitic acid.
Finally, mass spectral data obtained for the model roasting was compared to those obtained
early on FT-ICR-MS data on roasted coffee extract and the presence of tentative
transesterification products was thus confirmed in roasted coffee beans.
5.2 Materials and methods
5.2.1 Chemicals and materials
All the chemicals (analytical grade) were purchased from Sigma-Aldrich (Bremen,
Germany). The range of small organic acids was selected to purchase by their reported
occurrence in coffee and solid state. Roasted coffee (Robusta) samples (powder) were
obtained from the Kraft Foods Bremen (Germany).
128
OH
OHO
O
HO OH
OO
OH
O
O
HOOH
OO
HO
OH
O
HO
OOH
OO
HO
O OH
OH
HO
O OH
OH
O
OH
HO
O OH
OH
O
OH
HO
HO
OH
O
OH
O
HO
HO
OH
O
OH
O
OH
OH O
OH
OHHO
1 2 3
4 5 6 7
8 9 10 11
Oxalic acid Malonic acid Succinic acid Glutaric acid
Adipic acid Citric acid Malic acid Tartaric acid
(-)-qunic acid 5-Caffeoylquinic acid Caffeic acid
Figure 5.1 Structures of all the reactants involved in the model roasting experiments
5.2.2 Model roasting
Two different roasting experiments were carried out taking 5-CQA (2), quinic acid and
caffeic acid as substrates and oxalic acid (4), malonic acid (5), succinic acid (6), glutaric
acid (7), adipic acid (8), citric acid (9), malic acid (10) and dextrotartaric acid (11) (Figure
1) as reactants.
5 mg (0.014 mmol) of 5-CQA was heated with equimolar quantity of each organic acid (4-
11) separately. 75 mg (0.212 mmol) of 5-CQA was heated with 0.125 equivalents of each
organic acid collectively. 10 mg (0.055 mmol) of caffeic acid was heated with equimolar
amount of each organic acid separately and then 100 mg (0.55 mmol) of caffeic acid was
heated with 0.125 equivalents of each organic acid collectively. Quinic acid (1) (10 mg,
0.052 mmol) was heated with equimolar amount of each organic acid separately and then
100 mg (0.55 mmol) of quinic acid (1) was heated with 0.125 equivalents of each organic
acid collectively. Lastly, quinic acid (5 mg, 0.026 mmol) was heated with equimolar
amounts of linoleic acid and palmitic acid separately and then quinic acid (50 mg, 0.26
129
mmol) was heated with 0.5 equivalents of linoleic and palmitic acid collectively.
All of the samples were heated in a Buechi Glass Oven B-585 at 200 oC for 12 minutes
and then dissolved in methanol for LC-MS/MSn analysis.
5.2.3 Aqueous extract of roasted coffee
Four different samples according to their origin of roasted coffee powder (5 g of each)
were extracted with water by Soxhlet extraction using distilled water for 5 h. The extract
was treated with Carrez reagents (1 mL of reagent A plus 1 mL of reagent B) to
precipitate colloidal material and subsequently filtered through a Whatman no. 1 filter
paper. The water was removed in vacuo and the residue was stored at -20 oC until
required, thawed at room temperature, dissolved in methanol (6 g/L), filtered through
membrane filter and used for LC-MS.
5.2.4 Roasted coffee samples for ESI-FT-ICR-MS analysis
The samples were prepared by the procedure already reported by Jaiswal et al.1
LC-MSn
The LC equipment (Agillent 1100 series, Bremen, Germany) comprised a binary pump, an
auto sampler with a 100 µL loop and a DAD detector with a light-pipe flow cell
(recording at 320 and 254 nm and scanning from 200 to 600 nm). This was interfaced with
an ion-trap mass spectrometer fitted with an ESI source (Bruker Daltonics HCT Ultra,
Bremen, Germany) operating in full scan, auto MSn mode to obtain fragment ion m/z.
Tandem mass spectra were acquired in Auto-MSn mode (smart fragmentation) using a
ramping of the collision energy. Maximum fragmentation amplitude was set to 1 Volt,
starting at 30% and ending at 200%. MS operating conditions (negative mode) had been
optimized using 5-caffeoylquinic acid with a capillary temperature of 365 oC, a dry gas
flow rate of 10 L/min and a nebulizer pressure of 10 psi.
LC-TOF-MS
High Resolution LC-MS experiments were carried out using the same HPLC equipped
with a MicrOTOF Focus mass spectrometer (Bruker Daltonics, Bremen, Germany) fitted
with an ESI source. An internal calibration was achieved with 10 mL of 0.1 M sodium
formate solution injected through a six port valve prior to each chromatographic runs.
Calibration was carried out using the enhanced quadratic calibration mode.
HPLC
Separation was achieved on a 150 x 3 mm i.d. column containing diphenyl 5 µm, with a 5
130
mm x 3 mm i.d. guard column (Varian, Darmstadt, Germany). Solvent A was
water/formic acid (1000:0.05 v/v) and solvent B was methanol. Solvents were delivered at
a total flow rate of 500 µL/min. The gradient profile was from 10% B to 70% B linearly in
70 min followed by 10 min isocratic, and a return to 10% B at 90 min and 10 min isocratic
to re-equilibrate.
ESI-FT-ICR mass spectrometry
Ultra high resolution mass spectra were acquired using a Bruker solarix Fourier transform
Ion Cyclotron Resonance mass spectrometer (FTICR-MS) with a 12 T refrigerated
superconducting cryo-magnet. The instrument was equipped with a Dual electrospray ion
source with ion funnel technology. Spectra of the coffee samples were acquired in
electrospray ionization positive and negative ion mode using direct infusion with a syringe
pump with a flow rate of 120 μl/h. Stock solutions (1 mg/mL in MeOH) of the samples
were diluted 1:40 in 50:50 MeOH: water for negative ion mode and 1:20 in 49.95:49.95
MeOH : water and 0.1% formic acid for positive ion mode. The drying gas temperature of
the ion source was set to 200 °C. The instrument was externally calibrated with arginine
cluster using a 10 μg/ml arginine solution containing 50% methanol. Spectra were
acquired with 4 MW resulting in a resolving power of 500000 at m/z 400. Data were zero
filled once to a data size of 8 MW. A single sine apodization was performed prior to
Fourier transformation of the time domain signal. After spectra acquisition the mass
spectra were internally calibrated with known polyphenolic pseudo-molecular ions. The
ion accumulation time was set to 0.1 s and 300 scans were accumulated and added up for
each mass spectrum in a mass range between m/z 200 and 3000.
131
OH
O
O
O
HO
OH
O
O
OH
OH
OH
O OH
O
O
HO
OH
O
O
OH
OH
O
O
O
HO
OH
O
O
O
13 14 15
O
OH
OH
O
HO
OH
O
O
OH
OH
O
OH
O
O
16
OH
O
HO
OH
O
O
O
O
OH
O
O
17
O
OH
OO
HO O
O
O
O OH
O
O
OH
O
18 19
OH
O
HO
OH
O
O
O
O
O OH
O
20
OH
O
HO
O
O
O
OH
OH
OH
21
HO
O
OH
O
O
O
HO
OH
O
HO
OH
O
O
OH
OH
23
OH
O
O O
HO O
O
OH
O
OH
O
24
22
OH
O
O O
HO O
O
OH
O
OH
OHO O
OH
25
O
OO
O
OH
OH
OO
HO OH
O
OH
OH
O
O
O
HO
OH
O
O
O
H
12
O
OHOOH
O
HO
O
O
O
O
OH
O
OHO
Figure 5.2 Tentative structures of transesterification products (Continued)
132
OH
O
OO
HO OH
O
OH
O
O
O
OH
OH
29
OH
O
HO
OH
O
O
OH
OH
O
O
O
OH
OH
28
OH
OH
O O
HO OH
O
OO
OH
O
OH
26 27
O
OHO
OH
OH
OH
O O
HO O
O
HOO
O
O
OH
O
O
OH
OO
HO O
O
OH
O
O OH
OH
O
30
OH
O
OO
HO O
O
OH
O
OH
O
O
31
OH
OO
HO O
O
OH
O
O
OH
O
32
OH
O
OO
HO OH
O
OH
O
O
O
OH
OH
33
OH
O
O
O
HO
OH
O
O
O
OH
O
OH
OH
34
OH
O
O
O
HO
OH
O
O
OH
OH
O
OH
OH
OH
35
O
O
HO
O O
OHO
HO
36
OH
HO
OH
O
O
HO
O
OH
38
OHO
O
OH
OH
O
HO
O
37
HO
O
OH
OHHO
O
O
O
HO
HO
O
O
HO
O
O
O
HO
39 40
Figure 5.2 Tentative structures of transesterification products (Continued)
133
HO
O
O
HO
O
O O
OH
41
HO
O
OH
OHHO
O OO
OH
HO
O
O
HO
O OO
OH
43
HO
O
OH
OHHO
HO
O
OH
O O
O
HO
HO
O
O
HO
HO
O
OH
O O
O
HO
HO
O
O
HO
O
OH
O O
O
HO
HO
O
OH
OHHO
O
OH
O O
O
HO
44 45
45*
HO
O
OH
OHHO
OH
O
HO
O
O
HO
O
O
HO
OH
O
HO
O
O
HO
O
O
HOO
HO
O
O
49
OH
OHO
O
O
OH
HO
O
O
HO
42
46
47 48
50
OHO
O
O
O
O
OHO
OHO
51
OHO
OH
O
O
52
OHO
O
O
O
O
53
OHO
O
O
O
OHO
OHO O
54
OH
OH
OO
OH
O
OH55
OH
O
OO
HO
O
OH
O
OH
O
56
OHO
O
O
O
HO
OH
57
OH
O
OO
HO
OH
O
O
OH
O
58
OH
O
OO
HO
OH
O
OH
O
O
OH
59
Figure 5.2 Tentative structures of transesterification products (Continued)
134
O
O
71
O
OHO
OH
OH
O
O
OHO
OH
OHO
72
OHO
O
OH
OH
OHO
HO
O
61
OHO
OHO
OOH
O
60
O
OH
HO
OH
OO
H
HO
62OH
O
O
O
OH
O
O
OH
H
63
OH
OH
OO
HO O
O
OH
O
O
OH
O
64
OH
O
OH
OHHO
O
O
HO OH
65
OH
O
OH
OHO
HO
O
HO
OH66
OH
O
O
OHHO
HO
O
OH
HO67
O
O
OH
O
HO
O
HO
OH68
O
O
O
HO
HO
O
OH
HO69
O
O
OH
HO
O
O
HO OH
70
Figure 5.2 Tentative structures of transesterification products of 1, 2 and 3 from LC-MSn,
LC-TOF-MS and FT-ICR-MS data (regiochemistry of the condensation products is randomly
selected).
135
5.3 Results and discussion
Aim of this work is to achieve a better understanding of the chemical transformations
occurring at very fine level in chlorogenic acids profile during food processing, which
gives rise to interesting melanoidine reaction products. Analysis of melanoidines fraction
in roasted coffee is a complicated venture, which was undertaken in this contribution with
the help of the so called ‘Domino-Tandem MS’ approach proposed by Kuhnert.1 This
approach combines two powerful modern mass spectrometry tools developed in recent
times- high resolution MS and LC-tandem MS. High resolution MS provides ultimate
resolution in the molecular weight dimension and provides lists of molecular formulas
corresponding to melanoidine compounds, whereas LC-tandem MS provides resolution at
isomeric level combined with gas phase isolation of targeted ions, adding structural
information via fragment spectra.
In the first step of the work, potential transesterification products of the respective
substrates with each of the organic acid were postulated with the help of basic chemical
reactivity principles. Roasting conditions were simulated to acquire the condensation
products of 5-CQA (2), quinic acid (1) and caffeic acid (3) with the range of selected
small organic acids, which were reported to be present in various green and roasted coffee
samples in the literature.
Secondly, identification of these transesterification products was undertaken in the
samples generated by the model roasting experiments and in the roasted coffee samples.
Tandem LC-MS was employed to obtain further structural information and likely reaction
mechanisms involved in the formation of the transesterification products , which may have
formed in the model roasting experiments as well as in the coffee samples. The analytical
strategy involving high resolution FT-ICR-MS measurements as direct infusion
experiments to obtain ultimate resolution in the molecular weight dimension and generate
molecular formula lists of all compounds present in the coffee sample was also
incorporated.
The absence of the reference standards is a limitation encountered by presented work in
the area of the assignment of the regio- and stereo-chemistry of the focused group of
melanoidine fraction. However, employment of the combined analytical strategies in this
work has enabled us to open up to the new possibilities of exploring roasted coffee
melanoidine fraction for condensation products of acidic profile.
136
5.3.1 Transesterification of 5-CQA (2) in model roasting and in roasted coffee samples
Figure 5.3 shows the representative UV chromatogram of the model roasting experiment
sample generated by heating 5-CQA (2) with succinic acid (6). In the model roasting
experiment where oxalic acid and 5-CQA were heated together we found four peaks, out
of which three were identified as the isomers of caffeoyllactones as, formoyl-
caffeoylquinide (12a-12c) and the fourth peak was assigned as formoyl-caffeoylshikimate
(63) in LC-TOF-MS (Table 5.1). Malonic acid when heated with 5-CQA generated five
peaks of isomeric esters, two of caffeoyllactones and three of caffeoylquinic acids as
follows, acetoyl-caffeoylquinide (14a, 14b) and acetoyl-caffeoylquinic acid (15a-15c).
Succinic acid gave eight peaks containing four isomers of succinoyl-caffeoylquinic acid
(16a-16d), two isomers of succinoyl-caffeoylquinide (17a-17b) and two isomers of di-O-
succinoyl-caffeoylquinic acid (18a, 18b). In case of glutaric acid with 5-CQA, we
observed nine peaks generated by, six isomers of gluteroyl-caffeoylquinic acid (19a-19f)
and three isomers of gluteroyl-caffeoylquinide (20a-20c). Adipic acid+5-CQA formed six
isomeric compounds, out of which four were identified as the isomers of adipoyl-
caffeoylquinic acid (21a-21d) and two as the isomers of adipoyl-caffeoylquinide (22a,
22b). Citric acid when heated with 5-CQA formed the highest number of
transesterification products totaling to sixteen, which contained gluteroyl-caffeoylquinide
(20b), three isomers of citroyl-caffeoylquinic acid (23a-23c), four isomers of 24, which is
possibly an ester of degraded form of citric acid with quinic acid lactone, four isomers of
citroyol-caffeoylquinde (25a-25d) and four isomers of methy-citroyl-caffeoylquinide
(26a-26d). Fourteen peaks were observed in the EIC’s of respective m/z in negative mode
of malic acid+ 5-CQA roasting. Four were generated by the isomers of maloyl-
caffeoylquinic acid (28a-28d), two by isomeric methyl-maloyl-caffeoylquinic acid (29a,
29b), three by the isomers of maloyl-caffeoylquinide (30a-30c), three by isomeric methyl-
maloyl-caffeoylquinide (31a-31c) and two peaks were generated by fumaroyl-
caffeoylquinide (32). Lastly, tartaric acid showed five peaks in model roasting with 5-
CQA contributed by, two isomers of tarteroyl-caffeoylquinide (34a, 34b) and three
isomers of tarteroyl-caffeoylquinic acid (35a-35c).
In the model roasting experiment of 5-CQA with the range of organic acids, we identified
the total of 67 transesterification products in LC-TOF-MS listed in Table 5.1. Most of the
transesterification products are formed with the lactones of the chlorogenic acid. Formoyl-
caffeoyl-shikimate (63) and fumaroyl-caffeoyl-1,5-quinide (32b) were identified as only
137
two shikimate based dehydrated transesterification products arising from chlorogenic acid.
The fragmentation pattern of these two compounds is shown in Table 5.2. In these, we
can see that m/z 335, which is identical for the lactones and shikimates in MS2; fragments
into m/z 161 and m/z 179 in MS3 only in the case of 63 and 32b respectively.
138
Table 5.1 Transesterification products of 4-11 with 5-CQA (2) identified with LC-TOF-MS in the samples generated by heating each acid
separately with 5-CQA
No. Compounds
involved
Retention
time (min)
Condensation
product No.
Mol. Formula Theoretical m/z
(M-H)
Experimental m/z
(M-H)
Error
(ppm)
1 Oxalic acid+5-CQA 35.3 12a C17H15O9 363.0722 363.0726 1.1
2 36.3 63 C17H15O9 363.0722 363.0724 0.7
3 38.1 12c C17H15O9 363.0722 363.0706 4.2
4 51.3 12d C17H15O9 363.0722 363.0732 2.8
5 Malonic acid+5-CQA 36.6 14a C18H17O9 377.0878 377.0886 2.2
6 39.8 14b C18H17O9 377.0878 377.0874 1.1
7 25.3 15a C18H19O10 395.0984 395.0967 4.2
8 33.0 15b C18H19O10 395.0984 395.0976 1.8
9 37.1 15c C18H19O10 395.0984 395.0972 3.0
10 Succinic acid+5-CQA 27.3 16a C20H21O12 453.1038 453.1024 3.1
11 28.4 16b C20H21O12 453.1038 453.1036 0.6
12 32.2 16c C20H21O12 453.1038 453.1036 0.7
13 33.6 16d C20H21O12 453.1038 453.1043 1.0
139
14 38.7 17a C20H19O11 435.0933 435.0942 2.0
15 41.4 17b C20H19O11 435.0933 435.0949 3.7
16 49.3 18a C24H23O14 535.1093 535.1070 4.3
17 50.6 18b C24H23O14 535.1093 535.1109 2.9
18 Glutaric acid+5-CQA 31.2 19a C21H23O12 467.1195 467.1204 2.0
19 32.6 19b C21H23O12 467.1195 467.1201 1.2
20 34.0 19c C21H23O12 467.1195 467.1209 3.1
21 35.0 19d C21H23O12 467.1195 467.1203 1.8
22 35.8 19e C21H23O12 467.1195 467.1200 1.0
23 39.1 19f C21H23O12 467.1195 467.1185 2.2
24 41.7 20a C21H21O11 449.1089 449.1094 1.0
25 44.0 20b C21H21O11 449.1089 449.1101 2.6
26 44.8 20c C21H21O11 449.1089 449.1097 1.8
27 Adipic acid+5-CQA 35.1 21a C22H25O12 481.1351 481.1344 1.6
28 38.0 21b C22H25O12 481.1351 481.1359 1.5
29 38.8 21c C22H25O12 481.1351 481.1357 1.1
140
30 39.7 21d C22H25O12 481.1351 481.1351 0.2
31 44.9 22a C22H23O11 463.1246 463.1254 1.8
32 47.9 22b C22H23O11 463.1246 463.1268 4.8
33 Citric acid+5-CQA 43.0 20b C21H21O11 449.1089 449.1096 4.2
34 26.6 23a C22H23O15 527.1042 527.1056 2.6
35 28.6 23b C22H23O15 527.1042 527.1050 1.4
36 29.0 23c C22H23O15 527.1042 527.1053 2.0
37 42.8 24a C21H19O11 447.0933 447.0924 1.9
38 43.6 24b C21H19O11 447.0933 447.0928 1.1
39 44.0 24c C21H19O11 447.0933 447.0920 2.9
40 45.3 24d C21H19O11 447.0933 447.0935 0.4
41 33.4 25a C22H21O14 509.0937 509.0933 0.8
42 36.1 25b C22H21O14 509.0937 509.0927 2.0
43 40.4 25c C22H21O14 509.0937 509.0929 1.5
44 41.0 25d C22H21O14 509.0937 509.0924 2.6
45 42.2 26a C23H23O14 523.1093 523.1080 2.5
141
46 42.7 26b C23H23O14 523.1093 523.1088 1.0
47 44.1 26c C23H23O14 523.1093 523.1079 2.6
48 45.1 26d C23H23O14 523.1093 523.1082 2.1
49 Malic acid+5-CQA 20.8 28a C20H21O13 469.0988 469.0983 0.9
50 22.3 28b C20H21O13 469.0988 469.0991 0.6
51 23.4 28c C20H21O13 469.0988 469.0999 2.4
52 26.3 28d C20H21O13 469.0988 469.0994 1.4
53 33.5 29a C21H23O13 483.1144 483.1141 0.6
54 35.8 29b C21H23O13 483.1144 483.1142 0.8
55 32.5 30a C20H19O12 451.0882 451.0887 1.1
56 33.7 30b C20H19O12 451.0882 451.0899 3.7
57 36.5 30c C20H19O12 451.0882 451.0894 2.7
58 37.4 31a C21H21O12 465.1038 465.1038 0.1
59 38.8 31b C21H21O12 465.1038 465.1036 0.5
60 40.5 31c C21H21O12 465.1038 465.1031 1.6
61 46.1 32 C20H17O11 433.0776 433.0771 1.3
142
62 33.5 33 C21H23O13 483.1144 483.1141 0.6
63 Tartaric acid+5-CQA 33.4 34a C20H19O13 467.0831 467.0830 0.2
64 37.3 34b C20H19O13 467.0831 467.0827 1.0
65 16.8 35a C20H21O14 485.0937 485.0945 1.7
66 18.5 35b C20H21O14 485.0937 485.0954 3.5
67 23.7 35c C20H21O14 485.0937 485.0953 3.4
Table 5.2 Transesterification products of 4-11 with 5-CQA (2) identified with LC-MSn in the samples generated by heating each acid separately
with 5-CQA
No. Product
No.
Parent ion
(M-H)
Characteristic m/z of ions in negative ion mode
1 12a 363 MS2→ 335 (100), 173 (5); MS3 → 173 (100), 161 (11), 179 (19); MS4 → 110 (100), 92 (76), 136 (35), 71 (35)
2 63 363 MS2→ 335 (100), 317 (46); MS3 → 161 (100), 173 (75), 179 (70) ; MS4 →132 (100)
3 12c 363 MS2→ 335 (100); MS3 → 173 (100), 211 (65), 255 (45), 291 (34)
4 14a 377 MS2→ 335 (100), 317 (72); MS3 → 173 (100); MS4 →155 (100), 110 (58)
5 14b 377 MS2→ 335 (100), 317 (80); MS3 → 173 (100); MS4 → 110 (100), 92 (69), 71 (64), 129 (46)
6 14c 377 MS2→ 335 (100), 317 (66), 255 (29), 179 (37); MS3 → 173 (100), 161 (20), 179 (40); MS4 →110 (100), 137 (55),
71 (58), 129 (52), 86 (27)
7 14d 377 MS2→ 335 (100), 289 (5), 179 (9); MS3 → 173 (100), 179 (45), 161 (31); MS4 →155 (100), 110 (92), 92 (40)
143
8 17a 435 MS2→ 335 (100); MS3 → 173 (100), 179 (16), 161 (15); MS4 →110 (100), 92 (78), 71 (55), 155(34)
9 17b 435 MS2→ 335 (100); MS3 → 173 (100), 179 (42), 161 (43), 155 (8); MS4 →110 (100), 154 (65), 86 (47)
10 17c 435 MS2→ 335 (100); MS3 → 255 (100), 211 (83), 179 (60), 229 (30), 161 (58); MS4 →211 (100)
11 18 535 MS2→ 335 (100), 435 (36); MS3 → 173 (100); MS4 →154 (100), 110 (71), 92 (97)
12 20a 449 MS2→ 335 (100); MS3 → 173 (100), 179 (18), 161 (12); MS4 →110 (100), 92 (60), 155 (42)
13 20b 449 MS2→ 335 (100), 287 (20); MS3 → 173 (100), 179 (18), 161 (12); MS4 →155 (100), 127 (76), 110 (65), 92 (64)
14 20c 449 MS2→ 335 (100), 287 (34), 161 (20); MS3 → 255 (100), 210 (98), 173 (73), 291 (40), 228 (30); MS4 →210 (100)
15 19 467 MS2→ 305 (100), 353 (83), 335 (12), 191 (40), 406 (63); MS3 → 191 (100)
16 21a 481 MS2→ 353 (100), 319 (58), 191 (20); MS3 → 191 (100); MS4 →127 (100), 110 (88), 173 (68), 92 (79)
17 21b 481 MS2→ 319 (100), 353 (6), 460 (20); MS3 → 191 (100); MS4 →127 (100)
18 22a 463 MS2→ 335 (100), 300 (52), 161 (40); MS3 → 173 (100); MS4 →110 (100), 92 (76), 137 (48), 81 (43)
19 22b 463 MS2→ 301 (100), 335 (17), 161 (14); MS3 → 145 (100), 173 (84), 137 (71), 127 (57); MS4 →83 (100), 127 (22)
20 22c 463 MS2→ 301 (100), 335 (17), 161 (14); MS3 → 145 (100), 173 (87), 127 (37); MS4 →127 (100), 83 (22)
21 25 509 MS2→ 335 (100), 481 (17), 191 (27); MS3 → 255 (100), 291 (27), 227 (33), 211 (38)
22 24a 447 MS2→ 335 (100); MS3 → 173 (100); MS4 →110 (100), 154 (47), 100 (45), 92 (44)
23 24b 447 MS2→ 335 (100); MS3 → 173 (100), 255 (77), 211 (40), 291 (25)
24 30a 451 MS2→ 335 (100); MS3 → 173 (100), 179 (15); MS4 →110 (100), 92 (93), 71 (100), 85 (50), 155 (48)
25 30b 451 MS2→ 335 (100); MS3 → 173 (100), 179 (25), 161 (12), 134 (6); MS4 →155 (100), 110 (77), 92 (70), 71 (72)
26 32a 433 MS2→ 335 (100), 317 (20), 389 (32); MS3 → 173 (100), 179 (15); MS4 →110 (100), 92 (41), 71 (27), 59 (75)
27 64 433 MS2→ 335 (100), 389 (37); MS3 → 179 (100), 317 (12); MS4 →255 (100), 211 (63), 179 (96), 173 (53)
28 27a 671 MS2→ 335 (100), 353 (32), 191 (58); MS3 → 173 (100); MS4 → 110(100), 155 (16), 93 (42), 81 (27)
144
29 27b 671 MS2→ 509 (100), 353 (82), 191 (14); MS3 → 353 (100), 191 (47), 335 (10); MS4 → 191(100)
Table 5.4 Transesterification products of 4-11 with 5-CQA (2) identified with LC-TOF-MS in the roasted coffee samples
No. Compounds
involved
Retention
time (min)
Condensation
product No.
Mol. Formula Theoretical m/z
(M-H)
Experimental m/z
(M-H)
Error
(ppm)
1 Oxalic acid+5-CQA 44.4 13 C18H17O12 425.0725 425.0709 3.9
2 Malonic acid+5-CQA 25.4 15a C18H19O10 395.0984 395.0990 1.7
3 33.2 15b C18H19O10 395.0984 395.0980 1.1
4 38.9 15c C18H19O10 395.0984 395.1002 4.7
5 Succinic acid+5-CQA 28.2 17c C20H19O11 435.0933 435.0916 3.9
6 32.6 17d C20H19O11 435.0933 435.0914 4.4
7 Glutaric acid+5-CQA 31.6 19a C21H23O12 467.1195 467.1184 2.4
8 Adipic acid+5-CQA 29.3 21a C22H25O12 481.1351 481.1344 1.6
9 Citric acid+5-CQA 29.8 27 C32H31O16 671.1618 671.1622 0.7
145
Although only one condensation product was identified involving 5-CQA and adipic acid
in LC-MSn, which is adipoyl-caffeoylquinic acid (21a) in Figure 5.2, we found many
condensation compounds having chlorogenic acid skeleton in LC-TOF-MS (Table 5.1).
We were able to identify 29 transesterification products with their structural information
by LC-MSn (Table 5.2, Figure 5.2).
Figure 5.3 UV chromatogram at 254 nm of the model roasting experiment sample generated
by heating 5-CQA (2) with succinic acid (6)
Difference in the roasting conditions may have an effect on the integrity of the smaller
organic acids more than the acids having higher molecular weight. In the model roasting
experiments, oxalic acid (4) condenses with 5-CQA in its degraded form giving rise to the
formates of the chlorogenic acid lactones and shikimates such as formoyl-caffeoyl-1,5-
quinide (12) and Formoyl-caffeoyl-shikimate (63). But, the FT-ICR data of the roasted
coffee confirms the presence of oxaloyl-caffeoylquinic acid (13), which is an oxalate of 5-
CQA. Malonic acid was found to be condensed in its degraded form to produce acetates
with either 5-CQA or with 5-CQA lactones such as, acetoyl-caffeoyl-1,5-quinide (14) and
acetoyl-caffeoylquinic acid (15) in model roasting and in roasted coffee samples as well.
Model roasting showed three transesterification products with succinic acid, succinoyl-
caffeoylquinic acid (16), succinoyl-caffeoyl-1,5-quinide (17) and di-O-succinoyl-
caffeoylquinic acid (18), all of which are identified in roasted coffee samples analyzed by
FT-ICR-MS and LC-TOF-MS. Six isomers of compound 19 (glutaric acid+ caffeoylquinic
acid) and three isomers of 20 (glutaric acid+ caffeoyl-1,5-quinide) were observed in LC-
UV Chromatogram, 254 nm
0
5
10
15
Intens. [mAU]
5 10 15 20 25 30 35 Time [min]
146
TOF-MS and LC-MSn. Fragmentation pattern of 19 in tandem MS suggests that unlike
most of the other transesterification products, glutaric acid unit is condensed at quinic acid
part rather than the caffeoyl moiety. The fragmentation scheme is shown in Figure 5.4.
O
OO
O
OH
OH
OO
HO OH
O
OH
O
OO
OHO
HO OH
O
OH
+
OH
OH
OH
OO
HO O
O
OH
MS2
MS3
OHHO
HO O
O
OH
Parent ion at m/z 467 Base peak at m/z 305 Sec. peak at m/z 353
Base peak at m/z 191
19
Figure 5.4 Fragmentation scheme for compound 19 (glutaric acid+ caffeoylquinic acid)
Reaction of glutaric acid with 5-CQA gives isomeric products 19 and 20, which both are
glutaric acid+caffeoylquinic acid derivatives are observed to be present in the roasted
coffee also. The fragmentation patterns of some of the isomers of the transesters of adipic
acid, which are adipoyl-caffeoylquinic acid and adipoyl-cafeeoyl-1,5-quinides (21b, 22b
and 22c) are fully consistent with the fact that the transesterification is occurring at the
quinic acid moiety similar to the previous case. As shown earlier, aliphatic dicarboxylic
acids fragment preferentially if compared to hydroxycinnamic acid if esterified to quinic
acid. However, as found here, this earlier finding cannot be generalized. (Figure 5.5).
From the FT-ICR-MS data of the roasted coffee, presence of the products 21 and 22 in
coffee can be confirmed, unfortunately we cannot comment on the position of the adipic
acid transesterification. On the other hand, LC-TOF-MS data of the roasted coffee
samples suggests that the esterification of the adipic acid is taking place at the caffeoyl
moiety of 5-CQA (Table 5.3 and 5.4).
147
Table 5.3 Compounds transesterified with 5-CQA identified in FT-ICR-MS data of roasted
coffee samples
No. Compounds
involved
Condensation
product No.
Mol. Formula Theoretical m/z
(M-H)
Experimental m/z
(M-H)
Error
(ppm)
1 Oxalic
acid+5-CQA
13 C18H17O12 425.072550 425.072438 0.3
2 Malonic
acid+5-CQA
14 C18H17O9 377.087618 377.087608 0.5
3 15 C16H16O8 395.098370 395.098254 0.3
4 Succinic
acid+5-CQA
16 C20H21O12 453.103850 453.103741 0.2
5 17 C20H19O11 435.093285 435.091115 5.0
6 18 C24H23O14 535.109329 535.106990 4.4
7 Glutaric
acid+5-CQA
19 C21H23O12 467.119500 467.119502 0.0
8 20 C21H21O11 449.108935 449.108940 0.0
9 Adipic
acid+5-CQA
21 C22H25O12 481.135150 481.135189 -0.1
10 22 C22H23O11 463.124585 463.124635 -0.1
11 Citric acid+5-
CQA
23 C22H23O15 527.104244 527.102236 3.0
12 24 C21H19O11 447.093285 447.093547 -0.6
13 Malic acid+5-
CQA
28 C20H21O13 469.123403 469.123705 -0.6
14 29 C21H23O13 483.114414 483.114466 0.1
15 Tartaric
acid+5-CQA
34 C20H19O13 467.083114 467.080960 4.6
The highest number of the transesterification products in model roasting experiment is
contributed by citric acid transesters (Table 5.1). We observed 15 condensation products
with citric acid followed by the number of transesters generated by malic acid with 5-
148
CQA. In contradiction to this fact, the roasted coffee analysis by LC-TOF-MS and FT-
ICR-MS we were able to identify only four transesterification products, which observed to
be sourced from citric and malic acid together in FT-ICR-MS and one in LC-MS-TOF
(Table 5.3 and Table 5.4). This can be attributed to the degradation of the citric acid
during roasting process, which gives rise to the glutaric acid and succinic acid. Evidently,
in roasted coffee, transesters arising from glutaric acid succinic acid totals to five, which
is comparatively higher. Additionally, the identification of a gluteroyl-caffeoyl-1,5-
quinide (20b) in LC-TOF-MS analysis (Table 5.1) of the sample generated by heating
citric acid with 5-CQA proves that the degraded products of citric acid such as, gluatric
acid generate transesterification products in model roasting conditions. LC-MSn does not
provide as a reliable identification technique in the case of analysis of citric acid
condensation products due to its inability to provide the molecular formula considering
the fact that the molecular weight of the citric acid and quinic acid is the same. This on
the other hand, clears the confusion in the structural assignment for m/z 671 by
eliminating the possibility of the involvement of any citric acid moiety. We did not
observe any tartates in the tandem MS but, we identified tarteroyl-caffeoyl-1,5-quinide
(34) in roasted coffee in FT-ICR-MS analysis (Table 5.3).
149
Figure 5.5 Fragmentation patterns for m/z 481 (21) and m/z 463 (22)
If we observe the trend of the transesterification product formation when all the acids are
heated in collective equimolar quantity with 5-CQA from Table 5.5 and Table 5.6 we see
that the majority of the products are formed by the esters generated by glutaric and
succinic acid (product numbers 16-20). This elevated selectivity towards
transesterification products sourced from succinic and glutaric acid can be a result of the
increased quantity of these acids in the mixture due to the degradation of the citr ic acid
during model roasting experiment. This assumption is also supported by the fact that only
one citrate transester is identified at m/z 509 in Table 5.6.
191 353 461
319 -MS2(481), 35.3min
191 -MS3(481->319), 35.3min
-MS4(481->319->191), 35.4min
0
100
Intens.
[%]
0
100
0
100
50 100 150 200 250 300 350 400 450 m/z
127
161 335
301 -MS2(463), 48.9min
173 145
-MS3(463->301), 48.9min
-MS4(463->301->145), 49.0min
0
100
Intens.
[%]
0
100
0
100
50 100 150 200 250 300 350 400 450 m/z
83
150
Table 5.5 Transesterification products of 4-11 with 5-CQA (2) identified with targeted LC-
MSn in the samples generated by heating all of the acids collectively with 5-CQA
No. Product
No.
Parent
ion (M-H)
Characteristic m/z of ions in negative ion mode
1 18 535 MS2→ 335 (100), 435 (36); MS3 → 173 (100); MS4 →154 (100), 110
(71), 92 (97)
2 19 467 MS2→ 305 (100), 353 (83), 335 (12), 191 (40), 406 (63); MS3 → 191
(100)
3 27a 671 MS2→ 335 (100), 353 (32), 191 (58); MS3 → 173 (100); MS4 →
110(100), 155 (16), 93 (42), 81 (27)
4 27b 671 MS2→ 509 (100), 353 (82), 191 (14); MS3 → 353 (100), 191 (47), 335
(10); MS4 → 191(100)
Table 5.6 Transesterification products of 4-11 with 5-CQA (2) identified with LC-TOF-MS
in the samples generated by heating all of the acids collectively with 5-CQA
No. Retention
time (min)
Condensation
product No.
Mol. Formula Theoretical m/z
(M-H)
Experimental
m/z (M-H)
Error
(ppm)
1 38.5 14a C18H17O9 377.0878 377.0888 2.3
2 39.9 14b C18H17O9 377.0878 377.0876 0.6
3 25.4 15a C18H19O10 395.0984 395.0985 0.4
4 33.8 15b C18H19O10 395.0984 395.0989 1.4
5 37.3 15c C18H19O10 395.0984 395.0995 2.8
6 32.4 16c C20H21O12 453.1038 453.1031 1.6
7 39.9 17c C20H19O11 435.0933 435.0943 2.3
8 41.4 17b C20H19O11 435.0933 435.0995 3.9
9 30.9 19a C21H23O12 467.1195 467.1199 0.8
10 32.8 19b C21H23O12 467.1195 467.1179 3.4
11 34.1 19c C21H23O12 467.1195 467.1183 2.5
12 35.3 19d C21H23O12 467.1195 467.1203 1.8
151
13 40.0 19f C21H23O12 467.1195 467.1196 0.2
14 40.9 20a C21H21O11 449.1089 449.1105 3.5
15 44.1 20b C21H21O11 449.1089 449.1099 2.1
16 44.9 20c C21H21O11 449.1089 449.1102 2.9
17 43.0 20d C21H21O11 449.1089 449.1095 1.2
18 35.1 21a C22H25O12 481.1351 481.1350 0.3
19 38.1 21b C22H25O12 481.1351 481.1369 3.7
20 38.9 21c C22H25O12 481.1351 481.1361 1.9
21 45.0 22a C22H23O11 463.1246 463.1247 0.3
22 48.0 22b C22H23O11 463.1246 463.1262 3.5
23 41.0 25d C22H21O14 509.0937 509.0954 3.4
24 46.2 32 C20H17O11 433.0776 433.0793 4.0
25 16.7 35a C20H21O14 485.0937 485.0935 0.3
26 18.4 35b C20H21O14 485.0937 485.0945 1.8
5.3.2 Transesterification of quinic acid (1) in model roasting and in roasted coffee
samples
In the model roasting experiment of quinic acid (1) with the range of all organic acids (4-11)
37 isomeric transesterification products were identified by LC-TOF-MS listed in the Table
5.7. In the EIC’s at the respective m/z in negative mode of the structures we speculated to be
forming in this experiment, we found that in case of oxalic acid heated with quinic acid; only
one peak of oxaloylquinide was observed (36). Two peaks of two esters of quinic acid with
malonic acid were identified to be the esters of quinic acid (37, 38). Succinic acid was found
to be forming eight esterification products giving eight peaks, which were assigned to three
isomeric esters of quinic acid (39) and five isomeric esters if quinic acid lactone (40). Glutaric
acid when heated with QA was observed to be forming only one peak generated by an ester of
quinic acid lactone (41). Adipic acid+ QA showed six peaks assigned to an ester of quinic
acid (42) and five isomeric esters of quinic acid lactone (43). Surprisingly, citric acid
generated only three peaks when heated with QA, out of which one was assigned to be an
152
ester of quinic acid (44) and the rest two as esters of quinic acid lactone (45, 46). Malic acid
formed the highest number of esters with QA showing twelve peaks generated by two
isomeric esters of quinic acid (47) and the rest ten of the isomeric esters were identified as
esters of quinic acid lactone (48, 49). Four peaks were observed in the case of tartaric
acid+QA, all of which were assigned to be the isomers of esters of quinic acid lactone (50).
Figure 5.6 shows the representative total ion chromatogram in the negative mode of the
model roasting experiment sample generated by heating QA (1) with glutaric acid (7). Similar
to the case of model roasting experiment with 5-CQA, here we also observe that the majority
of the transesterification products are formed with dehydration at the quinic acid part.
Unfortunately due to the lack of the analytical data by LC-MSn, at the moment we cannot
comment on whether the dehydrated products are shikimates or lactones. Hence, in the
Figure 5.2, we have presented them as lactones for the purpose of simplification.
Figure 5.6 Total ion chromatogram in the negative mode of the model roasting experiment
sample generated by heating QA (1) with glutaric acid (7)
In the model roasting experiment, we identified oxaloylquinide, which is oxalic acid
condensed with the quinic acid lactone (36) (Table 5.7) but, in the roasted coffee sample
oxalic acid appears in the form of its degraded product condensed with quinic acid to give
formoylquinic acid (62) (Table 5.8). Except for the tartrates of the quinic acid or quinic acid
lactones, all the other acids were found as the transesterification products in roasted coffee
LC-MS analysis. Due to the unavailability of the structural information provided by the
tandem MS, the position of the dehydration taking place i.e. either at QA part or at citric acid
part in structure 45 is open for alternate interpretation. Twelve isomeric esters were identified
QA + Glutaric acid, TIC -
0.5
1.0
6 x10
Intens.
5 10 15 20 25 30 35 Time [min]
153
in the model roasting of malic acid with QA, which is highest number of the esters sourced
from a single acid if compared to of other acids. Continuing the trend, malates of QA were
found to have highest selectivity compared to other acids attributed to the number of the
maloyl-quinic acid derivatives formed shown in Table 5.9.
154
Table 5.7 Transesterification products of 4-11 with quinic acid (1) identified with LC-TOF-MS in the samples generated by heating each acid
separately with quinic acid (QA)
No. Compounds
involved
Retention
time (min)
Condensation
product No.
Mol. Formula Theoretical m/z
(M-H)
Experimental m/z
(M-H)
Error
(ppm)
1 Oxalic acid+ QA 5.2 36 C9H9O8 245.0303 245.0308 2.3
2 Malonic acid+ QA 8.9 37a C9H13O7 233.0667 233.0677 4.5
3 5.3 38 C10H13O9 277.0565 277.0556 3.2
4 Succinic acid+ QA 6.6 39a C11H15O9 291.0722 291.0717 1.6
5 18.4 39b C11H15O9 291.0722 291.0714 2.7
6 28.1 39c C11H15O9 291.0722 291.0715 2.1
7 7.4 40a C11H13O8 273.0616 273.0607 3.2
8 10.9 40b C11H13O8 273.0616 273.0607 3.4
9 22.3 40c C11H13O8 273.0616 273.0604 4.2
10 23.0 40d C11H13O8 273.0616 273.0604 4.4
11 24.0 40e C11H13O8 273.0616 273.0608 2.9
12 Glutaric acid+ QA 16.4 41 C12H15O8 287.0772 287.0788 5.4
13 Adipic acid+ QA 14.1 42 C13H19O9 319.1035 319.1017 5.5
155
14 14.9 43a C13H17O8 301.0929 301.0916 4.4
15 16.2 43b C13H17O8 301.0929 301.0920 2.9
16 17.9 43c C13H17O8 301.0929 301.0922 2.2
17 18.6 43d C13H17O8 301.0929 301.0919 3.2
18 23.3 43e C13H17O8 301.0929 301.0914 4.8
19 Citric acid+ QA 7.0 44 C13H17O12 365.0725 365.0712 3.6
20 7.5 45 C13H15O11 347.0620 347.0602 5.2
21 12.1 46 C13H13O10 329.0514 329.0496 5.5
22 Malic acid+ QA 5.0 47a C10H15O10 307.0671 307.0670 0.1
23 4.3 47b C10H15O10 307.0671 307.0664 2.1
24 3.9 48a C11H13O9 289.0565 289.0563 0.7
25 4.4 48b C11H13O9 289.0565 289.0561 1.3
26 4.7 48c C11H13O9 289.0565 289.0563 1.8
27 6.0 48d C11H13O9 289.0565 289.0557 2.8
28 3.6 49a C11H11O8 271.0459 271.0457 0.8
29 3.9 49b C11H11O8 271.0459 271.0464 1.5
156
30 4.7 49c C11H11O8 271.0459 271.0457 1.0
31 9.0 49d C11H11O8 271.0459 271.0458 0.5
32 9.6 49e C11H11O8 271.0459 271.0460 0.4
33 11.5 49f C11H11O8 271.0459 271.0453 2.5
34 Tartaric acid+ QA 3.9 50a C11H13O10 305.0514 305.0505 3.1
35 4.7 50b C11H13O10 305.0514 305.0517 1.0
36 5.6 50c C11H13O10 305.0514 305.0510 1.3
37 7.0 50d C11H13O10 305.0514 305.0519 1.5
Table 5.11 Transesterification products of 4-11 with CA (3) identified with LC-TOF-MS in the roasted coffee samples
No. Compounds
involved
Retention
time (min)
Condensation
product No.
Mol. Formula Theoretical m/z
(M-H)
Experimental m/z
(M-H)
Error
(ppm)
1 Oxalic acid+ CA 14.0 58 C20H13O10 413.0514 413.0530 3.8
2 Adipic acid+ CA 19.0 57 C15H15O7 307.0823 307.0816 2.5
3 Glutaric acid+ CA 41.4 56 C23H19O10 455.0984 455.0984 0.1
4 Succinic acid+ CA 45.3 54 C17H15O10 379.0671 379.0665 4.5
5 Malic acid+ CA 28.4 59 C22H17O11 457.0776 457.0759 6.7
157
Table 5.8 Compounds transesterified with quinic acid (QA) identified in FT-ICR-MS data of
roasted coffee samples
No. Compounds
involved
Condensation
product No.
Mol. Formula Theoretical m/z
(M-H)
Experimental m/z
(M-H)
Error
(ppm)
1 Oxalic
acid+QA
62 C8H11O7 219.051026 219.051024 0.0
2 Malonic
acid+QA
37 C9H13O7 233.066676 233.066700 0.1
3 Glutaric
acid+QA
41 C12H15O8 287.077241 287.077105 0.5
4 Adipic
acid+QA
43 C13H17O8 301.092891 301.092844 0.2
5 Citric
acid+QA
44 C13H17O12 365.072550 365.072629 0.2
6 45 C13H15O11 347.061985 347.062109 0.4
8 Malic
acid+QA
49 C11H11O8 271.045941 271.047040 4.1
Aliphatic esters of the quinic acid and its lactones can be termed as chlorogenic acids
according to the modern definition (717, 726). Accordingly, we have found ten new
chlorogenic acids in the analysis of the roasted coffee samples (Table 5.8 and Table 5.4)
done in this work by LC-TOF-MS and FT-ICR-MS. Malonic acid generates two esters with
quinic acid (QA) in the form of acetylquinic acid (37) and malonoylquinic acid (38), both of
which were found to be present in the roasted coffee (Table 5.4). Seven isomers of
succinoylquinic acid (39) and succinoyl-quinide (40) were identified in model roasting, out of
which only one isomer (39a) was identified in roasted coffee by LC-TOF-MS. Glutaroyl-
quinide (41) was identified in the roasted coffee in Table 5.8 similar to the observation from
the model roasting experiment. Furthermore, adipoyl-qunide (43), citroylquinic acid (44),
citroyl-qunide (45), fumaroylquinic acid (49) and formoylquinic acid (62) were identified in
FT-ICR-MS data (Table 5.8) and adipoyl-qunide (43) and maloyl-quinide (48) were
identified in LC-TOF-MS data (Table 5.4).
158
Table 5.9 Transesterification products of 4-11 with quinic acid (1) identified with LC-TOF-
MS in the samples generated by heating all of the acids collectively with quinic acid (QA)
No. Retention
time (min)
Condensation
product No.
Mol.
Formula
Theoretical
m/z (M-H)
Experimental
m/z (M-H)
Error
(ppm)
1 7.4 40a C11H13O8 273.0616 273.0607 3.1
2 16.4 41a C12H15O8 287.0772 287.0761 4.0
3 12.2 41b C12H15O8 287.0772 287.0763 3.2
4 14.2 42 C13H19O9 319.1035 319.1023 3.6
5 18.7 43d C13H17O8 301.0929 301.0913 5.4
6 23.2 43e C13H17O8 301.0929 301.0914 4.7
7 8.5 45 C13H15O11 347.0620 347.0635 4.5
8 4.5 48b C11H13O9 289.0565 289.0559 2.2
9 4.7 48c C11H13O9 289.0565 289.0557 2.8
10 3.6 49a C11H11O8 271.0459 271.0472 4.5
11 3.9 49b C11H11O8 271.0459 271.0461 0.7
12 4.7 49c C11H11O8 271.0459 271.0455 1.6
13 9.7 49e C11H11O8 271.0459 271.0450 3.5
14 4.7 50b C11H13O10 305.0514 305.0500 4.6
5.3.3 Transesterification of caffeic acid (3) in model roasting and in roasted coffee
samples
We report ten new esters of the caffeic acid (52-61) identified in roasted coffee samples by
LC-TOF-MS and FT-ICR-MS (Table 5.10, Table 5.11 and Figure 5.2), out of which one
was identified as a dimer of caffeic acid, three were found to be esters bound to a di-caffeoyl
moiety (56, 58 and 59), three di-esters as oxalate (51), acetate (53) and succinate (54) were
identified and four were identified as mono-esters of caffeic acid.
Model roasting experiment results lead us to concur that caffeic acid produces the least
number of condensation products compared to 5-CQA and QA. We observed mono- and di-
159
acetate of caffeic acid (52, 53) as esters of degraded malonic acid along with compound 55,
which also found to be a peculiar product of a model roasting experiment. It appears to be a
dimer of caffeic acid possessing two possible positional isomers; both eluting at different
retention times hence can be differentiated (Figure 5.7). However, 55 could not be identified
in roasted coffee samples in the data acquired from both of the techniques. Interestingly, we
observed the esters of the dimers of the caffeic acid with oxalic acid, glutaric acid and malic
acid in roasted coffee samples (56, 58 and 59).
Table 5.10 Compounds transesterified with caffeic acid (CA) identified in FT-ICR-MS data
of roasted coffee samples
No. Compounds
involved
Condensation
product No.
Mol. Formula Theoretical m/z
(M-H)
Experimental m/z
(M-H)
Error
(ppm)
1 Malonic
acid+CA
52 C11H9O5 221.045547 221.045583 0.2
2 53 C13H11O6 263.056112 263.056078 0.1
3 Succinic
acid+CA
54 C17H15O10 379.067070 379.067254 0.5
4 Glutaric
acid+CA
60 C14H13O7 293.066676 293.066690 0.0
5 Adipic
acid+CA
57 C15H15O7 307.082326 307.082341 0.0
6 Dextrotartaric
acid+CA
61 C13H11O9 311.040856 311.040386 1.5
In addition to the esters of the small organic acids, we observed further molecular
formulas in the FT-ICR-MS data, which confirms the presence of the esters between the
free fatty acids and quinic acid in roasted coffee. Linoleoylquinic acid (71) (at m/z
453.285958, C25H41O7 with 0.4 ppm error) and palmitoylquinic acid (72) (at m/z
429.285901, C23H41O7 with 0.3 ppm error) were identified. LC-TOF-MS also confirms the
presence of these two products in roasted coffee samples. In contradiction to our
expectation, no corrosponding lactones such as palmitoylquinide or linoleoylquinide were
found either in model roasting or in roasted coffee samples.
160
Figure 5.7 EIC at m/z 341
5.4 Conclusions
In conclusion, we have identified 67 isomeric transesters between 5-CQA and small
organic acids in simulated roasting experiment, out of which 16 compounds were
observed to be present in roasted coffee samples. We established the relationship of the
appearance of higher number of the esters generated by glutaric acid and succinic acid to
the degradation of citric acid in roasting process. In this work we report ten new
chlorogenic acids in the form of aliphatic esters of quinic acid. We have found that free
quinic acid undergoes further esterification in coffee roasting by identifying the esters of
all the small organic acids incorporated in this study except for the tarteroylquinic acid
derivatives. Elucidation of the fate of the free caffeic acid in coffee melanoidines was
achieved by reporting various products formed such as dimers of the caffeic acid, esters of
the dimers of caffeic acid along with formates and acetates of caffeic acid. Confirmation
of the presence of the esters of free fatty acids and quinic acid will help in the further
investigation of the foam of the espresso coffee.
EIC 341.0000 -
0
500
1000
1500
2000
Intens.
34 36 38 40 42 44 46 48 Time [min]
55a
55b
161
References
1. Jaiswal, R.; Matei, M.F.; Golon, A.; Witt, M.; Kuhnert, N. Understanding the fate of
chlorogenic acids in coffee roasting using mass spectrometry based targeted and non-targeted
analytical strategies. Food Funct. 2012, 3, 976-984.
2. Clifford, M. What factors determine the intensity of coffee’s sensory attributes. Tea Coffee
Trade J 1987, 159, 35-39.
3. Maier, H. The acids of coffee. Proceedings 12th Asic College 1987, 229-237.
4. Mabrouk, A.F.; Deatherage, F.E. Organic acids in brewed coffee. Food Technol. (Chicago,
IL, U. S. ) 1956, 10, 194-197.
5. Lentner, C.; Deatherage, F.E. Organic acids in coffee in relation to the degree of roast.
Food Res. 1959, 24, 483-492.
6. Nakabayashi, T. Chemical studies on the quality of coffee. VI. Changes in organic acids
and pH of roasted coffee. Nippon Shokuhin Kogyo Gakkaishi 1978, 25, 142-146.
7. Engelhardt, U.H.; Maier, H.G. Acids in coffee. XI. The proportion of individual acids in
the total titratable acid. Z Lebensm Unters Forsch 1985, 181, 20-23.
8. Woodman, J.S.; Giddey, A.; Egli, R.H. Carboxylic acids of brewed coffee.
Int. Colloq. Chem. Coffee, 3rd 1968, 137-143.
9. Hughes, W.J.; Thorpe, T.M. Determination of organic acids and sucrose in roasted coffee
by capillary gas chromatography. J. Food Sci. 1987, 52, 1078-1083.
10. Engelhardt, U.H. Nichtflüchtige Säuren im Kaffee. Technische Universität Carolo-
Wilhelmina zu Braunschweig: 1984
11. Clarke, R.; Vitzthum, O. Coffee: recent developments. Wiley. com: 2008
12. Clifford, M. Chemical and physical aspects of green coffee and coffee products, In Coffee,
Anonymous ; Springer: 1985; pp. 305-374.
13. Scholze, A.; Maier, H. Quantitative Bestimmung von Säuren in Kaffee mittels Kapillar-
Isotachophorese. Lebensm Chem Gerichtl Chem 1982, 36, 111-112.
14. Engelhardt, U.H.; Maier, H.G. Säuren des Kaffees. Zeitschrift für Lebensmittel-
Untersuchung und Forschung 1985, 181, 206-209.
15. Scholze, A. Quantitative Bestimmung von Säuren im Kaffee durch Kapillar-
Isotachophorese. Technische Universität Carolo-Wilhelmina zu Braunschweig: 1983
16. Balzer, H.H. Chemistry I: Non-volatile compounds. Acids in coffee. Coffee 2001, 18-32.
17. Carisano, A.; Gariboldi, L. Gas chromatographic examination of the fatty acids of coffee
oil. J. Sci. Food Agric. 1964, 15, 619-622.
162
18. Speer, K.; Kölling-Speer, I. The lipid fraction of the coffee bean. Brazilian Journal of
Plant Physiology 2006, 18, 201-216.
19. Wajda, P.; Walczyk, D. Relation between acid value of extracted fatty matter and age of
green coffee beans. J. Sci. Food Agric. 1978, 29, 377-380.
20. Kurzrock, T.; Kölling-Speer, I.; Speer, K. Identification of dehydrocafestol fatty acid
esters in coffee. 1998, 27
21. Crozier, A.; Jaganath, I.B.; Clifford, M.N. Dietary phenolics: chemistry, bioavailability
and effects on health. Nat. Prod. Rep. 2009, 26, 1001-1043.
22. Maeaettae, K.R.; Kamal-Eldin, A.; Toerroenen, A.R. High-Performance Liquid
Chromatography (HPLC) Analysis of Phenolic Compounds in Berries with Diode Array and
Electrospray Ionization Mass Spectrometric (MS) Detection: Ribes Species. J. Agric. Food
Chem. 2003, 51, 6736-6744.
23. Kampmann, B.; Maier, H.G. Säuren des Kaffees. Zeitschrift für Lebensmittel-
Untersuchung und Forschung 1982, 175, 333-336.
24. Hucke, J.; Maier, H.G. Quinic acid lactone in coffee. Z Lebensm Unters Forsch 1985,
180, 479-484.
25. Galli, V.; Barbas, C. Capillary electrophoresis for the analysis of short-chain organic acids
in coffee. Journal of Chromatography A 2004, 1032, 299-304.
26. Ginz, M.; Balzer, H.H.; Bradbury, A.G.; Maier, H.G. Formation of aliphatic acids by
carbohydrate degradation during roasting of coffee. European Food Research and
Technology 2000, 211, 404-410.
27. Scholz-Boettcher, B.M.; Ernst, L.; Maier, H.G. New stereoisomers of quinic acid and their
lactones. Liebigs Ann. Chem. 1991, 1029-1036.
28. Bähre, F. Neue nichtflüchtige Säuren im Kaffee. Papierflieger: 1997
163
CHAPTER 6: Which spectroscopic technique allows best differentiation of
coffee varieties: Comparing principal component analysis using data
derived from CD-, NMR-, IR- spectroscopy and LC-MS in the analysis of
the chlorogenic acid fraction in green coffee beans
6.1 Introduction
In order to investigate parameters like geographic origin, varieties, adulterations, processing
conditions, sensory properties, beneficial health effects, shelf-life or any other desirable or
undesirable property of a food, a detailed knowledge of its composition and chemistry is
required and therefore becomes foremost a problem of analytical chemistry.
Once the chemical constituents of food have been elucidated comparison of the chemical
profile of different samples allows differentiation between samples and identification of
variations that are of interest to both producer and consumer.
In the last decade statistical methods aimed at data reduction, have become the method of
choice to undertake such a task, with multi-variant statistical methods, in particular principal
component analysis (PCA), becoming increasingly popular.3
The main philosophy of PCA is to reduce a large data set obtained from of a large number of
samples by using a selected spectroscopic method, in order to extract the most important
variations between the samples without any loss of information. These variations are termed
principal components of the samples, whereby each principal component is by definition
orthogonal to the next. Ideally, variations between sample groups can be identified and
through the variation of a spectroscopic parameter linked to a set of unique marker molecules.
PCA is mainly employed as an unsupervised pattern recognition technique providing
visualization of a multivariate dataset, thereby revealing trends, observations and outliers.
This visualization is achieved by the transformation of variables into a covariance based
coordinate system with the principal components as axis, thereby creating a two dimensional
representation termed score plot, from which a grouping or pattern of sample groups can be
extracted. Next to the score plot a so called loadings plot provides information about the
origin of the variances at a molecular level.
The success story of PCA has started with Nicholson’s work using high resolution NMR data
to identify disease related biomarkers from urine or plasma samples.4 Using PCA NMR data
of large patient groups was successfully compared and unique biomarkers for certain diseases
identified.4
164
PCA using a wide variety of analytical techniques, including NMR-, IR-, Raman
spectroscopy, HPLC, GC, or GC-MS, has been employed as an established statistical method
in other areas of research including metabolomics, food analysis and medical research.
An important question, which to our knowledge has never been answered, is which
spectroscopic technique is most suitable for the differentiation of a given set of food samples.
Methods used vary in their practicability and information content. For example, IR- or Raman
analysis provides rapid measurements, omitting sample preparation and using portable
inexpensive instrumentation, within minutes a measurement and hence ideally a reliable result
that allows distinction between the samples. However, distinction between the samples is
frequently based on overlapping peaks corresponding not to an individual molecular marker
but rather to a large group or family of molecules present in the sample. Techniques like MS
or NMR avoid this limitation; however, require extensive sample preparation, costly
sophisticated equipment resulting in satisfactory information on the structures of individual
markers being present in the sample.
Due to the particular complexity of food, all of the techniques mentioned above, have severe
limitations with respect to type of materials amenable to investigation, resolution, sensitivity
and information provided. To our knowledge, this is the first time to use the performance of a
wide selection of spectroscopic techniques, to evaluate their individual usefulness and
potential in PCA analysis. In this contribution we report of the use of four different analytical
techniques in the characterisation of a single food material. A direct comparison of the
analytical techniques used is important to show that not all techniques succeed in sample
differentiation at the same level. It is for the first time that the comparison of this kind has
become possible. The ability of a certain technique to differentiate sample groups is directly
linked to the nature of its chemical constituents and the techniques ability to provide
analytical useful information on these constituents. As a food material we have chosen green
coffee bean samples for the following three reasons: Firstly, we have in our research group
acquired an intimate knowledge of the secondary metabolite profile and phytochemistry of
this material, having over the last years identified around 100 different secondary metabolites
in the green coffee bean, the large majority being chlorogenic acids.7-9 This moderate amount
of secondary metabolites ensures additionally that the large majority of signals in any
spectroscopic data sets can be reliably assigned to well characterised compounds. Secondly,
coffee is an important commercial commodity, indeed after water and black tea the third most
consumed beverage on this planet with an annual production of 4.5 Mt and a market value in
excess of 5 Billion US$ of the raw material alone. Thirdly, green coffee beans are produced in
165
two varieties Caffea arabica and Caffea canephora (otherwise known as Robusta coffee)
whose distinction and adulteration forms an important problem for the coffee industry. It
should, however, be noted that distinction of intact green coffee beans by visual inspection is
rather straightforward due to significant morphological differences between Robusta and
Arabica coffee beans. Only in the case of processed coffee, either roasted or grounded a
distinction based on chemical composition is required to which the methods presented here
can be applied.
Supposedly high quality coffee blends consist typically of 100% Arabica coffee beans. Lower
quality, cheaper blends may have some proportion of Robusta beans, or they may consist
entirely of Robusta. Arabica beans produce allegedly a superior taste in the cup, being more
flavourful and complex than their Robusta counterparts. Robusta beans in contrast tend to
produce a bitterer brew, with a musty flavour and stronger body. Obviously, this difference in
sensory properties could be related to the individual phytochemical profile of the two coffee
varieties and could be characterised by PCA.
Metabolomics, phytochemical profiling using PCA based methods have been frequently
applied to the problem of distinguishing green Arabica from Robusta coffee beans.
Admittedly distinction between green bean Arabica and Robusta samples does not constitute a
difficult scientific challenge as mentioned earlier. However, the distinction has been
frequently used as a benchmark for analytical chemistry methodology and should be viewed
as such. Briandet and Downey have used IR and NIR spectroscopy to study the differences
between the two varieties.10,11 NIR has been further used by Esteban-Diez and Lyman to
distinguish Arabica from Robusta green coffee beans.12,13 Wang et al. could show that as well
Kona coffee could be distinguished from other varieties using FTIR spectroscopy.14 In all of
this work, distinction between varieties was possible due to PCA analysis, however, due to the
nature of the spectroscopic technique used, only spectroscopic bands corresponding to groups
of compounds rather than individual phytochemical constituents could be identified.
Rubayizy15 could show using Raman spectroscopy that levels of the terpene Kahweol and
lipid content allows distinction between Arabica and Robusta green coffee beans. Materny
and co-workers have demonstrated that Raman microscopy can be employed directly on a
single green coffee bean to allow distinction between these two varieties, based on signals
corresponding to lipids and chlorogenic acids.16 Valdenebro et al. could show that the
geographic origin of green coffee beans can be identified using sterol profiles analysed by
GC-MS.17 Korhonova et al. found using GC-MS based PCA that differences in volatile
fractions exist between Arabica and Robusta beans.18 Mendonca and Alonso Salces were able
166
to distinguish Arabica and Robusta green coffee beans based on PCA data using HPLC
analysis of chlorogenic acid profiles. We have recently reported on PCA analysis of green
coffee beans using Raman spectroscopy and microscopy19, 20 and LC-MS methods, as well
discussing the importance of scaling and normalisation procedures in identifying meaningful
variations in PCA.21
6.2 Materials and methods
All the chemicals (analytical grade) were purchased from Sigma-Aldrich (Bremen, Germany)
and used as is. 28 different types of Arabica green coffee beans from different origins were
purchased from the main supplier of coffee (Münchhausen, Bremen and supermarkets in
Bremen Germany), and 10 different types of Robusta green coffee beans were obtained as a
generous offer from D.R. Wakefield & Co. Ltd., London, England.
6.2.1 Statistical analysis
Principal component analysis (PCA) was performed using the robust commercial software
package Unscrambler (v 9.7; CAMO A/S). In order to discriminate between extracted CGA
from Arabic and Robusta green coffee beans, a qualitative classification was performed by
PCA. PCA is one of the most common multivariate analysis methods used to reduce the
dimensionality of large data sets by finding combinations of variables that describe the major
trends in the data 1. The first PC carries most information or most explained variance and the
second PC carries the maximum residual information, which is not taken into account by the
first PC.
The similarity between the samples can be displayed by the PCA scores’ scatter plot of two
PCs, which shows the distribution of the samples in a new frame plot according to their
scores. The loading plot describes the relationship between variables and a specified principal
component. The loading value of each variable on a specific PC reflects how much the
variable contributed to that PC.
6.2.1.1 Methanolic extract of coffee beans
10 g of each sample of different green Robusta and Arabica coffee beans was frozen using
liquid nitrogen before grinding. The methanolic extract was prepared by Soxhlet extraction
using aqueous methanol (70%) for 5 hr. The extract was treated with Carrez reagent to
precipitate colloidal material, and filtered through Machery-Nagel MN-615 folded filter
paper. The methanol was removed in a rotary evaporator at reduced pressure. The aqueous
residue was kept in a deep freezer at -80°C for 1.5 hr, followed by lyophilisation under 0.94
mbar for 24 hrs. using Christ Alpha 1-4 LSC in order to remove the water from the extracted
167
CGA under most protective conditions. The extracts of CGA were stored at -20°C until
required. Before use for LC-MS, these extracts were thawed at room temperature, dissolved in
methanol (60mg/10mL), and filtered through a membrane filter.
6.3 Experimental
1H-NMR spectroscopy
All coffee extracts samples were thawed at room temperature, dissolved in DMSO-d6 (10 mg
of each coffee extract in 0.6 ml DMSO-d6), sonicated for three minutes. NMRs were
measured within one hour after sonication. 1H NMR spectra were acquired on a JEOL ECX-
400 spectrometer operating at 400 MHz at room temperature, using a 5 mm probe.
Circular Dichroism spectroscopy
CD spectra were obtained using Jasco J-810 spectrometer. All coffee extracts samples were
dissolved in DMSO (12.5 mg in 2 ml DMSO). Measurement conditions were kept as, Band
width-1 nm, Response-0.5 sec, Sensitivity-standard, Wavelength range 280 to 400 nm, Data
pitch-1 nm, Scanning speed-1000 nm/min, Accumulation-2. All measurements were done at
room temperature.
IR-spectroscopy
IR spectra for all coffee extracts samples were obtained using Bruker vector 33 ATR
spectrometer. Transmittance spectra were recorded against the wavelength range of 400 cm-1
to 4000 cm-1. Spectra were recorded directly of the dry coffee extracts samples without any
solvent.
LC-TOF-MS
The LC equipment (Agilent 1100 series, Karlsruhe, Germany) comprised a binary pump, an
auto sampler with a 100 μL loop, and a DAD detector with a light-pipe flow cell (recording at
320 and 254 nm and scanning from 200 to 600 nm). This was interfaced with a MicrOTOF
Focus mass spectrometer (Bruker Daltonics, Bremen, Germany) fitted with an ESI source and
internal calibration was achieved with 10 mL of 0.1 M sodium formate solution injected
through a six port valve prior to each chromatographic run. Calibration was carried out using
the enhanced quadratic mode.
HPLC
Separation was achieved on a 150 x 3 mm i.d. column containing diphenyl 5 μm, with a 5 mm
x 3 mm i.d. guard column (Varian, Darmstadt, Germany). Solvent A was water/formic acid
168
(1000:0.005 v/v) and solvent B was methanol. Solvents were delivered at a total flow rate of
500 μL/min. The gradient profile was from 10% B to 70% B linearly in 60 min followed by
10 min isocratic, and a return to 10% B at 90 min and 10 min isocratic to re-equilibrate.
LC-MSn
The LC equipment (Agilent 1100 series, Karlsruhe, Germany) comprised a binary pump, an
auto sampler with a 100 µL loop, and a DAD detector with a light-pipe flow cell (recording at
320 and 254 nm and scanning from 200 to 600 nm). This was interfaced with an ion-trap mass
spectrometer fitted with an ESI source (Bruker Daltonics HCT Ultra, Bremen, Germany)
operating in full scan, auto MSn mode to obtain fragment ion m/z. As necessary, MS2, MS3
and MS4 fragment-targeted experiments were performed to focus only on compounds
producing a parent ion at m/z 397, 559, and 573. Tandem mass spectra were acquired in Auto-
MSn mode (smart fragmentation) using a ramping of the collision energy. Maximum
fragmentation amplitude was set to 1 Volt, starting at 30% and ending at 200%. MS operating
conditions (negative mode) had been optimized using 5-caffeoylquinic acid with a capillary
temperature of 365 oC, a dry gas flow rate of 10 L/min, and a nebulizer pressure of 10 psi.
Data processing
LC-MS data were processed using Data Analysis 4.0 (Bruker Daltonics, Bremen). Raw
calibrated LC-MS data were further processed by Profile Analysis 2.0 (Bruker Daltonics,
Bremen) and if required further processed using Origin 7.0 and Matlab. Buckets were created
in an m/z value range between 300 and 900, unless stated otherwise, with a bucket size of 60 s
and 1 Da. Kernels were defined as 20 s and 0.2 Da.
6.4 Results and discussion
The aim of this contribution is to assess the value of PCA analysis in terms of the data
obtained from four different spectroscopic techniques (IR CD, NMR and MS) and the ability
of the PCA analysis to differentiate between samples, taking green coffee beans as an
example. Detailed questions that require addressing are whether it is possible to distinguish
green coffee beans according to several parameters including variety of the coffee (Arabica or
Robusta), geographical origin of the coffee, growth conditions (e.g. altitude) or processing
conditions. How should the PCA parameters and methods be chosen in order to achieve
optimal distinction? Should a distinction by PCA be possible? How can such a distinction be
rationalised on a molecular level?
For each PCA scores plot a PCA analysis produces a so called loading plot, in which the most
important data points that are responsible for the distinction are displayed. Ideally, a PCA
169
analysis provides a list of unique molecular markers, unique to each sample group,
distinguishing two sample groups. What is the nature of these points in the loading plot? Are
they molecular markers identified by PCA unique to a sample or not and if not how should
PCA be carried out to identify unique molecular markers?
In order to address all of these questions we have analyzed a series of aqueous methanolic
extracts of 38 different green coffee bean samples by high resolution LC-ESI-TOF-MS in the
negative ion mode, CD (circular dichroism), IR (Infrared) and NMR spectroscopy (Table
6.1). For the extraction process we used an optimized extraction method, if compared to
previous work,1 using a mild Soxhlet method followed by protein removal with Carrez
reagent and subsequent freeze drying to yield bright yellow to orange powders. A total of 38
commercial green bean coffee samples, 10 Robusta samples and 28 Arabica samples of
different geographic origins were extracted.
6.4.1 LC-MS
LC-MS conditions used were as described earlier.19 In addition to the high resolution mass
measurements we carried out LC-ESI-tandem-MS measurements using an ion trap mass
spectrometer to be able to assign individual compounds not only on the basis of retention time
and high resolution m/z value, but as well to use fragmentation data for correct structure
assignment.
Similar to previous work, around 50-100 well resolved chromatographic peaks could be
identified in each chromatogram and peaks assigned to individual distinct compounds, in the
majority chlorogenic acids. A list of selected compounds identified is given in Table 6.2. A
typical chromatogram of a Robusta sample is shown in Figure 6.1.
170
Figure 6.1 Representative chromatogram of green coffee extract of sample No. 33 (Tanzania
Robusta), a) TIC in negative ion mode; b) UV-VIS chromatogram monitored at 320 nm
6.4.2 Circular Dichroism spectroscopy
The compounds under investigation within the green bean coffee extracts are almost
exclusively chlorogenic acids and their individual profile is assumed to be suited for
investigation by CD spectroscopy. CGAs are chiral non-racemic compounds of natural origin
that carry depending on their structure one to three chromophoric moieties. Mono-acyl quinic
acids produce simple CD spectra with a single maximum or minimum, whereas diacyl quinic
acids produce more complex CD spectra with a characteristic Cotton effect, due to the
interaction of the two distinct cinnamate chromophores. Therefore, PCA analysis using CD
spectroscopy should be ideally suited to assess variations in the relative ratio of monoacyl to
diacyl quinic acids. Typical CD spectra of 3-CQA and 3,5-diCQA are shown in Figure 6.2.
To the best of our knowledge no PCA was ever carried out using CD spectroscopy, despite
the fact that most natural products are chiral and many natural products possess diagnostic
chromophores.
Two characteristic CD spectra of a typical chlorogenic acid extract from green coffee beans
are as well shown in Figure 6.2. From the spectra, the individual regions characteristic for
single chromophore and dichromophoric systems can be appreciated.
TANZROBUSTA_111.D: TIC -All MS
TANZROBUSTA_111.D: UV Chromatogram, 318-322 nm0.0
0.2
0.4
0.6
0.8
1.0
1.2
7x10
Intens.
0
200
400
600
800
1000
Intens.
[mAU]
10 20 30 40 50 Time [min]
a
b
171
Figure 6.2 CD spectra of 3-CQA and 3,5-diCQA
A non-targeted (unsupervised) PCA analysis was carried out, in which the full data set was
processed. Once the principal components were calculated, an inspection of the various PCA
score plots allowed the identification of groups of samples. By inspection of the
characteristics of each individual data point in the groups in the score plot a conclusion can be
drawn with respect to the nature of these groups. Data points can thus be labelled according to
the groups identified.
-40
-30
-20
-10
0
10
20
30
40
50
0 50 100 150 200 250 300 350 400 450
3-CQAC
D (
md
eg)
-12
-10
-8
-6
-4
-2
0
2
4
6
8
100 150 200 250 300 350 400 450
3,5-diCQA
CD
(m
deg
)
172
Table 6.1 Origins, nature and grouping of green bean coffee samples analysed and included
in PCA analysis
Sample No. Origin / Type Arabica / Robusta Group
1 Tanzania Arabica A1
2 Guatemala SHG Arabica A1
3 Peru Bio Arabica A1
4 Nicaragua Maragogype Arabica A1
5 Kenya AA Arabica A1
6 Athiopien Wild Forest Bio Arabica A1
7 Athiopien Yivgachette Arabica A1
8 Athiopien Mokka Sidamo 2 Arabica A1
9 Reizaow Arabica A1
10 Coffeein Free Arabica A1
11 Costarica 2 Arabica A1
12 Brasilien 1 Arabica A1
13 Brasilien 2 Arabica A1
14 Maragogype Arabica A2
15 Malawi Pamwamba Arabica A2
16 Panama Boquete Arabica A2
17 Kenia 1 Arabica A2
18 Honduras Bio Arabica A2
19 Kameruls Arabica A2
20 Nicaragua Mataglpa Arabica A2
21 Costarica 1 Arabica A2
22 Columbia Exulso Arabica A2
23 Papua Neuguinea Arabica A3
24 Athiopien Mokka Sidamo 1 Arabica A3
173
25 Costarica 3 Arabica A3
26 Ethiopien Arabica A3
27 Indian Perl Mountain Arabica A3
28 Brazilien Santos Arabica A3
29 Indian 1 Robusta R
30 India Cherry AB Robusta R
31 Uganda Robusta R
32 India Parchment Robusta R
33 Tanzania Robusta R
34 Indonesia 1 Robusta R
35 Togo 1 Robusta R
36 Cameron Robusta R
37 Indonesia 2 Robusta R
38 India Cherry A Robusta R
Table 6.2 Numbering, nomenclature and high resolution MS data of selected secondary
metabolites identified in green bean coffee samples 21, 22
No. Name Mol.
formula
Theor. m/z (M-
H)
Exp. m/z (M-
H)
Error
(ppm)
1 3-O-caffeoylquinic acid C16H18O9 353.0878 353.0881 -0.7
2 4-O-caffeoylquinic acid C16H18O9 353.0878 353.0884 -1.6
3 5-O-caffeoylquinic acid C16H18O9 353.0878 353.0892 -3.9
4 3-O-feruloylquinic acid C17H20O9 367.0929 367.1047 -3.4
5 4-O-feruloylquinic acid C17H20O9 367.0929 367.1038 -0.8
6 5-O-feruloylquinic acid C17H20O9 367.0929 367.1045 -2.9
7 3-O-p-coumaroylquinic acid C16H18O8 337.0929 337.0931 -0.5
8 4-O-p-coumaroylquinic acid C16H18O8 337.0929 337.0921 2.4
9 5-O-p-coumaroylquinic acid C16H18O8 337.0929 337.0921 2.4
10 3-O-dimethoxycinnamoylquinic acid C18H22O9 381.1191 381.1202 -2.8
174
11 4-O-dimethoxycinnamoylquinic acid C18H22O9 381.1191 381.1191 -2.5
12 5-O-dimethoxycinnamoylquinic acid C18H22O9 381.1191 381.1202 -2.8
13 3-O-sinapoylquinic acid C18H22O10 397.1140 397.1125 3.8
14 4-O-sinapoylquinic acid C18H22O10 397.1140 397.1150 -2.5
15 5-O-sinapoylquinic acid C18H22O10 397.1140 397.1140 -4.9
16 3,4-di-O-caffeoylquinic acid C25H24O12 515.1195 515.1190 1.0
17 3,5-di-O-caffeoylquinic acid C25H24O12 515.1195 515.1172 4.5
18 4,5-di-O-caffeoylquinic acid C25H24O12 515.1195 515.1170 4.9
19 3,4-di-O-feruloylquinic acid C27H28O12 543.1508 543.1512 -0.8
20 3,5-di-O-feruloylquinic acid C27H28O12 543.1508 543.1514 -1.1
21 4,5-di-O-feruloylquinic acid C27H28O12 543.1508 543.1539 -3.4
25 3-O-feruloyl-4-O-caffeoylquinic acid C26H26O12 529.1351 529.1343 1.7
26 3-O-caffeoyl-4-O-feruloylquinic acid C26H26O12 529.1351 529.1351 -0.1
27 3-O-feruloyl-5-O-caffeoylquinic acid C26H26O12 529.1351 529.1373 -4.0
28 3-O-caffeoyl-5-O-feruloylquinic acid C26H26O12 529.1351 529.1367 -3.0
29 4-O-feruloyl-5-O-caffeoylquinic acid C26H26O12 529.1351 529.1351 0.1
30 4-O-caffeoyl-5-O-feruloylquinic acid C26H26O12 529.1351 529.1349 0.5
31 3-O-dimethoxycinnamoyl-4-O-
caffeoylquinic acid
C27H28O12 543.1508 543.1488 3.6
32 3-O-dimethoxycinnamoyl-5-O-
caffeoylquinic acid
C27H28O12 543.1508 543.1491 3.1
33 4-O-dimethoxycinnamoyl-5-O-
caffeoylquinic acid
C27H28O12 543.1508 543.1526 -3.4
34 3-O-dimethoxycinnamoyl-4-O-
feruloylquinic acid
C27H28O12 543.1508 543.1508 -4.1
35 3-O-dimethoxycinnamoyl-5-O-
feruloylquinic acid
C27H28O12 543.1508 543.1515 -1.4
36 4-O-dimethoxycinnamoyl-5-O-
feruloylquinic acid
C27H28O12 543.1508 543.1525 -3.1
37 3-O-p-coumaroyl-4-O-caffeoylquinic
acid
C25H24O11 499.1246 499.1227 3.7
38 3-O-caffeoyl-4-O-p-coumaroylquinic
acid
C25H24O11 499.1246 499.1247 -0.2
39 3-O-p-coumaroyl-5-O-caffeoylquinic C25H24O11 499.1246 499.1248 -0.5
175
acid
40 3-O-caffeoyl-5-O-p-coumaroylquinic
acid
C25H24O11 499.1246 499.1247 -0.2
41 4-O-caffeoyl-5-O-p-coumaroylquinic
acid
C25H24O11 499.1246 499.1246 -4.9
42 4-O-p-coumaroyl-5-O-caffeoylquinic
acid
C25H24O11 499.1246 499.1249 -0.6
43 3-O-p-coumaroyl-4-O-feruloylquinic
acid
C26H26O11 513.1402 513.1389 2.6
44 3-O-p-coumaroyl-5-O-feruloylquinic
acid
C26H26O11 513.1402 513.1141 -2.9
45 4-O-p-coumaroyl-5-O-feruloylquinic
acid
C26H26O11 513.1402 513.1406 -0.7
49 3-O-sinapoyl-5-O-caffeoylquinic acid C27H28O13 559.1457 559.1481 -4.2
50 3-O-sinapoyl-4-O-caffeoylquinic acid C27H28O13 559.1457 559.1472 -2.6
51 3-O-(3,5-dihydroxy-4-
methoxy)cinnamoyl-4-O-feruloylquinic
acid
C27H28O13 559.1457 559.1458 -0.2
52 4-O-sinapoyl-3-O-caffeoylquinic acid C27H28O13 559.1457 559.1457 0.9
53 3-O-sinapoyl-5-O-feruloylquinic acid C28H30O13 573.1614 573.1641 -4.7
54 4-O-sinapoyl-5-O-feruloylquinic acid C28H30O13 573.1614 573.1599 -2.5
55 4-O-sinapoyl-3-O-feruloylquinic acid C28H30O13 573.1614 573.1634 -3.5
56 4-O-trimethoxycinnamoyl-5-O-
caffeoylquinic acid
C28H30O13 573.1614 573.1611 0.4
57 3-O-trimethoxycinnamoyl-5-O-
caffeoylquinic acid
C28H30O13 573.1614 573.1623 -1.7
58 3-O-trimethoxycinnamoyl-5-O-
feruloylquinic acid
C29H32O13 587.1770 587.1748 3.8
59 3-O-trimethoxycinnamoyl-4-O-
feruloylquinic acid
C29H32O13 587.1770 587.1766 0.7
60 4-O-trimethoxycinnamoyl-5-O-
feruloylquinic acid
C29H32O13 587.1770 587.1764 1.0
61 3-O-dimethoxycinnamoyl-4-O-feruloyl-
5-O-caffeoylquinic acid
C37H36O15 719.1981 719.2001 -2.7
62 3,4,5-tri-O-caffeoylquinic acid C34H29O15 677.1512 677.1522 -3.5
63 3,5-di-O-caffeoyl-4-O-feruloylquinic
acid
C35H31O15 691.1668 691.1647 3.1
176
64 3-O-feruloyl-4,5-di-O-caffeoylquinic
acid
C35H31O15 691.1668 691.1711 -6.2*
65 3,4-di-O-caffeoyl-5-O-feruloylquinic
acid
C35H31O15 691.1668 691.1647 3.1
66 3-O-caffeoyl-4,5-di-O-feruloylquinic
acid
C36H33O15 705.1825 705.1851 -3.8
67 3,4-di-O-feruloyl-5-O-caffeoylquinic
acid
C36H33O15 705.1825 705.1833 -1.1
68 3,4-di-O-caffeoyl-5-O-sinapoylquinic
acid
C36H33O16 721.1774 721.1795 -2.9
69 3-O-sinapoyl-4,5-di-O-caffeoylquinic
acid
C36H33O16 721.1774 721.1766 1.1
The PCA score and loading plots of the obtained CD data is shown in Figure 6.3 using 34
spectra of reasonable quality. While the majority of Robusta coffee samples cluster in one
region and the majority of Arabica sample cluster in a second region differentiated on the PC
1 axis, a total of four Arabica samples fall within the plot area of the Robusta samples.
Therefore it must be concluded that a reliable distinction, in the absence of further undue data
manipulation, is not possible using CD data.
Figure 6.3 The PCA score and loading plots of the obtained CD spectral data (Continued)
-3
-2,5
-2
-1,5
-1
-0,5
0
0,5
1
1,5
2
-15 -10 -5 0 5 10 15
Robusta
ArabicaPC1
PC2
177
Figure 6.3 The PCA score and loading plots of the obtained CD spectral data
6.4.3 Infrared spectroscopy
As mentioned previously, vibrational spectroscopy has in the past been successfully used to
differentiate between Robusta and Arabica green coffee bean samples using both IR and
Raman techniques. Loading plots revealed that distinction is mainly based on ester C=O
absorptions of CGAs and lipids.
A typical ATR-IR spectrum is shown in Figure 6.4. The PCA score and loading plots of 20
samples providing good quality IR spectra is shown in Figure 6.5.
As can be seen from the score plot in Figure 6.5, a reliable distinction between Arabica and
Robusta samples is possible in the PC 1 dimension with loading plots confirming differences
of absorption in the ester carbonyl region of the spectra.
178
Figure 6.4 ATR-IR spectrum of Panama Boguete Arabica extract
Figure 6.5 The PCA score and loading plots of the obtained IR spectral data (Continued)
-1
-0,8
-0,6
-0,4
-0,2
0
0,2
0,4
0,6
0,8
1
1,2
-3 -2 -1 0 1 2 3 4 5
PC1
PC2
Arabica
Robusta
179
Figure 6.5 The PCA score and loading plots of the obtained IR spectral data
6.4.4 1H NMR spectroscopy
NMR based methods have been used previously on several occasions for the distinction
between Arabica and Robusta green bean coffee samples.22-25 1H-NMR spectra of all 38
samples have been acquired in DMSO-d6 and a representative spectrum is shown in Figure
6.6.
The spectrum is dominated by signals corresponding to chlorogenic acids, saccharides,
caffeine and small organic acids. The resulting PCA score plot is shown in Figure 6.7.
180
Figure 6.6 1H-NMR spectra of Tanzania Robusta in DMSO-d6
Figure 6.7 The PCA score plot of the obtained NMR spectral data
-2
-1,5
-1
-0,5
0
0,5
1
1,5
2
2,5
-1,5 -1 -0,5 0 0,5 1 1,5 2 2,5 3 3,5 4
Robusta
PC1
PC2
Arabica
181
Again the Robusta green bean samples cluster nicely in one region of the plot, whereas the
Arabica samples are scattered all over the score plot making reliable distinction not possible.
A closer analysis of the spectra reveals that CGA signals suffer from relative large chemical
shift variations (> 0.2 ppm) presumably due to the well reported CGA-caffeine interactions.
Non-covalent complexes stabilised by π-π interactions are formed between caffeine and
CGAs in aqueous solutions resulting in concentration dependant aromatic solvent induced
chemical shift changes and hence spectroscopic changes affecting the success of PCA
analysis. This problem can be overcome by incorporating different NMR techniques such as
13C NMR spectroscopy as shown by Wei et al.31 and DOSY NMR as a virtual separation
technique, which has been proved to discriminate the complex natural mixtures to the
regioisomeric level by using matrix assisted approach as shown by Gresley et al.. 32
6.5 Conclusions
In conclusion we have shown that PCA analysis using data sets obtained by different
spectroscopic methods of identical coffee extracts show distinctly different success in
distinguishing groups of samples. In the particular case of chlorogenic acid extracts from
Arabica and Robusta coffee the use of LC-MS data and IR spectroscopy allowed
straightforward distinction of sample groups. In contrast Circular Dichroism and NMR
spectroscopy failed to give a satisfactory level of distinction. From these results it becomes
obvious that one size does not fit all and that in order to carry out successful multivariant
statistical analysis great care and consideration needs to be taken to choose the correct
experimental technique. In this particular case NMR spectroscopy failed to achieve
differentiation of samples since the many chlorogenic acid derivatives present in green coffee
beans are structurally very similar and hence do not produce NMR signals, in which the
individual components are well resolved. Secondly in NMR spectroscopy non-covalent
interactions of CGAs with caffeine and an accompanying concentration dependant change of
chemical shifts hinders a reliable analysis. From this we propose that prior to carrying out
PCA analysis the ability of the spectroscopic method used to differentiate and resolve the
compounds of interest must be assessed. Secondly it should be established whether
interactions between individual components result in variations of experimental parameters.
182
References
1 J. W. Drynan, M. N. Clifford, J. Obuchowicz and N. Kuhnert, Natural Product Reports,
2010, 27, 417-462.
2 R. Jaiswal, T. Sovdat, F. Vivan and N. Kuhnert, Journal of Agricultural and Food
Chemistry, 2010, 58, 5471-5484.
3 S. Wold, K. Esbensen and P. Geladi, Chemometrics and Intelligent Laboratory Systems,
1987, 2, 37-52.
4 J. K. Nicholson, J. C. Lindon and E. Holmes, Xenobiotica, 1999, 29, 1181-1189.
5 D. Krug, G. Zurek, B. Schneider, C. Bassmann and R. Muller, Lc Gc Europe, 2007, 41-42.
6 D. Krug, G. Zurek, B. Schneider, R. Garcia and R. Muller, Analytica Chimica Acta, 2008,
624, 97-106.
7 M. N. Clifford, K. L. Johnston, S. Knight and N. Kuhnert, Journal of Agricultural and
Food Chemistry, 2003, 51, 2900-2911.
8 M. N. Clifford, S. Knight, B. Surucu and N. Kuhnert, Journal of Agricultural and Food
Chemistry, 2006, 54, 1957-1969.
9 M. N. Clifford, S. Marks, S. Knight and N. Kuhnert, Journal of Agricultural and Food
Chemistry, 2006, 54, 4095-4101.
10 R. Briandet, E. K. Kemsley and R. H. Wilson, Journal of Agricultural and Food
Chemistry, 1996, 44, 170-174.
11 G. Downey, R. Briandet, R. H. Wilson and E. K. Kemsley, Journal of Agricultural and
Food Chemistry, 1997, 45, 4357-4361.
12 D. J. Lyman, R. Benck, S. Dell, S. Merle and J. Murray-Wijelath, Journal of Agricultural
and Food Chemistry, 2003, 51, 3268-3272.
13 I. Esteban-Diez, J. M. Gonzalez-Saiz, C. Saenz-Gonzalez and C. Pizarro, Talanta, 2007,
71, 221-229.
14 J. Wang, S. Jun, H. C. Bittenbender, L. Gautz and Q. X. Li, Journal of Food Science, 2009,
74, C385-C391.
15 A. B. Rubayiza and M. Meurens, Journal of Agricultural and Food Chemistry, 2005, 53,
4654-4659.
16 R. M. El-Abassy, P. Donfack and A. Materny, Food chemistry, in press.
17 M. S. Valdenebro, M. Leon-Camacho, F. Pablos, A. G. Gonzalez and M. J. Martin,
Analyst, 1999, 124, 999-1002.
18 M. Korhonova, K. Hron, D. Klimcikova, L. Mueller, P. Bednar and P. Bartak, Talanta,
2009, 80, 710-715.
183
19 N. Kuhnert, R. Jaiswal, M. F. Matei, T. Sovdat and S. Deshpande, Rapid Communications
in Mass Spectrometry, 2010, 24, 1575-1582.
20 R. A. van den Berg, H. C. J. Hoefsloot, J. A. Westerhuis, A. K. Smilde and M. J. van der
Werf, Bmc Genomics, 2006, 7.
21 R. Jaiswal and N. Kuhnert, Rapid Communications in Mass Spectrometry, 2010, 24, 2283-
2294.
22 R. Jaiswal, M. A. Patras, P. J. Eravuchira and N. Kuhnert, Journal of Agricultural and
Food Chemistry, 2010, 58, 8722-8737.
23 J. Baggenstoss, L. Poisson, R. Kaegi, R. Perren and F. Eschert, Journal of Agricultural and
Food Chemistry, 2008, 56, 5847-5851.
24 S. Gal, P. Windemann and E. Baumgartner, Chimia, 1976, 30, 68-71.
25 I. M. Kamal, V. Sobolik, M. Kristiawan, S. M. Mounir and K. Allaf, Innovative Food
Science & Emerging Technologies, 2008, 9, 534-541.
26. M. N. Clifford, W. Zheng and N. Kuhnert, Phytochemical Analysis, 2006, 17, 384-
393.
27 R. A. van den Berg, C. M. Rubingh, J. A. Westerhuis, M. J. van der Werf and A. K.
Smilde, Analytica Chimica Acta, 2009, 651, 173-181.
28. D. Perrone, A. Farah, C. M. Donangelo, T. de Paulis and P. R. Martin, Food Chemistry,
2008, 106, 859-867.
29. M. N. Clifford, J. Kirkpatrick, N. Kuhnert, H. Roozendaal and P. R. Salgado, Food
Chemistry, 2008, 106, 379-385.
30. M. N. Clifford, W. G. Wu and N. Kuhnert, Food Chemistry, 2006, 95, 574-578.
31. F. Wei, K. Furihata, M. Koda, F. Hu, R. Kato, T. Miyakawa and M. Tanokura, Journal of
Agricultural and Food Chemistry 2012, 60 (40), 10118-10125.
32. A. Gresley, J. Kenny, C. Cassar, A. Kelly, A. Sinclair and M. Fielder, Food
Chemistry 2012, 135 (4), 2879-2886.
184
Conclusions
In this study it was observed that the acyl migration phenomena occur before dehydration
takes place at the quinic acid moiety. Acyl migration is facilitated in presence of the liquid
media as compared to the roasting process. Therefore, the lower temperature roasts like the
‘cinnamon roast’ produce a large number of acyl migration products than higher temperature
roasts which generate dehydration products like lactones and shikimates high in numbers.
Esters present on C3 position of the quinic acid are prone to hydrolysis of the ester bond than
undergoing acyl migration in any experimental condition. The amount of esters present on C3
and C4 positions of quinic acid moiety in a cup of coffee after roasting and brewing processes
is highly contributed by C5 positioned esters in case of mono-caffeoylquinic acids content. In
contrast to this observation we found that acyl migration to C5 position from C3 and C4 is only
possible in base hydrolysis i.e. it is highly pH dependant. 1,3-diCQA and 5-CQA were
observed to be more stable than the rest of the reference standards in both roasting and
brewing conditions.
In this contribution, muco-quinic acid, scyllo-quinic acid, epi-quinic acid and cis-quinic acid
were selectively synthesized. Their behavior in LC-MSn along with commercially available (-
)-quinic acid was studied. For the first time it was observed that these diastereoisomers are
distinguishable on the basis of their fragmentation behavior as well as their chromatographic
elution order. In this study, it was observed that muco-quinic acid, scyllo-quinic acid and epi-
quinic acid are present in hydrolyzed Guatemala roasted coffee sample as possible products of
roasting. Non selective isomerization of (-)-quinic acid using acetic acid/conc. H2SO4 was
performed from which, epi-quinic acid, scyllo-quinic acid and (-)-quinic acid could be
identified using newly assigned fragmentation schemes and retention times characteristic to
the specific compound.
In this work, the further condensation of the CGAs and their decomposition products with the
non-volatile fraction of the total acid content of the roasted coffee samples is reported.
Selected small organic acids were heated individually with 5-caffeoyl quinic acid to check if
simulated roasting conditions facilitate the formation of the transesterification products. Same
experimental conditions were used incorporating caffeic acid and quinic acid as well. Also, 5-
caffeoylquinic acid, caffeic acid and quinic acid were heated in presence of the mixture of all
the organic acids separately to check, which of the organic acid show greater affinity towards
the formation of the condensed esters. All the samples acquired from these experiments were
analyzed by high resolution ESI-TOF-MS. Four green coffee samples were also roasted in the
185
conditions described earlier and then analyzed by ESI-FT-ICR-MS to identify the
transesterification products in roasted coffee samples. Ten new chlorogenic acid derivatives
were identified in the roasted coffee samples.
PCA analysis using data sets obtained by different spectroscopic methods of identical coffee
extracts show distinctly different success in distinguishing groups of samples. In the particular
case of chlorogenic acid extracts from Arabica and Robusta coffee the use of LC-MS data and
IR spectroscopy allowed straightforward distinction of sample groups. In contrast, Circular
Dichroism and NMR spectroscopy failed to give a satisfactory level of distinction. From these
results it becomes obvious that one size does not fit all and that in order to carry out
successful multivariant statistical analysis great care and consideration needs to be taken to
choose the correct experimental technique. In this particular case NMR spectroscopy failed to
achieve differentiation of samples since the many chlorogenic acid derivatives present in
green coffee beans are structurally very similar and hence do not produce NMR signals, in
which the individual components are well resolved. Secondly in NMR spectroscopy non-
covalent interactions of CGAs with caffeine and an accompanying concentration dependent
change of chemical shifts hinders a reliable analysis. It is established from this work that prior
to carrying out PCA analysis the ability of the spectroscopic method used to differentiate and
resolve the compounds of interest must be assessed. Secondly it should be established
whether interactions between individual components result in variations of experimental
parameters.
186
List of publications and manuscripts
1. Eravuchira, P.J.; El-Abassy, R.M.; Deshpande, S.; Matei, M.F.; Mishra, S.; Tandon, P.;
Kuhnert, N.; Materny, A. Raman spectroscopic characterization of different regioisomers of
monoacyl and diacyl chlorogenic acid. Vib. Spectrosc. 2012, 61, 10-16.
2. Jaiswal, R.; Deshpande, S.; Kuhnert, N. Profiling the chlorogenic acids of Rudbeckia hirta,
Helianthus tuberosus, Carlina acaulis and Symphyotrichum novae-angliae leaves by LC-MSn.
Phytochem. Anal. 2011, 22, 432-441.
3. Kuhnert, N.; Jaiswal, R.; Matei, M.F.; Sovdat, T.; Deshpande, S. How to distinguish
between feruloyl quinic acids and isoferuloyl quinic acids by liquid chromatography/tandem
mass spectrometry. Rapid Commun. Mass Spectrom. 2010, 24, 1575-1582.
4. Kuhnert, N.; Dairpoosh, F.; Jaiswal, R.; Matei, M.; Deshpande, S.; Golon, A.; Nour, H.;
Karakose, H.; Hourani, N. Hill coefficients of dietary polyphenolic enzyme inhibitiors: can
beneficial health effects of dietary polyphenols be explained by allosteric enzyme denaturing?
J Chem Biol 2011, 4, 109-116.
5. Jaiswal, R.; Deshpande, S.; Kuhnert, N.; An investigation of the Hydroxycinnamate Profile
of Six Galium Plants from the Rubiaceae Family. Phytochemical Anal. 2013, (Submitted).
6. Kuhnert, N.; Karakoese, H.; Jaiswal, R.; Deshpande, S. Investigating the Photochemical
Changes of Chlorogenic Acids Induced by UV Light in Model Systems and in Agricultural
Practice with Stevia rebaudiana Cultivation as an Example. 2013, (Manuscript)
7. Deshpande, S.; El-Abassy, R.M.; Jaiswal, R.; Eravuchira, P.J.; Kammer, B.; Materny, A.;
Kuhnert, N. Which spectroscopic technique allows best differentiation of coffee varieties:
Comparing principal component analysis using data derived from CD-, NMR-, IR-
spectroscopy and LC-MS in the analysis of the chlorogenic acid fraction in green coffee
beans. Analytical Meth. 2013, (Submitted)
8. Deshpande, S.; Jaiswal, R.; Matei, M.; Kuhnert, N. Acyl migration in mono- and di-
caffeoylquinic acids under basic and aqueous acidic conditions and dry roasting conditions.
2013, (Manuscript)
9. Deshpande, S.; Matei, M.; Jaiswal, R.; Bassem, B.; Kuhnert, N. Synthesis, structure and
tandem MS investigation of diastereomers of quinic acid. 2013, (Manuscript)
187
10. Deshpande, S.; Kuhnert, N. Transesterification of chlorogenic acids with small organic
acids present in the coffee bean. 2013, (Manuscript)
188
Curriculum Vitae
Dr. Sagar Anil Deshpande
Seefahrtstrasse 7,
Bremen 28759,
Germany
Telephone (Germany): +49-17684146335
Telephone (India): +91-7719995630
Email: [email protected]
Nationality: Indian
Birth Place: Pune, India
Marital Status: Married
Skills Synthesis, isolation, identification and structure elucidation of natural
products.
Analysis of complex mixtures, food materials, beverages, food
processing and economically important plants.
Expertise in analytical and preparative HPLC, LC-TOF-MS, FT-ICR-
MS, LC-MS/MS, LC-Tandem-MS, MALDI-TOF, NMR, IR, CD and UV-
Vis spectrometers.
Education Jacobs University Bremen
24th January 2014
Doctor of Philosophy (PhD) in Chemistry with Prof. Dr. Nikolai
Kuhnert
Dissertation: “Mass Spectrometry Based Investigation of
Chlorogenic Acid Reactivity and Profile in Model Systems and
Coffee processing”, ten publications till date.
189
Jacobs University Bremen
September 2010
Master of Science (MSc) in Nanomolecular Science, with the
completion of Thesis entitled, “Analysis of Plant Secondary
Metabolites with Different Analytical Techniques” with Prof. Dr.
Nikolai Kuhnert.
N. Wadia College, Pune University
October 2007
Master of Science (MSc) in Organic Chemistry, with the
completion of Thesis entitled, “Methoxycarbonylation of Amines
with Organic Carbonates Catalyzed by Lead Compounds” with Mr.
Sunil Joshi (Scientist E II), Homogeneous Catalysis Division,
National Chemical Laboratory (NCL) Pune.
S. P. College, Pune University
May 2005
Bachelor of Science (BSc) in Chemistry.
Research
Experience
PhD research, Jacobs University Bremen
September 2010– January 2014
Synthesis of natural products, especially diastereomers of (-)-quinic acid,
chlorogenic acids, chlorogenic acid lactones and hydroxycinnamoyl-
shikimates; synthesis of natural products in three to eight steps, using
protecting group strategies; synthetic compounds characterization by melting
point, IR, NMR, HRMS, Tandem-MS, XRD. Monitoring chemical
transformations in chlorogenic acid profile in green coffee during coffee
processing incorporating analysis of green and roasted coffee by LC-TOF-MS,
FT-ICR-MS and LC-Tandem-MS. Structural elucidation using LC-tandem MS.
Analysis of stevia, arnica, burdock and gardenia LC-TOF-MS and LC-Tandem-
MS for their polyphenol contents.
MSc laboratory rotations and thesis research, Jacobs University
Bremen
August 2008– September 2010
Organic and organometallic synthesis of chlorogenic acids. NMR, IR,
fluorescence, UV-visible and CD spectroscopy, Preparative HPLC, tandem and
high resolution mass spectrometry coupled with LC/GC. Principal Component
Analysis of green coffee bean extracts.
MSc thesis research, National Chemical Laboratory, Pune
May 2006– January 2007
Catalytic synthesis of carbamates, optimization of the process through
screening of substrates and catalysts.
Other Experience Teaching assistant, Advanced integrated organic and analytical
chemistry laboratory, Jacobs University Bremen
September 2013– December 2013
190
Supervising and grading BSc students for laboratory performance, delivering
introductory talks for experiments including the directions for instrument
handling.
Supervision of BSc research projects, Jacobs University Bremen
May 2012- July 2012
Training, directly supervising and partially grading students for BSc laboratory rotations and theses. Resident associate for Bremische and Jacobs University Bremen
May 2012 – Present
Representative of Jacobs University and member of Bremische staff. Ensuring
harmony between students living off campus and locals. Providing assistance
and solutions for tenancy related issues.
Awards Full merit based Jacobs University stipend for MSc studies
August 2008 – September 2010
PhD stipend for the third party funded project from Kraft Foods
September 2010 – August 2013
Language skills Marathi: Mother tongue English: Fluent
Hindi: Fluent German: Basic knowledge (A2)
Computer skills ChemDraw and ChemSketch.
SciFinder, Reaxys (Beilstein), Web of Knowledge, RefWorks and EndNote.
Data Analysis (4.0) and Profile Analysis (Bruker Daltonics).
Microsoft Office and Microsoft Operating Systems (Windows all versions).
References
Prof. Dr. Nikolai Kuhnert, FRSC Prof. Dr. Michael N. Clifford (Emeritus)
Room 117, Research III Centre for Nutrition and Food Safety
Chemistry, School of Engineering and Science School of Biomedical and Molecular Sciences
Campus Ring 8 University of Surrey
Jacobs University Bremen Guildford, Surry GU2 7XH, UK
Bremen 28759, Germany Email: [email protected]
Email: [email protected] (To be contacted by email only)
Phone: 0049-4212003120
191
List of publications and manuscripts
1. Eravuchira, P.J.; El-Abassy, R.M.; Deshpande, S.; Matei, M.F.; Mishra, S.; Tandon, P.; Kuhnert,
N.; Materny, A. Raman spectroscopic characterization of different regioisomers of monoacyl and diacyl
chlorogenic acid. Vib. Spectrosc. 2012, 61, 10-16.
2. Jaiswal, R.; Deshpande, S.; Kuhnert, N. Profiling the chlorogenic acids of Rudbeckia hirta,
Helianthus tuberosus, Carlina acaulis and Symphyotrichum novae-angliae leaves by LC-MSn.
Phytochem. Anal. 2011, 22, 432-441.
3. Kuhnert, N.; Jaiswal, R.; Matei, M.F.; Sovdat, T.; Deshpande, S. How to distinguish between
feruloyl quinic acids and isoferuloyl quinic acids by liquid chromatography/tandem mass
spectrometry. Rapid Commun. Mass Spectrom. 2010, 24, 1575-1582.
4. Kuhnert, N.; Dairpoosh, F.; Jaiswal, R.; Matei, M.; Deshpande, S.; Golon, A.; Nour, H.; Karakose,
H.; Hourani, N. Hill coefficients of dietary polyphenolic enzyme inhibitiors: can beneficial health
effects of dietary polyphenols be explained by allosteric enzyme denaturing? J Chem Biol 2011, 4, 109-
116.
5. Jaiswal, R.; Deshpande, S.; Kuhnert, N.; An investigation of the Hydroxycinnamate Profile of Six
Galium Plants from the Rubiaceae Family. Phytochemical Anal. 2013, (Submitted).
6. Kuhnert, N.; Karakoese, H.; Jaiswal, R.; Deshpande, S. Investigating the Photochemical Changes
of Chlorogenic Acids Induced by UV Light in Model Systems and in Agricultural Practice with Stevia
rebaudiana Cultivation as an Example. 2013, (Manuscript)
7. Deshpande, S.; El-Abassy, R.M.; Jaiswal, R.; Eravuchira, P.J.; Kammer, B.; Materny, A.; Kuhnert,
N. Which spectroscopic technique allows best differentiation of coffee varieties: Comparing principal
component analysis using data derived from CD-, NMR-, IR- spectroscopy and LC-MS in the analysis
of the chlorogenic acid fraction in green coffee beans. Analytical Meth. 2013, (Submitted)
8. Deshpande, S.; Jaiswal, R.; Matei, M.; Kuhnert, N. Acyl migration in mono- and di-caffeoylquinic
acids under basic and aqueous acidic conditions and dry roasting conditions. 2013, (Kraft Foods
approval pending)
9. Deshpande, S.; Matei, M.; Jaiswal, R.; Bassem, B.; Kuhnert, N. Synthesis, structure and tandem
MS investigation of diastereomers of quinic acid. 2013. (Kraft Foods approval pending)
10. Deshpande, S.; Kuhnert, N. Transesterification of chlorogenic acids with small organic acids
present in the coffee bean. 2013. (Kraft Foods approval pending)
Conferences and Workshops
Presented a poster in 5th International Conference on Polyphenols and Health
(ICPH) in Sitges, Barcelona in 2011.
Presented a poster in GDCh-Wissenschaftsforum Chemie 2011 in Bremen, 04 - 07
September 2011
Presented a poster at the National Chemical Laboratory, Pune, India. This was a part of 6TH
International Annual Symposium on Catalysis (CAMURE 6) January 2007
Mass Spectrometry Workshop, Jacobs University Bremen, Bremen, Germany, July 24 – 27,
2013