Phytochemical Characterization of Stevia rebaudiana

Phytochemical Characterization of

Stevia rebaudiana by

Hande Karaköse

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 (supervisor)

Professor of Organic and Analytical Chemistry, Jacobs University Bremen

Prof. Dr. Gerd-Volker Röschenthaler

Professor of Chemistry, Jacobs University Bremen

Dr. Adam Le Gresley

Doctor of Organic Chemistry, Kingston University London

Date of Defense: December 21, 2012

School of Engineering and Science

Declaration of Authorship

I, Hande Karaköse, hereby declare that the thesis I am submitting is entirely my own

original work unless where clearly indicated otherwise. I have used only the sources, the

data and the support that I have clearly mentioned. This PhD thesis has not been

submitted for conferral of degree elsewhere.

Bremen, November 30, 2012

Signature

Table of Contents

Acknowledgments.......................................................................................................... i

Abbreviations ................................................................................................................ ii

List of Figures .............................................................................................................. iv

List of Tables .............................................................................................................. vii

Abstract ...................................................................................................................... viii

Chapter Page

I. INTRODUCTION ......................................................................................................1

II. REVIEW OF LITERATURE....................................................................................4

2.1. Steviol Glycosides in Stevia rebaudiana ..........................................................4

2.2. Pharmacology, Toxicology and Regulations ....................................................5

2.3. Pharmacokinetics of Stevioside: Absorption, Metabolism, Excretion .............9

2.4. Biosynthesis of the Steviol Glycosides ...........................................................10

2.5. Analysis of Steviol Glycosides of Stevia rebaudiana.....................................14

2.6. Phenolic Acids ................................................................................................16

2.7. Proteomics of Stevia rebaudiana ....................................................................21

2.8. Lipid Analysis .................................................................................................25

III. RESEARCH OBJECTIVE ....................................................................................30

Chapter Page

IV. Steviol Glycosides Analysis by LC-MS ................................................................31

4.1. Overview .........................................................................................................31

4.2. Materials & Methods ......................................................................................31

4.2.1. Extraction method ................................................................................31

4.2.2. LC-MS analysis of steviol glycosides ...................................................31

4.2.3. HPLC conditions ...................................................................................32

4.2.4. Calibration curve of steviol glycoside standards ..................................33

4.2.5. Method Validation ................................................................................33

4.2.6. Solid phase extraction (SPE) of steviol glycosides ...............................33

4.3. Results & Discussion ......................................................................................34

4.3.1. Identification of steviol glycosides .......................................................36

4.3.2. Method Validation ................................................................................39

4.3.3. Comparison to SPE sample clean up ....................................................40

4.3.4. Quantification of steviol glycosides ......................................................42

4.4. Conclusion ..................................................................................................... 44

V. Polyphenols in Stevia rebaudiana .........................................................................45

5.1. Overview .........................................................................................................45


5.2.1. Sample preparation ...............................................................................45

5.2.2. LC-MS analysis of polyphenols ............................................................45

5.2.3. Calibration curve of standard compounds ............................................46

5.2.4. Hydrolysis of flavonoid glycosides ......................................................46

5.2.5. Statistical analysis .................................................................................46


5.3.1. Characterization of chlorogenic acids ...................................................50

5.3.2. Characterization of flavonoid glycosides ..............................................54

5.3.3. Quantification of chlorogenic acids and flavonoid glycosides .............56

3.3.1. Sample variation ........................................................................57

3.3.2. Flavonoid quantification ............................................................65

3.3.3. Principal component analysis (PCA) .........................................67

5.3.4. Statistical evaluation of quantification data of polyphenols in stevia ...69

3.4.1. Statistical spread of data .............................................................69

3.4.2. Correlations .................................................................................70

3.4.3. Analysis of variance (ANOVA) ..................................................76

5.4. Conclusion. .....................................................................................................80

Chapter Page

VI. Lipid Analysis of Stevia ........................................................................................81

6.1. Overview .........................................................................................................81


6.2.1. Extraction method ................................................................................81

6.2.2. Methyl ester formation ..........................................................................81

6.2.3. GC-FID conditions................................................................................81

6.2.4. GC-MS conditions ................................................................................82

6.2.5. Calibration curve of FAME ..................................................................82

6.2.6. MALDI-TOF MS ..................................................................................82


6.4. Conclusion ......................................................................................................93

VII. Proteomics of Stevia .............................................................................................94

7.1. Overview .........................................................................................................94


7.2.1. Extraction of proteins ............................................................................94

7.2.2. Protein analysis .....................................................................................95

7.2.3. MALDI-TOF MS conditions ................................................................97


7.3.1. SDS results ............................................................................................98

7.3.2. 2D-SDS .................................................................................................99

7.3.3. MALDI-TOF MS results ....................................................................100

7.4. Conclusion ....................................................................................................105

VIII. Summary ...........................................................................................................106

IX. References ............................................................................................................107

APPENDIX ................................................................................................................113

Publications ...............................................................................................................149

Curriculum Vitae

i

ACKNOWLEDGMENTS

This research has been completed with the support of a large number of people. I would like to

express my gratitude to them.

First of all, sincere thanks to my supervisor Prof.Nikolai Kuhnert for his guidance and expert

advices during my study.

I wish to acknowledge the support of European Union (Project DIVAS) and Jacobs University

Bremen for the full scholarship and funding of the research. I am grateful also to Dr. Kienle at

the University of Hohenheim for the valuable discussions.

A special thanks to my friends, Agnieszka Golon and Rohan Shah for their assistance during the

experimental work of the study. Also, thanks to my colleagues at Jacobs University for providing

a pleasant working atmosphere and Anja Müller for the technical assistance.

Finally, my warmest thanks belong to my parents Dilek and Nejdet and my brother Çağatay, for

their confidence in me and for being always supportive and interested in my work.

Phytochemical Characterization of Stevia rebaudiana

Hande Karaköse ii

Jacobs University Bremen

ABBREVIATIONS

HPLC high performance liquid chromatography

MS mass spectrometry

LC-MS liquid chromatography coupled with mass spectrometry

GC-MS gas chromatography coupled with mass spectrometry

GC-FID gas chromatography coupled with flame ionization detector

EI-MS electron impact ionization mass spectrometry

TIC total ion chromatogram

EIC extracted ion chromatogram

BPC base peak chromatogram

MS2/MS

3 tandem mass

ESI-MS electrospray ionization mass spectrometry

m/z mass-charge ratio

UV ultra-violet

MALDI matrix assisted laser ionization

TOF-MS time of flight mass spectrometry

HR-MS high resolution mass spectrometry

HILIC hydrophilic interaction chromatography

CGAs chlorogenic acids

CQA caffeoylquinic acid

FQA feruloylquinic acid

MeOH methanol

ACN acetonitrile

DXS deoxyxylose-5-phosphate synthase

CDPS copalyl diphosphate synthase

KS kaurene synthase

KO kaurene oxidase

KAH kaurenoic acid hydroxylase

UGTs UDP-glycosyltransferases

RT Retention time

RebA rebaudioside A


Hande Karaköse iii


RebC rebaudioside C

PCA principal component analysis

ANOVA analysis of variance

TCA trichloroacetic acid

IEF isoelectric focusing

HCCA α-cyano-4-hydroxycinnamic acid

FAME fatty acid methyl esters

FTICR Fourier transform ion cyclotron resonance

APCI atomic pressure chemical ionization

SPE solid phase extraction

S/N signal to noise ratio

LOD limit of detection

LOQ limit of quantification

RSD % relative standard deviation %

K7g kaempferol-7-O-glycoside

Q3g quercetin-3-O-glycoside

KS kolmogorov-smirnov test

DTT dithiothreitol

TFA trifluoroacetic acid

MVA mevalonic acid pathway

MEP 2-C-methyl-D-erythritol 4-phosphate/1-deoxy-D-xylulose 5-phosphate pathway

APS ammonium persulfate


Hande Karaköse iv


LIST OF FIGURES

Figure 1.General structure of steviol glycosides and representative structure of rebaudioside A

Figure 2.Structures of few artificial sweeteners

Figure 3.Structures of steviol metabolites

Figure 4.Hypothetical excretion route of stevioside

Figure 5.Steviol glycoside biosynthesis via the MEP pathway

Figure 6.Alternative MVA pathway

Figure 7.Examples for hydroxybenzoic and hydroxycinnamic acids

Figure 8.General structure of quinic acid and one of chlorogenic acids as an example

Figure 9.Generic structure of major classes of flavonoids

Figure 10.Strategies for MS based protein identification

Figure 11.Peptide fragmentation nomenclature

Figure 12.Examples of lipid categories

Figure 13.Total ion chromatogram in negative ion mode using C18 column of methanolic Stevia

rebaudiana extract showing phenolics (chlorogenic acids, flavonoids) and steviol glycosides

Figure 14.Base peak chromatogram of steviol glycosides obtained using HILIC column

Figure 15.Mechanism of fragmentation in tandem MS spectra of rebaudioside A and

rebaudioside E illustrating how isomeric compounds can be distinguished by tandem MS

Figure 16.Tandem MS spectra of rebaudioside A (above) and rebaudioside E (below) in

negative ion mode

Figure 17.Tandem MS spectra of rebaudioside D in negative ion mode

Figure 18.Total ion chromatograms for comparison of different amounts of material I in SPE

cleanup procedure

Figure 19.Total ion chromatograms for comparison of SPE cleanup of the stevia extract with

materials I and II cartridges

Figure 20.Radar plot of steviol glycoside concentrations varying between seven varieties

(average values taken within +/- 3σ) and in comparison to non-EU samples

Figure 21.Radar plot of steviol glycoside concentrations varying between all origins (average

values taken within +/- 3σ) and in comparison to non-EU samples

Figure 22.Base peak chromatogram in negative ion mode using C18 column of methanolic

Stevia rebaudiana extract showing chlorogonic acids, flavonoids and steviol glycosides


Hande Karaköse v


Figure 23.Structures of caffeoylquinic acids and flavonoid glycosides

Figure 24.Extracted ion chromatogram of m/z 353 of three mono-caffeoylquinic acids 3-

caffeoylquinic acid, 5-caffeoylquinic acid, and 4-caffeoylquinic acid (from left to right) in

negative ion mode

Figure 25.Consecutively MS, MS2

and MS3

spectra of 3-caffeoylquinic acid in negative ion

mode

Figure 26.Consecutively MS, MS2

and MS3

spectra of 4-caffeoylquinic acid in negative ion

mode

Figure 27.Consecutively MS, MS2, MS

3 and MS

4spectra of 3,5-dicaffeoylquinic acid in negative

ion mode

Figure 28.Consecutively MS, MS2, MS

3 and MS

4spectra of 4,5-dicaffeoylquinic acid in negative

ion mode

Figure 29.Chemical structure of four flavonoid aglycones identified in Stevia rebaudiana leaves

Figure 30.Extracted ion chromatogram of m/z 447.0 in negative ion mode

Figure 31.An example of tandem MS spectra for compound 1, revealing its identity as

kaempferol glucopyranoside

Figure 32.Fragmentation illustration on luteolin-7-glucoside

Figure 33.Radar plot of individual chlorogenic acid concentrations varying between seven

varieties (average values taken within +/- 3σ) and in comparison to non-EU samples

Figure 34.Radar plot of mono- and di-acyl quinic acids concentrations varying between seven


Figure 35.Bar plot of total mono- and di-acyl quinic acids concentrations varying between seven


Figure 36.Map showing the origins of stevia cultivation within the project

Figure 37.Radar plot of mono- and di-acyl quinic acids concentrations varying between all

origins (average values taken within +/- 3σ) and in comparison to non-EU samples

Figure 38.Bar plot of total mono- and di-acyl quinic acids concentrations varying between all

origins (average values taken within +/- 3σ) and in comparison to non-EU samples

Figure 39a.PCA analysis of phenol profile of 35 stevia leaf LC-MS datasets

Figure 39b.PCA analysis of phenol profile of 40 stevia leaf LC-MS datasets

Figure 40.Histogram of 5-CQA, showing the normal distribution of the dataset


Hande Karaköse vi


Figure 41.Graph showing the correlation between a) 3,5-diCQA/4,5-diCQA b) 3-CQA/5-CQA

c) 5-CQA/4,5-diCQA and d) 4-CQA/3,5-diCQA

Figure 42.Linear dependency of cis-5-CQA with 5-CQA and two isomers of cis-4,5-diCQAs

Figure 43.Amount of 5-CQA in mg/100g dry leaves from three harvests from location A (TCV)

and location B (Amfilikeia) during 2011

Figure 44.log of trans/cis-5-CQA concentrations against the number of sunshine hours in the

month for a total of ten harvests from six locations

Figure 45.GC-MS chromatogram of total lipid extracts from Stevia rebaudiana leaves from

sample (Uconor, Var.4)

Figure 46.GC-MS chromatogram of FAME standard mixture

Figure 47.Representative EI-MS spectra obtained from GC-MS measurement of stevia extract

Figure 48.Structures of fatty acids in Stevia rebaudiana extract

Figure 49.Fatty acid profile of average stevia leaf in % X:Y denominates the number of carbon

atoms in the fatty acid (X) and the number of double bonds in the fatty acid (Y)

Figure 50.MALDI-MS spectrum of total lipid extract in positive ion mode using 2,5-DHB as a

matrix

Figure 51.Chemical structures of terpenes identified in Stevia rebaudiana leaves

Figure 52.GC chromatogram of methylesterified steviol and stevia extract

Figure 53.Extraction procedure of proteins

Figure 54.SDS gel for sample number 8 TCV harvest I, loaded on gel at different concentrations

1 mg/mL and 0.5 mg/mL

Figure 55.2D-SDS separation of stevia total protein extract. 7cm strip of pH 4-7, where spot 1

and 2 are at 55 kDa, and spot 5 at 15kDa

Figure 56.MALDI-TOF MS spectra and mass list of trypsin digested 2D-SDS spot and Mascot

search result showing the sequence information for RuBisCO enzyme with the score of 45%

Figure 57.Mascot search result of MALDI spectra

Figure 58.MS/MS de novo sequencing of the m/z 1230. Series of y and b fragments are labeled

Figure 59.Structure and fragmentation of m/z 1230 based on denovo sequencing


Hande Karaköse vii


LIST of TABLES

Table 1.High resolution mass spectrometrical data of stevia terpene glycosides in negative ion

mode from LC-TOF MS analysis

Table 2.Enzymes involved in biosynthesis of steviol glycosides

Table 3.Steviol glycosides values from 166 samples

Table 4.Chromatographic and MS data on flavonoid glycosides and CGAs present in stevia

Table 5.Average values (taken within +/- 3σ) for chlorogenic acids in seven different varieties

Table 6.Average values (taken within +/- 3σ) for chlorogenic acids between origins

Table 7.Average values (taken within +/- 3σ) for chlorogenic acids between harvests

Table 8.Comparison of average values (taken within +/- 3 σ) for chlorogenic acids between three

harvests of same variety and origin

Table 9.Flavonoid glycosides average values for two major flavonoids in samples between

origins determined by LC-MS directly from extracts without hydrolysis

Table 10.Flavonoid glycosides average values between varieties

Table 11.Values for flavonoids quercetin, kaempferol, luteolin and apigenin determined after

hydrolysis of total polyphenol fractions using HCl/MeOH, determined by LC-MS

Table 12.Descriptive statistics of caffeoylquinic acids

Table 13.Correlation coefficients of mono and di-CQAs

Table 14.Correlation coefficients of cis isomers according to Spearman’s rule

Table 15.Results of test of homogeinity of variances

Table 16.ANOVA results for effect of origin on stevia CGA content

Table 17.Test of homogeneity of variances

Table 18.ANOVA results for effect of variety on stevia CGA content

Table 19.Total lipid values in weight % from 46 samples

Table 20.Quantities of polyunsaturated fatty acids

Table 21.Retention time and NIST scores of some terpenes identified in stevia extract

Table 22.Amount and properties of chosen stevia leaves for protein extraction

Table 23.Preparation of separation and stacking gel for 2D SDS PAGE


Hande Karaköse viii


ABSTRACT

Stevia rebaudiana (Bertoni) is from the Asteraceae family of plants with significant economic

value due to its high content of natural zero calorie steviol glycoside sweeteners in its leaves.

The leaves contain ent-kaurene glycosides, comprising stevioside, rebaudioside A, B, C, D, E, F

and dulcoside A. Rebaudioside A and stevioside are the most abundant diterpene glycosides

(steviol glycosides) in the leaves.

The phytochemical characterization of stevia leaves is playing an important role in a future EU

consumption of stevia as a novel food. For this purpose, the chemical composition of stevia

(phenols and steviol glycosides detailed, with lipids and proteins in representative cases) was

studied and methods have been developed for quantitative and qualitative analysis.

Stevia leaves cultivated in more than ten locations inside and outside of Europe with seven

different varieties, corresponding to total of 166 stevia samples, were extracted and their

chemical composition was profiled and quantified by LC-MS for steviol glycosides and

polyphenols (chlorogenic acids and flavonoids). Profiling, identification and quantification of

terpenoids and lipids were achieved by using GC, GC-MS and MALDI-TOF techniques. In

addition, protein extraction and analysis was carried out to identify potentially allergenic proteins

in stevia leaves. Protein separation and isolation was achieved with 2-dimensional

electrophoresis (2DE) and MALDI-TOF MS analysis was performed for the identification of

individual proteins.

Furthermore, as stevia may cultivated within various regions of the EU with different soil and

climatic conditions it is important to know whether an EU-common specification will be

achieved and how stevia leaves from regions outside EU can be distinguished on a scientific

basis. For this purpose, principal component analysis (PCA) was performed based on the LC-MS

dataset of stevia phenols. In addition, effect of growth origin and variety on stevia secondary

metabolite profile was analyzed statistically by ANOVA (analysis of variance).

1

1. INTRODUCTION

Stevia rebaudiana (stevia or S.rebaudiana) is native to Paraguay and belongs to the Asteraceae

family of plants. Stevia and its sweet taste was botanically described by M.S. Bertoni in 18991.

The high content of natural, zero-calorie sweeteners in its leaves makes stevia of a significant

economic value in the food industry in many applications as a sweetener. Interest in stevia

products has dramatically increased recently due to its approval by European and US legislating

authorities. Stevia is likely to become a major source of high-potency sweetener for the growing

natural food market in the future.

The majority of the annual stevia production of an estimated 4000 t is produced in China and

South America. The stevia crop has been shown to be highly adaptable to cultivation in many

other parts of the world. S. rebaudiana occurs naturally on acid soils of pH 4 – 5, but will also

grow on soils with pH levels of 6.5 – 7.5 making it an interesting alternative to plants cultivated

on poor soils such as tobacco2.

Stevia contains ent-kaurene glycosides, comprising stevioside, rebaudioside A, B, C, D, E, F and

dulcoside A (Figure 1, Table 1), which give the leaves its characteristic taste of 200-400 times

sweeter than sucrose. Stevioside has a sweetening power 300 times that of sucrose, and

rebaudioside A is 400 times sweeter than sucrose3. Rebaudioside A and stevioside are the most

abundant compounds; steviolbioside and rebaudioside B are believed to be formed by partial

hydrolysis during the extraction process4. The rest of the steviol glycosides (e.g. dulcoside A,

rebaudioside C) are at trace levels. In addition to being a natural sweetener, steviol glycosides

have functional and sensory properties superior to those of many other high-potency sweeteners.

Stevia leaves can be used in their natural state (fresh or dried form), due to its high sweetening

intensity. Only small quantities are needed for comparison with white sugar. It does not increase

the blood sugar level therefore; it can be used by diabetics without adverse glycemic responses.

The human fecal microflora hydrolyzes stevioside and rebaudioside A to their common aglycon

steviol in 10 and 24 h, respectively but steviol is not degraded by the human body 5.

In addition to diterpene glycosides, a number of secondary plant metabolites have been identified

from S. rebaudiana including labdane-type diterpenes, triterpenoids and steroids, phenolic acids

(flavonoid glycosides and chlorogenic acids), and oil components. From S. rebaudiana, ten


Hande Karaköse 2


labdane-type diterpenoids were identified, including austroinulin,isoaustroinulin6, sterebins (A -

H)7, 8

. A triterpenoid, lupeol 3-palmitate, was also separated from stevia9. As plant sterols, β-

sitosterol, stigmasterol and campesterol were identified from S. rebaudiana10

.

The presence of chlorogenic acids (CGAs) and flavonoid glycosides in stevia leaves gives the

plant additional health benefits, and it could as well affect its organoleptic properties.

CGAs are a large family of esters formed between quinic acid and certain trans -

hydroxycinnamic acids, most commonly caffeic, p-coumaric, and ferulic acid. Similar to

chlorogenic acids the presence of flavonoid compounds adds a health benefit to the usage of

stevia leaves in food products. Flavonoids are a class of secondary metabolites that are produced

ubiquitously in fruits and vegetables. By definition flavonoids are compounds with a C6-C3-C6

structure comprising two aromatic ring, one fused as a benzopyran.

The secondary metabolites of interest in the present study were; steviol glycosides, chlorogenic

acids, flavonoid glycosides, lipids, volatile terpenes and proteins. The main objective of this

project was to provide a scientific basis for a future EU specification for stevia. The steviol

glycoside and polyphenol profile and quantities of stevia samples cultivated in different

European and non-European countries with seven different botanical varieties, harvested at three

different times were obtained by analyzing stevia leaf extracts using a HPLC-TOF MS system.

The identification of the compounds was achieved by analyzing tandem mass spectra and high

resolution mass spectrometry (HR-MS) and for selected unknown phenolic compounds

spectroscopic MS rules previously developed in our laboratory was used to elucidate structures.

Stevia leave proteins were purified separated and sequenced with an aim to identify potentially

allergenic proteins using 2D gel electrophoresis and MALDI-TOF MS technique. Lipids and

volatile terpenes were determined by subjecting non-polar solvent extracts of stevia leaves to

GC-MS and MALDI-TOF MS. Identification of the compounds was achieved using NIST

library and comparison of retention times and GC-MS data of fatty acid methyl ester standard

mixtures.

Statistical analysis (PCA, correlation studies and ANOVA) served for differentiating the non-EU

and EU cultivated stevia samples and for studying the relations between each components and

growth conditions of stevia.


Hande Karaköse 3


O

OH

HO

HO O

HO

O

CH3

CH3

O

CH2

O

OH

O OH

OO

O

HO

OH OH

OH

OH

OH

HO

COORH3C

CH3

OR1

CH2

General structure ofsteviol glycosides

Rebaudioside A

O O

CH3

CH3

O

CH2

O

OH

O OH

OHO

HO

OH OH

HO

Stevioside

OH

O

OH

HO

HO

HO

Figure 1.General structure of steviol glycoside and representative structure of rebaudioside A.



Compound R R1 Molecular

Formula

Experimental

m/z (M-H+)

-

Theoretical

m/z (M-H+)

-

Relative Error

(ppm)

Steviol

Steviolbioside

H

H

H

glc2 - 1glc

C20H30O3

C32H50O13

317.0819

641.3181

317.0717

641.3179

9.0

0.4

Rubusoside Glc glc C32H50O13 641.3166 641.3179 2.0

Stevioside Glc glc2 - 1glc C38H60O18 803.3751 803.3707 5.5

Rebaudioside A Glc glc32 -1glc

1glc

C44H70O23 965.425 965.4235 1.6

Rebaudioside B H glc32 -1glc

1glc

C38H60O18 803.368 803.3707 2.8

Rebaudioside C

(Dulcoside B)

Glc glc32 -1rham

1glc

C44H70O22 949.427 949.4286 1.7

Rebaudioside D glc2-1glc glc32 -1rham

1glc

C50H80O28 1127.4726 1127.4763 3.3

Rebaudioside E glc2-1glc glc2-1glc C44H70O23 965.4199 965.4235 3.7

Rebaudioside F Glc glc32 -1xyl

1glc

C43H68O22 935.4097 935.4129 3.5

Dulcoside A Glc glc2 - 1rham C38H60O17 787.3732 787.3758 3.3


Hande Karaköse 4


2. REVIEW OF LITERATURE

2.1. Steviol Glycosides in Stevia rebaudiana

The most commercially important compounds from leaves of Stevia rebaudiana Bertoni

(S.rebaudiana/stevia) are the sweet tasting ent-kaurene diterpenoid glycosides. Two main

glycosides are stevioside and rebaudioside A. There are other related compounds including

Rebaudioside B-E, Dulcoside A and C which occur as minor components. Summaries of the

compounds from stevia are shown in Table 1.

Diterpene glycosides from S. rebaudiana contain a common aglycone called steviol (13-

hydroxy-ent-kaur-16-en-19-oic acid), and differ only in the glycosidic constituents attached at

C-13 and/or C-19.

Stevioside is the main sweet tasting glycoside in stevia (5-10 %) and was reported to be 250-300

times sweeter than sucrose. Rebaudioside A (2-4%) is the second most abundant ent-kaurene and

sweetest compound in stevia, its sweetness is 400 times more than sucrose. It was reported to

have a more pleasant taste and it is more water soluble than stevioside. Rebaudioside B, D, and E

may be also present in minor quantities; however, it is suspected that rebaudioside B is a

byproduct of the isolation technique11

. The two main compounds stevioside and rebaudioside,

primarily responsible for the sweet taste of stevia leaves, were first isolated by two French

chemists, Bridel and Lavielle (1931)12

.

The diterpene, steviol (Table 1) is the aglycone of stevia glycosides. Diterpene glycosides form

with the formation of ester bond between glucose molecule and carboxyl group of steviol and

replacing of hydroxyl hydrogen with combinations of glucose, rhamnose and xylose.

Stevioside has two linked glucose molecules at the hydroxyl site, whereas rebaudioside A has

three glucoses, with the central glucose of the saccharate connected to the central steviol

structure. Rebaudioside C and Dulcoside A possess a rhamnose sugar, whereas Rebaudioside F

possesses one xylose unit in its structure.

After sensory panel testing, Rebaudioside A was reported to have the least bitterness of all the

steviol glycosides in the stevia plant. Glycosides are molecules that contain glucose and other

non-sugar substances called aglycones. The taste receptor of tongue reacts to the glucose in the

glycosides, thus steviol glycosides with more glucose (e.g. rebaudioside A) taste sweeter than

those with less glucose (e.g.stevioside)13

. The bitter receptors of the tongue react to the

aglycones, or to polyphenols in the case of stevia leaves usage.

http://en.wikipedia.org/wiki/Rebaudioside_A

http://en.wikipedia.org/wiki/Glycoside

http://en.wikipedia.org/wiki/Aglycone


Hande Karaköse 5


2.2. Pharmacology, Toxicology and Regulations

A sugar substitute is a food additive that enhances the effect of sugar in taste. There are natural

and synthetic sugar substitutes. Those that are not natural are, in general, called artificial

sweeteners. Food additives must be approved by the FDA, which publishes a Generally

Recognized as Safe (GRAS) list of additives. The majority of sugar substitutes approved for use

are artificially-synthesized compounds. Sugar substitutes are used for reasons, including weight

loss, dental care, diabetes and hypoglycemia. Sugar substitutes which are commonly used in

foods are; aspartame, cyclamate, saccharin and sucralose. Starting with aspartame, it was

produced from two amino acids: aspartic acid and phenylalanine. It is about 200 times sweeter

than sucrose. The safety of aspartame has been studied extensively including animal studies,

clinical and epidemiological research14

. Hypotheses of adverse health effects have focused on the

three metabolites of aspartame, which are aspartic acid, methanol and phenylalanine and further

breakdown products including formic acid and formaldehyde15

. Aspartame is rapidly hydrolyzed

in the small intestines. Even with ingestion of very high doses of aspartame (over 200 mg/kg), no

aspartame is found in the blood due to the rapid breakdown16

. Furthermore, people with the

genetic disorder phenylketonuria should avoid aspartame since they have a decreased ability to

metabolize naturally occurring essential amino acid phenylalanine. The acceptable daily intake

(ADI) value for aspartame is determined as 40 mg/kg of body weight 17

.

Sucralose is a chlorinated sugar which is 600 times sweeter than sucrose. FDA approved usage

of sucralose after reviewing 110 studies in humans and animals18

. However, some adverse

effects were observed at doses that significantly exceeded the estimated daily intake which is 1.1

mg/kg/day 19

.

Saccharin was produced first in 1878 by a chemist working on coal tar derivatives. Studies in

laboratory rats during the early 1970s linked saccharin with the development of bladder cancer in

rodents. As a consequence, all food containing saccharin was labeled with a warning20

. However,

in 2000, the warning labels were removed because rodents, unlike humans, have a unique

combination of high pH, high calcium phosphate, and high protein levels in their urine, which

leads to formation of microcrystals that damages the bladder and over-produced cells to repair

the damage leads to tumor formation. Since this does not occur in humans, the conclusion was

there is no cancer risk21, 22

. In the European Union, saccharin is also known by the E number

(additive code) E954.

http://en.wikipedia.org/wiki/Metabolite#Metabolites

http://en.wikipedia.org/wiki/Aspartic_acid

http://en.wikipedia.org/wiki/Methanol

http://en.wikipedia.org/wiki/Phenylalanine

http://en.wikipedia.org/wiki/Hydrolyzed

http://en.wikipedia.org/wiki/Small_intestines

http://en.wikipedia.org/wiki/Phenylketonuria

http://en.wikipedia.org/wiki/Coal_tar

http://en.wikipedia.org/wiki/E_number


Hande Karaköse 6


Beginning in the 1970s with saccharin until the present day, artificial sweeteners have generated

a lot of controversy. The chemical nature of many artificial sweeteners (Figure 2) do not present

a good public image; e.g. saccharin was first produced from a coal-tar derivative23

, aspartame

breaks down into formaldehyde upon digestion15

, it also presents a health hazard to people born

with phenylketonuria, and sucralose is manufactured by the selective chlorination of sucrose24

.

With the scientific evidence on a particular sweetener, which is often inconclusive, and with

many interests at stake, including the food additive approval process, and potential political and

economic consequences, the results of these disagreements have not been entirely consistent or

logical. Aspartame, for example, gained FDA approval over vocal opposition from certain public

health advocates, while stevia extract, a substance which arguably presents health risks, cannot

have FDA approval and avoids a complete ban only by classification as a “dietary supplement”

rather than as a food additive25

.

NH

OH

OCH3

O

O

O

NH2

OCl

HOOH

O

HO

O

OHHO

Cl

Cl

SNH

O

OO

aspartame sucralose saccharin

Figure 2. Structures of few artificial sweeteners.

Stevia was used extensively by the Guarani Indians for more than 1,500 years in Paraguay and

Brazil26

. Stevia was first used as a sweetener in Japan in the 1970s, and it was a natural substance

that had been in use before 1958s with no apparent ill effects. However, the FDA banned stevia

as unsafe food additive in 1991 after receiving an anonymous industry complaint, and restricted

its import27

. The stated reason of FDA was that toxicological information on stevia was

inadequate to demonstrate its safety.

Health controversies about stevia started with the study of Pezzuto in 198528

, which reported that

steviol, a breakdown product of stevioside and rebaudioside A is a mutagen in the presence of a

liver extract of pre-treated rats. But this finding was criticized and stated that it might be worth

exploring the possibility that the mutagenicity of steviol (as in the experiments of Pezzuto et

http://en.wikipedia.org/wiki/Phenylketonuria


Hande Karaköse 7


al.,1986)29

is due to an impurity and the very high dose used in the experiments30

. Metabolism

of steviol in rat liver is complex and some metabolites detected are shown in Figure 3. The major

metabolite is 15-α-hydroxysteviol, which is non-mutagenic both in the presence and absence of a

metabolic activating system. Other metabolites are 7-β-hydroxysteviol, 17-hydroxyisosteviol and

ent-16-oxo-17-hydroxybeyeran-19-oic acid29, 31

. The mutagenic substance was proposed to be

15-oxosteviol. But, this compound was not detected as a metabolite of steviol and it was reported

to be bactericidal and weakly mutagenic30

. Nevertheless, other bacterial mutagenic assays failed

to demonstrate steviol mutagenic activity32

. The nature of mutagenic metabolite thus remained in

doubt.

Stevia remained banned until 1994, when forced under the Dietary Supplement Health and

Education Act, the FDA revised the decision on stevia and permitted it to be used as a dieatary

supplement. Over the following years studies on the toxicology and adverse effects of stevia

showed contradictory results.

CH2

R1

R2

CH3HOOC

H3C

OH

CH2

CH3HOOC

H3C

OH

O

R1=OH R2=H; 15α-hydroxysteviol 15-oxosteviol

R1=H R2=OH; 7β-hydroxysteviol

O

CH3HOOC

H3C

OHCH2

O

CH3HOOC

H3C

CH2OH

Steviol-16,17-oxide 17-hydroxyisosteviol

Figure 3.Structures of steviol metabolites.

Toskulkao et. al reported stevioside and steviol to have very low acute oral toxicity in the mouse,

rat and hamster 33

. Xili et al.34

have performed a combined chronic and carcinogenicity study, in

Wistar rats and in this study however stevioside administration in the diet showed no

carcinogenic effects in the rat. Through the review of many other toxicological studies 35-38

on


Hande Karaköse 8


stevia and its compounds, the EFSA committee concluded in 1999 that steviol, one metabolite of

stevioside, that is produced by the human microflora is genotoxic and induces developmental

toxicity. Therefore, the European Commission in 1999 banned stevia and its products in foods in

the European Union pending further research.

The European Food Safety Authority reevaluated the safety of steviol glycosides and expressed

its opinion on 10 March 2010. The Authority established an Acceptable Daily Intake (ADI) for

steviol glycosides, expressed as steviol equivalents, of 4 mg/kg (BW/day). The European

Commission allowed the usage of steviol glycosides as a food additive, establishing maximum

content levels for different types of foods and beverages on 11 November 201139

.

Regarding the effect of stevia in diabetes, a 2011 study by Misra et. al. on diabetes induced to

rats by injection of alloxan, have shown that leaf extract of S. rebaudiana (200 and 400 mg/kg)

produced a significant decrease in the blood glucose level, without producing condition of

hypoglycemia after treatment 26

. In addition, a 2009 review indicated that stevioside and related

compounds have anti-hyperglycemic, anti-hypertensive, anti-inflammatory, anti-tumor, anti-

diarrheal, diuretic, and immunomodulatory actions40

. The effect of stevioside and steviol on

glucose absorption was investigated by Toskulkao et al. and it was reported that 1mM steviol

inhibits glucose absorbtion, whereas 5 mM has no inhibitory effect. The inhibition of glucose

absorption by steviol was related to steviol concentration and incubation time 41

. However, the

announced acceptable daily intake of steviol glycosides (4 mg/kg BW/day) would yield a

maximum plasma concentration of steviol of approx. 20 μM if stevioside is completely

converted to steviol. This concentration is far below the reported value to inhibit intestinal

glucose absorption. Therefore, more studies should be conducted using ADI amount to

reevaluate the effect of steviol on glucose absorption. However, it is worth pointing out that

stevioside does not interfere with glucose absorption40

.

Stevioside, other related steviol glycosides, or stevia leaves themselves have been used

commercially in many countries, especially in Asia, as food additives for sweetening a variety of

products without any side effects. Moreover, phytochemicals (especially polyphenolics and

steviol glycosides) of stevia were reported to exhibit significant pharmacological activities.

http://en.wikipedia.org/wiki/European_Commission

http://en.wikipedia.org/wiki/European_Union

http://en.wikipedia.org/wiki/European_Food_Safety_Authority



http://en.wikipedia.org/wiki/Hyperglycemic

http://en.wikipedia.org/wiki/Anti-hypertensive

http://en.wikipedia.org/wiki/Anti-inflammatory

http://en.wikipedia.org/wiki/Anti-tumor

http://en.wikipedia.org/wiki/Anti-diarrheal

http://en.wikipedia.org/wiki/Anti-diarrheal

http://en.wikipedia.org/wiki/Diuretic

http://en.wikipedia.org/wiki/Immunomodulatory


Hande Karaköse 9


2.3. Pharmacokinetics of Stevioside: Absorbtion, Metabolism, Excretion

Absorption and metabolism studies on steviol glycosides showed that the uptake of stevioside by

the intestinal tract is extremely low due to its high molecular size and hydrophlicity40, 42, 43

.

Stevioside is not degraded by the enzymes of the intestinal tract, gastric juice or digestive

enzymes from animals and humans 5, 42, 44

. However, bacterial intestinal flora of humans is able

to convert stevioside to steviol and Bacteroides sp. is responsible for this conversion in the lower

gastrointestinal tract of both rat and human 5. Koyoma et. al

44. investigated the metabolism of

stevia by incubating stevioside, rebaudioside A and steviol with pooled human faecal

homogenates obtained from healthy volunteers for 0.8 and 24 h under anaerobic conditions.

Stevioside, rebaudioside A were completely hydrolysed in 24 h, and no degradation of steviol

was observed. The author proposed a metabolic pathway for rebaudioside A, which suggests that

majority of rebaudioside A is hydrolyzed via stevioside to steviol and minority via rebaudioside

B to steviol. Steviol was not further metabolized in human intestinal microflora being

inconsistent with the study of Pezzuto et.al28

reporting the oxidation of steviol to hydroxysteviol,

or to 15-oxo-steviol (Figure 3).

Another study in 10 healthy volunteers showed that after 3 days of consumption of stevioside

(every day 3 times 250 mg capsules with 8 h intervals), steviol glucoronide is the only excretion

product of stevioside in urine. There was no detection of free steviol in urine. Moreover, after

enzymatic hydrolysis of urine by β-glucuronidase/sulfatase, steviol was the only aglycone and

there was no indication of steviol sulfates 42

. The excretion route proposed by Geuns et al. is

presented in Figure 4.

O-Glc-Glc

H

HH3C CO2Glc

H3C

OH

H

HH3C CO2H

H3C

OH

H

HH3C C

H3C

O

O O

H

H

CO2H

H

H

HO OH

bacteria

colon

liver

stevioside steviol steviol glucuronide

Figure 4.Hypothetical excretion route of stevioside42

.


Hande Karaköse 10


Another study in humans reported that 72 h after oral stevioside ingestion, steviol glucuronide

excretion in urine and free steviol in feces accounting for 62% and 5.2% of the total dose of

stevioside administered respectively40, 45

.

As conclusion from the reviewed literature, steviol glucorunonide is the main metabolite of

stevioside consumption and urinary excretion is responsible for the disposal from the body.

2.4. Biosynthesis of the Steviol Glycosides

Biosynthesis of steviol glycosides are still subject of discussion. There are two main proposed

pathways for steviol glycosides; 2-C-methyl-D-erythritol 4-phosphate/1-deoxy-D-xylulose 5-

phosphate pathway (MEP/DOXP pathway) (Figure 5) and mevalonic acid pathway (MVA)

(Figure 6).

MEP pathway of isoprenoid (terpenoids) biosynthesis is a metabolic pathway which leads to the

formation of isopentenyl pyrophosphate (IPP) (9) and dimethylallyl pyrophosphate (DMAPP)

(10) in the plastids of the plants. MEP pathway is an additional alternative pathway to mevalonic

acid pathway (MVA) for formation of isoprenoids (terpenoids). MVA reactions take place in

cytosol whereas MEP reactions occur in plastids. Pyruvate (1) and glyceraldehyde-3-phosphate

(2) are converted by DOXP synthase to 1-deoxy-D-xylulose-5-phosphate (3) and by DOXP

reductase to 2-C-methyl-D-erythritol 4-phosphate (4) (MEP). After subsequent reaction steps,

the end products IPP (9) and DMAPP (10), which are precursors of terpenoids, are formed.

Synthesis of all higher terpenoids occurs via formation of geranyl pyrophosphate (GPP) and

geranylgeranyl pyrophosphate (11) (GGPP).

In the proposed MEP pathway, steviol was synthesized from kaurene (13) 46

. The plant gene for

the first step in the MEP pathway is deoxyxyulose-5-phosphate (DXP) synthase (DXS), which

leads to the synthesis of DXP from pyruvate (1) and glyceraldehyde 3-phosphate (2). Once

synthesized, DXP can either be used for the production of vitamins like thiamin or in the MEP

pathway for isoprenoid synthesis.

The DXS amino acid sequence is highly conserved among plant species, which enabled Totte´ et

al. (2003)47

to design primers for RT-PCR and clone the DXS gene from Stevia. The author

suggested that steviol was synthesized via mevalonic acid pathway (MVA), involving mevalonic

acid in the biosynthesis of steviol, but no direct proof was given to support it. Brandle et. al

(2002)48

sequenced 5548 expressed sequence tags (ESTs) from stevia leaf cDNA library. The

http://en.wikipedia.org/wiki/Isopentenyl_pyrophosphate

http://en.wikipedia.org/wiki/Dimethylallyl_pyrophosphate


Hande Karaköse 11


ESTs were classified according to their function in primary or secondary metabolism and many

genes specific to MEP pathway but not the MVA pathway were identified, which concludes that

the source of IPP for diterpenes is through the MEP pathway.

Steviol glycosides share four common steps in biosynthetic pathway with gibberellic acid

formation. After oxidation of ent-kaurene at the C-19 position to ent-kaurenoic acid, the

pathways to the steviol glycosides and the gibberellins diverge. Steviol is produced by

hydroxylation of ent-kaurenoic acid at the C-13 position. Steviol is then glycosylated by series of

UDP-glucosyltransferases (UGTs). UGTs are highly regiospecific and recognize particular

substructure of the acceptor molecule rather than the molecule in its entirety49

. The MEP

pathway of steviol glycosides and the enzymes involved in this pathway is presented in Table 2

and Figure 5.

Table 2.Enzymes involved in biosynthesis of steviol glycosides

Enzyme abbreviation Enzyme

DXS deoxyxyulose-5-phosphate synthase

DXR deoxyxyulose-5-phosphate reductoisomerase

CMS 4-diphosphocytidyl-2-C-methyl-D-erythritol synthase

CMK 4-diphosphocytidyl-2-C-methyl-D-erythritol kinase

MCS 4-diphosphocytidyl-2-C-methyl-D-erythritol 2,4-cyclodiphosphate synthase

HDS 1-hydroxy-2-methyl-2(E)-butenyl 4-diphosphate synthase

HDR 1-hydroxy-2-methyl-2(E)-butenyl 4-diphosphate reductase

GGDPS geranylgeranyl diphosphate synthase

CPS copalyl diphosphate synthase

KS kaurene synthase

KO kaurene oxidase

KAH kaurenoic acid 13-hydroxylase


Hande Karaköse 12


COO-

O

OPO32-H

OH

O

OPO32-

OH

OH

O

OH

OPO32-

OH

OH

O P2O52-

OO

OH OH

N

N

O

NH2

OH

OHOH

O P2O52-

OO

OH OH

N

N

O

NH2

O

OHOH

PO32-

O

O

OHOH

PO2-

PO2-

O

O

OH

P2O62-

OP2O63- OP2O6

3-

OP2O63-

OP2O63-

H

H

COOH

H

H

COOH

H

H

OH

COOH

H

H

O glc

COOH

H

H

O glc glc

COO

H

H

O glc glc

glc COO

H

H

O glc glc

glc

glc

+

+

DXS

Pyruvate glyceraldehyde-3-phosphate

1-deoxyxylulose-5-phosphate

DXR

CMS

2-C

4-diphosphocytidyl-2-C-methyl-D-erythritol

CMKMCS

4-diphosphocytidyl-2-C-methyl-D-erythritol-2-phosphateHDR

isopentenyl diphosphate dimethylallyldiphosphate1-hydroxy-2methyl-(E)butenyl-4-diphosphate

2-C-methyl-D-erythritol-2,4-cyclodiphosphate

HDS

(-)-copalyl diphosphate

KS

CDPS

GGDPS

geranylgeranyl diphosphate

KOKAH

(-)-kaurenoic acidsteviol

UGT85C2 (-)-kaurene

steviolmonoside steviolbioside stevioside rebaudioside A

UGT

Figure 5.Steviol glycoside biosynthesis via the MEP pathway50

.


Hande Karaköse 13


SCoA

O

SCoA

O O

SCoAO-O

HO

O

SCoAO-O

HO

O

O-O

HOP

O

O-

OH

O

O-O

HOP

O

O-

OP

O

OH-OO

P

O

O-

OP

O

OH-O

OP

O

O-

OP

O

-OOH

AACT HMGS HMGR

MK

PMKPPMD

IDI

1 2 34

56

7

8

Figure 6.Alternative MVA pathway. Enzymes of the MVA pathway are as follows: AACT,

AcAc-CoA thiolase; HMGS, HMG-CoA synthase; HMGR, HMG-CoA reductase; MK,

mevalonate kinase; PMK, phosphomevalonate kinase; PPMD, diphospho-mevalonate

decarboxylase. 1, Ac-CoA; 2, AcAc-CoA; 3, HMG-CoA; 4, MVA; 5, mevalonate 5-phosphate;

6, mevalonate 5-diphosphate. Both MPE and MVA pathways lead to the formation of compound

8, dimethylallyl diphosphate; 7, isopentenyl diphosphate. The interconversion of IPP into

DMAPP is catalyzed by IDI, isopentenyl diphosphate isomerase51

.


Hande Karaköse 14


2.5. Analysis of Steviol Glycosides of Stevia rebaudiana

A wide range of analytical techniques have been used to determine the diterpenoid glycosides in

stevia. These techniques include thin layer chromatography (TLC) 52

, capillary electrophoresis

(CE) 53

, near infrared reflectance 54

, and enzymatic methods 55

. But the most used and most

efficient method is high performance liquid chromatography (HPLC or LC). The use of the

hyphenated technique coupled with mass spectrometry (LC-MS) in the analysis of plant extracts

provides important advantages because of the combination of the separation capabilities of LC

and the power of MS as an identification and confirmation method.

In many modern HPLC separations, prepacked columns are used and many types are available

from the manufacturers. However, it is possible to carry out most separations using silica column

for non-polar compounds or reversed phase C18 bonded phase column for polar compounds. The

solvent systems used in the analytical HPLC usually include gradient elutions using solvents of

aqueous acetic, formic or phosphoric acids with methanol or acetonitrile as an organic modifier.

The pH and ionic strength of the mobile phase are known to influence the retention of phenolics

in the column depending on protonation, dissociation, or a partial dissociation. A change in pH

which increases the ionization of a sample could reduce the retention in a reversed phase

separation. Thus, small amounts of acetic (2– 5%), formic, phosphoric or trifluoroacetic acid

(0.1%) are included in the solvent system to enhance ionization of phenolic and carboxylic

groups and hence to improve peak shapes, resolution and reproducubility of chromatographic

runs.

However, for steviol glycosides chromatographic separations in HPLC are not so straight

forward due to the structural similarity of steviol glycosides. Especially for isomer pairs of

stevioside/rebaudioside B and rebeaudioside A/rebaudioside E with the same molecular formula,

resulting in very close retention times in LC thus, resulting in selectivity problems due to peak

overlap and irreproducibility. Therefore, it is still a challenge to achieve efficient separation and

identification for steviol glycoside extracts.

Detection and separation of steviol glycosides on liquid chromatography (LC) were performed

employing amino (NH2)3, 56-60

, C18 61, 62

, and hydrophilic interaction chromatography (HILIC) 63

,

64 columns in combination with mostly UV or MS detection.


Hande Karaköse 15


Among those, two-dimensional LC 65, 66

and ultra high pressure liquid chromatography (UHPLC)

64, 67 systems were used as well.

Liquid chromatography with amino columns provides good separation of steviol glycosides and

have selectivity for the isomers of stevioside/rebaudioside B and rebaudioside A/rebaudioside E

but they are having disadvantage of reproducibility and long equilibration times 57

.

C18 column exhibits longer retention time and more robustness if compared to amino columns

but also poor selectivity for the separation of stevioside and rebaudioside A. Gradient elution is

not enough to overcome this problem, thus two dimensional systems either with connection of

two C18 columns 68

or C18 with amino column 66

were used.

Hydrophilic interaction chromatography is still new and a useful technique for the retention of

more polar analytes with increased selectivity if compared to reversed phase chromatography.

The interaction of the analytes is believed to be with the water rich layer forming on the surface

of the polar stationary phase against the water poor mobile phase. HILIC can offer a tenfold

increase in sensitivity over reversed-phase chromatography for detection of polar compounds

with mass spectrometry, due to more volatile organic solvent 69

. Some papers 63, 64

describe the

use of HILIC column for stevia extract. In those studies, steviol glycosides were separated with

isocratic elution using 5–20% water in acetonitrile with buffer or formic acid and the robustness

of the separation against changes of buffer concentration and percentage of water differ 64

.

Methods that use UV detection for steviol glycoside quantification are most popular; however

suffer from a series of disadvantages. Detection is typically carried out at 200-210 nm using the

carboxylic acid and olefinic chromophores. These wavelengths are very close to the UV cutoff of

acetonitrile as a solvent and particularly problematic if gradient elution is used. Additionally

many further stevia constituents from the matrix and other impurities absorb at these

wavelenghths. On no occasion was the absence or presence of co-eluting impurities established

in any published steviol glycoside UV method.

Despite some efforts in the development of methods aimed at the identification and

quantification of steviol glycosides until today no validated and certified method exists.

Furthermore no interlaboratory trials were ever conducted on steviol glycoside quantification

allowing a reliable assessment of method validity. Accordingly despite many contributions

http://en.wikipedia.org/wiki/Reversed-phase_chromatography


Hande Karaköse 16


published in the field of steviol glycoside analysis, there is still urgent need for critical

assessment of published methods and the development of a generally accepted standard method.

2.6. Phenolic Acids

Polyphenols are secondary metabolites that constitute one of the most widespread groups of

compounds in plants. They are derivatives of the pentose phosphate, shikimate and

phenylpropanoid pathways in plants70

. Polyphenolic compounds contribute to pigmentation of

flowers, fruits, leaves or seeds and play important role in the growth, reproduction and adaptative

strageties of plants71

. In food, phenolics contribute to the bitterness, astringency, color, flavor,

odor, and oxidative stability of products. Moreover, health-protecting capacity and antinutritional

properties of plant phenolics are of great importance to producers, processors and consumers72

.

The antioxidant activity of the dietary polyphenolics is considered to be much greater than that

of the essential vitamins, therefore contributing significantly to the health benefits of fruits73

Phenolic compounds are present in almost all foods of plant origin. Fruits, vegetables, and

beverages are the main sources for these compounds in the human diet. The level of phenolics in

plant sources depend on cultivation techniques, cultivar, growing conditions, ripening process, as

well as processing and storage conditions. In addition, the content of some phenolics may

increase under stress conditions such as UV radiation, infection by pathogens and parasites,

wounding, air polution and exposure to extreme temperatures74

.

Fruit and beverages such as coffee, tea and red wine constitute the main sources of polyphenols.

Certain polyphenols such as quercetin are found in all plant products (fruit, vegetables, cereals,

leguminous plants, fruit juices, tea, wine, infusions, etc), whereas others are specific to particular

foods (flavanones in citrus fruit, isoflavones in soya, phloridzin in apples). In most cases, foods

contain complex mixtures of polyphenols, which are often poorly characterized 75

.

Onions are rich sources of flavonoids76

. Flavonols, the predominant phenolics, are located

mostly in the tomato skin. Cherry tomatoes contained a much higher level of flavonols than

larger size tomato cultivars76, 77

. Anthocyanins are located in the red onion skin and the outer

fleshy layer78

.

The main group of polyphenols includes simple phenols, phenolic acids (benzoic and cinnamic

acid derivatives), coumarins, flavonoids, stilbenes, hydrolyzable and condensed tannins, lignans,

and lignins.


Hande Karaköse 17


Phenolic acids consist of two subgroups; the hydroxybenzoic and hydroxycinnamic acids.

Hydroxybenzoic acids include gallic, p-hydroxybenzoic, protocatechuic, vanillic and syringic

acids, which in common have the C6–C1 structure. Hydroxycinnamic acids, on the other hand,

are aromatic compounds with a three-carbon side chain (C6–C3), with caffeic, ferulic, p-coumaric

and sinapic acids (Figure 7) being the most common 71, 79

.

OH

R2R1

COOH OH

OH

COOH

COOH

R1

HO

hydroxybenzoic acid protocatechuic acid hydroxycinnamic acid

HO

HO COOH

HO

COOH

HO

H3CO COOH

caffeic acid p-coumaric acid ferulic acid

Figure 7.Examples for hydroxybenzoic and hydorxycinnamic acids.

Caffeic acid is the major representative of hydroxycinnamic acids and occurs in foods mainly as

chlorogenic acid (5-caffeoylquinic acid). Chlorogenic acids (CGAs) are a family of esters

formed between one or more residues of certain trans-cinnamic acids and quinic acid (1L-1

(OH),3,4/5-tetrahydroxycyclohexane carboxylic acid) which have axial hydroxyls on carbons 1

and 3 and equatorial hydroxyls on carbons 4 and 5. During processing, trans isomers may be

partially converted to cis isomers 80, 81

(Figure 8). The main classes of CGAs found in nature are

the caffeoylquinic acids (CQA), dicaffeoylquinic acids (diCQA), and, less commonly,

feruloylquinic acids (FQAs), each group with at least three isomers82

. CGAs are antioxidant

components produced by plants in response to environmental stress conditions such as infections

by microbial pathogens, mechanical wounding, and excessive UV or visible light levels83

Chlorogenic acids make up 5-10% of the weight of coffee beans and plays a significant role in

coffee color and aroma formation84

.


Hande Karaköse 18


OR1

C

OR3

OR4

OR5O

HO

6

1

OH

C

OH

O

OHO

HO

OH

OH

O

quinic acid trans -4-caffeoylquinic acid

OH

C

OH

O

OHO

HO

O

OHOH

cis-4-caffeoylquinic acid

Figure 8.General structure of quinic acid and one of chlorogenic acids as an example.

In addition to being found in coffee, these compounds are also found at significant levels in plant

foods such as apples, pears, tomato, potato, and eggplant85

. Coffee is a major source of

chlorogenic acid in the human diet; daily intake in coffee drinkers is 0.5–1 g; coffee abstainers

will usually ingest < 100 mg/d 86

.

In the last few years, CGAs has been the subject of several investigations in their potentially

beneficial effects in humans involving their antioxidant activity, among other beneficial effects.

Several pharmacological activities of CGAs including antioxidant activity, the ability to increase

hepatic glucose utilization,87-94

inhibition of the HIV-1 integrase,95-97

antispasmodic activity,98

and inhibition of the mutagenicity of carcinogenic compounds99

have been revealed by in vitro,

in vivo, and human intervention studies so far. CGAs and their metabolites display additional

highly favorable pharmacokinetic properties.100-102

Flavonoids are low molecular weight compounds, consisting of fifteen carbon atoms, arranged in

a C6–C3–C6 configuration. Essentially the structure consists of two aromatic rings, A and B,

joined by a 3-carbon bridge, usually in the form of a heterocyclic ring, C. The aromatic ring A is

formed via glucose metabolism with condensation of malonyl-coenyme A (CoA) catalyzed by

chalcone synthetase, while ring B and C is derived from phenylalanine through the shikimate

pathway, which is converted to cinnamic acid and to coumaric acid. Coumaric acid CoA and

three malonyl CoAs are condensed in a single enzymatic step to form naringenin chalcone. The

C-ring closes and becomes hydrated to form 3-hydroxyflavonoids (e.g., catechins), 3,4-diol

flavonoids (e.g., quercetin), and procyanidins103

.

Variations in the heterocyclic ring C give rise to the major flavonoid classes, i.e., flavonols,

flavones, flavanones, flavanols (or catechins), isoflavones, flavanonols, and anthocyanidins


Hande Karaköse 19


(Figure 9), while individual compounds within a class differ in the pattern of substitution of the

A and B rings. These substitutions may include oxygenation, alkylation, glycosylation, acylation,

and sulfation 79, 104

.

O

A C

B

1'6'

5'

4'3'

2'

2

3

45

7

8

6

O

O

O

O

OH

O

O

O

OH

O+

OH

Flavone Flavonol Flavanone

Flavanol Antocyanidin

Generic structure

Figure 9.Generic structure of major classes of flavonoids.

Within different subclasses of flavonoids, further differentiation is based on the number, position

and nature of substituent groups attached on the rings. Mostly they are sugars, such as glucose,

galactose, rhamnose, arabinose, xylose and rutinose. Flavonoid glycosides have many isomers

with the same molecular weight but different aglycone and sugar component at different

positions attaching on the aglycone ring 72, 105, 106

. Flavonoid glycosides as well are commonly

encountered in plant material and following ingestions these glycosides are hydrolysed by the

human microbial gut flora into their aglycones, which are subsequently absorbed and show

significant bioavailability.


Hande Karaköse 20


Flavonoids play different roles in the ecology of plants. Due to their attractive colors, flavones,

flavonols, and anthocyanidins may act as visual signals for pollinating insects. Because of their

astringency, catechins and other flavanols can represent a defense system against harm of insects

to the plant. Furthermore flavonoids protect plants from UV radiation of sun with their favorable

UV-absorbing properties104

.

Apart from their roles in plants, flavonoids play important role in human diet. Flavonoids are

important antioxidants (hydrogen-donating radical scavengers) due to their high redox potential,

which allows them to act as reducing agents, hydrogen donors, and singlet oxygen quenchers.

The antioxidant property of flavonoids may protect tissues against oxygen free radicals and lipid

peroxidation. Thus, flavonoids might contribute to the prevention of atherosclerosis, cancer and

chronic inflammation107

. In addition, they have a metal chelating potential, which play an

important role in oxygen metabolism and are essential for many physiological functions 108

. The

proposed binding sites for trace metals to flavonoids are the catechol moiety in ring B, the 3-

hydroxyl, 4-oxo groups in the heterocyclic ring, and the 4-oxo, 5-hydroxyl groups between the

heterocyclic and the A rings. However, the major contribution to metal chelation is due to the

catechol moiety, as exemplified by the more pronounced bathochromic shift produced by

chelation of copper to quercetin compared to that of kaempferol (similar in structure to quercetin

except that it lacks the catechol group in the B ring)104

.

Flavonoids and phenolic acids have protective role in carcinogenesis, inflammation,

atherosclerosis, thrombosis and have high antioxidant capacity. Furthermore, flavonoids have

been reported as aldose reductase inhibitors blocking the sorbitol pathway that is linked to many

problems associated with diabetes106

.

Phenolic acids in stevia were analyzed by HPLC on a C18 column by Kim et. al 109

and the main

phenolic compounds found were pyrogallol, 4-methoxybenzoic acid, p-coumaric acid, 4-

methylcatechol, sinapic and cinnamic acid. The flavonoids detected in stevia leaves belong to the

subgroups of flavonols and flavones. They were identified using two-dimensional UHPLC-DAD

and LC-MS/MS and spectroscopic methods (1H and

13C NMR, IR, and 2D NMR)

110, 111.

There is currently great interest in phenolic acids research due to the possibility of improved

public health through diet, where preventative health care can be promoted through the

consumption of fruit and vegetables. Therefore, the presence of such compounds with proven


Hande Karaköse 21


health benefits in stevia extracts would affect the the organoleptic properties of stevia based

products and would add health aspects to the use of stevia as sweenetening agent.

2.7. Proteomics of Stevia rebaudiana

Determination of amino acid content and potential allergenic proteins in food is very significant

and necessary research in the food formulation processes. Currently, there has been no published

data related to stevia allergens and there is only one published paper for the protein analysis in

stevia112

. Therefore, screening and quantitative analysis of stevia proteins and any potential

allergen is crucial. Detailed and comprehensive characterization of plant-derived food allergens

can be carried out using proteomics. In proteomics, after the separation and purification of the

protein, proteins are identified by mass spectrometry. Proteomic technologies using 2D-PAGE

and immunoblotting are then applied in the identification of new allergens113

.

In the past, protein determination was carried out by mRNA analysis, but later it was found that

there was no correlation with protein content as gene expression is regulated post-

transcriptionally and translation from mRNA cause differences 114, 115

. Most proteins are

chemically modified through post-translational modifications, mainly through the addition of

carbohydrate and phosphate groups. Such modifications play an important role in modulating the

function of many proteins. The most common post-translational modifications include

glycosylation, phosphorylation, ubiquitination, methylation, acetylation, and lipidation115

.

The major methods to study proteins include, high quality separation of proteins in two

dimensions (2D-SDS PAGE), characterization of separated proteins by mass spectrometry and

information collection using bioinformatic tools and databases115

. Matrix-assisted laser

desorption/ionization (MALDI) and electrospray ionization mass spectrometry (ESI) are widely

used techniques for proteomic studies.

Sample preparation is the most important step in the analysis of proteins from plants due to the

low protein content relative to other systems and the large quantities of polysaccharides, lipids,

phenolics and other secondary metabolites. Pretreatment of samples for 2D electrophoresis

involves solubilization, denaturation and reduction to completely break up the interactions

between the proteins and removal of all interfering compounds (phenolic compounds, nucleic

acids) to ensure efficient separation115

.

The most common protein extraction protocol is based on precipitating proteins from

homogenized tissue or cells with trichloracetic acid (TCA) in acetone. An alternative protocol is

http://en.wikipedia.org/wiki/Matrix-assisted_laser_desorption/ionization

http://en.wikipedia.org/wiki/Matrix-assisted_laser_desorption/ionization


Hande Karaköse 22


based on the solubilization of proteins in phenol, followed by their precipitation with ammonium

acetate in methanol. No one protocol is necessarily more appropriate than the other; studies have

suggested that they are complementary116

. In order to evaluate the effectiveness of a given

extraction protocol, protein quantification is needed. Bradford, Lowry and BCA assays are the

most common colorimetric methods.

Identification of proteins proceeds with separation of proteins from the extract with 2D-SDS

PAGE and subsequent digestion of the individual separated proteins or digestion of the entire

protein mixture followed by separation of the resulted peptides117

.

2D SDS PAGE separates proteins in two dimensions; the first dimension in a pH gradient

according to their isoelectric point (pI), and in the second dimension, the proteins is separated

according to their molecular weight.

The first dimension of electrophoresis involves denaturing isoelectric focusing using

immobilized pH gradient gels (IPG). IPG strips have a gradient of charge imbedded in

acrylamide. IPG strips improve the reproducibility and reliability and overcome pH gradient

instability. Strips come in a variety of pH ranges and lengths (from 7 to 24 cm). Samples can be

applied on the strips by cup loading or by in-gel rehydration. In cup loading method, the strips

are pre-rehydrated with rehydration buffer and the samples are applied into the loading cup at

either acidic or basic end. In in-gel rehydration, the sample in lysis buffer is diluted with the

rehydration buffer. The IPG matrix absorbs the proteins. Isoelectric focusing (IEF) is carried out

on a first dimension electrophoresis unit consisting of five phases of stepped voltage from 500 to

3500 V (Multiphor) or 500 to 8000 V (IPGPhor)115

. After completion of the first dimension the

proteins are separated according to their mass in second dimension using SDS-PAGE. Separation

of proteins in second dimension is based on differences in their electrophoretic mobility due to

differences in their size. SDS is a very effective solubilising agent for a wide range of proteins.

The majority of proteins bind SDS at a ratio of 1.4 g SDS / 1 g protein to form negatively

charged complexes118

. Proteins are transferred electrophoretically from the IEF strip into a

narrow starting zone prior to entering the main separating gel. This concentrates the proteins and

results in very sharp bands or spots. Once the protein samples have entered the separating gel,

the negatively charged protein-SDS complexes continue to move towards the anode. As they

pass through the separating gel the proteins are resolved on the basis of their size because of the

molecular sieving properties of the gel. After 2D electrophoresis, the protein spots can be


Hande Karaköse 23


visualized by staining with either coomassie brilliant blue stain, which will detect proteins

present in amount greater than 100 ng or with silver staining for amounts in ng range 119

.

The end point of any proteomics expression is to identify and characterize the proteins. Edman

degradation was the standard method for protein sequencing for the last 25 years120

. Other

traditional approaches for protein identification include the use of antibodies to perform Western

blots. However, this method has restricted use due to antibodies non-specific binding and the

availability of antibodies to all proteins121

.

Development in mass analysis techniques for mass spectrometry (MS) and the ability to correlate

MS data of proteins to sequences in databases have opened up new possibilities in protein

sequencing.

Protein identification via MS can be carried out in the form of whole-protein analysis (‘top-

down’ approach) or analysis of enzymatically or chemically produced peptides (‘bottom-up’

approach). To date, one of the most common methods of identifying proteins is through peptide-

mass fingerprinting (PMF). The proteins are digested with a proteolytic enzyme such as trypsin,

to produce a set of tryptic fragments unique to each protein115

. Summary of other MS-based

proteomic strategies is presented in Figure 10.

Figure 10.Strategies for MS-based protein identification122

.


Hande Karaköse 24


Mass spectrometry analysis of peptides and proteins relies exclusively on soft ionization

techniques that create intact gas-phase ions from biomolecules. The creation of intact molecular

ions enables accurate measurement of molecular weight. The electrospray ionization (ESI) and

matrix-assisted laser desorption/ionization (MALDI) techniques are widely used in proteomic

studies. The most common mass analyzers used with MALDI are time-of-flight (TOF) mass

spectrometers.

MALDI-TOF allows the analysis of high molecular weight compounds with high sensitivity and,

soft ionization with little or no fragmentation. MALDI uses a solid matrix to co-crystallize with

peptides/proteins on a sample plate and a laser light as its ionizing beam. Ionization occurs when

these matrix molecules absorb the energy provided by a laser (usually 337 nm). Release of the

energy causes a rapid thermal expansion of matrix and analyte into the gas phase. Proton transfer

from analyte to matrix may result in charge reduction to the singly charged ion observed in the

gas-phase123

The matrix is typically a small energy absorbing molecule such as α-cyano-4-

hydroxycinnamic acid (HCCA) or 2,5,-dihydroxybenzoic acid (2,5-DHB). The molecular weight

values of the trypsinized peptides or intact proteins obtained by MALDI-TOF are then used to

identify the predicated proteins using web-based search engines such as MASCOT.

In cases where the protein is not present in the database, the proteins may be analyzed by tandem

mass spectrometry (ESI-MS). The fragment ions observed in MS/MS spectrum is analyzed to

derive the order of the amino acids in the tryptic peptides or in the intact protein. This method is

known as de novo sequencing. If the fragment ion carries its charge on the N-terminal, the ion is

categorized as a, b or c. If the charge is on the C-terminal of the fragment the type of the ion can

be x, y or z (Figure 11). The difference in the mass between adjacent y- or b-ions corresponds to

that of an amino acid. This can be used to identify the amino acid and, hence the peptide

sequence115

.


Hande Karaköse 25


N C C N C

R1

H

O R2

H

C

O

N C

R3

H

C

O

HN C

R4

H

C

O

OH

H

H

H H

H+

N C C N C

R1

H

O R2

H

C

O

N C

R3

H

C

O

HN C

R4

H

C

O

OH

H

H

H H

H+

Figure 11.Peptide fragmentation momenclature.

2.8. Lipid Analysis

Lipids play an important role in physiology and pathophysiology of living systems. All plant

cells produce fatty acids from acetyl-CoA by a common pathway localized in plastids124

. Fatty

acyls (FAs) are group of molecules synthesized by chain-elongation of an acetyl-CoA primer

with malonyl-CoA or methylmalonyl-CoA groups. Structures with a glycerol group are

represented by two categories: the glycerolipids (GLs), composed mainly of mono-, di- and tri-

substituted glycerols, and the glycerophospholipids (GPs), which are defined by the presence of

a phosphate (or phosphonate) group esterified to one of the glycerol hydroxyl groups. Other

compounds including fatty chains [e.g., fatty acids, fatty alcohols and inter-fatty esters (waxes)]

are also considered in this category. The sterol lipids (STs) and prenol lipids (PRs) share a

common biosynthetic pathway via the polymerization of dimethylallyl

pyrophosphate/isopentenyl pyrophosphate but have obvious differences in terms of their eventual

Y3

Y2 Y1

b3 b2 b1

a1 a2 a3

x3

a1

z3

c1

x2

a2

z2

c2

x1

a3

c3

z1


Hande Karaköse 26


structure and function. Another well-defined category comprises sphingolipids (SPs), which

contain a long-chain base as their core structure. Saccharolipids (SLs) comprise lipids in which

fatty acyl groups are linked directly to a sugar backbone. The final category comprises

polyketides (PKs), which are a diverse group of metabolites from plant and microbial sources125

(Figure 12). Although they are different in their chemical composition, they all share one

characteristic, which is solubilization in non-polar solvents, such as chloroform and hexane.

OH

O

O O

O

HO

O

P

O

O

OHN+

HO

H

H

H H

H

O

OP

HO

O

O P

O

HOO

O

N

NH

O

O

OHOH

HNO

O

HO

HO

O

OH

O OH

O

HO

O

OH

HNH

O

OHH

OH

O

O

O

H

H

O O

O

Fatty acyls: hexadecanoic acid

Glycerophospholipids: 1-hexadecanoyl-2-(9Z-octadecenoyl)-sn-glycero-3-phosphocholine

Sterol lipids:cholest-5-en-3-b-ol

Saccharolipids:UDP-3-O-(3R-hydroxy-tetradecanoyl)-a-D-N-acetylglucosamine

Glycerolipids: 1-hexadecanoyl-2-(9Z-octadecenoyl)-sn-glycerol

Sphingolipids: N-(tetradecanoyl)-sphing-4-enine

Prenol lipids: 2E,6E-farnesol

Polyketides: aflatoxin B1

Figure 12.Examples of lipid categories.

Lipids are one of the most important metabolites of the organism. Essential fatty acids and fat-

soluble vitamins, which are required by organism, can be supplied from lipids. Terpenes and

steroids like vitamin A, D, E, K, cholic acid, and steroid hormones are related with nutrition,


Hande Karaköse 27


metabolism, and regulation function. Lipids on the surface of the organisms protect its surface

from mechanical damages and heat losing, and for the cells, it is closely related to cell

recognition 126

. Apart from their biological functions, fatty acids, triacylglycerols and

phospholipids are the primary classes of lipids of interest in foods.

In general, lipid analysis involves three basic steps which are; 1) extraction of lipids from the

sample, 2) analytical separation, 3) identification and quantification of lipids.

Lipids are mainly extracted from cells, plasma, and tissues. The components obtained depend on

the method of extraction used, especially the solvent. All lipids have a polar head and a nonpolar

tail. Therefore, mixture of chloroform and methanol in a two-step extraction was chosen to

obtain better dissolution of lipids. This approach was developed in 1950s by Folch127

. Lipids of

all major classes could be recovered via chloroform/methanol extraction, typically according to

the Folch, or Lees, and Sloane Stanley or Bligh and Dyer protocols, in which they are mostly

enriched in the chloroform phase 128

.

The most widely used method for the extraction of solid samples is Soxhlet extraction. Purified

lipid extracts are susceptible to oxidation. They should be dissolved in a non-polar solvent (e.g.,

hexane or chloroform) and stored at −20°C in a glass container in a nitrogen atmosphere. They

can be stored in refrigeration temperature (0–4°C) for short-term 125

.

Several separation techniques have been used for the determination of lipids. Long-chain fatty

acids have been determined by gas chromatography (GC) or liquid chromatography (LC).

Supercritical fluid chromatography (SFC) 129

and thin layer chromatography (TLC) 130

was also

utilized for lipid separation.

Traditionally, lipids have been analyzed using gas chromatographic (GC) separation with flame-

ionization detection (FID) or mass spectrometry (MS) detection. Compounds must be thermally

stable with high vapor pressure to be volatilized during the injection in to the GC. Therefore

lipids have to be converted into derivatives with lower boiling points, such as alcoholic esters.

Lipids can be analyzed after hydrolysis, derivatization, or pyrolysis with GC technique. Fatty

acid methyl ester derivatization is the most common method used for analysis with GC

technique. Transesterification is one mechanism that can be employed to form FAMEs from

fatty-acid esters in foods. Alkali- or acid-catalyzed transesterification procedures can be used to

form FAMEs in a methanolic medium. The separation of FAMEs is usually achieved on highly

polar liquid phases and the analytes are separated according to their chain length and degree of


Hande Karaköse 28


saturation. The identification of the fatty acids in the sample can be achieved with comparison of

the retention times of FAME reference compounds. 125

. Although the derivatization solves the

problem of sample volatility, problems with ester formation may include the incomplete

conversion of fatty acids to FAME, loses of highly volatile short chain fatty acids or formation of

contaminations which can overlap with the FAME peak in the GC chromatogram131

.

The wide choice of mobile and stationary phase makes selectivity extremely powerful in HPLC.

Depending on the mobile phases and stationary phases used, there are two modes in LC; which

are normal phase (NP) and reversed phase (RP). RP-LC methods have been developed utilizing

both aqueous and non-aqueous solutions for lipid analysis, whereas in NP-LC, non-polar

solvents are used for separation. RP-LC separates lipids according to their fatty acyl composition

and in NP-LC separation occurs on the basis of their class of compounds125

. Lipids with

molecular weights of 100 – 2000 Da can be detected by LC-MS126

.

Electrospray ionization-mass spectrometry (ESI-MS) has been successfully applied to lipid

analysis. Especially, the combination of chromatographic techniques with MS provides the

technical support for the analysis of lipids and accelerates the emergence of the lipidomics.

Lipidomics, focuses on the global analysis of lipids and their metabolites. The concept

lipidomics was raised by Han et al. in 2003126, 132

.

Strategies currently used in lipidomics include direct-infusion ESI–MS and ESI–MS/MS, LC

coupled with ESI–MS or MS/MS, and MALDI combined with Fourier transform ion cyclotron

resonance MS (MALDI–FTICR–MS) or time-of-flight–MS (MALDI–TOF–MS). In addition, for

some classes of lipids, LC coupled with APCI–MS was also used133

. The analytes can be directly

injected to MS without prior separation and soft ionization. ESI-MS has the advantage that the

structural identification of lipids is more straightforward using different MS/MS experiments

such as precursor ion scan, product ion scan and neutral loss scan. However, ion suppression can

be the major complication in the direct injection MS experiments. MALDI-MS is a laser-based,

soft-ionization method that is often used for analysis of large proteins, but has also been used

successfully with lipids. Important advantages of MALDI–MS in lipid analysis are the speed of

analysis and simplicity of operation: the analytes are ionized under relative soft conditions by

laser desorption using an ultraviolet-absorbing matrix. One important disadvantage of MALDI is

the presence of a lot of background in the lower mass range due to the matrix molecules. In

addition, MALDI–MS is generally less quantitative compared to ESI–MS technique133

.


Hande Karaköse 29


Currently, a wide range of analytical techniques are used for analyzing the lipids, however none

of them provides a global lipid profiling. Therefore, the more different techniques we use, the

wider aspect we can have about our sample.


Hande Karaköse 30


3. RESEARCH OBJECTIVE

The project aims to carry out the chemical profiling of plant called Stevia rebaudiana. Stevia

produces, as its main secondary metabolite diterpene glycosides (steviol glycosides), which are

natural sweeteners. As a sweetener it has two advantages. Firstly, as a terpene it does not cause

allergic reactions unlike most peptide based sweeteners. Secondly, stevia grows almost on any

soils, in particular on fields where tobacco used to be grown. Due to pH of soil become acidic

after tobacco cultivation, very few plants have this specialty and the stevia cultivation in

European Union would offer the tobacco farmers an alternative crop. Most importantly, the

advantages and health benefits of natural sweeteners from stevia make it promising as a novel

food in near future.

Within the EU project (DIVAS) Stevia rebaudiana was cultivated in variety of locations in the

Mediterranean region of Europe over a period of two years using seven different botanical

varieties of stevia. The leaves were harvested between two and three times annually producing a

total of 166 different samples of Stevia rebaudiana leaves. The objective of the project was to

provide scientific basis for a future specification of stevia as novel food in Europe. For this

purpose, chemical specification of stevia leaves has been carried out. Methods have been

developed and used to study the major classes of secondary metabolites. Analysis of the

chemical composition of stevia leaves will allow definition of upper and lower limits of all

relevant stevia plant constituents that are appropriate to chemical analysis including proteins,

lipids and secondary plant metabolites comprising polyphenols and terpene glycosides.

As stevia may be cultivated within various regions of EU with different soil and climatic

conditions it is important to know whether an EU-common specification will be achieved and

how stevia leaves from regions outside EU can be distinguished on a scientific basis. For this

purpose, dataset obtained from stevia leaves from various parts of Europe were analyzed

statistically.


Hande Karaköse 31


4. STEVIOL GLYCOSIDES ANALYSIS BY LC-MS

4.1. Overview

Steviol glycosides extracted from Stevia rebaudiana leaves were subjected to LC- tandem MS

and LC-TOF MS for characterization. The separation of steviol glycosides were compared using

C18 reverse phase column and hydrophilic interaction chromatography (HILIC) column.

Additionally tandem MS data of steviol glycosides is presented using an ion trap instrument,

taking advantage of low energy collision induced fragmentation and multi stage fragmentation

up to MS4 to identify steviol glycosides. LC-TOF method using HILIC column was validated

and used for quantification of steviol glycosides in 166 stevia leaves extracts harvested in

Europe.

4.2.Materials & Methods

4.2.1. Extraction Method

Extraction of steviol glycosides, as well as phenolic acids, lipids and volatile terpenes were

achieved using soxhlet extraction system. Soxhlet conditions were optimised with respect to

solvent volume, extraction cycles and time. Prior to the steviol glycoside extraction a chloroform

extraction was carried out to remove the lipid fraction. In addition to the steviol glycosides the

methanolic fraction contained phenolics in quantitative amounts used for quantification.

Extraction of Steviol glycosides and Phenolic acids: Two grams of S. rebaudiana leaves were

immersed in liquid nitrogen, ground in a hammer mill, and extracted first with 150 mL of

chloroform in a Soxhlet apparatus (Buchi B-811 extraction system) for 2 h and then with 150 mL

of methanol for another 2 h. Solvents were removed from the methanolic extract in vacuo, and

extracts were stored at - 20 oC until required.

4.2.2. LC-MS Analysis of Steviol glycosides

Compound identification was carried out using high resolution mass spectrometry (HR-MS) and

tandem MS using an ion trap mass spectrometer. A HR-MS using an ESI-TOF-MS experiment

allowed determination of molecular formulae based on the accurate mass measurements.

Molecular formulas were in general accepted if an error below 5 ppm was experimentally

observed, as accepted by all peer reviewed chemistry journals.


Hande Karaköse 32


LC-TOF MS: The LC equipment (Agilent 1100 series, Bremen, Germany) comprised a

binary pump, an autosampler with a 100 μL loop, and a diode array detector with a light-pipe

flow cell (recording at 210 nm and scanning from 200 to 600 nm). This was interfaced with a

MicroTOF Focus mass spectrometer (Bruker Daltonics) fitted with an ESI source. The MS

parameters were: nebulizer 1.6 bar, dry gas 12.0 L/min, dry temperature 220 0C. The

MicroTOF was operated in negative ion mode and the mass range was 150 – 1200 m/z.

Internal calibration was achieved with 10 mL of 0.1 mol/L sodium formate solution injected

through a six-port valve prior to each chromatographic run. Calibration was carried out using

the enhanced quadratic calibration mode.

LC-MSn

(tandem MS): The LC equipment (Agilent 1100 series, Bremen, Germany)

comprised a binary pump, an autosampler with a 100 μL loop, and a diode array detector with

a light-pipe flow cell (recording at 210 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 Auto-MSn mode to obtain fragment ions 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 V, starting at

30% and ending at 200%. MS operating conditions (negative mode) were capillary

temperature of 365 oC, a dry gas flow rate of 10 L/min, and a nebulizer pressure of 50 psi.

4.2.3. HPLC conditions

HILIC conditions: Separation was achieved on a 4,6 x 150 mm Dionex Acclaim Mixed Mode

Wax-1 column with 5 μm particle size. Solvent A was 10 mM ammonium formate buffer at pH

3 and solvent B was acetonitrile (ACN). Solvents were delivered at a total flow rate of 0.5

mL/min and the column temperature was set to 40 oC. 5 μL of samples in 80% ACN/water were

injected in to LC-MS system, unless stated otherwise. The isocratic profile was 85 %ACN and

15% water (10 mM ammonium formate buffer).

Reverse phase (C18) conditions: Separation was on a 250 x 3 mm C18 column (Varian Pursuit

XRS) with 5 μm particle size. Solvent A was water/formic acid (1000+0.005 v/v), and solvent B

was acetonitrile (ACN). Solvents were delivered at a total flow rate of 0.5 mL/min and the

column temperature was set to 25 oC. 5 μL of samples were injected in to LC-MS system,


Hande Karaköse 33


unless stated otherwise. The gradient profile was 10 to 80% B in 60 min and a return to 10% B

at 65 min and 5 min isocratic to re-equilibrate.

4.2.4. Calibration Curve of Steviol Glycoside Standards

Stock solutions of the standard compounds of stevioside, rebaudioside A and steviolbioside were

prepared in 80% ACN/water. A series of standard solutions was injected (5μL) into the LC-MS

system. The areas of the peaks of each standard from extracted ion chromatograms (EIC) were

used to make the respective standard curves.

4.2.5. Method Validation

Selectivity of the method was determined by comparing the chromatograms of leaf extract and

reference compounds. Precision was determined by intra and inter-day measurements with three

different concentration of standard solution of stevioside and rebaudioside A on the HILIC

column and evaluated by the relative standard deviation (%RSD). Accuracy of the method was

determined by spiking two different stevia leaf extracts with three different amounts of

stevioside and rebaudioside A, separately and RSD was calculated. Quantification was achieved

by applying calibration curve equations obtained by the least square method.

4.2.6. Solid Phase Extraction (SPE) of Steviol glycosides

Extraction of stevia leaves: 0.8 g of stevia pulverized leaves were sonicated and heated with 30

mL of ACN/water (70:30 v/v) for 15 minutes. Then, the extract was filtrated through a 0.45 μm

filter.

SPE Material I Cartridges were filled with the steviaclean stationary phase specially produced

for Stevia rebaudiana (Knauer GmbH) in the amounts of 0.2 g, 0.4 g and 0.6 g. Each was

condinitioned with water (1mL) and 3 mL of ACN/water (90:10 v/v). 1mL of stevia extract was

loaded on the cartridge. The steviol glycosides were eluted with 2 mL of ACN/water (90:10 v/v).

The eluate was filtered and subjected to HPLC analysis with amino column (Knauer, Eurospher

100 NH2, 5 μm, 150 x 3mm). 5 μL of samples were injected in to LC-MS system. Solvents were

delivered at a total flow rate of 1.0 mL/min and the column temperature was set to 35 oC. UV

detection was at 210 nm. 134


Hande Karaköse 34


SPE Material II Bond Elut C18 cartridges were conditioned with 3 mL of ACN/water (90:10

v/v). 1 mL of stevia extract was loaded on the cartridges. The steviol glycosides were eluted with

ACN/water (90:10 v/v). The eluate was subjected to HPLC analysis with amino and HILIC

columns.

4.3. Results & Discussion

An optimized analytical method was developed for steviol glycoside analysis in stevia leaf

methanol extracts using LC-MS.

Typically for steviol glycoside extraction, solvent extraction using hot water or acetonitrile is

employed followed frequently by further solid phase extraction (SPE) sample clean up. The

challenge of steviol glycoside extraction lies in the different solubilities of steviol glycosides in

aqueous and organic solvents. Good solubility of all steviol glycosides has been reported for

water, however co-extraction of phenolic constituents and carbohydrates exacerbate separation

problems and therefore analysis.

For this reason, first, optimization to use different organic solvents for steviol glycoside

extraction from dried leaf material using Soxhlet extraction was performed. Despite reports that

rebaudioside A shows moderate methanol solubility, we first optimized for Soxhlet extraction

using methanol. Prior to Soxhlet extraction 2 g of dried leaves were treated with liquid nitrogen

and crushed and milled using a blade mill. Extraction times, extraction cycles and extraction

volume were optimized by multiple extraction experiments and it was concluded that using 2 g

of leaf material allowed for a reproducible amount of steviol glycosides being extracted from

notoriously heterogenic plant material.

For development of the LC-MS method, a standard reversed phase C18 column to a HILIC

column was compared. C18 columns have the advantage that steviol glycosides and all phenolic

constituents can be analysed and quantified present in stevia leaves, however retention times are

long and selectivity is not satisfactory. Co-elution of rebaudioside A with stevioside was

observed using a C18 column (Figure 13). At retention times up to 25 min. a total of twelve

chlorogenic acids and nine flavanone glycosides were detected from the analysis of

commercially obtained stevia leaves (please refer to article attached in appendix).


Hande Karaköse 35


Figure 13.Total ion chromatogram in negative ion mode using a C-18 column of methanolic

Stevia rebaudiana extract showing phenolics (chlorogenic acids, flavonoids and steviol

glycosides).

A total of eight steviol glycosides could be observed at retention times from 30 to 42 minutes

with rebaudioside A and stevioside co-eluting. Efficient separation and resolution of this isomer

pair and rubusoside/steviolbioside were achieved with HILIC column using acetonitrile/water

(10 mM ammonium formate) as solvent in the HPLC method. In contrast to C18 column, the

elution order is inverted on the HILIC column. The more glucose units attached to the backbone

structure resulted in later retention times on HILIC. Thus, it was not possible to detect steviol on

the HILIC column. A base peak chromatogram (BPC) is presented in Figure 14 showing baseline

separation of all naturally occurring steviol glycosides. Characterization of steviol glycosides

was achieved by ion-trap mass spectrometry with selected ion monitoring (SIM), and

confirmation of elemental composition was provided by ESI-TOF measurements (Table 1).

Figure 14.Base peak chromatogram of steviol glycosides obtained using HILIC column.

RebA

0.00

0.25

0.50

0.75

1.00

1.25

7 x10

0 5 10 15 20 25 30 35 40 45 Time [min]

stevioside

Chlorogenic acids and flavonoids Diterpene glycosides

Intens

.

Steviolbioside

Stevioside

RebA RebC

Dulcoside

A

Rubusoside


Hande Karaköse 36


4.3.1. Identification of Steviol Glycosides

Compound identification was carried out using high resolution mass spectrometry (HR-MS)

followed by tandem MS using an ion trap mass spectrometer. HR-MS values for all steviol

glycosides are given in Table 1.

COORH3C

CH3

OR1

CH2

General structure ofsteviol glycosides



Compound R R1 Molecular

Formula

Experimental

m/z (M-H+)

-

Theoretical

m/z (M-H+)

-

Relative Error

(ppm)

Steviol

Steviolbioside

H

H

H

glc2 - 1glc

C20H30O3

C32H50O13

317.0819

641.3181

317.0717

641.3179

9.0

0.4

Rubusoside Glc glc C32H50O13 641.3166 641.3179 2.0

Stevioside Glc glc2 - 1glc C38H60O18 803.3751 803.3707 5.5

Rebaudioside A Glc glc32 -1glc

1glc

C44H70O23 965.425 965.4235 1.6


1glc

C38H60O18 803.368 803.3707 2.8

Rebaudioside C

(Dulcoside B)

Glc glc32 -1rham

1glc

C44H70O22 949.427 949.4286 1.7


1glc

C50H80O28 1127.4726 1127.4763 3.3


Rebaudioside F Glc glc32 -1xyl

1glc

C43H68O22 935.4097 935.4129 3.5

Dulcoside A Glc glc2 - 1rham C38H60O17 787.3732 787.3758 3.3


Hande Karaköse 37


Representative tandem MS spectra and fragmentation pathways are shown in Figure 15 and

Figure 16 for rebaudioside A ([M-H+]

- m/z 965) and rebaudioside E ([M-H

+]

- m/z 965). These

two isomers were differing in their MS2 fragmentation, resulting with m/z 803 peak [M-H

+-glc]

-

with the loss of one glucose unit ([M-H2O]- m/z 162) for rebaudioside A, and m/z 641 peak with

the loss of two glucose units for rebaudioside E ([M-H+-2glc]

-). It is most probable that the both

rebaudiosides are losing the glucose first from carboxylic acid moiety due to the increased

stability of the resulting resonance stabilized anion. In general steviol glycosides could be

characterized up to MS4. Rebaudioside A loses one glucose unit each in MS

3 and MS

4 resulting

with the m/z 479 and m/z 317 peaks corresponding to [M-H+-3glc]

- and [M-H

+-4glc]

- ions,

respectively.

Figure 15.Mechanism of fragmentation in tandem MS spectra of Rebaudioside A and

Rebaudioside E illustrating how isomeric compounds can be distinguished by tandem MS.

O

OH

HO

O O

HO

O

CH3

H3C

O

CH2

O

OH

HO OH

O

O

OH

OH

OH

OH

MS2 -324

Rebaudioside EO

HO

HO

OH

HO

O

OH

HO

HO O

HO

O

CH3

H3C

O

CH2

O

OH

O OH

OO

O

HO

OH OH

OH

OH

OH

OH

HO

MS2 -162

MS3 -162

MS4 -162

Rebaudioside A


Hande Karaköse 38


Figure 16.Tandem MS spectra of Rebaudioside A (above) and Rebaudioside E (below) in

negative ion mode.

Tandem mass spectra for rebaudioside D is presented in Figure 17. The structure losses one

glucose unit ([M-H2O]- m/z 162) in MS

2 resulting with the [M-H

+-glc]

- ion with m/z 787 and in

MS3 [M-H

+-2glc]

- with m/z 625. In MS

4 rebaudioside D loses one rhamnose sugar unit ([M-

H2O]- m/z 146) resulting with the [M-H

+-2glc - rham]

- with m/z 479 ion in negative mode.

Further tandem MS data of steviol glycosides are presented in appendix A.

Figure 17.Tandem MS spectra of Rebaudioside D in negative ion mode.

949.6 985.6

-MS

787.4 -MS2

479.1

625.2 -MS3

317.1 479.1

-MS4

0

1 7 x10

Intens.

0.0

0.5

0 2 4 6 x10

0

1 6 x10

200 400 600 800 1000 m/z

322.9 479.1 803.4

641.2 -MS2

317.0

479.1 -MS3

317.0 -MS4

0

2

5 x10 Inten

s.

0.0 0.5

1.0

0

1

2

4

200 400 600 800 1000 m/z

965.6 1001.6

-MS,

803.4 -MS2

317.1 413.1 479.1

641.2 -MS3

317.1

479.1 -MS4

0 1 2

7 x10

Intens

.

0.0 0.5

1.0 1.5

0.0

0.5

0

1

200 400 600 800 1000 m/z


Hande Karaköse 39


4.3.2. Method Validation

The validation procedure involved determination of limit of quantitation (LOQ), determination

of linear range for quantitation, repeatability studies including multiple injection and day to day

repeatability. Additionally, inter sample repeatability experiments were carried out.

Sensitivity & Selectivity

Retention times of the reference compounds of rebaudioside A and stevioside were compared to

the chromatograms obtained from the leaf extracts, apart from that, the accurate masses were

obtained for each peak in the chromatogram from HR-MS measurements.

The calibration curve was linear in the range of 10 – 500 μg/mL for stevioside and 5 – 500

μg/mL for rebaudioside A and steviolbioside. The equations of calibration curves obtained by the

least square method were as follows;

Stevioside: y = 35999x - 845055 R² = 0.9925

Rebaudioside A: y = 39360x + 287715 R² = 0.9935

Steviolbioside: y= 246929x –100000 R2 = 0.9985

where y is the peak area from the LC chromatogram and x is the μg/mL for rebaudiosideA and

stevioside.

Precision

Precision was calculated based on intra and inter-day (n=3) repeatability. Standard solution of

rebaudioside A at the concentrations of 5, 10, 100 and 500 μg/mL and 10, 50, 300 and 350

μg/mL for stevioside were measured on three different days on the HILIC column and the results

were evaluated by calculating the %RSD. The repeatability of the inter-day measurements was in

the range of 4.1 - 6.7 % for rebaudioside A and 1.9 – 5.4 % for stevioside.

Intra-day measurements were evaluated by calculating the %RSD of three injections of each

concentration of 50, 200 and 300 μg/mL of rebaudioside A and 50, 250, 350 μg/mL of

stevioside. Intra-day precision was in the range of 2.4 – 5.6 % for rebaudioside A and 0.9 – 3.7

% for stevioside.


Hande Karaköse 40


Limit of detection was determined as the concentration of the component providing signal to

noise ratio (S/N) of three and for limit of quantification, the concentration resulting in a S/N of

10. According to that, for stevioside LOD was obtained as 2.5 ppm and LOQ as 10 ppm. For

rebaudioside A, LOD was 5 ppm and LOQ was obtained as 10 ppm. For steviol glycoside LOD

was 2 ppm and LOQ was 5 ppm.

Accuracy

The accuracy of the method was determined by calculating the relative error observed in a

standard addition experiment. Stevia leaf extracts were spiked with different amounts of

stevioside and rebaudioside A separately. The relative errors was in the range of 0.043 – 0.074

μg/mL for rebaudioside A and 0.056 – 0.14 μg/mL for stevioside

Comparison to UV data quantification

Calibration curves were also obtained from UV measurements (210 nm) for stevioside. The

sensitivity of the method was less compared to results of LC-MS using EIC. The linearity range

was 50 – 500 μg/mL. S/N of 2:1 was achieved with concentration of 75 ppm. The equation of

calibration curve obtained by the least square method was y = 2.8799x – 90.019 (R2 = 0.9972).

The %RSD of triple injections of 50, 250, 350 μg/mL of the standard was respectively, 19.1, 9.9,

and 3.4.

4.3.3. Comparison to SPE sample clean up

Conventional methods for steviol glycoside quantification frequently employ SPE sample

pretreatment followed by UV based quantification. Two problems might arise here, which have

never been addressed: Firstly, do after SPE treatment analytes co-elute in the steviol glycosides

not observed but co-quantified by UV and secondly, do SPE materials retain steviol glycosides.

Two different SPE stationary phases reported in the literature were tested for presample

treatment of stevia extracts prior to HPLC-MS analysis. SPE with material I was resulting with

decrease in the peaks in the LC chromatogram especially with rebaudioside A, with the increased

amount of material I in the cartridge (Figure 18). SPE cleanup with material II did not have a

dramatic effect on the HPLC analysis with both amino and HILIC columns and gave the similar

results if compared with material I based procedure (Figure 19).


Hande Karaköse 41


Consequently, great care is required when using SPE based protocols since retention of

rebaudioside A, the most lipophilic steviol glycoside, might lead to non-satisfactory accuracy.

Figure 18.Total ion chromatograms for comparison of different amounts of material I in SPE

cleanup procedure.

Figure 19.Total ion chromatograms for comparison of SPE cleanup of the stevia extract with

materials I and II cartridges.

0.2 g Material I

0.6 g Material I

stevioside

Rebaudioside C Rebaudioside A

0.2 g Material II

0.2 g Material I


Hande Karaköse 42


Additionally, it was shown that for both SPE materials I and II if using identical HPLC-

conditions to those reported in the literature using an MS detector co-elution of several analytes

with Rebaudioside A, E and Dulcoside were observed (appendix B, page 122). The nature of the

co-eluting analytes remains largely unknown, however, it can be anticipated that some of them

display UV absorption at 210 nm, again leading to non-satisfactory accuracy. These critical

assessments of SPE based LC-UV quantification methods clearly show that this type of method

require urgent improvement and amendments and are inferior to LC-MS based methods without

SPE sample pretreatment.

4.3.4. Quantification of Steviol Glycosides

Steviol glycoside levels were quantified in all 166 samples made available within the project.

Quantification of steviol glycosides were performed using the validated method with HILIC

column. For three selected steviol glycosides (rebaudioside A, stevioside and steviol), calibration

curves were obtained using six-point calibration from the extracted ion chromatogram (EIC) of

LC-TOF measurement. The quantities of other steviol glycosides were calculated relatively

according to stevioside values for each sample. Please refer to section 4.3.2 for the calibration

curve data.

Data analysis reveals that there are distinct differences between the steviol glycoside profile in

Stevia rebaudiana leaves in the seven different varieties analyzed and distinct differences

between Stevia rebaudiana leaves from different origins. The average amounts and range

(min&max values) of quantity of each steviol glycoside in all 166 samples is presented in Table

3. Detailed quantification data can be found in appendix B.

Table 3.Steviol glycosides values from 166 samples

Steviol glycoside Average (mg/100 g leaves) Range (mg/100 g leaves)

Rebaudioside A 1.017 79 - 5336

Stevioside 6071 252 - 17509

Dulcoside A 239 5 - 680

Rubusoside 111 5 - 459

Rebaudioside C 282 26 - 820

Total 9036 554 - 18067


Hande Karaköse 43


Additionally, from the obtained quantification data, the samples were analyzed on a radar plot

based on their variety and origin (Figure 20 and 21). Within the project seven defined botanical

varieties of Stevia rebaudiana were cultivated and their steviol glycoside profile was determined.

The distinct differences in the stevioside profile of different variety of stevia leaves can be easily

recognized. From the radar plot (Figure 20), it can be seen that variety 5 and non-EU samples

have the maximum value of stevioside, whereas variety 7 having the minimum value for

stevioside but maximum value for rebaudioside A.

Stevia rebaudiana was within this project cultivated in nine different locations within the EU.

Additionally samples from outside the EU were available for comparison. In Figure 21, EU

cultivated Stevia rebaudiana shows higher concentrations of stevioside (Amiflikeia and

Argentinie having the maximum end values). However for rebaudioside A, Conaga cultivated

stevia shows higher end concentration values.

Figure 20.Radar plot of steviol glycoside concentrations varying between seven varieties

(average values taken within +/- 3 σ) and in comparison to non-EU samples. Concentrations are

given on radial axis in g/100g dry leaf material. Outer numbers are indicating the 7 varieties and

non-EU samples; numbers inside the plots are indicating the concentrations.

0.000

2.000

4.000

6.000

8.000

10.000 1

2

3

4

5

6

7

Non-EU

RebA

Stevioside

Dulcoside A

Rubusoside

RebC


Hande Karaköse 44


Figure 21.Radar plot of steviol glycoside concentrations varying between all origins (average

values taken within +/- 3 σ) and in comparison to non-EU samples. Concentrations are given on

radial axis in g/100g dry leaf material.

4.4. Conclusion

In conclusion, steviol glycosides were successfully analyzed and quantified using LC-MS

technique. Stevia leaves can be analyzed using a variety of different LC-MS methods. While LC-

MS on a C18 column allows analysis of steviol glycosides, however suffering from selectivity

problems, analysis on a HILIC column coupled to ESI-TOF detection allows separation and

quantification of all known naturally occurring steviol glycosides. Linear range, sensitivity and

reproducibility were excellent. Using both high resolution MS and tandem MS on an ion trap

instrument reliable structure confirmation can be carried out based on characteristic MSn

fragment spectra of all steviol glycosides.

Distinct differences in the quantity of steviol glycosides within the stevia samples were observed.

From the data it was observed that variety 7 is having the maximum concentration for

rebaudioside A but minimum value for stevioside concrentration. EU origin cultivated stevia

samples have the maximum concentration for stevioside but average value for rebaudioside A

concentration.

0.000

2.000

4.000

6.000

8.000

10.000

12.000 TCV

Uconor

Agrinion

Toumpa

Portugal

Amfilia

Argentinie

Granada

Turkei

Amiflikeia

APTTB

Conaga

RebA

Stevioside

Dulcoside A

Rubusoside

RebC


Hande Karaköse 45


5. POLYPHENOLS in STEVIA REBAUDIANA

5.1. Overview

The hydroxycinnamate derivatives of S. rebaudiana have been investigated qualitatively and

quantitatively by LC-MSn. Chlorogenic acids and flavonoid glycosides of Stevia rebaudiana

from different origins in all around Europe with seven different botanical varieties were profiled

and quantified. The correlation study between CGAs, differences between stevia samples and

effect of origin and variety on the CGA profile were tested statistically from the obtained dataset.

5.2. Materials & Methods

5.2.1. Sample Preparation

Two grams of S. rebaudiana leaves was immersed in liquid nitrogen, ground in a hammer mill,

and extracted first with 150 mL of chloroform in a Soxhlet apparatus (Buchi B-811 extraction

system) for 2 h for removal of lipid fraction and then with 150 mL of methanol for another 2 h.

Solvents were removed from the methanolic extract in vacuo, and extracts were stored at - 20 oC

until required.

5.2.2. LC-MS Analysis of Polyphenols

LC-TOF MS: The LC equipment (Agilent 1100 series, Bremen, Germany) comprised a binary

pump, an autosampler with a 100 μL loop, and a diode array detector with a light-pipe flow cell

(recording at 254 nm and scanning from 200 to 600 nm). This was interfaced with a MicroTOF

Focus mass spectrometer (Bruker Daltonics) fitted with an ESI source. The MS parameters were:

nebulizer 1.6 bar, dry gas 12.0 L/min, dry temperature 220 0C. The MicroTOF was operated in

negative ion mode and the mass range was 150 – 1200 m/z. Internal calibration was achieved

with 10 mL of 0.1 mol/L sodium formate solution injected through a six-port valve prior to each

chromatographic run. Calibration was carried out using the enhanced quadratic calibration mode.

LC-MSn: The LC equipment (Agilent 1100 series, Bremen, Germany) comprised a binary

pump, an autosampler with a 100 μL loop, and a diode array detector with a light-pipe flow cell

(recording at 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 Auto-MSn mode to obtain fragment ions 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 V, starting at 30% and ending at 200%. MS operating


Hande Karaköse 46


conditions (negative mode) were capillary temperature of 365 oC, a dry gas flow rate of 10

L/min, and a nebulizer pressure of 50 psi.

HPLC: Separation was achieved on a 250 x 3 mm C18 column (Varian Pursuit XRS) with 5 μm

particle size. Solvent A was water/formic acid (1000+0.005 v/v), and solvent B was acetonitrile

(ACN). Solvents were delivered at a total flow rate of 0.5 mL/min and the column temperature

was set to 25 oC. 5 μL of samples were injected in to LC-MS system, unless stated otherwise.

The gradient profile was 10 to 80% B in 60 min and a return to 10% B at 65 min and 5 min

isocratic to re-equilibrate.

5.2.3. Calibration Curve of Standard Compounds

Most abundant chlorogenic acid derivatives (3-CQA, 4-CQA, 5-CQA, 3,5-diCQA, 4,5 diCQA)

and two flavonoid glycosides (quercetin-3-glycoside and kaempferol-7-glycoside) were chosen

for calibration curves.

Stock solutions of the standard compounds were prepared in 80% ACN/water. A series of

standard solutions was injected (5μL) into the LC-MS system. The areas of the peaks of each

standard from extracted ion chromatograms (EIC) were used to make the respective standard

curves.

5.2.4. Hydrolysis of Flavonoid Glycosides

5 mg crude extract was dissolved in 2 ml 2M HCl and heated at 90 0C for 40 min. Sample was

then directly used for LC-MS or diluted with MeOH.

5.2.5. Statistical Analysis

Statistical analyses of the data were performed using IBM SPSS 20. The distributions of the

variables were tested for normality using the Kolmogorov-Smirnov test. Associations between

the variables were investigated using both parametric (Pearson’s correlation) and non-parametric

(Spearman’s correlation) techniques. Results were interpreted using the widely accepted 5%

level of significance.

To test whether there were differences on each chlorogenic acid with respect to its origin or

variety, separate one-way ANOVA analyses was employed, followed by two post-hoc tests:

Fisher’s Least Significant Difference (LSD) as the least conservative test where equal variances

are assumed and Games-Howell test where non-equal variances are assumed for the multiple pair


Hande Karaköse 47


wise comparison tests. All empirical results were interpreted using the widely accepted 5% level

of significance (p < 0.05).


Stevia extract was analyzed by LC-MSn in the negative ion mode using an ESI ion-trap mass

spectrometer, allowing assignments of compounds to regioisomeric level, and also by HR-MS

using ESI-TOF in negative ion mode connected to LC, that allowed determination of molecular

formulae based on the accurate mass measurements. Molecular formulas were in general

accepted if an error below 5 ppm was experimentally observed, as accepted by all peer reviewed

chemistry journals. In a second experiment tandem MS experiments were carried out and the

observed fragmentation patterns compared to those of authentic reference materials (either

obtained commercially or from our own laboratory). After obtaining multi-dimensional

information of four parameters on chromatographic retention times, UV-spectra (UV-VIS DAD

detector coupled to LC-MS system), HR-MS and tandem-MS data comparison to authentic

reference material did allow compound identification. Peak assignments of CGAs have been

made on the basis of structure diagnostic hierarchical keys previously developed81, 135

.

Chlorogenic acids and flavonoids were quantified using an established reversed phase LC-MS

method on a C-18 column using ESI-TOF-MS in the negative ion mode. All required analytes

showed baseline separation with exception of the pair 3,4- and 4,5 dicaffeoyl quinic acid. A

typical chromatogram using a reversed phase C-18 column of a stevia extract showing polar

polyphenols at early retention times and more lipophilic steviol glycosides at later retention

times is shown in Figure 22. Abbreviations and numbering of CGAs and flavonoids are

presented in Figure 23, Table 4.

In all cases quantitation was carried out using extracted ion chromatograms only. Additionally

flavonoids were quantified using kaempferol-7-glucoside and quercetin-3-glucoside as reference

standards, resulting for flavonoids in relative values rather than absolute values.


Hande Karaköse 48


Figure 22.Base peak chromatogram in negative ion mode using a C18 column of methanolic

Stevia rebaudiana extract showing phenolics (CGAs, flavonoids and steviol glycoside).

Table 4.Chromatographic and MS data on flavonoid glycosides and CGAs present in stevia

leaves

Compound Number Compound* m/z [M-H+]

-

1 3-caffeoylquinic acid (3CQA) 353



4 Rutin 609

5 Quercetin-galactoside 463

6 Kaempferol-glucopyranoside

Quercetin-rhamnoside 447

7 Quercetin-fructoside; Luteolin-glucuronide 461

8 Quercetin-pentoside 433

9 kaempferolxylosylglucoside

Naringin 579

10 Apigenin-galactoside 431

11 3,5-di-caffeoylquinic acid (3,5diCQA) 515

12 4,5-dicaffeoylquinic acid(4,5diCQA) 515

13 Quercetin-diglucoside-rhamnoside 771

14 Kaempferol-glucosylrhamnosyl-glucoside/galactoside 755

15

Kaempferol-rhamnopyranosyl-glucopyranoside(rutinoside) isomers

Quercetin-dirhamnoside

Apigenin-diglucoside/galactoside

593

16 Quercetin-trisaccharide 741

17 Kaempferol 3-rhamnopyranosyl-rhamnopyranosyl-glucopyranoside 739

*Compounds named for flavonoid glycosides are only possible structures, which were not identified or confirmed.

0.0

0.5

1.0

1.5

2.0

2.5

6 x10

5 10 15 20 25 30 35 40

Steviol glycosides CGAs & Flavonoid glycosides

6 5 8

9

1 4

7

10

11

12

13

14

2

3

Time [min]

Intens.


Hande Karaköse 49


Figure 23.Structures of caffeoylquinic acids and flavonoid glycosides.

OH

OH

OOH

HOOC

O OH

O

OHOH

HOOC

O

OH

3

O

OH

OHOH

HOOC

O

OH

OH

5 OH

OH

O

OOH

HOOC

O

OH

OH

O

O

OHOH

HOOC

O

OH

OH

43O

OH

HO

4

O

OH

OH

5

O

OH

OOH

HOOC

O

OH

OH

5

O

OH

OH

3

3-CQA 5-CQA

3,5-diCQA 3,4-diCQA 4,5-diCQA

O

O

OHOH

HOOC

O

OH

OH

4

O

5

HO

HO

cis-4,5-diCQA

4-CQA

HO O

O

OOH

OHOH

OHO

OHOH

OH

Quercetin-3-O-beta-D-glucoside

O

O OO

OH

O

HO

OH

OH

OH

OOH

HO

OH

OHOH

OH

OOH

HO

OH

OOH

HO

Kaempferol

Quercetin Luteolin Apigenin

O O

OH

OOH

OHO

HOOH

OH

Kaempferol-7-O-beta-D-glucoside

OH

OH

OH


Hande Karaköse 50


5.3.1. Characterization of Chlorogenic acids

Mono-caffeoylquinic acids (mono-CQA) and di-caffeoylquinic acids (di-CQAs) were identified

in stevia using the hierarchial keys previously developed81

. It is possible to discriminate between

the isomers of caffeoylquinic acids and feruloylquinic acids by using LC-MSn. The

fragmentation pattern depends on the stereochemical relationships between the substituents on

the quinic acid moiety. Four peaks were detected at m/z 353.1 and assigned as well-known 3-

CQA, trans-5-CQA, and 4-CQA and cis-5CQA. Three dicaffeoylquinic acid isomers were

identified by their parent ion m/z 515.2 and were assigned as 3,5-diCQA, 3,4-diCQA, and 4,5-

diCQA using the hierarchial keys81, 135

. Two further peaks present as minor components showed

fragmentation patterns similar to that of 4,5-diCQA, which were identified as cis isomers of 4,5-

diCQA. Figures 24 - 26 show selected data including an extracted ion chromatogram for

monocaffeoyl quinic acids and two tandem mass spectra for selected regioisomeric stevia

caffeoyl quinic acids as typical secondary metabolites. All other fragmentation patterns are

provided in appendix C.

Twenty-four hydroxycinnamic acid derivatives of quinic and shikimic acid were detected in the

work using stevia leaves not cultivated in this project and the results have been published in the

course of this project (Please refer to the article attached in appendix for detailed insight).


Hande Karaköse 51


OH

HOOC

O

OH

OH

OH

OH

O

1 3

OH

HOOC

OH

OH

O

1

O

OH

5

OH

HOOC

OH

O

OH

OH

O

OH

HO

Figure 24.Extracted ion chromatogram of m/z 353 of three mono-caffeoylquinic acids, from left

to right: 3-Caffeoylquinic acid (1), 5-Caffeoylquinic acid (2) and 4-Caffeoylquinic acid (3) in

negative ion mode.

All monoacyl CGA (m/z ~ 353) gave the expected parent ion (monoacyl CGA - H+) (Figure 25

and 26) in tandem MS analysis. 3-CQA and 5-CQA produce an MS2 base peak at m/z ~ 191

corresponding to [quinic acid-H+]- (Q1 ion) and in MS

3 it fragments to Q2 ion with m/z 85.1 and

[quinic acid – H2O - H+]- (Q3 ion) at m/z 172.8 (Figure 25). 3-CQA might be discriminated by its

MS2 peak at m/z ~135 and by slight difference of the intensity of m/z ~178 in MS

2. It is easy to

distinguish the 4-substititued CGA by it is dehydrated quinic acid moiety which gives MS2 base

peak at m/z ~173.

EIC 353.0 -All MS

0.0

0.5

1.0

1.5

2.0

8 Inte

ns.

0 10 20 30 40 50

1

2

3

Cis-2

x10

Time

[min]


Hande Karaköse 52


Figure 25.Consecutively MS, MS2and MS

3 spectra of 3-Caffeoylquinic acid in negative ion

mode.


3 spectra of 4-Caffeoylquinic acid in negative ion

mode.

The diacyl CGAs behave similarly, giving the parent ion [diacyl CGA – H+]- at m/z 515.

DiCQAs loses the caffeic acid moiety in MS2 yielding a [diacyl CGA – cinnamoyl – H

+]

- and

these ions are identical to the parent ions obtained from monoCQAs. 4,5-diCQA give dehydrated

quinic acid moiety as base peak at m/z ~173 in MS3 as previously seen for 4-CQA. This ion was

not observed for 3,5-diCQA, instead MS3 base peak was at m/z 191 as previously observed for 3-

CQA (Figure 27 and 28). Thus, 3,5-diCQA can be distinguished easily from the 4-acylated

caffeoylquinic acid isomers.

472.8

352.9

374.9

-MS

172.7 -MS2(352.9)

71.3 154.7

93.0 -MS3(353.1->172.7)

0 1 2 3 7 x10

Intens.

0 2 4 6

6 x10

0.0

0.5

1.0

5 x10

200 400 600 800 1000 m/z

472.9 729.2

352.9 -MS

134.8

190.7 -MS2(352.9)

172.8 85.1 -MS3(353.1->190.6)

0 1 2 3 7 x10

Intens.

0.0

0.5

1.0

7 x10

0

2

4

6 4 x10

200 400 600 800 1000 m/z

OH

HOOC

OH

OH

O

OH

OH

O

Q1

OH

HOOC

OH

O

Q2 Q3

CH

CHHO

HO Caffeic m/z 135


Hande Karaköse 53



3 MS

4 spectra of 3,5-dicaffeoylquinic acid in negative

ion mode.


3 MS

4 spectra of 4,5-dicaffeoylquinic acid in negative

ion mode.

1031.4

515.0 -MS

172.7 254.8 298.9

352.9 -MS2(515.0)

134.8

172.7 -MS3(515.3->352.9)

93.0

0 1

8 x10 Intens.

0.0

0.5

8 x10

0

1

7 x10

0 2 4 5 x10

200 400 600 800 1000 m/z

352.9 613.0

515.0

1031.3

-MS

190.7

352.9 -MS2(515.0)

134.7

190.7 -MS3(515.3->352.9)

85.1 126.8

172.7

-MS4(515.3->353.1->190.7)

0

1 8 x10

Intens.

0.0

0.5

8 x10

0

2

7 x10

0 1 2 5 x10

200 400 600 800 1000 m/z

-MS4(515.3->353.1->172.8)


Hande Karaköse 54


5.3.2. Characterization of Flavonoid Glycosides

Flavonoid glycosides are an important group of plant natural products in which different type of

sugars are linked to an aglycone. Determination of the identity and position of linkage to the

aglycone of sugars by mass spectrometry alone is challenging. From the mass spectra of

flavonoid glycosides using tandem MS we can obtain molecular mass, structure of the aglycone

(from its m/z), number of sugar rings and their configuration. Negative ion mode was selected for

the analyses because previous results suggested that negative mode was more sensitive than

positive ion mode.

A total of twelve peaks in the chromatogram corresponding to flavonoid glycosides were

identified. All compounds could be identified as belonging to this class of compounds due to

their characteristic fragmentation patterns in tandem MS showing neutral losses of sugar

moieties following by characteristic fragment spectra of the aglycones. The nature of the

aglycones was further substantiated by hydrolysis of the total phenol extract followed by LC-MS

analysis revealing that four flavonoid aglycones quercetin, kaempferol, luteolin and apigenin

(Figure 29 and Table 4) are present in Stevia rebaudiana leaves.136

O O O O

OH

O

HO

OH

OH OH

OOH

HO

OH

OH

OH

OH

OOH

HO

OH

OOH

HO

KaempferolMw 286

QuercetinMw 302

LuteolinMw 286

ApigeninMw 270

Figure 29.Chemical structure of four flavonoid aglycones identified in Stevia rebaudiana leaves.

Although the aglycone and the glycane were identified for an observed m/z, the accurate

structure of the flavonoids glycoside could not be determined because identity and the site of

connection of monosaccharide cannot be determined by LC-MS. The possible structures of

compounds were determined by comparison of the mass spectral data obtained with literature

data. However, the detailed chemical structure of these flavonoids could not be established

unambiguously by LC-MS. None of the compounds present were shown to be identical to any of

the twelve reference standards used commercially or to reference compounds available in our

laboratory. A preparative LC isolation and full structure elucidation of the compounds was


Hande Karaköse 55


outside the scope and timeline of the project but needs further attention. A table showing the

chromatographic and MS characteristics of the flavonoids was given previously in Table 4 and

extracted ion chromatogram and tandem mass spectra for selected compound is shown in Figure

30 and 31. All other tandem mass spectral data of flavonoid glycosides are presented in appendix

C.

The possible fragmentation pattern and ion nomenclature of flavonoid glycosides is illustrated on

luteolin-7-O-rutinoside in Figure 32.

Figure 30.Extracted ion chromatogram of m/z 447.0 in negative ion mode.

Figure 31.An example of tandem MS spectra for compound 1, revealing its identity as

kaempferol glucopyranoside.

447.0 -MS

284.8 -MS2(447.0)

174.7 216.7

-MS3(447.3->286.8)

0 50

100 Intens.

[%]

0 50

100

0 50

100

100 200 300 400 500 600 700 m/z

1

EIC 447.0 -All MS

0.00

0.25

0.50

0.75

1.00

1.25 8 x10 Intens.

5 1

0 1

5 2

0 2

5 3

0

1

3

2

Time

[min]


Hande Karaköse 56


O

OH

OO

OH

OH

OH

CH2OH

O

OH

OH

Figure 32.Fragmentation illustration on luteolin-7-glucoside137

The most useful fragmentation for flavonoid aglycone identification are the cleavage of two C—

C bonds of the C-ring, resulting in structural information for A and B ions (Figure 32). These

ions can be rationalized by retro-Diels–Alder (RDA) reactions and are the most diagnostic

fragments for flavonoid identification since they provide information on the number and type of

substituents in the A- and B-rings. The flavonoid aglycone fragment ions can be designated

according to the nomenclature proposed by Ma et al 138, 139

.

5.3.3. Quantification of Chlorogenic acids & Flavonoid Glycosides

Phenolics including chlorogenic acids and flavonoids were quantified using an established

reversed phase LC-MS method on a C-18 column using ESI-TOF-MS in the negative ion mode.

Standard solutions were analyzed using the same chromatographic method as used for stevia leaf

extracts as indicated before (chapter 5.2.2). The calibration curves were obtained by the external

standard method on six levels of concentration of reference compounds. For quantitation the six

most abundant chlorogenic acid derivatives (3-CQA, 4-CQA, 5-CQA. 3,4-diCQA, 3,5-diCQA,

4,5 diCQA) were chosen and calibration curves were obtained with excellent linearities. In all

cases quantitation was carried out using extracted ion chromatograms only. Cis isomers of 5-

CQA and 4,5-diCQA were quantified based on the corresponding calibration curves of trans

isomers, resulting in relative values of cis isomers (please refer to appendix D for quantification

data, page 140). Additionally flavonoids were quantified using kaempferol-7-glycoside (k7g) and

B0

A0

Z0

X0

A1

Y0

B1


Hande Karaköse 57


quercetin-3-glycoside (q3g) as reference standards, resulting for flavonoids in relative values

rather than absolute values (appendix D).

The linearity and the equations of calibration curves obtained by the least square method were as

follows:

Reference compound Linearity (μg/mL) Equation R2

3-CQA 1-100 82912x+346549 0.99

5-CQA 5-500 49062x+907693 0.99

4-CQA 1-150 129889x+395550 0.99

3,5-diCQA 10-200 138064x+7000000 0.98

4,5-diCQA 10-400 138243x+4000000 0.98

K7g 50-1000 26105x+6000000 0.98

Q3g 2-150 97003x+2000000 0.98

Where y is the peak area from the LC chromatogram and x is the μg/mL for the reference

compound.

5.3.3.1. Sample Variation

From the obtained quantification data, a series of statistical analysis was carried out. As a first

step, for each variety, origin and harvest average values and standard deviations were

determined. Additionally, minimum and maximum values for each sample subgroups are given

in the tables.

Within the project seven defined botanical varieties of Stevia rebaudiana were cultivated and

their phenolic profile was determined. From the data variations between different batches and

average values averaged over all samples from a single variety can be compared. Additionally

variations in single compound quantities, quantities of groups of compounds (e.g. mono-caffeoyl

quinic acids, dicaffeoyl quinic acids) or ratios of two single compounds can be compared.

From the data for example it can be seen that the average concentration of all monocaffeoyl

quinic acids remains rather constant over all varieties (2.123 - 2.686 g/100g), whereas a more

spread of data is observed for dicaffeoyl quinic acids (1.484 – 2.432 g/100g). Varieties 5, 6, 7

and 3 show on average increased levels of chlorogenic acids compared to varieties 2. Average

values are given in Table 5, for detailed quantification data please refer to appendix D. All EU


Hande Karaköse 58


Stevia rebaudiana varieties show considerable higher levels of chlorogenic acids and flavonoids

if compared to some samples obtained from outside the EU.

Variations can be displayed in a radar plot shown in Figure 33, 34, 35. Here the lower

chlorogenic acid content in non-EU samples as well as in variety 2 can be appreciated.

Table 5.Average values (taken within +/- 3 σ) for chlorogenic acids in seven different varities

Variety

3CQA

(g/100g

leaves)

4CQA

(g/100g

leaves)

5CQA

(g/100g

leaves)

Total mono

(g/100g

leaves)

3,5-diCQA

(g/100g

leaves)

4,5-diCQA

(g/100g

leaves)

TotaldiCQA

(g/100g

leaves)

1 0.279 0.096 1.921 2.262 0.925 1.159 1.969

2 0.267 0.094 1.804 2.208 0.880 1.208 1.807

3 0.290 0.114 2.267 2.650 1.101 1.213 2.022

4 0.294 0.118 2.258 2.616 1.101 1.259 2.020

5 0.279 0.105 2.108 2.466 1.365 1.313 2.432

6 0.261 0.105 2.100 2.686 1.309 1.396 2.272

7 0.178 0.113 2.262 2.507 1.277 1.488 2.108

Non-EU 0.195 0.092 1.836 2.123 0.881 0.700 1.484


Hande Karaköse 59


Figure 33.Radar plot of individual chlorogenic acid concentrations varying between seven

varieties (average values taken within +/- 3 σ) and in comparison to non-EU samples.

Concentrations are given on radial axis in g/100g dry leaf material. Outer numbers indicating the

7 varieties and non-EU samples; numbers inside the plot are indicating the average

concentrations of individual chlorogenic acids.

Figure 34.Radar plot of total mono- and di-acyl quinic acids (chlorogenic acids) concentrations

varying between seven varieties (average values taken within +/- 3 σ) and in comparison to non-

EU samples. Concentrations are given on radial axis in g/100g dry leaf material. Outer numbers

indicating the 7 varieties and non-EU samples; numbers inside the plots are indicating the

concentrations.

0.000

0.500

1.000

1.500

2.000

2.500 1

2

3

4

5

6

7

Non-EU

3-CQA

4-CQA

5-CQA

3,5 diCQA

4,5-diCQA

0

0.5

1

1.5

2

2.5

3 1

2

3

4

5

6

7

Non-EU

monoCQA

diCQA


Hande Karaköse 60


Figure35. Bar plot of total mono- and di-acyl quinic acids (chlorogenic acids) concentrations

varying between seven varieties (average values taken within +/- 3 σ) and in comparison to non-

EU samples. Concentrations are given on radial axis in g/100g dry leaf material.

Variations by Origin

Stevia is capable of growing almost anywhere under poor soil conditions, in particular where

tobacco used to grown. There are only few plants possessing this feature. Therefore, stevia can

serve as an alternative crop to the tobacco farmers and discourage tobacco cultivation inside EU.

Stevia rebaudiana was within this project cultivated in nine different locations (Figure 36) within

the EU (e.g. Conaga,TCV, Italy; Granada, Spain; and Toumpa, Agrinio, Amfikleia, Greece).

Additionally samples from outside the EU were available for comparison (e.g. Paraguay,

Argenitine). Stevia leaves obtained from these origins were analyzed for studying the effect of

growth conditions (e.g. sun, soil, and climate) on the metabolite profile.

According to the literature polyphenol concentrations are due to their physiological function as

UV protection agents a direct function of growth altitude and climatic conditions, in particular

sunshine hours. Accordingly variations of chlorogenic acid concentrations between different

origins should be expected.

0.000 0.500 1.000 1.500 2.000 2.500 3.000

1

2

3

4

5

6

7

Non-EU

di CQA

mono CQA


Hande Karaköse 61


Figure 36.Map showing the origins of stevia cultivation within the project.

Indeed the data reveal significant variations in CGA concentrations varying from 3.090 -1.637

g/100g for total monocaffeoyl quinic acids and 2.890 - 1.144 g/100g for dicaffeoyl quinic acids.

EU cultivated Stevia rebaudiana shows concentrations of CGAs nicely sandwiched between

extreme values at both ends observed in samples from outside the EU (e.g. highest for

Argentinian samples with an average value of 2.890 g/100g dicaffeoyl quinic acids and APTTB

and Portugal samples with a lowest average value of 1.448 g/100g and 1.144 g/100g

respectively). Again a radar plot shown in Figure 37 was used to display variations between

different origins. Average values are given in Table 6.


Hande Karaköse 62


Figure 37.Radar plot of total mono- and di-acyl quinic acids (chlorogenic acids) concentrations

varying between all origins (average values taken within +/- 3 σ) and in comparison to non-EU

samples. Concentrations are given on radial axis in g/100g dry leaf material.

A bar plot shown in Figure 38 allows further direct comparison between samples of different

origins.

Figure 38.Bar plot of total mono- and di-acyl quinic acids (chlorogenic acids) concentrations

varying between all origins (average values taken within +/- 3 σ) and in comparison to non-EU

samples. Concentrations are given on radial axis in g/100g dry leaf material.

0.000

1.000

2.000

3.000

4.000 TCV

Uconor

Agrinion

Toumpa

Portugal

Amfilia

Argentinie

Granada

Turkei

Amiflikeia

APTTB

Conaga

mono CQA

di CQA

0.000 0.500 1.000 1.500 2.000 2.500 3.000 3.500

TCV

Uconor

Agrinion

Toumpa

Portugal

Amfilia

Argentinie

Granada

Turkei

Amiflikeia

APTTB

Conaga

di CQA

mono CQA


Hande Karaköse 63


Table 6.Average values (taken within +/- 3 σ) for chlorogenic acids between origins

Variation between harvests

Stevia rebaudiana was harvested three times within this project. From the data for example it can

be seen that the average concentration of all dicaffeoyl quinic acids and monocaffeoyl quinic

acids remains rather constant over first and second harvests, whereas a significant decrease can

be observed in the third harvests for monocaffeoyl quinic acids. Overall, for individual

compounds and for total amounts, second harvests are having the highest values 2.595g /100g for

monocaffeoyl quinic acids and third harvests are for dicaffeoylquinic acids 2.242 g/100 g (Table

7). However, decrease in the quantities of CGAs with the later harvest could be observed if the

three harvests are compared while the origin and the variety of stevia are kept constant (Table 8).

In particular mono-CQAs are showing higher decreases from harvest I to harvest III, whereas

diCQAs values are rather constant within chosen three origins with same variety presented in

Table 8.

Origin 3CQA

(g/100g leaves)

5CQA

(g/100g leaves)

4CQA

(g/100g leaves)

Total mono

(g/100g leaves)

3,5-diCQA

(g/100g leaves)

4,5-diCQA

(g/100g leaves)

TotaldiCQA

(g/100g leaves)

TCV 0.224 2.015 0.101 2.336 0.909 0.987 1.757

Uconor 0.323 2.060 0.119 2.575 1.373 1.480 2.079

Agrinion 0.278 2.078 0.104 2.604 1.233 1.339 2.460

Toumpa 0.269 2.450 0.123 2.780 1.351 1.255 2.471

Portugal 0.202 1.580 0.079 1.861 0.555 0.589 1.144

Amfilia 0.342 2.208 0.110 2.674 0.898 1.043 1.941

Argentinie 0.284 2.446 0.122 2.883 1.435 1.520 2.890

Granada 0.197 1.855 0.093 2.145 0.686 0.703 1.389

Turkei 0.104 1.837 0.092 2.093 0.727 0.859 1.586

Amiflikeia 0.441 2.646 0.133 3.090 1.302 1.254 2.335

APTTB 0.206 1.363 0.068 1.637 0.746 0.703 1.448

Conaga 0.310 2.525 0.126 2.956 0.993 1.410 2.328


Hande Karaköse 64


Table 7.Average values (taken within +/- 3 σ) for chlorogenic acids between harvests

Harvest

3CQA

(g/100g

leaves)

4CQA

(g/100g

leaves)

5CQA

(g/100g

leaves)

Total

monoCQA

(g/100g

leaves)

3,5-diCQA

(g/100g

leaves)

4,5-diCQA

(g/100g

leaves)

TotaldiCQA

(g/100g

leaves)

I

0.306

0.110 2.096 2.562 1.004 1.164 1.967

II

0.276

0.113 2.251 2.595 1.209 1.182 2.142

III

0.145

0.091 1.828 2.064 1.226 1.155 2.242

Table 8.Comparison of average values (taken within +/- 3 σ) for chlorogenic acids between three

harvests of same variety and origin

Sample no. Origin Variety Harvest 3CQA 5CQA 4CQA Totalmono 3,5diCQA 4,5diCQA TotaldiCQA

71 Agrinion 3 I 0.413 3.266 0.163 3.843 1.008 1.282 2.290

123 Agrinion 3 II 0.247 2.626 0.131 3.004 1.570 1.342 2.913

162 Agrinion 3 III 0.168 1.372 0.069 1.608 0.974 0.938 1.912

98 Toumpa 3 I 0.491 4.099 0.205 4.795 1.208 1.535 2.743

127 Toumpa 3 II 0.175 2.490 0.124 2.789 1.650 1.271 2.921

126 Toumpa 3 III 0.187 1.961 0.098 2.246 1.684 1.567 3.251

137 TCV 3 I 0.441 3.153 0.158 3.752 1.578 1.266 2.844

139 TCV 3 II 0.267 2.692 0.135 3.094 1.854 1.155 3.010

161 TCV 3 III 0.161 1.777 0.089 2.026 0.987 1.077 2.064


Hande Karaköse 65


5.3.3.2. Flavonoid quantification

A selection of the two major flavonoids identified within the chromatogram was quantified as

their glycosylated derivatives using calibration curves for closely related compounds quercetin-

3-glycoside and kaempferol-7-glycoside for all 166 samples. Additionally, for ten selected

samples a flavonoid hydrolysis was carried out using acidic methanol with subsequent

quantification of the four flavonoid aglycones quercetin, apigenin, luteolin and kaempferol

(please refer to chapter 5.3.2 for identification). Relative values for flavonoids based on LC-MS

data are contained within Tables 9 and 10. Data for the flavonoid hydrolysis are given in Table

11.

Table 9.Flavonoid glycosides average values for two major flavonoids in samples between

origins determined by LC-MS directly from extracts without hydrolysis. Reference compounds

used were quercetin-3-glucoside and kaempferol-7-glucoside. Total flavone value designates

addition of all intensities of all flavonoid signal in chromatogram referenced to kaempferol-7-

glucoside.

Origin

Kaempferol-7-glucoside

(g/100g leaves)

Quercetin-3-glucoside

(g/100g leaves)

Total flavones

(g/100g leaves)

TCV 1.653 0.089 3.630

Uconor 2.762 0.103 6.397

Agrinion 3.431 0.047 7.865

Toumpa 3.602 0.063 8.170

Portugal 3.792 0.051 9.122

Amfilia 4.001 0.073 8.142

Argentinie 4.461 0.053 9.698

Granada 3.094 0.062 7.491

Turkei 2.885 0.088 7.444

Amiflikeia 2.956 0.066 6.984

APTTB 3.782 0.052 9.730

Conaga 2.506 0.101 5.726


Hande Karaköse 66


Table 10.Flavonoid glycosides average values between varieties

Table 11.Values for flavonoids quercetin, kaempferol, luteolin and apigenin determined after

hydrolysis of total polyphenol fraction using HCl/MeOH, determined by LC-MS. n.d.

indicatesthat value was outside linear range of method140

.

Sample

no Origin Variety Harvest Year

Kaempferol

(mg/100g)

Quercetin

(mg/100g)

Luteolin

(mg/100g)

Apigenin

(mg/100g)

21 TCV 1 II 28.09.2010 82.5 75 12 4

32 Uconor 2 I 11.08.2010 n.d 78 8 3

41 Uconor 3 I 11.08.2010 70 638 9 n.d.

47 Toumpa 3 II 10.09.2010 108 463 n.d. 5

61 Amfilia 3 II 15.09.2010 90 455 n.d n.d.

87 Uconor 7 I 13.07.2011 190 737.5 14 n.d.

89 Uconor 4 I 2011 65 353 n.d. n.d.

94 Uconor 5 I 13.07.2011 88 417.5 n.d. n.d.

114 Uconor 6 II 2011 n.d 445 7 2

115 Conaga 3 I 2011 73 195 8 4

123 Agrinion 3 II 2011 73 318 n.d. n.d.

Variety

Kaempferol-7-glucoside

(g/100g leaves)

Quercetin-3-glucoside

(g/100g leaves)

Total flavones

(g/100g leaves)

1 2.860 0.073 7.224

2 2.090 0.099 5.401

3 2.955 0.062 7.412

4 3.095 0.104 6.640

5 2.904 0.076 8.083

6 2.850 0.045 7.682

7 2.164 0.155 5.432

Non-EU 3.064 0.097 7.638


Hande Karaköse 67


If individual figures for flavonoids from LC-MS data are compared to those obtained by

hydrolysis an obvious differences in values is apparent, even if the g/100 g dry leaves values are

corrected for the increased molecular weight of the flavonoids. A direct comparison must be

viewed with great care since firstly the literature hydrolysis method has been validated only for

flavonoid glucosides. However, in the previous identification section we have shown that

compounds present in stevia are hexosides but not glucosides, and can hence show a greatly

reduced hydrolysis kinetic. Secondly, the quantitative data without hydrolysis are based on the

use of reference standards that are chemically different from the compounds present in the leaf.

Again it can be anticipated that ion enhancement effects in LC-MS might contribute to higher

levels of flavonoids determined here. However, the values obtained are highly useful since they

give a detailed insight into relative variations between flavonoids between different varieties and

origins and give a guideline towards absolute and accurate values.

5.3.3.3. Principal Component Analysis (PCA)

A principle component analysis (PCA) based on the LC-MS dataset of stevia phenols was carried

out to allow differentiation between different stevia varieties and geographic origins. Figure 39a

shows a representative analysis with score and loading plots where within the scores plot every

spot corresponds to a single leave sample with clear groupings apparent allowing distinction

between varieties based on their phenolic profile. The loading plot reveals in each data point a

pair of retention time and m/z ratio, therefore defining individual compounds whose quantities

can be used as phytochemical markers for variety distinction.

A second PCA analysis was carried out with an aim to distinguish EU cultivated samples from

non-EU cultivated samples. For this purpose 20 EU and 20 non EU samples were subjected to a

full PCA analysis. Score and loading plots are shown in Figure 39b.

From the score plot it can be seen that the samples fall in three groupings. Group A from South

American samples can clearly be distinguished from all other samples based on their high

diCQA content (from score plot).

A second group B contains exclusively European samples from the Uconor cooperative. The

final group C contains both EU and non-EU samples e.g. from Turkey, Ucraine and India, which


Hande Karaköse 68


group together. The sample distinction information from the loading plot suggests that a

combination of rebaudioside A concentrations and diCQA concentrations allows distinction here.

Figure 39a.PCA analysis of phenol profile of 35 stevia leaf LC-MS datasets. Score plot is on the

left with each colour representing a different stevia variety and loading plot on the right with

each data point corresponding to individual compounds allowing differentiation.

Figure39b.PCA analysis of phenol profile of 40 stevia leaf LC-MS datasets (red points non-EU

samples, blue points EU samples). Score plot is on the left with each colour representing a

different stevia variety and loading plot on the right with each data point corresponding to

individual compounds allowing differentiation.

-1.0 0.0 1.0 PC 2

-1

0

1

PC 4

-1.0 0.0 1.0 PC 2

-1

0

1

PC 4

-0.5 0.0 0.5 1.0 PC 1

-0.4

-0.2

0.0

0.2

PC 2

-0.5 0.0 0.5 1.0 PC 1

-0.4

-0.2

0.0

0.2

PC 2

A

B

C

3,5 diCQA

Reb A


Hande Karaköse 69


5.3.4. Statistical Evaluation of Quantification Data of Polyphenols in Stevia

From the obtained data, a series of statistical analysis was carried out. For each variety, origin

and harvest average values and standard deviations were determined. Additionally, the statistical

pattern and type of statistical distribution of the data was analyzed for each subgroup. The

correlation studies were performed by Pearson's correlation, with the significance value of p <

0.05.

Significant differences among stevia leaves for each variable were assessed with analysis of

variance (ANOVA).

5.3.4.1. Statistical Spread of Data

The distribution of the dataset is an essential step for examination of data in statistical analysis.

The most important and useful distribution of data is Gaussian (normal) distribution. A normal

distribution can be easily characterized by observing its symmetrical bell shaped curve on a

histogram (Figure 40). Skewness and kurtosis values (< 1) show also that the data is normally

distributed (Table 12). The Kolmogorov-Smirnov (KS) test was also used for the analysis of data

distribution. In this test, the significance value above 0.05 means the data is normally distributed.

Each mono and di-CQAs as well as total mono and di-CQAs quantities obtained from 166 stevia

samples showed normal distribution as judged by the KS test. In contrast quantitative data for

cis-5 CQA and cis-4,5 diCQA were found to exhibit a non-Gaussian distribution.

The KS test result, mean values, standard deviations, skewness and kurtosis of the curve for each

CGA is represented in Table 12. Histogram of 5-CQA is presented as an example in Figure 40.

Figure 40.Histogram of 5-CQA, showing the normal distribution of the dataset.


Hande Karaköse 70


Table 12.Descriptive statistics of caffeoylquinic acids

Statistics

3-CQA 5-CQA 4-CQA 3,5-diCQA 4,5-diCQA Total diCQA Total monoCQA

Mean 0.295 2.488 0.124 1.201 1.237 2.435 2.907

Std. Deviation 0.144 0.852 0.426 0.512 0.485 0.937 1.009

Skewness 0.411 0.151 0.151 0.049 0.476 0.111 0.137

Kurtosis -0.217 0.365 0.372 -0.848 0.170 -0.513 0.212

Minimum 0.002 0.193 0.010 0.173 0.210 0.311 0.205

Maximum 0.719 4.986 0.249 2.476 2.609 4.575 5.608

KS test, Asymp sig. (2 tailed) 0.521 0.951 0.916 0.468 0.746 0.584 0.817

5.3.4.2. Correlations

The correlation analysis was performed to measure the degree of relation between two chosen

CGAs. There are several different correlation methods, the most commons are Pearson

(parametric) test, which is used in the case of normally distributed populations and the Spearman

(nonparametric) test, in which there is no requirement for the assumption of normality or

homogeneity of variance. The main result obtained from these tests is the correlation coefficient

(r) and it ranges from -1.0 to +1.0. The closer “r” is to -/+1, the more closely the two variables

are related.

Correlations of CGAs grouped as mono to monoCQA, diCQA to diCQA and as well as

monoCQA to diCQAs were tested according to Pearson correlation. However, correlations of cis

derivatives were tested according to Spearman correlation due to their non-Gaussian data

distribution. Overall, correlation was observed for all CGAs with each other (Table 13).

However, the strong correlation was observed between the mono-CQAs and 3,5-diCQA with

4,5-diCQA (Pearson correlation of 0.762 and 0.791 respectively) and 5-CQA with 4,5-diCQA as

well as 4-CQA with 4,5-diCQA (Pearson correlation of 0.573 for both cases). The rest of the

correlations were slightly weaker. It needs to be pointed out that, all correlation coefficients

closer to +1, indicates that increasing quantity of e.g. 5-CQA leads to quantity increase in the

4,5-diCQA. The square of the correlation coefficient (r) is the percent of the variation in one

variable which is related to the variation in the other. In the case of correlation of 3,5-diCQA

with 4,5-diCQA 62% of the variance is related (or is correlated) (Figure 41).


Hande Karaköse 71


Table 13.Correlation coefficients of mono and di-CQAs

Correlations

3-CQA 5-CQA 4-CQA 3,5-diCQA 4,5-diCQA

3-CQA

Pearson Correlation (r) 1 0.762** 0.762** 0.416** 0.514**

Sig. (2-tailed) 0.000 0.000 0.000 0.005

R2 0.581 0.581 0.173 0.264

5-CQA

Pearson Correlation (r) 0.762** 1 1.000** 0.532** 0.573**

Sig. (2-tailed) 0.000 - 0.000 0.000 0.000

R2 0.581 - 1 0.283 0.328

4-CQA

Pearson Correlation (r) 0.762** 1.000** 1 0.532** 0.572**

Sig. (2-tailed) 0.000 0.000 - 0.000 0.000

R2 0.581 1 - 0.283 0.327

3,5-diCQA

Pearson Correlation (r) 0.416** 0.532** 0.532** 1 0.791**

Sig. (2-tailed) 0.000 0.000 0.000 - 0.000

R2 0.173 0.283 0.283 - 0.625

4,5-diCQA

Pearson Correlation (r) 0.514** 0.573** 0.572** 0.791** 1

Sig. (2-tailed) 0.000 0.000 0.000 0.000 -

R2 0.264 0.328 0.327 0.625 -

** Correlation is significant at the 0.01 level (2-tailed).

b) a)


Hande Karaköse 72


Figure 41.Graph showing the correlation between a) 3,5-diCQA/4,5-diCQA b) 3-CQA/5-CQA,

c) 5-CQA/4,5-diCQA and d) 4-CQA/3,5-diCQA.

Furthermore, another study was performed on whether the formation of cis-caffeoylquinic acids

plays a role in agricultural practice and whether they would serve as a useful marker for UV

exposure of plant tissues. Agricultural parameters including climatic conditions were obtained

through records of the nearest official weather station.

To investigate the biosynthetic origin of the cis-CQA derivatives, initially the statistical pattern

and type of statistical distribution of the quantitative data was analyzed. Both 5-CQA and 3,4 di-

CQA concentration show a Gaussian distribution profile over all samples analyzed as judged by

the Kolmogorov Smirnov test (section 5.3.4.1). In contrast quantitative data for cis-5 CQA and

cis-3,4 diCQA were found to exhibit a non-Gaussian distribution.

For a large data set of quantitative data for three mono- and four di-CQA concentrations,

correlation of all chlorogenic acids isomers correlate linearly with each other pointing to a

common stimulus of biosynthetic production in the plant (Figure 41).

The correlation linearity test was applied to cis derivatives of 5-CQA and 4,5-diCQ with their

trans derivatives and there was no linear concentration dependency (Figure 42). This result

indicates towards two distinct stimuli and pathways of their biosynthetic production.

In contrast concentrations of the two cis-4,5-diCQA isomers and concentration of cis-5-CQA and

the cis-4,5 diCQA derivatives show a linear concentration dependency (Table 14 & Figure 42).

This should be interpreted as all cis derivatives sharing the same external stimulus for

c) d)


Hande Karaköse 73


production. This stimulus is different from that of trans-CQA biosynthesis. Due to that the

stimulus of cis-CQA production is direct irradiation by UV light of exposed leaves (appendix E).

Figure 42.Linear dependency of cis-5-CQA with 5-CQA and two isomers of cis-4,5-diCQAs.

Table 14.Correlation coefficients of cis-isomers according to Spearman’s rule

Correlations

Cis-5CQA Cis-4,5diCQA1 Cis-4,5diCQA2

Spearman's rho

Cis-5CQA

Correlation Coefficient 1.000 0.265** 0.183*

Sig. (2-tailed) . 0.003 0.041

N 126 126 126

Cis-4,5diCQA1

Correlation Coefficient 0.265** 1.000 0.731**

Sig. (2-tailed) 0.003 . 0.000

N 126 126 126

Cis-4,5diCQA2

Correlation Coefficient 0.183* 0.731** 1.000

Sig. (2-tailed) 0.041 0.000 .

N 126 126 126

** Correlation is significant at the 0.01 level (2-tailed).

* Correlation is significant at the 0.05 level (2-tailed).

In a dataset comprising 120 samples of Stevia rebaudiana grown in different locations, linear

correlations of trans-mono and di-CQAs with each other indicates a common stimulus in

production. However, non-linear correlation between cis and trans-CQAs and linear correlation


Hande Karaköse 74


between the cis-CQAs convincingly demonstrates that the production of cis and trans CQA

derivatives must follow two distinctly different pathways, where trans-CQAs production is not

UV dependent stimulus.

To substantiate this hypothesis, available UV irradiation data with cis-CQA concentrations were

correlated. As a first set of data, changes of cis-CQA concentrations in different harvests were

studied. For two set of samples three harvests were carried out from June to September of crops

grown at the same locations. For the last harvest in September the number of sunshine hours

affecting the plants was always considerably lower if compared to the two early harvests from

June to August. Sunshine hours were available through the weather databases of the nearest

airport weather station (www.weather.org and www.weather.online.co.uk). Representative data

are shown in Figure 43 where in two bar charts the total cis-CQA concentration is given for three

harvests in two different locations.

Figure 43.Amount of 5-CQA in mg/100g dry leaves from three harvests from location A (TCV)

and location B (Amfilikeia) during 2011.

0

50

100

150

200

250

300

350

400

Harvest 1 A Harvest 1B Harvest 2A Harvest 2B Harvest 3A Harvest 3B

mg/100g


Hande Karaköse 75


A more global picture including all available data was obtained by the log of trans/cis-5-CQA

concentrations against the number of sunshine hours in the month prior to sample collection for

six locations were in 2010 and 2011 a total of ten harvests were collected. The data plotted in

this manner display a linear relationship with the quotient trans/cis-5-CQA concentration clearly

depending on the number of sunshine hours. The more sunshine hours the plant leaves were

exposed to the smaller quotient, so the higher the relative amount of cis-5-CQA concentrations

(Figure 44).

Figure 44.log of trans/cis-5-CQA concentrations against the number of sunshine hours in the

month for a total of ten harvests from six locations.

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250 300 350 400

log C(trans)/C (cis) vs sunshine hours in month


Hande Karaköse 76


5.3.4.3. Analysis of Variance (ANOVA)

One way ANOVA was performed to study the influence of origin and, in another test, variety on

the mono and di-CQA content of stevia leaves cultivated in various regions of EU with different

soil and climatic conditions. Post-hoc test was applied to find out which origin or varieties differ,

if any.

ANOVA is a statistical analysis used for identifying factors that are influencing a given dataset.

It is a statistical technique for comparing mean values for multiple independent populations.

ANOVA analysis selects the most discriminating variables by calculating an F factor which is

proportional to the ratio of the ‘within-group’ variance to the ‘between groups’ variance. The

higher this ratio, the more the groups are significantly different from each other141

.

The One-Way ANOVA, used in this study, compares the mean of one or more groups based on

one independent factor. However, there are two assumptions to be met within the dataset. First

one is, the data should be normally distributed (the data distribution was analyzed in chapter

5.3.4.1). The second assumption is that the variances of the groups to be compared should be

similar. This can be checked by Levene’s test (test of homogeneity of variances). If the

significance value is greater than 0.05 (found in the Sig. column, e.g. Table 15) then there is

homogeneity of variances and therefore, the assumption of homogeneity of variance is met. If the

Levene's test is significant (lower than 0.05), it would mean that there is no similar variances and

the assumption of homogeneity of variance is not met, therefore it would be necessary to use an

adjusted test such as the Welch statistic (the Robust Tests of Equality of Means) within ANOVA.

In general, when setting up the analysis, it is common and advantageous to select a test for either

situation since we do not know if the assumption is met or violated.

Influence of Origin: Levene’s test was performed to assess whether the assumption of

homogeneity of variance between groups is met within the data.

Test of homogeneity results revealed that for dependent variable 3-CQA and 4,5-diCQA

Levene’s test is not significant (F8/116 = 1.783, p=0.087), (F8/116 = 1.430, p=0.191),respectively.

However, for 5-CQA, 4CQA and 3,5-diCQA Levene’s test was significant at the level of 0.05

(Table 15), which states that the variances in the different groups of origins are different (the

groups are not equal variance on the dependent variable, that is origin in this case) and therefore,


Hande Karaköse 77


a modified procedure which do not assume equality of variance (Welch statistic) was also

applied to these data along with ANOVA.

Table 15. Results of test of homogeneity of variances

Levene

Statistic

df1 df2 Sig.

cqa3 1.783 8 116 0.087

cqa5 2.058 8 116 0.046

cqa4 2.073 8 116 0.044

totalmono 2.146 8 116 0.037

diCQA35 2.018 8 116 0.050

diCQA45 1.430 8 116 0.191

totaldi 1.733 8 116 0.098

*df: degrees of freedom

ANOVA results (Table 16) did not reveal significant differences between the average CGA

content of stevia leaves of different origins. The data only hints at some differences in the

average content of stevia leaves from different origin destinations. These, however, can be

considered only marginally significant at the 5% level (p= 0.054 for e.g. 4-CQA). Moreover,

from the ANOVA output, robust test of equality of means (Welch test) is considered for the

CQAs that are not meeting the assumption of homogeneity of variance. Like the ANOVA test, if

the significance value is less than 0.05 in Welch test, then it means there are statistically

significant differences between groups.


Hande Karaköse 78


Table 16.ANOVA results for effect of origin on stevia CGA content

From the results of Welch test and One-Way ANOVA, the conclusion would be that at least two

of the group means is significantly different from the each other. Beyond this, it is necessary to

conduct a post-hoc comparison test to see exactly which pairs of groups are significantly

different. There are varieties of post hoc tests available. The most commonly used LSD test is

applied for CQAs (3CQA, 4,5-diCQA) which are meeting the assumption of homogeneity of

variance. Games-Howell post hoc test is applied for the CQAs (5CQA, 4CQA, 3,5diCQA)

violating the assumption of homogeneity of variance (equal variances not assumed).

Multiple comparison (post hoc) results obtained from LSD test (Fisher's Least Significant

Difference) revealed that the 3-CQA amounts are significantly different at the 0.05 level between

the origins of Agrinion&Granada, Granada&TCV, Granada&Uconor and Granada & non-EU

origins. 4,5-diCQAs resulted to be significantly different at the 0.05 level for origin of Granada

and the all other origins except Portugal.

Multiple comparison (post hoc) results obtained from Games-Howell test revealed that there is

no significant difference at the 0.05 level between the origins for 5-CQA and 4-CQA. However,

significant difference observed for 3,5-diCQA between the origins of Amiflikeia and Toumpa

with the significance value of 0.008.

One Way Anova

Compound F Sig. (p)

3-CQA 1.022 0.138

5-CQA 1.850 0.054

4-CQA 1.853 0.054

Totalmono-CQA 1.985 0.064

3,5-diCQA 1.767 0.106

4,5-diCQA 1.727 0.127

Total-diCQA 1.881 0.089

Robust Test of Equality of Means

Compound F Sig. (p)

3-CQA Welch 1.823 0.132

5-CQA Welch 1.588 0.183

4-CQA Welch 1.590 0.182

Totalmono Welch 1.960 0.101

3,5-diCQA Welch 3.135 0.017

4,5-diCQA Welch 2.666 0.034

Total-diCQA Welch 3.120 0.017


Hande Karaköse 79


Influence of Variety: Test of homogeneity results revealed that for dependent variable 3-CQA

(F7/117 = 1.207, p=0.304), 3,5-diCQA (F7/117 = 0.545, p=0.799) and 4,5-diCQA (F7/117 = 1.193,

p=0.312) Levene’s test was not significant. However, for 5-CQA and 4-CQA Levene’s test was

significant at the level of 0.05 (p=0.018) (Table 17). Same as in the analysis of origins, Welch

test was also applied to data for the cases of violation of homogeneity of variances.

Table 17. Test of homogeneity of variances

Levene Statistic df1 df2 Sig.

cqa3 1.207 7 117 0.304

cqa5 2.542 7 117 0.018

cqa4 2.549 7 117 0.018

totalmono 2.242 7 117 0.036

diCQA35 0.545 7 117 0.799

diCQA45 1.193 7 117 0.312

totaldi 1.090 7 117 0.374

*df:degrees of freedom

ANOVA results did not reveal significant differences between the average CGA content of

stevia leaves of different varieties. Moreover, Welch test did not reveal any significant difference

at the level of 0.05. The results of ANOVA test and Welch test is presented in Table 18.

Table 18.ANOVA results for effect of variety on stevia CGA content

One Way ANOVA

Compound F Sig. (p)

3-CQA 0.849 0.549

5-CQA 0.459 0.862

4-CQA 0.464 0.859

Totalmono-CQA 0.463 0.860

3,5-diCQA 1.901 0.075

4,5-diCQA 1.446 0.194

Total-diCQA 1.577 0.149

Robust Test of Equality of Means

Compound F Sig. (p)

3-CQA Welch 0.926 0.500

5-CQA Welch 0.428 0.878

4-CQA Welch 0.433 0.874

Totalmono Welch 0.373 0.911

3,5-diCQA Welch 1.637 0.161

4,5-diCQA Welch 1.250 0.305

Total-diCQA Welch 1.402 0.238


Hande Karaköse 80


Post-hoc tests with Games-Howell method resulted with no pair-wise differences for 5-CQA and

4-CQA for each variety. However for 3-CQA with the LSD test, the mean difference between

variety 1-5 and variety 3-5 was significant at the 0.05 level (p=0.04). The varieties 4-6, 5-7 and

6-7 had significant difference (p=0.04) when 4,5-diCQA quantity in stevia leaves was taken as a

dependent variable. Variety 6 showed significant difference with varieties 2, 3, 4 and 7 for 3,5-

diCQA quantity in stevia leaves.

5.4. Conclusion

Mono-CQAs and di-CQAs including their corresponding cis-isomers and flavonoid glycoside of

120 samples (after removal of outliers and duplicates) of Stevia rebaudiana grown in different

locations and with different botanical varieties were profiled and quantified successfully. For the

first time a full quantitative data set from a large number of samples from different origins and

varieties, in which the full profile of all secondary metabolite quantities was determined in any

agricultural plant was obtained.

With the obtained dataset, differences between stevia leaves and influence of the origin of

cultivation and the botanical varieties on the CQAs profile were studied successfully with the

most common used statistical method ANOVA. Pair-wise comparisons of varieties and origins

for each CQA were achieved by less conservative statistical post hoc test (LSD) and Games-

Howell post hoc test to determine exact pair of variety/origin that are differentiating from each

other.

Correlation studies showed that the production of cis and trans CQA derivatives must follow two

distinctly different pathways and the stimulus of cis-CQA production is direct irradiation by UV

light exposed leaves. Therefore, cis-CQA concentrations may serve as useful markers of UV

exposure of plant material in agricultural practices.

Finally, with the PCA, it was possible to differentiate the EU and non-EU cultivated stevia

leaves according to their rebaudioside A and diCQAs profile.


Hande Karaköse 81


6. LIPID ANALYSIS of STEVIA

6.1. Overview

Lipid profile of stevia leaves was determined for 46 chosen samples by GC-MS. Total lipid

amounts of each sample were determined gravimetrically after chloroform extraction.

Identification was achieved for the fatty acids by comparison of retention times and mass spectra

with the commercially obtained fatty acid methyl esters (FAME) reference mixture. This study

was complemented by a MALDI-TOF-MS anaylsis for further lipid identification. Additionally,

a steam distillation extract was subjected to GC-MS analysis for the analysis of volatile terpenes.

Identification of the terpenes was achieved by software based NIST library search.


6.2.1. Extraction

Two grams of stevia leaves was immersed in liquid nitrogen, ground in a hammer mill, and

extracted with 150 mL of chloroform in a Soxhlet apparatus (Buchi B-811 extraction system) for

2 h. The solvent was removed in vacuo and total lipid amount was determined gravimetrically.

6.2.2. Methyl Ester Formation

Total lipid extract was dissolved in 2 mL of chloroform after gravimetric determination. 200 μL

of methanolic potassium hydroxide solution (2 mol/L) and 1 g of sodium hydrogen sulfate

monohydrate (NaHSO4) were added to 1 mL of lipid extract solution in chloroform to form the

methy esters of lipids for GC analysis.

6.2.3. GC-FID

GC analyses were performed on GC-2010 (Shimadzu, Kyoto, Japan) equipment with flame

ionization detector and split/splitless injector. Injector temperature was at 290 °C and samples

were injected using autosampler (1 µL) with split ratio of 1:10. Capillary columns was used Rxi -

5 ms (15 m × 0.25 mm, with film thickness of 0.25 µm) Restek. The temperature program was

raised from 80 °C (1min) up to 300 °C at rate 5°C/min, and the total run time was 50 mins.

Helium was used as carrier gas at flow rate of 5 mL/min. Detector temperature was set at 310 °C.

To form the flame, hydrogen gas flow, 40 mL/min, and air gas flow, 400 mL/min, were used.

GC solution 2.10 software was used for data collection, and calculation of all parameters.


Hande Karaköse 82


6.2.4. GC-MS

GC-MS analyses have been carried out with a Varian CP-3800 gas chromatograph (Palo Alto,

CA, USA) equipped with a split/splitless injector and coupled to a Saturn 2000 ion-trap Varian

mass spectrometer (MS). Data acquisition was performed using a Star Toolbar system (Varian).

Samples were injected manually with split ratio of 1:10 at 290 °C. The compounds were

analyzed on a 30 m, 0.25 mm I.D. fused-silica capillary column coated with a 0.1µm layer of

poly (5% phenyl/95% dimethylpolisiloxane) (Rxi-5Sil MS, Restek) using helium as the carrier

gas at flow rate of 1.3 mL/min, respectively. The oven temperature was heated from 80 °C (1

min) to 300 °C at the rate 5 °C/min and the total run time was 50 min. For the MS, the electron

multiplier was set to 1350 V and ionization was accomplished by electron impact (EI). The

transfer line temperature was set at 300 °C whilst 244 °C and 120 °C were the temperatures used

for the trap and the manifold, respectively. Mass spectra were recorded from m/z 40–600.

6.2.5. Calibration Curve of FAME

The quantitative analyses have been initiated by generation of calibration curves for FAME

standard mixture from the range of C14:0/C14:1 – C24:0/C24:1. Calibration curves were

generated by plotting peaks areas of FAME standard mixture (Marine Oil FAME Mix (20

components) from Restek) at different concentrations as function of peak areas. In doing so,

FAME standard solution was diluted with n-heptane to a series of standards with concentrations

of 5, 50, 100, 250, 500, 750, 1000, 1250, 1500, 2000, 3000 μg/L.

6.2.6. MALDI-TOF MS

As matrix solution 5g/L 2,5-dihydroxybenzoic acid (2,5-DHB) solution in acetonitrile containing

0.1% trifluoroacetic acid (TFA) was used due to robustness of DHB to impurities. 1 μl aliquot of

the organic extracts of stevia was mixed with 1 μl of the matrix solution on the maldi target

(MPT AnchorChip 600-384 target, Bruker Daltonics) and the matrix crystals were allowed to air-

dry.

MALDI-TOF spectra were acquired on an Autoflex II MALDI-TOF-TOF mass spectrometer

(Bruker Daltonics) equipped with a 337 nm nitrogen laser. The instrument was operated in the

reflector mode: source, 19.00 kV; lens, 8.95 kV; and reflector, 20 kV, using an optimized ion

extraction delay time of 80 ns. The laser frequency was set at 25 Hz with 50 laser shots per


Hande Karaköse 83


acquisition. The laser strength was kept about 40% above threshold to obtain optimum signal to

noise ratio. Spectra were obtained by summing, on average, 200 laser shots. Spectra were

acquired in the mass range 0–2500 amu. The instrument was externally calibrated in the

enhanced quadratic calibration mode prior to acquisition using a peptide tune-mix sample

(Bruker Daltonics).


46 Stevia rebaudiana samples were chosen for lipid analysis. Firstly, total lipids were

determined gravimetrically after extraction with a non-polar solvent. Since hexane extraction

provided values below 1 weight %, a Soxhlet chloroform extraction was selected for this

purpose. The total lipid sample was in a following step subjected to GC-MS analysis. Table 19

shows total lipid values obtained for 46 representative samples comprising at least three samples

from all seven varieties and samples from all origins. Secondly the total lipid fraction was

subjected to basic hydrolysis and derivatisation followed by GC-MS analysis to identify the

individual fatty acid spectrum of all seven stevia varieties and of representative samples from all

origins. Derivatisation chosen included the formation of fatty acid methylesters. Identification of

fatty acid methylesters was achieved through GC-MS by comparison of retention time and mass

spectra with a commercial certified reference compound mixture. A representative GC

chromatogram is shown in Figure 45 for total lipid profile of stevia after chloroform extraction

and in Figure 46, the GC chromatogram of FAME standard mixture can be observed142, 143.

Figure 47 presents the fragmentation of fatty acid by electron impact ionization obtained in GC-

MS measurement of stevia leaves chloroform extract.

Additionally a steam distillation extract was obtained and subjected to GC-MS analysis to profile

the volatile terpenes and to compare the lipid composition with chloroform extract.


Hande Karaköse 84


Figure 45.GC.MS chromatogram of total lipid extracts from Stevia rebaidiana leaves from

sample (Uconor, Var.4).

Figure 46.GC-MS chromatogram of FAME standard mixture.

C14:0

C14:1

C16:0

C16:1

C18:1 C18:1 C18:2 C18:3

C18:0

C20:0 C20:1 C20:2

C20:4 C20:3

C20:5

C22:0

C22:1

C22:6


Hande Karaköse 85


Figure 47.Representative EI-MS spectra obtained from GC-MS measurement of stevia extract.


Hande Karaköse 86


Table 19.Total lipid values in weight % from 46 samples

Sample no Origin Variety Harvesting Year %

119 Agrinion 4 II 2011 3.42

120 conaga 6 I 2011 7.71

121 Conaga 7 I 2011 4.19

122 Agrinion 6 II 2011 5.14

123 Agrinion 3 II 2011 5.45

124 Toumpa 7 I 2011 3.05

125 Toumpa 4 III 2011 4.43

126 Toumpa 3 III 2011 6.67

127 Toumpa 3 II 2011 5.59

128 Agrinion 4 II 2011 0.50

129 Toumpa 4 II 2011 4.32

132 Agrinion 3 II 2011 5.92

133 Toumpa 5 II 2011 4.26

134 Toumpa 6 II 2011 4.15

135 Toumpa 7 II 2011 3.92

136 TCV 5 I 30.06.2011 1.68

137 TCV 3 I 30.06.2011 5.48

138 TCV 6 I 30.06.2011 5.16

139 TCV 3 II 11.08.2011 7.23

140 TCV 6 II 17.08.2011 4.90

141 TCV 4 II 24.08.2011 4.09

142 TCV 7 I 18.08.2011 2.11

143 TCV 4 I 07.07.2011 3.42

144 TCV 5 II 17.08.2011 8.23

147 Amiflikeia 5 I 2011 3.25

148 Amiflikeia 6 I 2011 4.23

149 Amiflikeia 1 I 2011 4.27

154 Toumpa 5 I 2011 3.24

8 TCV 4 I 04.08.2010 1.52

49 Toumpa 3 I 30.07.2010 3.37

43 Agrinion 2 II 20.09.2010 2.63

75 Granada 3 I 09.09.2010 3.41

54 Amfilia 4 II 15.09.2010 2.71

30 Uconor 4 II 07.05.2010 0.54

53 Amfilia 4 I 04.08.2010 2.99

5 TCV 1 I 04.08.2010 0.89

16 TCV 2 I 04.08.2010 0.98

19 TCV 3 II 10.09.2010 6.74

6 TCV 2 I 04.08.2010 5.11

10 TCV 2 II 28.09.2010 7.05

14 TCV 2 I 04.08.2010 2.27

32 Uconor 2 I 11.08.2010 3.14

43 Agrinion 2 II 20.09.2010 3.69

37 Uconor 4 I 11.08.2010 2.60

34 Uconor 4 II 07.09.2010 5.13

36 Uconor 4 I 11.08.2010 1.46


Hande Karaköse 87


For determination of the fatty acid profile GC-FID was applied using the GC-MS data as a

reference point. Oleic acid was found to be the predominant fatty acid in stevia with an average

ratio of saturated to unsaturated fatty acids of 1:3.8. Among the saturated fatty acids 16:0

palmitic acid was the predominant acid with varying quantities of stearic acid present (24 - 2 %).

Saturated fatty acids with shorter and longer chain length were not identified. In terms of

unsaturated fatty acids 16:1 was identified next to three co-eluting isomers of 18:1, with oleic

acid being the predominant compound. Values stated for 18:1 acids represent a sum over all

three isomers. A series of polyunsaturated fatty acids were identified as well comprising 16:2,

18:3 and 18:2 acids with variable quantities. Values are given in Table 20.

Table 20.Quantities of polyunsaturated fatty acids

Sample Origin Variety Harvest Year % 16:2 %16:1 %16:0 %18:3 % 18:2 % 18:1 % 18:0

S10 TCV 2 II 28.09.2010 15.45 9.97 12.04 15.56 10.25 35.09 1.64

S6 TCV 2 I 04.08.2010 20.41 20.52 11.42 2.51 7.57 34.59 2.98

S14 TCV 2 I 04.08.2010 22.03 14.73 9.50 13.49 6.87 30.39 2.99

S34 Uconor 4 II 07.09.2010 9.29 1.71 12.26 6.00 11.29 52.06 7.39

S37 Uconor 4 I 11.08.2010 3.54 2.10 14.70 8.41 12.44 55.97 2.84

S5 TCV 1 I 04.08.2010 2.09 0.72 14.59 4.70 13.04 57.64 7.23

S10 TCV 2 II 28.09.2010 14.56 9.02 13.01 11.90 10.32 36.11 5.09

S16 TCV 2 I 04.08.2010 4.00 0.51 14.56 5.50 12.61 55.03 7.78

S19 TCV 3 II 10.09.2010 1.21 0.57 12.57 8.35 12.77 48.69 15.83

S32 Uconor 2 I 11.08.2010 2.18 1.11 14.86 6.80 11.80 53.53 9.71

S36 Uconor 4 I 11.08.2010 5.15 3.38 14.18 9.03 13.75 46.55 7.96

S43 Agrinion 2 II 20.09.2010 1.49 0.39 11.12 10.81 10.28 49.80 16.11

S45 Portugal 1 I 07.07.2010 1.42 0.68 10.35 11.98 10.40 40.32 24.86

S121 Conaga 7 I 2011 7.91 4.77 14.20 6.00 13.59 47.71 5.82

S122 Agrinion 6 II 2011 8.44 5.68 15.47 4.10 9.43 52.31 4.57

S123 Agrinion 3 II 2011 10.69 6.36 13.41 8.63 9.44 48.57 2.90

S124 Toumpa 7 I 2011 0.93 0.46 21.06 5.77 16.42 54.17 1.20

S127 Toumpa 3 II 2011 9.05 5.67 16.23 6.39 13.06 43.56 6.03

S136 TCV 5 I 30.06.2011 1.11 0.61 14.03 6.84 10.42 52.11 14.88

S143 TCV 4 I 07.07.2011 8.05 5.38 14.04 4.43 11.29 50.25 6.56

S147 Amiflikeia 5 I 2011 8.27 5.48 15.26 8.71 8.41 45.93 7.94

S148 Amiflikeia 6 I 2011 10.93 7.22 15.29 7.20 9.95 42.27 7.13

S149 Amiflikeia 1 I 2011 7.26 4.38 16.83 4.95 10.70 49.61 6.27

S154 Toumpa 5 I 2011 0.63 0.39 15.93 4.80 11.48 60.90 5.86

S120 Conaga 6 I 2011 1.71 0.55 14.79 5.37 11.58 58.93 7.07

S142 TCV 7 I 18.08.2011 2.23 1.31 17.64 6.13 15.83 51.55 5.30

Average 6.93 4.37 14.21 7.48 11.35 48.22 7.46


Hande Karaköse 88


It should be noted that in plants fatty acid composition is reported to be a consequence of

climatic conditions, in particular temperature, rather than botanical variation. The total lipid

content determined is well in line with values from other leafy dietary plants. For example values

for green tea has been determined as 3-5% total lipids, spinach for 4.5 % total lipids and the

botanically related Asteraceae plant lettuce at 4-6 % total lipids. The average fatty acid

distribution and their structures are shown in Figure 48 and Figure 49.

Figure 48.Structures of fatty acids in Stevia rebaudiana extract.

OH

O

16:0 Palmitic acid (hexadecanoic acid)

OH

O

16:1 Palmitoleic acid (hexadec-9-enoic acid)

HO

O

16:2 9,12-Hexadecadienoic acid

OH

O

18:0 Stearic acid (octadecanoic acid)

OH

O

18:1 Oleic acid (octadec-9-enoic acid)

HO

18:2 Oleic acid (9,12 - octadecadienoic acid)

O

OHO

18:3 Linoleic acid (9,12,15 - octadecatrienoic acid)


Hande Karaköse 89


Figure 49.Fatty acid profile of average stevia leaf in %. X:Y denominates the number of carbon

atoms in the fatty acid (X) and the number of double bonds in the fatty acid (Y). Data were

obtained from 46 samples after lipid hydrolysis, derivatisation and GC-MS analysis

Average fatty acid and lipid values determined were as follows:

• Total lipids: 4.52 % (0.89-8.23 %)

• Ratio saturated/unsaturated fatty acids: 1:3.8 (1:6-1:1.8)

This work was complemented by a MALDI-TOF-MS anaylsis of the lipid fraction allowing

identification of selected intact lipids, mainly triacylglycerides of oleic acid. A representative

MALDI-MS spectrum is shown in Figure 50.

0

10

20

30

40

50

60

% 16:2 %16:1 %16:0 %18:3 % 18:2 % 18:1 % 18:0 % X:Y


Hande Karaköse 90


Figure 50.MALDI-MS spectrum of total lipid extract in positive ion mode using 2,5-DHB as a

matrix. Main peak at m/z 872.3 corresponding to trioleic acid glyceride.

Finally direct GC-MS analysis of the lipid fraction was carried out supplemented by a GC-MS

analysis of a steam distillation extract of Stevia rebaudiana leaves. This analysis allowed

positive identification of twelve mono- and sesquiterpenes using a NIST library search. The

NIST library search identification of compounds was made if the database spectrum showed a

match with the experimental spectrum with a NIST score of 800 or above. Additionally a library

A database search was carried out and compounds accepted if a library A score of above 25 %

and a NIST score of above 800 was observed. A NIST score of above 800 is generally accepted

in the literature for positive compound identification. All compounds identified were additionally

present in both steam distillation extract and total lipid fraction. Selected structures identified are

shown below in Figure 51 and retention times with NIST score for few terpenes are presented in

Table 21.

It should be noted that previously an additional 25 volatile terpenes have been reported in Stevia

rebaudiana leaves. These compounds are shown in the appendix F, but their presence could not

be confirmed in this study144

.

229.4

326.7

361.5

478.8

533.8

593.9

872.3

0_O20\1: +MS

0

20

40

60

80

100

Intens.

[%]

200 400 600 800 1000 1200 1400 1600 1800 2000 m/z

H2C O (CH2)8 CH CH (CH2)7 CH3

HC O CO (CH2)7 CH CH (CH2)7 CH3

H2C O CO (CH2)7 CH CH (CH2)7 CH3

Ether type lipid: 1,2-dioleoyl-3-oleylglycerol (C57H106O5)


Hande Karaköse 91


O

myrtenal alpha pinenesabinene terpinene

O

cumin aldehyde

O

1,8 cineol

3-carene copaene humulene (caryophyllene)

gamma cadinene

-pinenelimonene

Figure 51.Chemical structures of terpenes identified in Stevia rebaudiana leaves.

Table 21.Retention time and NIST scores of some terpenes identified in stevia extract

Terpene RT (min) NIST score m/z

sabinene 9.5 838 136

α-pinene 10.4 920 136

3-carene 10.3 925 136

caryophyllene 10.5 727 204

γ-cadinene 11.2 887 204

copaene 12.8 848 204

In addition to the mono- and sesquiterpenes the GC-MS data show a series of triterpenes at

longer retention times. NIST search clearly indicates their identity as triterpenes with low NIST

scores for steroid and related structures, however, no match in the database allowed positive

compound identification. None of the terpenes identified was reported to show a toxicologically

problematic profile. Compounds have been reported in many additional dietary plants. General

volatile terpene levels in the plant are very low. The total FID or TIC (total ion current in MS


Hande Karaköse 92


data) are around 3-5 % of the total intensity of the lipid fraction, which allows an estimate 1.5 -

2.5 × 10-3

weight % of total volatile lipids.

Additionally the GC-MS data revealed the presence of giberillic acid (structure below) a well

known plant hormone in the lipid fraction. To confirm its existence, methylester formation of

reference compound of steviol (having the closest chemical structure) was performed and

subjected to GC-MS analyses (Figure 52). The retention times and mass spectral data were

compared between stevia extract and the steviol standard. Furthermore, increase in the intensity

was observed after standard steviol addition to the stevia extract.

Figure 52.GC chromatogram of methylesterified steviol and stevia extract.

O

OC

COOHHO

OH

giberillic acid

OR

CH3

RO2C CH3

steviol


Hande Karaköse 93


Further secondary metabolite search

The groups of Nash and Fleet recently reported on the isolation of novel indolizidine alkaloid

they named steviamine from Stevia rebaudiana leaves.145

Within the LC-MS data available ions

corresponding to its m/z ratio was searched, however, an ion corresponding to this mass at a

significant level was not found in any sample analyzed in neither positive nor negative ion mode

(above S/N 20). The absence of the ion can be confirmed later on in future studies with an

authentic reference samples requested from the Fleet group.

N

OHHO

HO

Steviamine (not found)

6.4. Conclusion

Lipid profile and quantification data for chosen stevia samples were successfully obtained.

Steam distillation and chloroform extracts of stevia leaves analyzed on GC-MS and GC-FID did

not have significant differences in lipid profile. Furthermore, volatile terpenes were identified by

NIST library search.


Hande Karaköse 94


7. PROTEOMICS of STEVIA

7.1. Overview

The main aim of stevia leave protein analysis was to purify, separate and sequence stevia leaf

proteins with an aim to identify potentially allergenic proteins. Isolation and purification of

stevia leaf proteins were achieved using 2D gel electrophoresis. The resulting 2D gels were

stained with coumassie blue staining to visualise and identify individual proteins. Susbsequent

trypsin digestion and MALDI-TOF MS analysis was carried out on selected proteins. For

MALDI-MS sample preparation anchor chip MALDI targets were used in conjunction with 2,5-

DHB (2,5-dihydroxy benzoic acid) matrices. Using data base search algorithms the obtained

mass spectrometrical data were used in attempt to sequence the proteins.


7.2.1. Extraction of Proteins

Fresh plant tissues were crushed in pestle and mortar in presence of liquid nitrogen in order

to prevent the degradation of protein by the release of protease enzyme. The leaves were

crushed in fine powder, and to precipitate the proteins, the powder was suspended in TCA

extraction buffer at -20 oC for overnight (Table 22). The proteins precipitated as white flakes

after the overnight incubation. The supernatant is collected and centrifuged at 5000g for 30

min. The supernatant was discarded, and the protein pellets settled at bottom was washed

with ice cold acetone and centrifuged again at 5000g for 15 min. The washing step was

repeated for 3 to 4 times and then this protein pellet was dried by passing nitrogen gas at

slow stream, after the drying process these pellets can be stored at -80°C for few months for

later identification processes (Figure 53).116

Table22. Amount and properties of chosen stevia leaves

Sample

Number Harvest Variety Origin

Tissue

(g)

Extraction buffer

(mL/g)

Approx.protein weighed

after extraction (mg)

8 I 4 TCV 10 20 20

10 I 2 TCV 10 20 15

36 I 4 Uconor 10 20 12

7 II 7 TCV 10 20 15


Hande Karaköse 95


Figure 53.Extraction procedure of proteins. Red arrow shows the process of TCA extraction and

Blue arrow follows the phenol extraction116

.

7.2.2. Protein Analysis

SDS-PAGE Protocol: The protein pellet (200μg) was resuspended in IEF buffer/sample buffer

(125 μL for 7 cm strip). 100 μL of IEF buffer with the protein sample was vortexed for 10 min at

10000 rpm and the supernatant was collected. The remaining volume of 25 μL of IEF buffer was

added on the remaining pellet and vortexed for 20 mins and centrifuged at 10000 rpm. The

supernatants were collected and added to the previous collected supernatant.

The collected supernatant (IEF buffer with protein sample) was added in the rehydration tray

uniformly and then the IPG strip (strip for isoelectric focusing, ph range was 3 - 10 and pH 4 - 7)

was placed in the rehydration tray and it was kept overnight for sample absorption in to the strip.


Hande Karaköse 96


For running the IPG strip, the voltage gradient of 250 volt for 30 min, and then to 3500 volts for

4 hours was used.

Equilibration buffer I:

6M urea, 0.375M Tris-HCl pH8.8, 2% SDS, 20% Glycerol, 2%(w/v) DTT.

Equilibration buffer II :

6M urea, 0.375M Tris-HCl pH8.8, 2% SDS, 20% Glycerol, 2.5%(w/v) iodoacetamide.

Sample buffer:

CHAPS (Roche), Pharmalyte (pH 3-10) (GE Healthcare Lifesciences), dithiothreitol (DTT)

(Sigma-Aldrich), Serdolit MB-1 (Serva), urea, Pefabloc® (VWR), Thiourea (Fluka)146

Rehydration buffer was purchased from Biorad ReadyPrep™ 2D starter kit.

2D SDS-PAGE Protocol: The gel casting chamber was filled from the bottom to a height of

about 2 cm below the top of the glass plates with separation gel and stacking gel (Table 23) on

the top. The gels were carefully overlaid with 1.0-1.5 mL buffer-saturated 2-butanol to allow for

complete polymerization. In order to ensure good contact between the IPG strip and the gel an

agarose solution was added. The agarose solution was kept at 70 °C and added first on top of the

gel. Immediately after, the equilibrated strip was placed on the gel. The IPG strip gel was then

subjected to electric field at 121volt and 45Amp for 1 to 2 hours for separation of proteins

according to their molecular weight. The gel was stained overnight in Coomassie brilliant blue

and later on destained with dd H2O. The protein bands were excised for destaining and trypsin

digestion147

.

Table 23.Preparation of separation and stacking gel for 2D SDS PAGE

A. 12.5 % Separating Gel B. 4% Stacking gel

Water 3.3 ml

30% Acrylamide 4.2 ml

1.5M Tris (pH 8.8) 2.5 ml

10% SDS 100 μl

10% APS 50 μl

TEMED 5.0 μl

Water 6.1 ml

30% Acrylamide 1.3 ml

0.5M Tris (pH 6.8) 2.5 ml

10% SDS 100 μl

10% APS 50 μl

TEMED 10 μl


Hande Karaköse 97


Destaining gel pieces excised from Coomassie stained 2D SDS gel: The protein bands were

excised from 2D-SDS and the each excised band was cut in to smaller pieces. The gel pieces

were transferred to a microcentrifuge tube and 100 μL of 100 mM ammonium

bicarbonate/acetonitrile (1:1, v/v) was added, incubated with occasional vortexing for 30 min.

500 μL of neat acetonitrile was added and incubated at room temperature with occasional

vortexing, after gel pieces become white and shrink, the acetonitrile was removed. The destained

gel pieces were then subjected to in-gel digestion. Alternatively, they can be stored at -20 0C for

a few weeks until they needed.

Trypsin Digestion: 50 μL of trypsin buffer (Promega gold mass spectrometry grade, Germany)

was added on the destained gel pieces and it was left in an ice bucket for 30 min, and more

trypsin buffer was added if all solution was absorbed. After 90 mins, 10 μL of ammonium

bicarbonate buffer was added to cover the gel pieces and to keep them wet during the enzymatic

cleavage. The gel pieces were kept in an air circulation thermostat for incubation overnight at 37

0C. The tubes were chilled to room temperature, and 1μL aliquot of supernatant was directly used

for MALDI-TOF MS analysis. For further analysis 10 μL of 0.1 % (v/v) TFA was added in to

the tube, vortexed and centrifuged at 10,000 rpm, aliquot was withdrawn and dried down in a

vacuum centrifuge and stored at -20 0C till it was needed for further MS/MS analysis.

7.2.3. MALDI-TOF MS

MALDI-TOF MS analysis was carried out on selected trypsin digested proteins. For MALDI-

MS sample preparation anchor chip MALDI targets (MPT AnchorChip 600-384 target, Bruker

Daltonics.) were used in conjunction with DHB (dihydroxy benzoic acid) matrices (Sigma

Aldrich). Using data base search algorithms the obtained mass spectrometrical data were used in

attempt to sequence the proteins.

As matrix solution 5g/L 2,5-dihydroxybenzoic acid (DHB) solution in acetonitrile containing

0.1% trifluoroacetic acid (TFA) was used since DHB is more robust to impurities. 1 μl aliquot of

the organic extracts of stevia was mixed with 1 μl of the matrix solution on the maldi target

(MPT AnchorChip 600-384 target, Bruker Daltonics) and the matrix crystals were allowed to air-

dry.


Hande Karaköse 98


MALDI-TOF spectra were acquired on an Autoflex II MALDI-TOF-TOF mass spectrometer

(Bruker Daltonics) equipped with a 337 nm nitrogen laser. The instrument was operated in the

reflector mode: source, 19.00 kV; lens, 8.95 kV; and reflector, 20 kV, using an optimized ion

extraction delay time of 80 ns. The laser frequency was set at 25 Hz with 50 laser shots per

acquisition. The laser strength was kept about 40% above threshold to obtain optimum signal to

noise ratio. Spectra were obtained by summing, on average, 200 laser shots. Spectra were

acquired in the mass range 0–7500 amu. The instrument was externally calibrated in the

enhanced quadratic calibration mode prior to acquisition using a peptide tune-mix sample

(Bruker Daltonics).

7.3. Results and Discussion

7.3.1. SDS Results

SDS gel was performed to separate the proteins according to their molecular weight, but SDS gel

resulted with only one big band which corresponds to molecular weight of 55 kDa (Figure 54)

and other bands appeared as a light background, which can be due to insufficient separation or

due to low concentration of proteins with low range molecular weight. The bands on SDS gel

was subjected to trypsin digestion and analyzed by MALDI-TOF. The results from MALDI were

not sufficient for characterization of stevia proteins, thus, 2D-SDS separation of proteins were

performed.

There was no significant difference observed in the protein profile of four chosen stevia samples,

judged by their SDS gel, 2D gel bands and MALDI-TOF analysis.

In addition, only for one stevia sample phenol extraction method was tested and SDS gel bands

were compared with TCA/acetone extraction. Essentially, there was no benefits obtained from

phenol extraction, though the procedure was more time consuming.


Hande Karaköse 99


Figure 54.SDS Gel, for sample number 8 TCV harvest I, loaded on gel at different

concentrations 1mg/mL and 0.5mg/mL.

7.3.2. 2D-SDS

Proteins extracted by TCA/acetone method was separated by 2D SDS−PAGE are shown in

Figure 55. Essentially, the same results were obtained when the samples were prepared by

phenol precipitation (data not shown). Resolved protein spots were more concentrated at 55kDa

band, thus, more interest was given on these protein spots for further analysis by MALDI-TOF.

The selected spots were cut and treated with trypsin digestion separately and subjected to

MALDI-TOF analysis.


Hande Karaköse 100


Figure 55.2D-SDS separation of stevia total protein extract. 7cm strip of pH 4-7, where spot 1

and 2 are at 55kDa, and spot 5 at ~15kDa

7.3.3. MALDI-TOF MS Results

Proteins isolated by 2-D electrophoresis were digested by trypsin, and the resulting peptides were

mass analyzed by MALDI-TOF. The masses obtained from the spectral data were compared with

expected values computed from sequence database entries according to the enzyme's cleavage

specificity. The results were scored, and the ranking suggests the protein being identified or not.

Enzyme cleavage specificity, number of detected cleavage peptides, and mass accuracy are the

critical parameters148

. Moreover, peptide mass tolerance was set to 50 ppm; the mode of

proteolytic digestion was chosen as ‘trypsin digestion’ the searching database used was NCBI

and the searching taxonomy was “other green plants”. These parameters play a very crucial role

in MALDI –TOF for an exact surveillance of related protein sequence and reducing the false

positive results. The list of peptide masses were transferred into the peptide mass fingerprint

(PMF) search program Mascot.

The result of a peptide mass search with Mascot contains a lot of information. First, the

probability based score is very important. A protein is identified with a score higher than 100.

Second, the full protein summary report should be considered. By clicking onto the accession

number of the first hit more detailed protein information is displayed. The nominal mass must be

in accordance with the experimental data obtained from the gel electrophoresis. If this is not the

case, protein fragments or adducts should be considered. Furthermore, the sequence coverage

70kD

A 55kD

A

40kD

A

35kD

A

25kD

A

1

5

2

3

4

6

9

10

00

9

11

00

9

8


Hande Karaköse 101


(SC) and the number of mass values matched (MM) are very important. The difference between

the number of mass values searched and the number of mass values matched should be as small

as possible149

.

The results in Figure 56, shows that the PMF spectra generated has a close proximity to

ribulose-1,5-bisphosphate carboxylase (RuBisCO) enzyme150

, the result has the score of 232

with the expect value of 5.3E-18 at 50 ppm, and the sequence coverage is 45% (Figure 56 & 57).

The sequence which appears in bold black represents the matched peaks. Each matched peak

(m/z) defines a particular type of amino acid sequence, which is identified by database search.

The MS/MS fragmentation was processed for the unmatched and matched peaks with respect to

RuBisCO enzyme (Figure 58).


Hande Karaköse 102


Protein sequence coverage: 45%

Matched peptides shown in bold black.

1 KDYKLTYYTP EYETKDTDIL AAFRVTPQPG VPPEEAGAAV AAESSTGTWT

51 TVWTDGLTSL DRYKGRCYGI EPVPGEDNQY IAYVAYPLDL FEEGSVTNMF

101 TSIVGNVFGF KALRALRLED LRIPTAYVKT FDGPPHGIQV ERDKLNKYGR

151 PLLGCTIKPK LGLSAKNYGR ACYECLRGGL DFTKDDENVN SQPFMRWRDR

201 FLFCAEAIYK AQAETGEIKG HYLNATAGTC EDMMKRAVFA RELGVPIVMH

251 DYLTGGFTAN TSLAHYCRDN GLLLHIHRAM HAVIDRQKNH GMHFRVLAKA

301 LRMSGGDHIH SGTVVGKLEG EREITLGFVD LLRDDFIETD RSRGIYFTQD

351 WVSLPGVLPV ASGGIHVWHM PALTEIFGDD SVLQFGGGTL GHPWGNAPGA

401 VANRVALEAC VQARNEGRDL ATEGNEIIRE ATKWSPELAA ACEVWKEIKF

451 EFQAMDTLDG DKDKDKKR

Figure 56.MALDI-TOF MS spectra and mass list of trypsin digested 2D-SDS spot (spot number

8) and Mascot search result showing the sequence information for RuBisCO enzyme with the

score of 45%. The sequence which appears in bold black represents the matched peaks (known in

the database and matched with experimental data) and sequences which are in black represents

the unmatched peaks (unknown peaks).


Hande Karaköse 103


Figure 57.Mascot search result of MALDI spectra, the score is 232 and expectation rate is 5.3e-

18; this data is generated by using Matrix science which acts as search engine. NCBI database

and other green plant taxonomy was selected, the peptide mass tolerance was kept at 50 ppm and

two partials. The gene bank accession number is “gi| 38146633|”.

MALDI TOF peptide mapping of the stevia protein yielded partial identification of the sequence

from MASCOT database search. To prove the sequence obtained from the database and to have

more information on the unknown peaks MS/MS fragmentation of chosen peaks with significant

intensity was performed.

The MS/MS fragmentation of one of the unmatched peak m/z 842 was performed and the peptide

was tried to be sequenced by de-novo sequencing and database search. The de-novo sequence for

this peptide resulted with the sequence of AVAETVPR, however when this sequence was

subjected for MASCOT search, the result was 100% sequence coverage for a hypothetical

protein without any correlation with Stevia rebaudiana. Furthermore, the MS/MS fragmentation

of one of the matched peak with m/z 1230 was performed (Figure 58) and again the peptide was

sequenced by de-novo sequencing (Figure 59) and the sequence obtained was subjected to

MASCOT database search and indeed, the suggested sequence which was DLATEGNEIIR

resulted in 100% sequence coverage with that of RuBisCO sequence belonging to Stevia

rebaudiana as it was shown in Figure 56, with the gene bank accession number of gi| 38146633|.


Hande Karaköse 104


Figure 58.MS/MS de novo sequencing of the m/z 1230. Series of y and b fragments are labeled.

Figure 59.Structure and fragmentation of m/z 1230 based on de novo sequencing.

y1

b10

y10

b1

y6

b2

y7


Hande Karaköse 105


The database information on plant proteomics is too limited. The only plant which has entire

genome had been sequenced is Arabidopsis thaliana. Apart from that genome of few plants, such

as rice maze tomato and wheat had been sequenced. Currently, there is no plant genetically

related to the Asteraceae family available in the databases. Therefore, only insufficient sequence

coverage is being obtained from the database searches. The missing parts of the sequence were

partially covered by studying the tandem mass spectra of unknown peaks and sequencing by de-

novo method. However, successful de novo sequencing requires full sequence coverage, thus

demanding better quality spectra than those typically used for data base searching and sequences

obtained by de novo needs confirmation and this can be done either by chemically synthesizing

the obtained sequence and comparing their mass spectrical information or by complete

sequencing of the Stevia rebaudiana genome.

7.4. Conclusion

This study was the first attempt for sequencing leaf proteins of a plant from Asteraceae family

and for Stevia rebaudiana. MALDI-TOF has given a breakthrough in this research and once

again proved to be a very crucial technique in field of proteomics, for the first time ever Stevia

proteins are being characterized. The total of 75 peaks were generated by MALDI-TOF out of

which 33 matched peaks yielded protein score of 232 and 45% of sequence coverage of

RuBisCO enzyme, and 42 of them were considered to be unmatched with the native sequence.

The satisfactory match obtained in the database with the only protein sequence established in

stevia available, clearly indicates that the method employed was valid. However, further studies

are necessary to have the entire sequence of proteins exist in stevia leaves.


Hande Karaköse 106


8. SUMMARY

In summary, a unique and novel dataset comprising the full analysis of around sixty secondary

metabolites from an agricultural plant using seven different varieties from nine different origins

were obtained. In addition to this, correlation studies between the phytochemical information and

climatic metadata obtained within the project on growth conditions provide a new insight in to

agricultural plant science in general.

In more detail, secondary metabolite profile of 166 Stevia rebaudiana leave samples is analyzed

and ten volatile terpenes have been identified and fatty acid profile and quantities have been

obtained.

First attempt to sequence leaf proteins of a plant from Asteraceae family and for Stevia

rebaudiana was performed. Proteins were separated and analyzed successfully and efficient

results were obtained.

Furthermore, around fourty phenolic secondary metabolites, of the class of chlorogenic acids and

flavonoid glycosides were identified and quantified. Additionally, ten steviol glycosides have

been analyzed and quantified.

For a total of ten compounds accurate quantitative data have been obtained for all 166 samples

and for a further fourty compounds relative concentration variations. The data allow a full

description of variations between plants of different varieties and of different origins. Both for

phenolics and steviol glycosides significant variations between origins, varieties and harvests

have been observed as well as variations between stevia samples cultivated in EU and outside.

As conclusion, the quantitative data allow a scientifically sound and state of the art specification

of stevia leaves for licenscing as a novel food in the EU.


Hande Karaköse 107


9. REFERENCES

1. M. S. Bertoni, Revista de Agronomia de l'Assomption, Editon edn., 1899, vol. 1:35. 2. C. C. Shock, in California Agriculture, University of California, Editon edn., 1982, vol. 36, pp. 4-5. 3. N. Kolb, J. L. Herrera, D. J. Ferreyra and R. F. Uliana, Journal of Agricultural and Food Chemistry,

2001, 49, 4538-4541. 4. I. Prakash, G. E. DuBois, J. F. Clos, K. L. Wilkens and L. E. Fosdick, Food and Chemical Toxicology,

2008, 46, S75-S82. 5. C. Gardana, P. Simonetti, E. Canzi, R. Zanchi and P. Pietta, Journal of Agricultural and Food

Chemistry, 2003, 51, 6618-6622. 6. N. A. Ibrahim, S. El-Gengaihi, H. Motawe and S. A. Riad, European Food Research and

Technology, 2007, 224, 483-488. 7. Y. Oshima, J. Saito and H. Hikino, Tetrahedron, 1986, 42, 6443-6446. 8. Y. Oshima, J.-I. SAITO and H. HIKINO, Phytochemistry, 1988, 27, 624-626. 9. K. Yasukawa, A. Yamaguchi, J. Arita, S. Sakurai, A. Ikeda and M. Takido, Phytotherapy Research,

1993, 7, 185-189. 10. A. D. Kinghorn, Stevia The Genus Stevia, CRC Press, 2001. 11. J. Brandle, Agriculture and Agri-Food Canada, Editon edn., 2004, Retrieved 2006. 12. M. Bridel and R. Lavielle, Academie des Sciences Paris Comptes Rendus (Parts 192), Editon edn.,

1931, pp. 1123–1125. 13. C. Hellfritsch, A. Brockhoff, F. Stähler, W. Meyerhof and T. Hofmann, Journal of Agricultural and

Food Chemistry, 2012, 60, 6782-6793. 14. H.U. Mitchell, Sweeteners and Sugar Alternatives in Food Technology, Wiley, 2008. 15. C. Trocho, R. Pardo, I. Rafecas, J. Virgili, X. Remesar, J. A. Fernández-López and M. Alemany, Life

Sciences, 1998, 63, 337-349. 16. B. A. Magnuson, G. A. Burdock, J. Doull, R. M. Kroes, G. M. Marsh, M. W. Pariza, P. S. Spencer,

W. J. Waddell, R. Walker and G. M. Williams, in Crit Rev Toxicol, United States, Editon edn., 2007, vol. 37, pp. 629-727.

17. A. G. Renwick, Food Additives and Contaminants, 2006, 23, 327-338. 18. V. L. Grotz and I. C. Munro, Regulatory Toxicology and Pharmacology, 2009, 55, 1-5. 19. I. M. Baird, N. W. Shephard, R. J. Merritt and G. Hildick-Smith, Food and Chemical Toxicology,

2000, 38, Supplement 2, 123-129. 20. Saccharin warning AP via Telegraph-Herald. 1973-05-22. Retrieved 2011-06-09. 21. J. Whysner and G. M. Williams, Pharmacology & Therapeutics, 1996, 71, 225-252. 22. E. Dybing, Toxicology, 2002, 181–182, 121-125. 23. R. L. Myers and R. L. Myers, The 100 most important chemical compounds: a reference guide,

Westport, Conn: Greenwood Press. 24. L. J. Navia, R. E. Walkup, N. M. Vernon, D. S. Neiditch, Production of sucralose without

intermediate isolation of crystalline sucralose-6-ester, US Patent 536/124, Mar 12, 1996. 25. L. Burnett, Artificial Sweetener controversies from saccharin to sucralose. leda.law.harvard.edu.

Retrieved 20 December 2010 26. H. Misra, M. Soni, N. Silawat, D. Mehta, B. Mehta and D. Jain, J Pharm Bioallied Sci, Editon edn.,

2011, vol. 3, pp. 242-248. 27. J. Hawke, in Nexus magazine, Editon edn., February–March 2003, Retrieved 20 December 2010,

vol. 10. 28. J. M. Pezzuto, C. M. Compadre, S. M. Swanson, D. Nanayakkara and A. D. Kinghorn, Proceedings

of the National Academy of Sciences, 1985, 82, 2478-2482. 29. P. J., Elsevier, New Trends in Natural Products Chemistry, Editon edn., 1986, pp. 371-385. 30. E. Procinska, B. A. Bridges and J. R. Hanson, Mutagenesis, 1991, 6, 165-167.


Hande Karaköse 108


31. H. R. Compadre CM, Nanayakkara NP, Pezzuto JM, Kinghorn AD., Biomed Environ Mass Spectrom, Editon edn., 1988, vol. 15, pp. 211-222.

32. M. Matsui, K. Matsui, Y. Kawasaki, Y. Oda, T. Noguchi, Y. Kitagawa, M. Sawada, M. Hayashi, T. Nohmi, K. Yoshihira, M. Ishidate and T. Sofuni, Mutagenesis, 1996, 11, 573-579.

33. C. Toskulkao, L. Chaturat, P. Temcharoen and T. Glinsukon, Drug Chem Toxicol, 1997, 20, 31-44. 34. L. Xili, B. Chengjiany, X. Eryi, S. Reiming, W. Yuengming, S. Haodong, H. Zhiyian, Food Chem

Toxicol, 1992, 30, 957-965. 35. Y. Aze, K. Toyoda, K. Imaida, S. Hayashi, T. Imazawa, Y. Hayashi, M. Takahashi, Eisei Shikenjo

Hokoku, 1991, 48-54. 36. K. Toyoda, H. Matsui, T. Shoda, C. Uneyama, K. Takada, M. Takahashi, Food and Chemical

Toxicology, 1997, 35, 597-603. 37. R. M. Oliveira-Filho, O. A. Uehara, C. A. Minetti and L. B. Valle, Gen Pharmacol, 1989, 20, 187-

191. 38. C. Wasuntarawat, P. Temcharoen, C. Toskulkao, P. Mungkornkarn, M. Suttajit, T. Glinsukon,

Drug Chem Toxicol, 1998, 21, 207-222. 39. Official Journal of the European Union, Editon edn., 11 November 2011, vol. 1131/2011, p. 26. 40. V. Chatsudthipong and C. Muanprasat, Pharmacology & Therapeutics, 2009, 121, 41-54. 41. C. Toskulkao, M. Sutheerawatananon, C. Wanichanon, P. Saitongdee and M. Suttajit, Journal of

Nutritional Science and Vitaminology, 1995, 41, 105-113. 42. J. M. C. Geuns, J. Buyse, A. Vankeirsbilck, E. H. M. Temme, F. Compernolle and S. Toppet, Journal

of Agricultural and Food Chemistry, 2006, 54, 2794-2798. 43. G. Brahmachari, L. C. Mandal, R. Roy, S. Mondal and A. K. Brahmachari, Archiv der Pharmazie,

2011, 344, 5-19. 44. E. Koyama, K. Kitazawa, Y. Ohori, O. Izawa, K. Kakegawa, A. Fujino, Food and Chemical

Toxicology, 2003, 41, 359-374. 45. A. Wheeler, A. C. Boileau, P. C. Winkler, J. C. Compton, I. Prakash, X. Jiang, D. A. Mandarino,

Food and Chemical Toxicology, 2008, 46, S54-S60. 46. N. Totté, L. Charon, M. Rohmer, F. Compernolle, I. Baboeuf and J. M. C. Geuns, Tetrahedron

Letters, 2000, 41, 6407-6410. 47. N. Totté, W. V. d. Ende, E. J. M. Van Damme, F. Compernolle, I. Baboeuf, J. M. C. Geuns,

Canadian Journal of Botany, 2003, 81, 517-522. 48. J. E. Brandle, A. Richman, A. K. Swanson and B. P. Chapman, Plant Molecular Biology, 2002, 50,

613-622. 49. T. Humphrey, A. Richman, R. Menassa and J. Brandle, Plant Molecular Biology, 2006, 61, 47-62. 50. J. E. Brandle and P. G. Telmer, Phytochemistry, 2007, 68, 1855-1863. 51. A. Hemmerlin, J. L. Harwood and T. J. Bach, Progress in Lipid Research, 2012, 51, 95-148. 52. J. Metivier and A. M. Viana, Journal of Experimental Botany, 1979, 30, 805-810. 53. P. Mauri, G. Catalano, C. Gardana and P. Pietta, ELECTROPHORESIS, 1996, 17, 367-371. 54. P. Nishiyama, M. Alvarez and L. G. E. Vieira, Journal of the Science of Food and Agriculture, 1992,

59, 277-281. 55. H. Mizukami, K. Shiiba and H. Ohashi, Phytochemistry, 1982, 21, 1927-1930. 56. M. Jamaluddin Ahmed and R. M. Smith, Journal of Separation Science, 2002, 25, 170-172. 57. U. Woelwer-Rieck, C. Lankes, A. Wawrzun and M. Wüst, European Food Research and

Technology, 2010, 231, 581-588. 58. H. C. Makapugay, N. P. D. Nanayakkara and A. D. Kinghorn, Journal of Chromatography A, 1984,

283, 390-395. 59. , P. Karásek, M. Roth, K Beneš K řík Č ký Analytical

and Bioanalytical Chemistry, 2007, 388, 1847-1857. 60. J. Liu and S. F. Y. Li, Journal of Liquid Chromatography, 1995, 18, 1703-1719.


Hande Karaköse 109


61. T něk Nep ı mp x A m ı nd ček Journal of Food Composition and Analysis, 2001, 14, 383-388.

62. V. J. Y. Minne, F. Compernolle, S. Toppet and J. M. C. Geuns, Journal of Agricultural and Food Chemistry, 2004, 52, 2445-2449.

63. e -Rieck, W. Tomberg and A. Wawrzun, Journal of Agricultural and Food Chemistry, 2010, 58, 12216-12220.

64. B. Zimmermann, U. Woelwer-Rieck and M. Papagiannopoulos, Food Analytical Methods, 2012, 5, 266-271.

65. F. Cacciola, P. Delmonte, K. Jaworska, P. Dugo, L. Mondello and J. I. Rader, Journal of Chromatography A, 2011, 1218, 2012-2018.

66. J. Pol, B. Hohnova and T. Hyotylainen, Journal of Chromatography A, 2007, 1150, 85-92. 67. C. Gardana, M. Scaglianti and P. Simonetti, Journal of Chromatography A, 2010, 1217, 1463-

1470. 68. J.M.C. Geuns, Validated techniques of analysis. In Stevia and Steviol Glycosides,

Euprint:Heverlee, Belgium, 2010; ISBN 978-90-742-53116, p61-113. 69. D. M. W.-D. Eric S. Grumbach , Pamela C. Iraneta , Bonnie Alden , Jeffrey R. Mazzeo, in LCGC

North America, Editon edn., 2004, vol. 22, p. 1010. 70. M. Murata, E. Tanaka, E. Minoura and S. Homma, Bioscience, Biotechnology, and Biochemistry,

2004, 68, 501-507. 71. L. Bravo, Nutrition Reviews, 1998, 56, 317-333. 72. M. Naczk and F. Shahidi, Journal of Chromatography A, 2004, 1054, 95-111. 73. H. Wang, G. Cao and R. L. Prior, Journal of Agricultural and Food Chemistry, 1996, 44, 701-705. 74. M. Naczk and F. Shahidi, Journal of Pharmaceutical and Biomedical Analysis, 2006, 41, 1523-

1542. 75. C. Manach, A. Scalbert, C. Morand, C. Rémésy and L. Jiménez, The American Journal of Clinical

Nutrition, 2004, 79, 727-747. 76. A. Crozier, M. E. J. Lean, M. S. McDonald and C. Black, Journal of Agricultural and Food

Chemistry, 1997, 45, 590-595. 77. A. J. Stewart, S. Bozonnet, W. Mullen, G. I. Jenkins, M. E. J. Lean and A. Crozier, Journal of

Agricultural and Food Chemistry, 2000, 48, 2663-2669. 78. L. Gennaro, C. Leonardi, F. Esposito, M. Salucci, G. Maiani, G. Quaglia and V. Fogliano, Journal of

Agricultural and Food Chemistry, 2002, 50, 1904-1910. 79. N. Balasundram, K. Sundram and S. Samman, Food Chemistry, 2006, 99, 191-203. 80. M. N. Clifford, Journal of the Science of Food and Agriculture, 2000, 80, 1033-1043. 81. M. N. Clifford, K. L. Johnston, S. Knight and N. Kuhnert, Journal of Agricultural and Food

Chemistry, 2003, 51, 2900-2911. 82. M. N. Clifford and J. R. Ramirez-Martinez, Food Chemistry, 1990, 35, 13-21. 83. V. Marques and A. Farah, Food Chemistry, 2009, 113, 1370-1376. 84. J. H. Chen and C.-T. Ho, Journal of Agricultural and Food Chemistry, 1997, 45, 2374-2378. 85. M. Lepelley, G. Cheminade, N. Tremillon, A. Simkin, V. Caillet and J. McCarthy, Plant Science,

2007, 172, 978-996. 86. M. R. Olthof, P. C. H. Hollman and M. B. Katan, The Journal of Nutrition, 2001, 131, 66-71. 87. Y. Kono, K. Kobayashi, S. Tagawa, K. Adachi, A. Ueda, Y. Sawa and H. Shibata, Biochimica Et

Biophysica Acta-General Subjects, 1997, 1335, 335-342. 88. M. R. Luzia, K. C. C. DaPaixao, R. Marcilio, L. C. Trugo, L. M. C. Quinteiro and C. A. B. DeMaria,

International Journal of Food Science and Technology, 1997, 32, 15-19. 89. A. W. Herling, H. J. Burger, D. Schwab, H. Hemmerle, P. Below and G. Schubert, American Journal

of Physiology-Gastrointestinal and Liver Physiology, 1998, 274, G1087-G1093.


Hande Karaköse 110


90. J. Shearer, A. Farah, T. de Paulis, D. P. Bracy, R. R. Pencek, T. E. Graham and D. H. Wasserman, Journal of Nutrition, 2003, 133, 3529-3532.

91. H. Hemmerle, H. J. Burger, P. Below, G. Schubert, R. Rippel, P. W. Schindler, E. Paulus and A. W. Herling, Journal of Medicinal Chemistry, 1997, 40, 137-145.

92. W. J. Arion, W. K. Canfield, F. C. Ramos, P. W. Schindler, H. J. Burger, H. Hemmerle, G. Schubert, P. Below and A. W. Herling, Archives of Biochemistry and Biophysics, 1997, 339, 315-322.

93. K. L. Johnston, M. N. Clifford and L. M. Morgan, American Journal of Clinical Nutrition, 2003, 78, 728-733.

94. M. D. del Castillo, J. M. Ames and M. H. Gordon, Journal of Agricultural and Food Chemistry, 2002, 50, 3698-3703.

95. W. E. Robinson, M. Cordeiro, S. AbdelMalek, Q. Jia, S. A. Chow, M. G. Reinecke and W. M. Mitchell, Molecular Pharmacology, 1996, 50, 846-855.

96. W. E. Robinson, M. G. Reinecke, S. AbdelMalek, Q. Jia and S. A. Chow, Proceedings of the National Academy of Sciences of the United States of America, 1996, 93, 6326-6331.

97. B. McDougall, P. J. King, B. W. Wu, Z. Hostomsky, M. G. Reinecke and W. E. Robinson, Antimicrobial Agents and Chemotherapy, 1998, 42, 140-146.

98. A. Trute, J. Gross, E. Mutschler and A. Nahrstedt, Planta Medica, 1997, 63, 125-129. 99. H. F. Stich, M. P. Rosin and L. Bryson, Mutation Research, 1982, 95, 119-128. 100. N. Hanlon, N. Coldham, A. Gielbert, N. Kuhnert, M. J. Sauer, L. J. Kingi and C. Ioannides, British

Journal of Nutrition, 2008, 99, 559-564. 101. A. Stalmach, W. Mullen, D. Barron, K. Uchida, T. Yokota, C. Cavin, H. Steiling, G. Williamson and

A. Crozier, Drug Metabolism and Disposition, 2009, 37, 1749-1758. 102. A. Farah, M. Monteiro, C. M. Donangelo and S. Lafay, Journal of Nutrition, 2008, 138, 2309-2315. 103. H. M. Merken and G. R. Beecher, Journal of Agricultural and Food Chemistry, 2000, 48, 577-599. 104. P.-G. Pietta, Journal of Natural Products, 2000, 63, 1035-1042. 105. E. de Rijke, P. Out, W. M. A. Niessen, Journal of Chromatography A, 2006, 1112, 31-63. 106. A z nić F Buc Ž M eš A M n B Nig ić Molecules, Molecules, Editon edn., 2009,

vol. 14, pp. 2466-2490. 107. P. C. H. Hollman and M. B. Katan, Food and Chemical Toxicology, 1999, 37, 937-942. 108. I. Ignat, I. Volf and V. I. Popa, Food Chemistry, 2011, 126, 1821-1835. 109. I.-S. Kim, M. Yang, O.-H. Lee and S.-N. Kang, LWT - Food Science and Technology, 2011, 44, 1328-

1332. 110. S. Ghanta, A. Banerjee, A. Poddar and S. Chattopadhyay, Journal of Agricultural and Food

Chemistry, 2007, 55, 10962-10967. 111. U. Wölwer-Rieck, Journal of Agricultural and Food Chemistry, 2012, 60, 886-895. 112. G. Li, R. Wang, A. J. Quampah, Z. Rong, C. Shi and J. Wu, Journal of Agricultural and Food

Chemistry, 2011, 59, 13065-13071. 113. J. M. González-Buitrago, L. Ferreira, M. Isidoro-García, C. Sanz, F. Lorente and I. Dávila, Clinica

Chimica Acta, 2007, 385, 21-27. 114. S. Rogers, M. Girolami, W. Kolch, K. M. Waters, T. Liu, B. Thrall and H. S. Wiley, Bioinformatics,

2008, 24, 2894-2900. 115. V. Dhingra, M. Gupta, T. Andacht and Z. F. Fu, International Journal of Pharmaceutics, 2005, 299,

1-18. 116. C. M. D. Tal Isaacson, Ramu S Saravanan,Yonghuae, Carmen Catala, Montserrat Saladie &

Jocelyn KC Rose, Nature protocol, Editon edn., 13 July 2006. 117. R. M. LoPachin, R. C. Jones, T. A. Patterson, W. Slikker Jr and D. S. Barber, NeuroToxicology,

2003, 24, 761-775. 118. J. A. Reynolds and C. Tanford, Proc Natl Acad Sci U S A, 1970, 66, 1002-1007. 119. V. Neuhoff, N. Arold, D. Taube and W. Ehrhardt, Electrophoresis, 1988, 9, 255-262.


Hande Karaköse 111


120. J. E. Shively, EXS, 2000, 88, 99-117. 121. D. Lin, D. L. Tabb, Biochimica et Biophysica Acta (BBA) - Proteins&Proteomics, 2003, 1646, 1-10. 122. X. Han, A. Aslanian and J. R. Yates Iii, Current Opinion in Chemical Biology, 2008, 12, 483-490. 123. R. Krüger, A. Pfenninger, I. Fournier, M. Glückmann and M. Karas, Analytical Chemistry, 2001,

73, 5812-5821. 124. J. B. Ohlrogge and J. G. Jaworski, Annual Review of Plant Physiology and Plant Molecular Biology,

1997, 48, 109-136. 125. A. Carrasco-Pancorbo, N. Navas-Iglesias and L. Cuadros-Rodríguez, TrAC Trends in Analytical

Chemistry, 2009, 28, 263-278. 126. C. Zhu, P. Hu, Q.-L. Liang, Y.-M. Wang and G.-A. Luo, Chinese Journal of Analytical Chemistry,

2009, 37, 1390-1396. 127. J. Folch, S. Arsove and J. A. Meath, Journal of Biological Chemistry, 1951, 191, 819-831. 128. V. Matyash, G. Liebisch, T. V. Kurzchalia, A. Shevchenko and D. Schwudke, Journal of Lipid

Research, 2008, 49, 1137-1146. 129. A. Nomura, J. Yamada, T. Yarita and T. Maeda, The Journal of Supercritical Fluids, 1995, 8, 329-

333. 130. T. White, S. Bursten, D. Federighi, R. A. Lewis and E. Nudelman, Analytical Biochemistry, 1998,

258, 109-117. 131. K. Eder, Journal of Chromatography B: Biomedical Sciences and Applications, 1995, 671, 113-

131. 132. X. Han and R. W. Gross, in J Lipid Res, United States, Editon edn., 2003, vol. 44, pp. 1071-1079. 133. C. Hu, R. van der Heijden, M. Wang, J. van der Greef, T. Hankemeier and G. Xu, Journal of

Chromatography B, 2009, 877, 2836-2846. 134. K. U. Borstel René, LCGC chromatography online, Application note, Editon edn., 2011. 135. M. N. Clifford, S. Knight and N. Kuhnert, Journal of Agricultural and Food Chemistry, 2005, 53,

3821-3832. 136. A. J. Day, F. Mellon, D. Barron, G. Sarrazin, M. R. A. Morgan, Free Radical Research, 2001, 35,

941-952. 137. A. Plazonic, F. Bucar, Z. Males, A. Mornar, B. Nigovic and N. Kujundzic, in Molecules, Switzerland,

Editon edn., 2009, vol. 14, pp. 2466-2490. 138. F. Cuyckens and M. Claeys, Journal of Mass Spectrometry, 2004, 39, 1-15. 139. Y.L. Ma, Q.M. Li, H. Van den Heuvel, M. Claeys, Rapid Communications in Mass Spectrometry,

1997, 11, 1357-1364. 140. J. M. Harnly, R. F. Doherty, G. R. Beecher, J. M. Holden, D. B. Haytowitz, S. Bhagwat, S. Gebhardt,

Journal of Agricultural and Food Chemistry, 2006, 54, 9966-9977. 141. C. J. Schwartz, The mixed-model ANOVA: The truth, the computer packages, the books. Part I:

Balanced data, The American Statistician, 1993. 142. L. Zelles and Q. Y. Bai, Soil Biology & Biochemistry, 1993, 25, 495-507. 143. G. Gutnikov, Journal of Chromatography B-Biomedical Applications, 1995, 671, 71-89. 144. A. Martelli, C. Frattini and F. Chialva, Flavour and Fragrance Journal, 1985, 1, 3-7. 145. A. L. Thompson, A. Michalik, R. J. Nash, F. X. Wilson, R. van Well, P. Johnson, G. W. J. Fleet, C. Y.

Yu, X. G. Hu, R. I. Cooper and D. J. Watkin, Acta Crystallographica Section E-Structure Reports Online, 2009, 65, O2904-U2232.

146. G. B. A. Görg, O. Drews, A. Köpf, C. Lück, G. Reil, W. Weiss, A Laboratory Manual, Technische Universität München, Editon edn., 2003.

147. A. Shevchenko, H. Tomas, J. Havlis, J. V. Olsen and M. Mann, Nat. Protocols, 2007, 1, 2856-2860. 148. V. Egelhofer, K. Büssow, C. Luebbert, H. Lehrach and E. Nordhoff, Analytical Chemistry, 2000, 72,

2741-2750.


Hande Karaköse 112


149. B. Thiede, W. Höhenwarter, A. Krah, J. Mattow, M. Schmid, F. Schmidt and P. R. Jungblut, Methods, 2005, 35, 237-247.

150. J. L. Panero and V. A. Funk, Molecular Phylogenetics and Evolution, 2008, 47, 757-782.


Hande Karaköse 113


APPENDIX

A. Tandem mass spectra of steviol glycosides in negative ion mode.

Stevioside m/z 803 [M-H+]

-

Steviolbioside m/z 641 [M-H+]

-

317.0

479.1 -MS2(641.2)

317.0 -MS3(641.6->479.4)

-MS4(641.6->479.4->317.1)

0

1

2

7 x10 Intens.

0.0

0.5

1.0 7 x10

-1

0

1

2

200 400 600 800 1000 m/z

641.2 -MS2(803.4)

317.0

479.1 -MS3(803.8->641.2)

317.0 -MS4(803.8->641.5->479.3)

0.0 0.5 1.0

7 x10 Intens.

0 2 4 6

6 x10

0

1

2 6 x10

200 400 600 800 1000 m/z


Hande Karaköse 114


Rebaudioside F m/z 935 [M-H+]

-

Dulcoside A m/z 787 [M-H+]

- m/z 823[M+Cl

-]-

Rubusoside m/z 641[M-H+]

-

317.1

479.1 -MS2(641.2)

317.0

0

2

7 x10 Intens.

0.0

0.5

1.0

7 x10

-1 0 1 2

200 400 600 800 1000 m/z

-MS4(641.6->479.4->317.2)

-MS3(641.6->479.4)

625.3 -MS2(823.4)

317.0

479.1 -MS3(824.0->625.2)

317.0 -MS4(824.0->625.5->479.3)

0.0

0.5

1.0 7 x10

Intens.

0

2

4 6 x10

0 2 4 6

5 x10

200 400 600 800 1000 m/z

773.4 -MS2(935.5)

317.1 413.1 479.1

611.2 -MS3(936.0->773.3),

317.0

479.1 -MS4(936.0->773.7->613.4)

0.0

0.5

1.0 6 x10

Intens.

0 1 2 3 5 x10

0 2 4 6 4 x10

200 400 600 800 1000 m/z


Hande Karaköse 115


B. Quantification data of steviol glycosides

Bar plot showing minimum, maximum and average values obtained from all 166 stevia samples

within within +/- 3 σ

0.000

2.000

4.000

6.000

8.000

10.000

12.000

14.000

16.000

18.000

20.000

RebA stevioside DulcosideA Rubusoside RebC sum

min

average

max

allsamples

Stdev(σ) 0.741 3.077 0.152 0.080 0.154 3.200

g/100g

leaves RebA Stevioside DulcosideA Rubusoside RebC sum

min 0.079 0.252 0.005 0.005 0.026 0.554

average 1.017 7.314 0.264 0.122 0.319 9.036

max 5.336 17.509 0.680 0.459 0.820 18.067

Quantity of Steviol Glycosides

Hande Karaköse 116


Sample No. Origin Variety Harvesting Year RebA

( g/100g)

Stevioside

(g/100g)

DulcosideA

( g/100g)

Rubusoside

( g/100g)

RebC

(g/100g)

Sum

( g/100g)

1 TCV 3 II 10.09.2010 0.596 6.168 0.140 0.128 0.196 7.229

2 TCV 4 II 14.09.2010 1.411 7.781 0.263 0.113 0.409 9.975

3 TCV 3 I 04.08.2010 1.203 7.980 0.299 0.086 0.383 9.952

4 TCV 1 I 04.08.2010 1.400 8.197 0.132 0.138 0.447 10.313

5 TCV 1 I 04.08.2010 1.592 7.608 0.103 0.101 0.518 9.921

6 TCV 2 I 04.08.2010 1.983 2.962 0.186 0.042 0.366 5.538

7 TCV 2 I 04.08.2010 1.959 2.863 0.104 0.049 0.476 5.450

8 TCV 4 I 04.08.2010 1.591 4.773 0.273 0.118 0.409 7.164

9 TCV 4 II 14.09.2010 1.007 5.390 0.300 0.128 0.297 7.122

10 TCV 2 II 28.09.2010 2.154 7.360 0.179 0.107 0.399 10.199

11 TCV 2 II 28.09.2010 1.607 6.803 0.210 0.101 0.308 9.030

12 Pojava 4 II 14.09.2010 0.565 6.316 0.235 0.083 0.211 7.411

13 TCV 3 I 04.08.2010 0.632 5.648 0.254 0.080 0.215 6.827

14 TCV 2 I 04.08.2010 2.058 3.585 0.073 0.038 0.345 6.100

15 TCV 1 I 01.09.2010 1.001 5.178 0.075 0.072 0.256 6.582

16 TCV 2 I 04.08.2010 1.569 3.314 0.073 0.035 0.503 5.494

17 TCV 3 I 04.08.2010 0.717 5.769 0.217 0.048 0.223 6.975

18 TCV 4 I 04.08.2010 0.800 5.322 0.169 0.047 0.248 6.586

19 TCV 3 II 10.09.2010 0.550 6.003 0.178 0.097 0.178 7.006

20 TCV 2 II 28.09.2010 1.622 4.889 0.078 0.051 0.316 6.956

21 TCV 1 II 28.09.2010 0.853 7.590 0.117 0.073 0.322 8.955

22 TCV 3 II 10.09.2010 0.362 7.808 0.308 0.136 0.154 8.768

23 TCV,Pojana 4 I 04.08.2010 0.640 6.297 0.165 0.048 0.254 7.404

24 TCV,pojana 3 I 04.08.2010 0.292 7.049 0.380 0.074 0.150 7.945

25 TCV 1 I 01.09.2010 0.923 5.300 0.065 0.075 0.256 6.619

26 TCV 4 II 14.09.2010 0.738 5.554 0.154 0.063 0.209 6.719

27 TCV 4 I 04.08.2010 0.657 5.138 0.158 0.041 0.196 6.190


Hande Karaköse 117


Sample No. Origin Variety Harvesting Year RebA Stevioside DulcosideA Rubusoside RebC sum

28 TCV 2 II 28.09.2010 1.669 4.870 0.266 0.058 0.820 7.684

29 TCV 3 II 10.09.2010 0.437 6.549 0.255 0.125 0.169 7.535

30 Uconor 4 II 07.05.2010 0.562 5.699 0.209 0.065 0.185 6.720

31 Uconor 3 I 11.08.2010 0.337 3.769 0.297 0.075 0.160 4.638

32 Uconor 2 I 11.08.2010 1.955 3.881 0.048 0.034 0.407 6.325

33 Uconor 2 II 07.05.2010 1.672 4.530 0.060 0.045 0.503 6.810

34 Uconor 4 II 07.09.2010 0.722 5.217 0.124 0.064 0.202 6.328

35 Uconor 4 I 11.08.2010 0.404 4.799 0.280 0.071 0.143 5.696

36 Uconor 4 I 11.08.2010 0.619 3.780 0.104 0.045 0.195 4.745

37 Uconor 4 I 11.08.2010 1.032 5.085 0.087 0.047 0.286 6.537

38 Uconor 3 I 11.08.2010 0.540 6.065 0.235 0.076 0.201 7.117

39 Uconor 4 I 11.08.2010 0.797 3.773 0.122 0.044 0.223 4.959

40 Uconor 2 I 11.08.2010 1.471 2.947 0.075 0.041 0.479 5.015

41 Uconor 3 I 11.08.2010 0.664 5.185 0.139 0.057 0.268 6.313

42 Uconor 3 II 07.09.2010 0.214 5.704 0.281 0.125 0.127 6.451

43 Agrinion 2 II 20.09.2010 1.561 4.908 0.173 0.080 0.325 7.047

44 Toumpa 1 I 30.07.2010 0.444 4.896 0.206 0.083 0.166 5.795

45 Portugal 1 I 07.07.2010 0.600 6.732 0.051 0.064 0.252 7.700

46 Amfilia 1 I 04.08.2010 0.266 5.301 0.240 0.078 0.133 6.018

47 Toumpa 3 II 10.09.2010 0.365 4.107 0.126 0.054 0.143 4.794

48 Agrinion 2 I 09.08.2010 1.048 4.093 0.133 0.053 0.242 5.568

49 Toumpa 3 I 30.07.2010 0.355 5.134 0.190 0.071 0.146 5.895

50 Agrinion 4 I 09.08.2010 0.789 3.533 0.064 0.026 0.251 4.664

51 Agrinion 4 II 20.09.2010 1.034 4.449 0.143 0.066 0.243 5.935

52 Portugal 4 I 26.06.2010 0.563 4.573 0.124 0.103 0.188 5.551

53 Amfilia 4 I 04.08.2010 0.434 4.741 0.200 0.071 0.155 5.601

54 Amfilia 4 II 15.09.2010 0.491 4.528 0.147 0.094 0.200 5.460

55 Argentinie 2009 - - 2009 0.514 5.264 0.289 0.220 0.216 6.504


118


56 Paragan2009 - - 2009 0.373 3.750 0.178 0.162 0.200 4.663

57 Argentinien2010 - - 2010 0.552 7.212 0.337 0.240 0.252 8.593

58 Amfilia 2 II 15.09.2010 1.125 5.026 0.119 0.107 0.302 6.680

59 Agrinion 1 II 20.09.2010 1.044 7.864 0.348 0.116 0.362 9.735

60 Amfilia 2 I 04.08.2010 1.114 6.732 0.265 0.083 0.323 8.518

61 Amfilia 3 II 15.09.2010 0.770 6.073 0.184 0.127 0.328 7.481

62 Toumpa 1 I 30.07.2010 0.444 4.396 0.567 0.211 0.210 5.829

63 Amfilia 1 II 15.09.2010 0.695 3.428 0.366 0.187 0.257 4.933

64 Portugal 3 I 26.06.2010 1.768 6.869 0.546 0.339 0.578 10.100

65 Toumpa 1 II 10.09.2010 0.703 3.745 0.455 0.173 0.366 5.442

66 Agrinion 1 I 09.08.2010 0.673 3.569 0.357 0.107 0.327 5.033

67 Amfilia 3 I 04.08.2010 1.100 7.131 0.190 0.097 0.366 8.884

68 Toumpa 2 II 10.09.2010 0.419 8.115 0.458 0.384 0.190 9.566

69 Agrinio 3 II 20.09.2010 0.910 9.167 0.488 0.151 0.274 10.990

70 Toumpa 4 II 10.09.2010 0.733 7.944 0.342 0.123 0.268 9.410

71 Agrinion 3 I 09.08.2010 0.693 7.053 0.327 0.123 0.292 8.488

72 Toumpa 4 I 30.07.2010 0.806 5.995 0.222 0.060 0.243 7.326

73 Granada 2 I 15.09.2010 1.728 6.826 0.073 0.110 0.486 9.224

74 Granada 4 I 15.09.2010 0.779 6.875 0.261 0.080 0.332 8.327

75 Granada 3 I 09.09.2010 0.484 7.916 0.406 0.099 0.225 9.130

76 Equador - - - 0.647 7.599 0.005 0.005 0.067 8.323

77 DZ - - - 0.319 7.042 0.048 0.050 0.082 7.541

78 Kenia 2010 - - 2010 0.615 7.866 0.019 0.051 0.123 8.674

79 Paraguay 2009 - - 2009 0.193 10.950 0.078 0.049 0.094 11.364

80 Argentinien 2010 - - 2010 0.185 12.898 0.110 0.055 0.100 13.349

81 Indien 2010 - - 2010 0.556 14.084 0.054 0.053 0.162 14.907

82 Argenitinien 2009 - - 2009 0.245 12.551 0.091 0.048 0.026 12.962

83 Hohenheim 2010 - - 2010 0.443 17.509 0.074 0.016 0.026 18.067


119

Sample No. Origin Variety Harvesting Year RebA Stevioside DulcosideA Rubusoside RebC Sum

84 Krim - - - 0.190 0.252 0.026 0.046 0.040 0.554

85 Turkei PS1 2010 - - 2010 0.500 12.696 0.028 0.021 0.214 13.282

86 Turkei PS2 2010 - - 2010 0.446 8.683 0.022 0.016 0.196 9.202

87 Uconor 7 I 13.07.2011 0.523 1.058 0.041 0.012 0.029 1.664

88 Agrinon 4 I 2011 0.721 2.416 0.049 0.011 0.031 3.228

89 Uconor 4 I 2011 0.079 4.269 0.299 0.107 0.188 4.943

90 Agrinion 3 I 2011 0.421 8.658 0.068 0.016 0.049 9.210

91 Agrinion 5 I 2011 0.396 3.335 0.177 0.061 0.272 4.241

92 Uconor 6 I 13.07.2011 0.455 2.968 0.124 0.066 0.191 3.804

93 Uconor 3 I 13.07.2011 0.151 3.060 0.306 0.056 0.075 3.648

94 Uconor 5 I 13.07.2011 0.639 2.590 0.137 0.041 0.170 3.577

95 Agrinion 6 I 2011 0.790 3.817 0.151 0.055 0.282 5.096

96 Toumpa 1 I 2011 0.576 5.840 0.372 0.182 0.263 7.234

97 Toumpa 2 I 2011 1.101 4.465 0.254 0.203 0.282 6.305

98 Toumpa 3 I 2011 0.501 5.638 0.356 0.171 0.168 6.833

99 Toumpa 4 I 2011 0.943 4.585 0.228 0.127 0.306 6.189

100 Amiflikeia 4 II 2011 1.603 6.729 0.563 0.103 0.530 9.529

101 Amiflikeia 5 II 2011 1.611 10.507 0.275 0.148 0.533 13.073

102 Amiflikeia 4 II 2011 1.753 10.559 0.403 0.141 0.452 13.309

103 Amiflikeia 3 II 2011 1.088 10.636 0.491 0.211 0.374 12.799

104 Amiflikeia - II 2011 0.802 11.020 0.306 0.146 0.305 12.578

105 Amiflikeia 6 II 2011 1.827 9.539 0.282 0.169 0.512 12.330

106 Amiflikeia 1 II 2011 1.053 10.517 0.457 0.233 0.337 12.597

107 APTTB 3 I 2011 1.029 10.030 0.420 0.149 0.307 11.935

108 APTTB 5 I 2011 1.101 9.921 0.350 0.154 0.333 11.859

109 APTTB 6 I - 1.282 8.761 0.268 0.164 0.358 10.833

110 APTTB 4 I - 1.078 8.955 0.286 0.140 0.398 10.856

111 Uconor 5 II 2011 1.133 8.765 0.337 0.108 0.444 10.786


120


112 Uconor 4 II 2011 1.589 7.906 0.180 0.086 0.417 10.178

113 Uconor 3 II 2011 0.646 8.769 0.375 0.207 0.266 10.263

114 Uconor 6 II 2011 1.110 10.151 0.297 0.101 0.473 12.133

115 Conaga 3 I 2011 0.694 9.683 0.413 0.100 0.285 11.175

116 Conaga 5 I 2011 1.856 8.011 0.247 0.055 0.429 10.599

117 Conaga 4 I 2011 0.997 10.457 0.395 0.098 0.384 12.331

118 Agrinion 5 II 2011 1.601 9.733 0.349 0.149 0.467 12.298

119 Agrinion 4 II 2011 1.173 9.625 0.351 0.148 0.379 11.676

120 Conaga 6 I 2011 2.222 8.436 0.346 0.064 0.587 11.654

121 Conaga 7 I 2011 4.901 1.148 0.680 0.021 0.680 7.429

122 Agrinion 6 II 2011 1.397 9.888 0.306 0.147 0.547 12.285

123 Agrinion 3 II 2011 1.140 10.231 0.490 0.181 0.434 12.477

124 Toumpa 7 I 2011 5.336 0.966 0.050 0.018 0.711 7.082

125 Toumpa 4 III 2011 1.000 10.844 0.405 0.159 0.326 12.734

126 Toumpa 3 III 2011 0.897 11.795 0.596 0.271 0.340 13.898

127 Toumpa 3 II 2011 0.743 10.671 0.524 0.249 0.330 12.517

128 Agrinion 4 II 2011 1.447 8.980 0.337 0.155 0.461 11.380

129 Toumpa 4 II 2011 0.934 9.641 0.335 0.191 0.376 11.477

130 Toumpa 3 II 2011 0.722 10.616 0.547 0.190 0.303 12.377

131 Toumpa 4 II 2011 1.235 9.892 0.349 0.146 0.466 12.089

132 Agrinion 3 II 2011 0.956 9.886 0.383 0.222 0.407 11.855

133 Toumpa 5 II 2011 0.770 10.641 0.427 0.194 0.387 12.420

134 Toumpa 6 II 2011 1.030 9.752 0.290 0.159 0.464 11.695

135 Toumpa 7 II 2011 4.590 1.515 0.017 0.023 0.599 6.745

136 TCV 5 I 30.06.2011 0.836 7.947 0.139 0.165 0.366 9.454

137 TCV 3 I 30.06.2011 0.525 9.503 0.411 0.224 0.281 10.943

138 TCV 6 I 30.06.2011 0.824 7.083 0.222 0.142 0.377 8.648

139 TCV 3 II 11.08.2011 0.642 9.780 0.498 0.184 0.246 11.350


Hande Karaköse 121



140 TCV 6 II 17.08.2011 0.922 8.626 0.302 0.116 0.537 10.504

141 TCV 4 II 24.08.2011 1.090 8.332 0.285 0.125 0.405 10.236

142 TCV 7 I 18.08.2011 3.729 1.099 0.019 0.031 0.613 5.491

143 TCV 4 I 07.07.2011 0.670 8.299 0.291 0.129 0.440 9.828

144 TCV 5 II 17.08.2011 1.215 10.728 0.410 0.127 0.594 13.073

145 Amiflikeia 3 I 2011 0.574 10.429 0.490 0.207 0.269 11.970

146 Amiflikeia 4 I 2011 1.167 10.350 0.310 0.195 0.422 12.444

147 Amiflikeia 5 I 2011 1.440 10.736 0.297 0.230 0.456 13.159

148 Amiflikeia 6 I 2011 1.233 10.463 0.355 0.225 0.430 12.706

149 Amiflikeia 1 I 2011 0.771 12.365 0.529 0.383 0.287 14.335

150 Amiflikeia 2 I 2011 1.175 11.315 0.397 0.334 0.425 13.646

151 Amiflikeia 4 I 2011 0.867 11.958 0.501 0.312 0.299 13.937

152 Amiflikeia 3 I 2011 0.711 10.987 0.623 0.459 0.313 13.093

153 Toumpa 3 I 2011 0.704 11.494 0.580 0.354 0.249 13.381

154 Toumpa 5 I 2011 0.902 10.066 0.316 0.160 0.387 11.831

155 Toumpa 6 I 2011 1.033 7.601 0.185 0.093 0.381 9.293

156 Toumpa 4 I 2011 1.233 10.828 0.336 0.296 0.441 13.133

157 Toumpa 4 I 2011 1.060 9.909 0.352 0.132 0.368 11.821

158 Toumpa 3 I 2011 0.720 10.251 0.379 0.184 0.289 11.823

159 TCV 6 III 2011 1.222 9.232 0.326 0.169 0.464 11.413

160 TCV 5 III 2011 1.069 10.200 0.389 0.191 0.493 12.341

161 TCV 3 III 2011 0.819 10.650 0.478 0.176 0.361 12.485

162 Agrinion 3 III 2011 1.239 10.607 0.425 0.142 0.495 12.908

163 TCV 4 III 2011 0.608 8.321 0.342 0.165 0.296 9.731

164 Agrinion 6 III 2011 1.720 12.198 0.455 0.198 0.713 15.283

165 Agrinion 5 III 2011 1.365 12.007 0.393 0.170 0.635 14.570

166 Agrinion 4 III 2011 1.778 12.302 0.540 0.224 0.704 15.549


Hande Karaköse 122


Co-elution of steviol glycosides in the UV-chromatogram at 210 nm (LC-MS measurement

with Knauer amino column):

Co-elution of rebaudioside A with stevioside or rebaudioside B:

Co-elution of rebaudioside C with dulcoside A:

196.0300.2

401.2 518.2

803.4

965.5

1001.4

-MS, 11.2min #670

0.0

0.2

0.4

0.6

0.8

1.0

5x10

Intens.

200 400 600 800 1000 m/z

179.1

247.0

300.2359.1

406.1

480.2

787.4

949.5

985.4

-MS, 7.6min #452

0

1

2

3

4x10

Intens.

200 400 600 800 1000 m/z

1

1

2

2

3


Hande Karaköse 123


Co-elution of stevioside with steviolbioside:

Extracted ion chromatogram of rebaudioside A (HILIC)

641.3

803.4

-MS, 6.1min #366

0

2

4

6

8

5x10

Intens.

200 400 600 800 1000 m/z

389.3

457.2

528.2

965.4

1011.4

1079.4

-MS, 27.0min #1612

0

1

2

3

5x10

Intens.

200 400 600 800 1000 m/z

3


Hande Karaköse 124


C. Tandem mass spectra of CGAs and flavonoid glycosides in negative ion mode

3-caffeoylquinic acid m/z 353 [M-H+]

-


-

190.7 -MS2(352.9)

93.0 172.7

126.8 -MS3(353.1->190.7)

83.4 -MS4(353.1->190.8->126.9)

0 2 4 6 7 x10

Intens.

0 1 2 3 5 x10

0.0

0.5

1.0 4 x10

200 400 600 800 1000 m/z

134.8

190.7 -MS2(352.9)

126.7 172.8 85.1

-MS3(353.1->190.6)

-MS4(353.1->190.7->85.4)

0.0

0.5

1.0 7 x10

Intens.

0 2 4 6 4 x10

-1

0

1

2

200 400 600 800 1000 m/z


Hande Karaköse 125



-

Cis-5-caffeoylquinic acid m/z 353 [M-H+]

-

3,5-dicaffeoylquinic acid m/z 515

190.7

352.9 -MS2(515.0)

134.7

190.7 -MS3(515.3->352.9)

85.1

126.8

172.7

-MS4(515.3->353.1->190.7)

0.00 0.25 0.50

8 x10 Intens.

0

1

2

7 x10

0

1

2 5 x10

100 200 300 400 500 600 700 m/z

190.7 -MS2(352.9)

85.0 172.7

126.8 -MS3(353.1->190.7)

108.9 -MS4(353.1->190.7->126.8)

0 2 4

6 x10 Intens.

0

2

4 4 x10

0

500

1000

200 400 600 800 1000 m/z

172.7 -MS2(352.9)

71.3 154.7

93.0 -MS3(353.1->172.7)

-MS4(353.1->172.7->93.1)

0.0

0.5

7 x10 Intens.

0.0

0.5

1.0

5 x10

-1 0 1

2

200 400 600 800 1000 m/z


Hande Karaköse 126


4,5-dicaffeoylquinic acid m/z 515

Cis-4,5-dicaffeoylquinic acid m/z 515

Cis-4,5-dicaffeoylquinic acid m/z 515

172.7

352.9 -MS2(515.0)

134.8

172.7 -MS3(515.3->353.0)

71.3 93.1

110.9

154.7 -MS4(515.3->353.1->172.8)

0.0 0.5 1.0

7 x10 Intens.

0

1

2 6 x10

0.0

0.5

1.0

4 x10

100 200 300 400 500 600 700 m/z

172.7 202.8

352.9 -MS2(515.0)

134.8

172.7 -MS3(515.2->352.9)

71.3

93.0

110.9 154.7

-MS4(515.2->353.0->172.8)

0.00 0.25 0.50 0.75

7 x10 Intens.

0.0 0.5 1.0 1.5 2.0

6 x10

0

1

2

4 x10

100 200 300 400 500 600 700 m/z

172.7 202.7 254.8 298.9

352.9 -MS2(515.0)

134.8

172.7 -MS3(515.3->352.9)

71.3

93.0 110.8

154.7

-MS4(515.3->353.1->172.8)

0.00 0.25 0.50

8 x10 Intens.

0.0 0.5 1.0 1.5

7 x10

0

2

4 5 x10

100 200 300 400 500 600 700 m/z


Hande Karaköse 127


Rutin m/z 609 [M-H+]

-

Quercetin-galactoside m/z 463 [M-H+]

-

178.6

300.8

299.8

-MS2(462.9)

106.9

270.8 178.7

-MS3(463.2->302.6)

150.7

-MS4(463.2->300.7->178.7)

0.0

0.5

7 x10 Intens.

0.00 0.25 0.50 0.75

6 x10

0

2

4

4 x10

200 400 600 800 1000 m/z

178.6 270.8

300.8

342.9

299.8

-MS2(609.0)

106.9

178.7 270.7

-MS3(609.4->300.6)

184.7

243.7 -MS4(609.4->300.6->271.1)

0.0

0.5

1.0 7 x10

Intens.

0.0

0.5

1.0 6 x10

0

1

2

4 x10

200 400 600 800 1000 m/z


Hande Karaköse 128


Kaempferol-glucopyranoside; Quercetin-rhamnoside; Kaempferol-glucopyranoside;

Quercetin-rhamnoside m/z 447

178.7

300.8

299.9

-MS2(447.0)

106.8

150.7 270.7

178.6 -MS3(447.2->300.7)

106.8

150.7 -MS4(447.2->300.7->178.8)

0.00 0.25 0.50 0.75

7 x10 Intens.

0 2 4 6 5 x10

0

2

4

4 x10

200 400 600 800 1000 m/z

254.7 326.9

283.8 -MS2(446.9)

150.7 226.7

254.7 -MS3(447.1->285.8)

226.8 -MS4(447.1->284.3->255.2)

0.0 0.5

1.0 6 x10

Intens.

0

1

2

5 x10

0 1000 2000 3000 4000

200 400 600 800 1000 m/z

284.8 -MS2(446.9)

242.7

673.9

198.7 -MS3(447.1->284.9)

170.8 -MS4(447.1->284.9->198.8)

0

1

7 x10 Intens.

0.0 0.5 1.0 1.5

5 x10

0 500

1000 1500 2000

200 400 600 800 1000 m/z


Hande Karaköse 129


Kaempferol-rhamnopyranosyl-glucopyranoside(rutinoside) isomers; Quercetin-

dirhamnoside; Apigenin-diglucoside/galactoside m/z 593

Quercetin pentoside m/z 433

178.7

299.8 -MS2(432.9)

178.7

270.7 -MS3(433.2->300.5)

186.7

242.7 -MS4(433.2->300.5->271.0)

0.0

0.5

7 x10 Intens.

0

1

2

6 x10

0

1

2

4 x10

200 400 600 800 1000 m/z

300.8 -MS2(432.9)

106.9

150.7 270.7

178.7 -MS3(433.1->300.7)

107.2

150.7 -MS4(433.1->300.8->178.7)

0.0

0.5

7 x10 Intens.

0.00 0.25 0.50 0.75

6 x10

0

2

4 4 x10

200 400 600 800 1000 m/z

284.8 -MS2(446.9)

150.7

254.7 -MS3(593.3->284.5)

162.6 210.7

-MS4(593.3->284.5->255.3)

0 2 4 6 6 x10

Intens.

0.00 0.25 0.50 0.75

6 x10

0.0 0.5 1.0 1.5

4 x10

200 400 600 800 1000 m/z


Hande Karaköse 130


Kaempferol-xylosyl-glucoside; Naringin m/z 579

Apigenin-galactoside m/z 431

268.7 -MS2(430.9)

148.8 267.6

224.8 -MS3(431.2->267.8)

168.7

196.7 -MS4(431.2->268.8->225.3)

0.0

0.5

7 x10 Intens.

0 2 4 6 4 x10

0 500

1000 1500

200 400 600 800 1000 m/z

178.6 342.9 414.9 489.0 560.9

299.8 -MS2(579.0)

150.7

270.7 -MS3(579.3->300.0)

198.6

226.7 -MS4(579.3->300.0->270.9)

0 1 2 3 6 x10

Intens.

0 2 4 6

5 x10

0.00 0.25 0.50 0.75

4 x10

200 400 600 800 1000 m/z

178.6 414.9

299.8 -MS2(579.0)

178.7

270.7 -MS3(579.3->300.1)

243.8 -MS4(579.3->300.1->270.9)

0

2

4 6 x10

Intens.

0 2 4 6 8 5 x10

0 500

1000 1500

200 400 600 800 1000 m/z


Hande Karaköse 131


Kaempferol-glucosylrhamnosyl-glucoside/galactoside m/z 755

Quercetin-trisaccharide m/z 741

Kaempferol 3-rhamnopyranosyl-rhamnopyranosyl-glucopyranoside m/z 739

254.7 283.8

326.9 393.0 443.0 473.0 642.0 692.0

575.1 -MS2(739.2)

162.8 212.7 256.7

282.9 308.9

339.0 393.0

428.9

547.1 338.1

-MS3(739.5->575.2)

262.8 295.8 -MS4(739.5->575.3->339.0)

0 2

4 4 x10

Intens.

0 1000 2000 3000

0

100

200

100 200 300 400 500 600 700 m/z

299.8 461.9

579.0 -MS2(741.2)

354.9 414.9

299.8 -MS3(741.5->578.5)

270.7 -MS4(741.5->579.2->300.0)

0

1

5 x10 Intens.

0 1 2 3

4 x10

0 250 500 750

200 400 600 800 1000 m/z

284.8 367.0 469.0

593.0 -MS2(755.2)

254.7 326.9

283.8

-MS3(755.7->593.3)

150.7

254.7 -MS4(755.7->593.3->284.5)

0 1 2 3 6 x10

Intens.

0 2 4 6 5 x10

0 2 4 6 4 x10

200 400 600 800 1000 m/z


Hande Karaköse 132


Quercetin-diglucoside-rhamnoside m/z 771

300.8 469.0

609.1 -MS2(771.2)

270.8 342.9

299.8

-MS3(771.6->608.6)

106.9

150.7

210.7

270.7 -MS4(771.6->609.4->300.9)

0 2 4

6 x10 Intens.

0.0 0.5 1.0 1.5

6 x10

0.0

0.5

1.0 5 x10

200 400 600 800 1000 m/z


Hande Karaköse 133


D. Quantification data of polyhenols in stevia

Minimum, average and maximum amounts (g/100g leaves) of polyphenols in all stevia samples (average

values taken within +/- 3 σ)

Bar plot showing minimum, maximum and average values obtained from all 166 stevia samples

within +/- 3 σ

0.000

2.000

4.000

6.000

8.000

10.000

12.000

14.000

16.000

18.000

min

ave

max

3CQA 5CQA 4CQA totalmono 3,5 4,5 totaldiCQA k7g q3g totalflavones

min 0.002 0.193 0.010 0.205 0.173 0.210 0.311 0.092 0.001 0.329

ave 0.310 2.481 0.124 2.915 1.203 1.241 2.442 2.854 0.084 6.993

max 2.828 4.986 0.249 5.608 2.476 2.609 4.575 6.662 0.611 16.415

stddev 0.243 0.855 0.043 1.010 0.511 0.487 0.938 1.259 0.086 3.102

Quantity of CQAs and Flavonoid glycosides

Hande Karaköse 134


Sample No. Origin Variety Harvesting Year 3cqa

( g/100g)

5cqa

( g/100g)

4cqa

( g/100g)

Totalmono

( g/100g)

3,5diCQA

( g/100g)

4,5diCQA

( g/100g)

Total diCQA

( g/100g)

k7g

( g/100g)

q3g

( g/100g)

Total flav.

( g/100g)

1 TCV 3 II 10.09.2010 0.139 1.847 0.092 2.078 0.804 0.624 1.428 0.787 0.087 1.641

2 TCV 4 II 14.09.2010 0.260 2.680 0.134 3.074 1.281 1.052 2.333 1.967 0.093 3.518

3 TCV 3 I 04.08.2010 0.325 2.900 0.145 3.369 0.512 1.155 1.667 1.541 0.085 3.041

4 TCV 1 I 04.08.2010 0.187 2.336 0.117 2.640 0.954 0.871 1.825 1.006 0.120 2.326

5 TCV 1 I 04.08.2010 0.232 2.197 0.110 2.538 0.735 0.814 1.549 0.609 0.136 1.529

6 TCV 2 I 04.08.2010 0.235 1.806 0.090 2.132 0.592 1.061 1.653 0.899 0.074 2.130

7 TCV 2 I 04.08.2010 0.198 1.961 0.098 2.256 0.670 0.959 1.630 0.987 0.107 2.208

8 TCV 4 I 04.08.2010 0.324 3.306 0.165 3.796 0.938 1.030 1.968 2.255 0.089 3.914

9 TCV 4 II 14.09.2010 0.223 2.911 0.146 3.280 1.413 0.997 2.410 2.290 0.073 4.009

10 TCV 2 II 28.09.2010 0.089 1.102 0.055 1.246 1.178 0.814 1.992 1.575 0.099 3.302

11 TCV 2 II 28.09.2010 0.051 0.593 0.030 0.673 0.514 0.488 1.002 0.582 0.110 1.180

12 Pojava 4 II 14.09.2010 0.221 2.103 0.105 2.429 0.778 0.993 1.771 1.735 0.080 2.696

13 TCV 3 I 04.08.2010 0.319 2.508 0.125 2.952 0.562 0.840 1.403 1.135 0.044 2.207

14 TCV 2 I 04.08.2010 0.253 1.876 0.094 2.223 0.537 0.899 1.437 0.767 0.110 1.507

15 TCV 1 I 01.09.2010 0.217 1.772 0.089 2.077 0.630 0.849 1.480 0.878 0.073 2.015

16 TCV 2 I 04.08.2010 0.232 2.058 0.103 2.393 0.557 0.791 1.348 0.772 0.145 1.627

17 TCV 3 I 04.08.2010 0.290 2.361 0.118 2.769 0.574 0.884 1.459 0.998 0.047 2.246

18 TCV 4 I 04.08.2010 0.317 2.057 0.103 2.477 0.519 0.853 1.372 1.416 0.059 2.481

19 TCV 3 II 10.09.2010 0.233 2.197 0.110 2.541 0.853 0.802 1.655 1.080 0.044 2.299

20 TCV 2 II 28.09.2010 0.136 0.729 0.036 0.901 0.907 1.046 1.953 1.125 0.053 2.522

21 TCV 1 II 28.09.2010 0.041 0.684 0.034 0.759 0.247 0.237 0.484 0.225 0.134 0.626

22 TCV 3 II 10.09.2010 0.276 2.361 0.118 2.755 0.922 1.130 2.052 0.471 0.090 1.103

23 TCV,Pojana 4 I 04.08.2010 0.279 2.615 0.131 3.024 0.492 0.795 1.287 1.210 0.142 2.022

24 TCV,pojana 3 I 04.08.2010 0.257 2.355 0.118 2.730 0.504 0.786 1.290 0.421 0.112 0.958

25 TCV 1 I 01.09.2010 0.198 1.615 0.081 1.894 0.473 0.653 1.126 0.450 0.107 1.097

26 TCV 4 II 14.09.2010 0.199 1.651 0.083 1.933 0.694 0.912 1.606 1.518 0.008 2.679

27 TCV 4 I 04.08.2010 0.282 2.223 0.111 2.617 0.435 0.694 1.128 1.303 0.079 2.252

28 TCV 2 II 28.09.2010 0.220 1.833 0.092 2.145 1.331 1.476 2.807 2.254 0.044 5.026


Hande Karaköse 135


Sample No. Origin Variety Harvesting Year 3cqa 5cqa 4cqa totalmono 3,5diCQA 4,5diCQA total diCQA k7g q3g Total flav.

29 TCV 3 II 10.09.2010 0.260 2.531 0.127 2.917 0.700 0.884 1.585 1.269 0.094 2.604

30 Uconor 4 II 07.05.2010 0.249 1.630 0.081 1.961 0.606 1.000 1.606 0.798 0.106 3.707

31 Uconor 3 I 11.08.2010 0.380 3.068 0.153 3.601 1.890 1.966 3.856 4.745 0.036 11.645

32 Uconor 2 I 11.08.2010 0.253 2.320 0.116 2.689 2.014 2.561 4.575 0.889 0.415 4.503

33 Uconor 2 II 07.05.2010 0.161 1.751 0.088 2.000 2.144 1.921 4.066 1.828 0.287 6.359

34 Uconor 4 II 07.09.2010 0.341 4.344 0.217 4.903 2.129 1.544 3.673 3.495 0.265 7.592

35 Uconor 4 I 11.08.2010 0.337 4.067 0.203 4.607 1.413 1.735 3.148 2.509 0.611 8.135

36 Uconor 4 I 11.08.2010 0.262 3.586 0.179 4.027 1.190 1.987 3.178 1.943 0.445 5.796

37 Uconor 4 I 11.08.2010 0.181 1.931 0.097 2.209 0.804 0.714 1.518 2.227 0.118 6.225

38 Uconor 3 I 11.08.2010 0.347 3.096 0.155 3.598 0.887 1.108 1.995 2.644 0.521 7.170

39 Uconor 4 I 11.08.2010 0.314 3.034 0.152 3.500 1.687 1.668 3.355 4.876 0.060 13.156

40 Uconor 2 I 11.08.2010 0.291 3.099 0.155 3.544 2.476 1.916 4.392 2.894 0.079 8.897

41 Uconor 3 I 11.08.2010 0.215 1.692 0.085 1.992 0.938 1.028 1.966 2.158 0.025 5.087

42 Uconor 3 II 07.09.2010 0.364 2.681 0.134 3.179 1.803 1.342 3.145 3.136 0.064 7.190

43 Agrinion 2 II 20.09.2010 0.386 2.634 0.132 3.152 1.695 1.681 3.376 3.204 0.067 6.682

44 Toumpa 1 I 30.07.2010 0.348 3.086 0.154 3.589 2.276 2.251 4.526 6.662 0.079 16.415

45 Portugal 1 I 07.07.2010 0.117 0.908 0.045 1.071 0.273 0.569 0.842 2.883 0.096 9.525

46 Amfilia 1 I 04.08.2010 0.719 4.135 0.207 5.061 1.333 1.391 2.725 3.026 0.047 7.179

47 Toumpa 3 II 10.09.2010 0.423 2.908 0.145 3.477 1.498 1.362 2.861 3.991 0.010 11.230

48 Agrinion 2 I 09.08.2010 0.456 2.808 0.140 3.404 0.406 1.328 1.734 2.551 0.025 6.536

49 Toumpa 3 I 30.07.2010 0.413 3.810 0.191 4.414 1.311 1.583 2.893 3.599 0.037 9.241

50 Agrinion 4 I 09.08.2010 0.542 3.346 0.167 4.056 1.122 1.551 2.673 4.589 0.019 8.169

51 Agrinion 4 II 20.09.2010 0.118 1.355 0.068 1.541 0.726 0.844 1.570 4.043 0.079 7.890

52 Portugal 4 I 26.06.2010 0.267 1.706 0.085 2.059 0.637 0.460 1.098 3.828 0.023 8.718

53 Amfilia 4 I 04.08.2010 0.284 3.049 0.152 3.485 0.452 0.704 1.157 3.423 0.146 5.716

54 Amfilia 4 II 15.09.2010 0.260 2.315 0.116 2.691 0.647 0.842 1.489 4.584 0.086 7.711

55 Argentinie 2009 - - 2009 0.284 2.403 0.120 2.808 1.442 1.641 3.083 3.250 0.008 7.656

56 Paragan2009 - - 2009 0.380 3.021 0.151 3.551 1.842 1.574 3.417 4.226 0.041 8.672


Hande Karaköse 136



57 Argentinien2010 - - 2010 0.288 2.138 0.107 2.533 1.357 1.427 2.784 3.420 0.009 8.073

58 Amfilia 2 II 15.09.2010 0.439 2.440 0.122 3.001 0.655 0.755 1.410 4.145 0.040 8.847

59 Agrinion 1 II 20.09.2010 0.123 1.341 0.067 1.531 1.035 1.245 2.280 3.638 0.043 7.423

60 Amfilia 2 I 04.08.2010 0.619 4.740 0.237 5.597 1.505 1.493 2.997 3.839 0.012 9.389

61 Amfilia 3 II 15.09.2010 0.407 2.143 0.107 2.658 0.914 1.110 2.024 3.792 0.113 8.905

62 Toumpa 1 I 30.07.2010 0.444 3.439 0.172 4.055 1.262 1.781 3.043 4.508 0.008 9.830

63 Amfilia 1 II 15.09.2010 0.317 1.932 0.097 2.346 0.642 0.714 1.356 4.884 0.112 15.769

64 Portugal 3 I 26.06.2010 0.223 2.125 0.106 2.454 0.754 0.738 1.492 4.666 0.035 13.528

65 Toumpa 1 II 10.09.2010 0.289 2.705 0.135 3.129 1.443 1.019 2.461 3.851 0.002 9.324

66 Agrinion 1 I 09.08.2010 0.374 3.483 0.174 4.031 1.249 1.487 2.737 4.356 0.078 9.066

67 Amfilia 3 I 04.08.2010 0.659 3.931 0.197 4.787 1.037 1.337 2.374 4.316 0.027 9.249

68 Toumpa 2 II 10.09.2010 0.366 3.010 0.151 3.527 1.053 0.914 1.967 2.577 0.058 6.943

69 Agrinio 3 II 20.09.2010 0.186 1.465 0.073 1.724 0.821 1.047 1.868 2.284 0.001 6.544

70 Toumpa 4 II 10.09.2010 0.442 3.975 0.199 4.615 1.331 0.868 2.200 3.638 0.088 7.843

71 Agrinion 3 I 09.08.2010 0.413 3.266 0.163 3.843 1.008 1.282 2.290 3.586 0.011 9.980

72 Toumpa 4 I 30.07.2010 0.591 3.768 0.188 4.547 1.522 2.381 3.903 2.278 0.087 4.869

73 Granada 2 I 15.09.2010 0.154 1.350 0.068 1.572 0.692 0.630 1.322 2.190 0.034 7.131

74 Granada 4 I 15.09.2010 0.160 1.962 0.098 2.220 0.369 0.559 0.928 4.286 0.063 8.816

75 Granada 3 I 09.09.2010 0.214 1.671 0.084 1.969 0.173 0.511 0.684 4.275 0.068 9.516

76 Equador - - - 0.015 0.308 0.015 0.338 0.325 0.210 0.535 0.099 0.135 0.350

77 DZ - - - 0.002 0.193 0.010 0.205 0.344 0.263 0.607 1.352 0.140 3.531

78 Kenia 2010 - - 2010 0.261 2.436 0.122 2.818 1.509 1.114 2.622 1.625 0.082 6.527

79 Paraguay 2009 - - 2009 0.450 2.772 0.139 3.361 2.033 1.895 3.928 4.301 0.036 10.225

80 Argentinien 2010 - - 2010 0.209 2.762 0.138 3.109 1.889 2.059 3.948 4.459 0.096 10.975

81 Indien 2010 - - 2010 0.122 1.322 0.066 1.510 0.790 1.757 2.547 2.175 0.122 8.091

82 Argenitinien 2009 - - 2009 0.354 2.488 0.124 2.966 2.019 2.526 4.546 6.294 0.005 13.673

83 Hohenheim 2010 - - 2010 0.123 1.133 0.057 1.313 0.266 0.450 0.716 3.298 0.030

7.328


Hande Karaköse 137



84 Krim - - - 0.033 1.183 0.059 1.275 0.549 0.238 0.311 2.969 0.191 7.441

85 Turkei PS1 2010 - - 2010 0.089 1.626 0.081 1.796 0.774 1.066 1.840 3.177 0.174 8.623

86 Turkei PS2 2010 - - 2010 0.120 1.917 0.096 2.132 0.496 0.900 1.395 2.246 0.294 5.762

87 Uconor 7 I 13.07.2011 0.365 3.645 0.182 4.193 1.786 2.045 3.831 5.159 0.026 9.928

88 Agrinon 4 I 2011 0.431 3.399 0.170 4.000 1.203 1.461 2.664 1.512 0.067 3.939

89 Uconor 4 I 2011 0.473 3.279 0.164 3.916 1.720 1.900 3.620 3.915 0.054 9.788

90 Agrinion 3 I 2011 0.398 3.040 0.152 3.590 1.245 1.436 2.681 3.913 0.056 10.088

91 Agrinion 5 I 2011 0.463 3.208 0.160 3.832 1.174 1.400 2.574 3.030 0.047 8.126

92 Uconor 6 I 13.07.2011 0.561 4.671 0.234 5.465 1.947 2.290 4.237 0.092 0.048 0.329

93 Uconor 3 I 13.07.2011 0.372 4.986 0.249 5.608 1.664 2.022 3.686 3.102 0.088 7.721

94 Uconor 5 I 13.07.2011 0.520 3.282 0.164 3.967 1.425 1.588 3.013 1.764 0.141 5.377

95 Agrinion 6 I 2011 2.828 1.225 0.061 4.114 1.592 1.956 3.548 2.220 0.111 6.236

96 Toumpa 1 I 2011 0.479 3.600 0.180 4.260 1.321 1.578 2.899 3.210 0.059 8.628

97 Toumpa 2 I 2011 0.495 3.203 0.160 3.858 0.909 1.137 2.046 3.659 0.061 10.518

98 Toumpa 3 I 2011 0.491 4.099 0.205 4.795 1.208 1.535 2.743 3.972 0.002 10.938

99 Toumpa 4 I 2011 0.461 2.879 0.144 3.484 1.883 2.052 3.935 3.709 0.091 7.981

100 Amiflikeia 4 II 2011 0.595 3.762 0.188 4.544 1.193 1.435 2.628 3.610 0.058 7.079

101 Amiflikeia 5 II 2011 0.434 2.756 0.138 3.328 0.968 1.410 2.378 3.248 0.163 7.859

102 Amiflikeia 4 II 2011 0.371 2.363 0.118 2.852 1.528 1.556 3.084 3.852 0.092 7.591

103 Amiflikeia 3 II 2011 0.479 2.814 0.141 3.433 1.623 1.790 3.413 3.248 0.022 7.671

104 Amiflikeia - II 2011 0.458 2.723 0.136 3.317 1.125 1.161 2.286 3.309 0.123 7.305

105 Amiflikeia 6 II 2011 0.451 2.574 0.129 3.154 1.099 1.166 2.264 2.884 0.073 7.167

106 Amiflikeia 1 II 2011 0.535 2.200 0.110 2.845 1.349 1.416 2.765 2.641 0.019 6.327

107 APTTB 3 I 2011 0.246 1.663 0.083 1.992 0.721 0.603 1.324 3.675 0.041 9.730

108 APTTB 5 I 2011 0.192 1.249 0.062 1.503 0.705 0.703 1.409 3.821 0.096 11.274

109 APTTB 6 I - 0.180 1.177 0.059 1.416 0.756 0.701 1.457 4.118 0.018 10.068

110 APTTB 4 I - 0.206 1.362 0.068 1.637 0.800 0.803 1.603 3.516 0.055 10.562

111 Uconor 5 II 2011 0.237 2.303 0.115 2.656 1.837 1.190 3.027 3.066 0.009 8.264


Hande Karaköse 138



112 Uconor 4 II 2011 0.293 2.063 0.103 2.459 1.406 1.275 2.681 3.578 0.035 7.598

113 Uconor 3 II 2011 0.300 3.108 0.155 3.564 1.553 1.525 3.077 2.840 0.054 6.933

114 Uconor 6 II 2011 0.301 2.170 0.109 2.580 1.482 1.228 2.711 3.110 0.025 8.264

115 Conaga 3 I 2011 0.356 2.621 0.131 3.107 0.901 1.506 2.408 3.041 0.030 7.221

116 Conaga 5 I 2011 0.250 2.422 0.121 2.792 1.216 1.422 2.638 2.195 0.097 6.593

117 Conaga 4 I 2011 0.354 2.309 0.115 2.778 1.045 1.524 2.569 3.227 0.067 7.027

118 Agrinion 5 II 2011 0.368 2.410 0.121 2.899 1.495 1.379 2.874 3.227 0.051 9.412

119 Agrinion 4 II 2011 0.358 3.144 0.157 3.659 1.433 1.514 2.947 3.564 0.068 7.308

120 Conaga 6 I 2011 0.326 3.301 0.165 3.792 1.295 2.068 3.362 3.428 0.088 10.269

121 Conaga 7 I 2011 0.262 2.747 0.137 3.147 0.507 1.189 1.696 0.637 0.225 2.064

122 Agrinion 6 II 2011 0.296 2.461 0.123 2.880 1.434 1.364 2.798 2.788 0.009 7.335

123 Agrinion 3 II 2011 0.247 2.626 0.131 3.004 1.570 1.342 2.913 4.336 0.019 9.887

124 Toumpa 7 I 2011 0.061 2.240 0.112 2.414 1.528 0.575 2.103 1.873 0.140 5.930

125 Toumpa 4 III 2011 0.148 1.884 0.094 2.126 1.311 1.310 2.621 3.680 0.060 8.463

126 Toumpa 3 III 2011 0.187 1.961 0.098 2.246 1.684 1.567 3.251 3.510 0.018 9.388

127 Toumpa 3 II 2011 0.175 2.490 0.124 2.789 1.650 1.271 2.921 3.769 0.013 9.366

128 Agrinion 4 II 2011 0.246 3.222 0.161 3.630 1.500 1.460 2.960 3.386 0.005 6.759

129 Toumpa 4 II 2011 0.155 2.863 0.143 3.161 1.725 1.426 3.151 4.079 0.142 8.424

130 Toumpa 3 II 2011 0.140 2.078 0.104 2.322 1.573 1.464 3.037 3.937 0.025 10.501

131 Toumpa 4 II 2011 0.112 2.289 0.114 2.515 1.321 1.288 2.609 4.684 0.057 9.631

132 Agrinion 3 II 2011 0.258 2.593 0.130 2.980 1.925 1.922 3.846 3.585 0.019 8.969

133 Toumpa 5 II 2011 0.134 2.375 0.119 2.628 1.732 1.454 3.186 3.827 0.092 11.186

134 Toumpa 6 II 2011 0.130 2.536 0.127 2.792 1.445 0.953 2.398 3.424 0.008 9.629

135 Toumpa 7 II 2011 0.116 2.285 0.114 2.516 0.850 0.657 1.507 1.442 0.179 4.449

136 TCV 5 I 30.06.2011 0.531 3.343 0.167 4.042 1.328 1.185 2.513 2.618 0.076 8.134

137 TCV 3 I 30.06.2011 0.441 3.153 0.158 3.752 1.578 1.266 2.844 2.870 0.034 7.671

138 TCV 6 I 30.06.2011 0.526 4.032 0.202 4.760 1.807 1.394 3.201 3.718 0.031 10.049

139 TCV 3 II 11.08.2011 0.267 2.692 0.135 3.094 1.854 1.155 3.010 3.503 0.068 8.909


Hande Karaköse 139



140 TCV 6 II 17.08.2011 0.289 2.399 0.120 2.808 1.679 1.520 3.200 3.105 0.062 8.506

141 TCV 4 II 24.08.2011 0.376 2.375 0.119 2.869 1.563 2.609 4.171 2.493 0.192 5.320

142 TCV 7 I 18.08.2011 0.084 1.777 0.089 1.950 1.715 1.412 3.127 1.711 0.208 4.790

143 TCV 4 I 07.07.2011 0.340 1.843 0.092 2.275 1.070 1.232 2.302 2.974 0.009 6.305

144 TCV 5 II 17.08.2011 0.385 2.323 0.116 2.824 1.933 2.474 4.407 2.781 0.031 7.833

145 Amiflikeia 3 I 2011 0.501 2.934 0.147 3.582 0.433 0.928 1.360 2.114 0.075 5.301

146 Amiflikeia 4 I 2011 0.554 3.595 0.180 4.329 1.182 0.989 2.170 2.013 0.128 4.423

147 Amiflikeia 5 I 2011 0.628 3.086 0.154 3.869 1.326 1.124 2.450 2.036 0.001 5.264

148 Amiflikeia 6 I 2011 0.524 3.398 0.170 4.092 1.353 0.735 2.089 2.035 0.044 5.307

149 Amiflikeia 1 I 2011 0.485 3.690 0.184 4.359 1.836 1.674 3.510 2.939 0.052 8.505

150 Amiflikeia 2 I 2011 0.399 2.798 0.140 3.337 2.183 1.078 3.262 2.966 0.048 7.316

151 Amiflikeia 4 I 2011 0.358 2.465 0.123 2.946 1.903 1.234 3.138 4.101 0.011 8.871

152 Amiflikeia 3 I 2011 0.335 2.154 0.108 2.596 1.842 1.119 2.961 3.351 0.086 8.778

153 Toumpa 3 I 2011 0.307 2.097 0.105 2.508 1.939 1.677 3.616 3.622 0.001 10.033

154 Toumpa 5 I 2011 0.242 3.266 0.163 3.672 1.694 1.281 2.975 2.500 0.099 6.702

155 Toumpa 6 I 2011 0.127 2.546 0.127 2.800 1.054 0.770 1.824 1.953 0.068 5.836

156 Toumpa 4 I 2011 0.242 2.434 0.122 2.797 2.178 1.794 3.972 4.547 0.092 9.592

157 Toumpa 4 I 2011 0.307 2.551 0.128 2.985 1.560 1.357 2.918 3.952 0.065 8.032

158 Toumpa 3 I 2011 0.171 2.763 0.138 3.072 1.485 0.871 2.356 3.479 0.095 8.160

159 TCV 6 III 2011 0.063 1.872 0.094 2.029 1.716 1.789 3.505 3.010 0.036 7.856

160 TCV 5 III 2011 0.147 1.552 0.078 1.777 1.622 1.176 2.798 2.464 0.091 6.536

161 TCV 3 III 2011 0.161 1.777 0.089 2.026 0.987 1.077 2.064 2.300 0.157 5.554

162 Agrinion 3 III 2011 0.168 1.372 0.069 1.608 0.974 0.938 1.912 3.132 0.041 8.298

163 TCV 4 III 2011 0.120 2.087 0.104 2.311 1.110 0.883 1.993 2.346 0.194 4.654

164 Agrinion 6 III 2011 0.148 2.041 0.102 2.291 1.489 1.147 2.636 4.011 0.006 10.699

165 Agrinion 5 III 2011 0.124 1.897 0.095 2.116 1.269 0.794 2.063 3.501 0.064 9.157

166 Agrinion 4 III 2011 0.179 1.837 0.092 2.108 0.982 0.865 1.847 4.156 0.118 8.261

Quantity of Trans & Cis-CQAs

Hande Karaköse 140


Sample No. Origin Variety Harvesting 3cqa

( g/100g)

5cqa

( g/100g)

4cqa

( g/100g)

Cis-5CQA

(g/100g)

3,5diCQA

( g/100g)

4,5diCQA

( g/100g)

Cis-4,5diCQA

(g/100g)

Cis-4,5diCQA

(g/100g)

1 TCV 3 II 0.139 1.847 0.092 0.313 0.804 0.624 0.002 0.022

2 TCV 4 II 0.260 2.680 0.134 0.516 1.281 1.052 0.011 0.037

3 TCV 3 I 0.325 2.900 0.145 1.153 0.512 1.155 0.009 0.035

4 TCV 1 I 0.187 2.336 0.117 0.060 0.954 0.871 0.009 0.045

5 TCV 1 I 0.232 2.197 0.110 0.518 0.735 0.814 0.006 0.024

6 TCV 2 I 0.235 1.806 0.090 0.590 0.592 1.061 0.009 0.030

7 TCV 2 I 0.198 1.961 0.098 0.593 0.670 0.959 0.007 0.028

8 TCV 4 I 0.324 3.306 0.165 1.123 0.938 1.030 0.016 0.031

9 TCV 4 II 0.223 2.911 0.146 0.657 1.413 0.997 0.006 0.026

10 TCV 2 II 0.089 1.102 0.055 0.121 1.178 0.814 0.006 0.018

11 TCV 2 II 0.051 0.593 0.030 0.023 0.514 0.488 0.023 0.057

12 Pojava 4 II 0.221 2.103 0.105 0.112 0.778 0.993 0.006 0.009

13 TCV 3 I 0.319 2.508 0.125 0.102 0.562 0.840 0.035 0.061

14 TCV 2 I 0.253 1.876 0.094 0.110 0.537 0.899 0.010 0.013

15 TCV 1 I 0.217 1.772 0.089 0.081 0.630 0.849 0.009 0.013

16 TCV 2 I 0.232 2.058 0.103 0.076 0.557 0.791 0.059 0.089

17 TCV 3 I 0.290 2.361 0.118 0.126 0.574 0.884 0.033 0.043

18 TCV 4 I 0.317 2.057 0.103 0.134 0.519 0.853 0.008 0.018

19 TCV 3 II 0.233 2.197 0.110 0.056 0.853 0.802 0.034 0.048

20 TCV 2 II 0.136 0.729 0.036 0.042 0.907 1.046 0.005 0.012

21 TCV 1 II 0.041 0.684 0.034 0.001 0.247 0.237 0.020 0.028

22 TCV 3 II 0.276 2.361 0.118 0.110 0.922 1.130 0.010 0.005

23 TCV,Pojana 4 I 0.279 2.615 0.131 0.100 0.492 0.795 0.051 0.078

24 TCV,pojana 3 I 0.257 2.355 0.118 0.101 0.504 0.786 0.013 0.024

25 TCV 1 I 0.198 1.615 0.081 0.065 0.473 0.653 0.037 0.051

26 TCV 4 II 0.199 1.651 0.083 0.073 0.694 0.912 0.007 0.018

27 TCV 4 I 0.282 2.223 0.111 0.113 0.435 0.694 0.014 0.013

28 TCV 2 II 0.220 1.833 0.092 0.070 1.331 1.476 0.010 0.014


Hande Karaköse 141


Sample No. Origin Variety Harvesting 3cqa 5cqa 4cqa cis 5CQA 3,5diCQA 4,5diCQA Cis 4,5-diCQA

(g/100g)

Cis 4,5-diCQA

(g/100g)

29 TCV 3 II 0.260 2.531 0.127 0.090 0.700 0.884 0.018 0.018

30 Uconor 4 II 0.249 1.630 0.081 0.075 0.606 1.000 0.004 0.004

31 Uconor 3 I 0.380 3.068 0.153 0.021 1.890 1.966 0.011 0.011

32 Uconor 2 I 0.253 2.320 0.116 0.037 2.014 2.561 0.023 0.023

33 Uconor 2 II 0.161 1.751 0.088 0.015 2.144 1.921 0.026 0.026

34 Uconor 4 II 0.341 4.344 0.217 0.043 2.129 1.544 0.035 0.035

35 Uconor 4 I 0.337 4.067 0.203 0.067 1.413 1.735 0.006 0.006

36 Uconor 4 I 0.262 3.586 0.179 0.019 1.190 1.987 0.008 0.008

37 Uconor 4 I 0.181 1.931 0.097 0.017 0.804 0.714 0.009 0.009

38 Uconor 3 I 0.347 3.096 0.155 0.023 0.887 1.108 0.016 0.016

39 Uconor 4 I 0.314 3.034 0.152 0.013 1.687 1.668 0.007 0.007

40 Uconor 2 I 0.291 3.099 0.155 0.018 2.476 1.916 0.016 0.016

41 Uconor 3 I 0.215 1.692 0.085 0.029 0.938 1.028 0.013 0.013

42 Uconor 3 II 0.364 2.681 0.134 0.198 1.803 1.342 0.017 0.017

43 Agrinion 2 II 0.386 2.634 0.132 0.138 1.695 1.681 0.045 0.045

44 Toumpa 1 I 0.348 3.086 0.154 0.006 2.276 2.251 0.026 0.026

45 Portugal 1 I 0.117 0.908 0.045 0.028 0.273 0.569 0.004 0.004

46 Amfilia 1 I 0.719 4.135 0.207 0.242 1.333 1.391 0.044 0.044

47 Toumpa 3 II 0.423 2.908 0.145 0.188 1.498 1.362 0.017 0.017

48 Agrinion 2 I 0.456 2.808 0.140 0.128 0.406 1.328 0.029 0.029

49 Toumpa 3 I 0.413 3.810 0.191 0.006 1.311 1.583 0.025 0.025

50 Agrinion 4 I 0.542 3.346 0.167 0.204 1.122 1.551 0.029 0.029

51 Agrinion 4 II 0.118 1.355 0.068 0.040 0.726 0.844 0.018 0.009

52 Portugal 4 I 0.267 1.706 0.085 0.048 0.637 0.460 0.009 0.008

53 Amfilia 4 I 0.284 3.049 0.152 0.113 0.452 0.704 0.009 0.009

54 Amfilia 4 II 0.260 2.315 0.116 0.085 0.647 0.842 0.040 0.018

55 Argentinie 2009 - - 0.284 2.403 0.120 0.038 1.442 1.641 0.005 0.017

56 Paragan2009 - - 0.380 3.021 0.151 0.108 1.842 1.574 0.022 0.024


Hande Karaköse 142



(g/100g)

Cis 4,5-diCQA

(g/100g)

57 Argentinien2010 - - 0.288 2.138 0.107 0.035 1.357 1.427 0.013 0.042

58 Amfilia 2 II 0.439 2.440 0.122 0.107 0.655 0.755 0.012 0.014

59 Agrinion 1 II 0.123 1.341 0.067 0.002 1.035 1.245 0.005 0.025

60 Amfilia 2 I 0.619 4.740 0.237 0.229 1.505 1.493 0.016 0.016

61 Amfilia 3 II 0.407 2.143 0.107 0.087 0.914 1.110 0.017 0.051

62 Toumpa 1 I 0.444 3.439 0.172 0.104 1.262 1.781 0.022 0.025

63 Amfilia 1 II 0.317 1.932 0.097 0.041 0.642 0.714 0.007 0.021

64 Portugal 3 I 0.223 2.125 0.106 0.042 0.754 0.738 0.008 0.012

65 Toumpa 1 II 0.289 2.705 0.135 0.127 1.443 1.019 0.006 0.012

66 Agrinion 1 I 0.374 3.483 0.174 0.155 1.249 1.487 0.038 0.032

67 Amfilia 3 I 0.659 3.931 0.197 0.078 1.037 1.337 0.013 0.004

68 Toumpa 2 II 0.366 3.010 0.151 0.093 1.053 0.914 0.010 0.003

69 Agrinio 3 II 0.186 1.465 0.073 0.036 0.821 1.047 0.008 0.020

70 Toumpa 4 II 0.442 3.975 0.199 0.008 1.331 0.868 0.015 0.016

71 Agrinion 3 I 0.413 3.266 0.163 0.174 1.008 1.282 0.021 0.056

72 Toumpa 4 I 0.591 3.768 0.188 0.093 1.522 2.381 0.042 0.114

73 Granada 2 I 0.154 1.350 0.068 0.043 0.692 0.630 0.009 0.017

74 Granada 4 I 0.160 1.962 0.098 0.058 0.369 0.559 0.001 0.004

75 Granada 3 I 0.214 1.671 0.084 0.024 0.173 0.511 0.007 0.013

76 Equador - - 0.015 0.308 0.015 0.005 0.325 0.210 #DIV/0! #DIV/0!

77 DZ - - 0.002 0.193 0.010 0.001 0.344 0.263 #DIV/0! #DIV/0!

78 Kenia 2010 - - 0.261 2.436 0.122 0.030 1.509 1.114 0.009 0.029

79 Paraguay 2009 - - 0.450 2.772 0.139 0.082 2.033 1.895 0.060 0.092

80 Argentinien 2010 - - 0.209 2.762 0.138 0.004 1.889 2.059 0.025 0.059

81 Indien 2010 - - 0.122 1.322 0.066 0.022 0.790 1.757 0.032 0.063

82 Argenitinien 2009 - - 0.354 2.488 0.124 0.022 2.019 2.526 0.011 0.030

83 Hohenheim 2010 - - 0.123 1.133 0.057 0.027 0.266 0.450 0.000 0.001


Hande Karaköse 143



(g/100g)

Cis 4,5-diCQA

(g/100g)

84 Krim - - 0.033 1.183 0.059 0.020 0.549 0.238 -0.015 0.000

85 Turkei PS1 2010 - - 0.089 1.626 0.081 0.023 0.774 1.066 0.008 0.040

86 Turkei PS2 2010 - - 0.120 1.917 0.096 0.054 0.496 0.900 0.019 0.018

87 Uconor 7 I 0.365 3.645 0.182 0.182 1.786 2.045 0.010 0.014

88 Agrinon 4 I 0.431 3.399 0.170 0.129 1.203 1.461 0.023 0.013

89 Uconor 4 I 0.473 3.279 0.164 0.177 1.720 1.900 0.011 0.018

90 Agrinion 3 I 0.398 3.040 0.152 0.009 1.245 1.436 0.007 0.012

91 Agrinion 5 I 0.463 3.208 0.160 0.158 1.174 1.400 0.015 0.017

92 Uconor 6 I 0.561 4.671 0.234 0.251 1.947 2.290 0.022 0.007

93 Uconor 3 I 0.372 4.986 0.249 0.208 1.664 2.022 0.038 0.013

94 Uconor 5 I 0.520 3.282 0.164 0.196 1.425 1.588 0.050 0.065

95 Agrinion 6 I 2.828 1.225 0.061 0.132 1.592 1.956 0.014 0.002

96 Toumpa 1 I 0.479 3.600 0.180 0.163 1.321 1.578 0.034 0.041

97 Toumpa 2 I 0.495 3.203 0.160 0.148 0.909 1.137 0.048 0.146

98 Toumpa 3 I 0.491 4.099 0.205 0.175 1.208 1.535 0.029 0.075

99 Toumpa 4 I 0.461 2.879 0.144 0.183 1.883 2.052 0.038 0.094

100 Amiflikeia 4 II 0.595 3.762 0.188 0.008 1.193 1.435 0.024 0.033

101 Amiflikeia 5 II 0.434 2.756 0.138 0.149 0.968 1.410 0.005 0.013

102 Amiflikeia 4 II 0.371 2.363 0.118 0.128 1.528 1.556 0.022 0.059

103 Amiflikeia 3 II 0.479 2.814 0.141 0.145 1.623 1.790 0.045 0.112

104 Amiflikeia - II 0.458 2.723 0.136 0.140 1.125 1.161 0.033 0.027

105 Amiflikeia 6 II 0.451 2.574 0.129 0.138 1.099 1.166 0.007 0.010

106 Amiflikeia 1 II 0.535 2.200 0.110 0.128 1.349 1.416 0.030 0.105

107 APTTB 3 I 0.246 1.663 0.083 0.064 0.721 0.603 0.003 0.001

108 APTTB 5 I 0.192 1.249 0.062 0.058 0.705 0.703 0.008 0.018

109 APTTB 6 I 0.180 1.177 0.059 0.053 0.756 0.701 0.010 0.034

110 APTTB 4 I 0.206 1.362 0.068 0.057 0.800 0.803 0.014 0.011

111 Uconor 5 II 0.237 2.303 0.115 0.122 1.837 1.190 0.022 0.040


Hande Karaköse 144



(g/100g)

Cis 4,5-diCQA

(g/100g)

112 Uconor 4 II 0.293 2.063 0.103 0.136 1.406 1.275 0.029 0.058

113 Uconor 3 II 0.300 3.108 0.155 0.146 1.553 1.525 0.029 0.026

114 Uconor 6 II 0.301 2.170 0.109 0.113 1.482 1.228 0.020 0.043

115 Conaga 3 I 0.356 2.621 0.131 0.138 0.901 1.506 0.028 0.018

116 Conaga 5 I 0.250 2.422 0.121 0.157 1.216 1.422 0.018 0.059

117 Conaga 4 I 0.354 2.309 0.115 0.088 1.045 1.524 0.043 0.042

118 Agrinion 5 II 0.368 2.410 0.121 0.135 1.495 1.379 0.021 0.043

119 Agrinion 4 II 0.358 3.144 0.157 0.125 1.433 1.514 0.033 0.074

120 Conaga 6 I 0.326 3.301 0.165 0.004 1.295 2.068 0.104 0.166

121 Conaga 7 I 0.262 2.747 0.137 0.177 0.507 1.189 0.023 0.021

122 Agrinion 6 II 0.296 2.461 0.123 0.103 1.434 1.364 0.025 0.032

123 Agrinion 3 II 0.247 2.626 0.131 0.126 1.570 1.342 0.012 0.011

124 Toumpa 7 I 0.061 2.240 0.112 0.099 1.528 0.575 0.013 0.007

125 Toumpa 4 III 0.148 1.884 0.094 0.094 1.311 1.310 0.055 0.048

126 Toumpa 3 III 0.187 1.961 0.098 0.082 1.684 1.567 0.019 0.064

127 Toumpa 3 II 0.175 2.490 0.124 0.108 1.650 1.271 0.018 0.007

128 Agrinion 4 II 0.246 3.222 0.161 0.197 1.500 1.460 0.048 0.095

129 Toumpa 4 II 0.155 2.863 0.143 0.113 1.725 1.426 0.028 0.024

130 Toumpa 3 II 0.140 2.078 0.104 0.112 1.573 1.464 0.036 0.033

131 Toumpa 4 II 0.112 2.289 0.114 0.100 1.321 1.288 0.014 0.022

132 Agrinion 3 II 0.258 2.593 0.130 0.106 1.925 1.922 0.032 0.080

133 Toumpa 5 II 0.134 2.375 0.119 0.006 1.732 1.454 0.011 0.009

134 Toumpa 6 II 0.130 2.536 0.127 0.145 1.445 0.953 0.025 0.015

135 Toumpa 7 II 0.116 2.285 0.114 0.155 0.850 0.657 0.020 0.004

136 TCV 5 I 0.531 3.343 0.167 0.121 1.328 1.185 0.027 0.027

137 TCV 3 I 0.441 3.153 0.158 0.135 1.578 1.266 0.033 0.029

138 TCV 6 I 0.526 4.032 0.202 0.164 1.807 1.394 0.033 0.029

139 TCV 3 II 0.267 2.692 0.135 0.147 1.854 1.155 0.008 0.015


Hande Karaköse 145



(g/100g)

Cis 4,5-diCQA

(g/100g)

140 TCV 6 II 0.289 2.399 0.120 0.131 1.679 1.520 0.029 0.091

141 TCV 4 II 0.376 2.375 0.119 0.084 1.563 2.609 0.048 0.116

142 TCV 7 I 0.084 1.777 0.089 0.112 1.715 1.412 0.023 0.038

143 TCV 4 I 0.340 1.843 0.092 0.120 1.070 1.232 0.022 0.028

144 TCV 5 II 0.385 2.323 0.116 0.004 1.933 2.474 0.024 0.039

145 Amiflikeia 3 I 0.501 2.934 0.147 0.192 0.433 0.928 0.001 0.024

146 Amiflikeia 4 I 0.554 3.595 0.180 0.206 1.182 0.989 0.021 0.034

147 Amiflikeia 5 I 0.628 3.086 0.154 0.199 1.326 1.124 0.017 0.007

148 Amiflikeia 6 I 0.524 3.398 0.170 0.189 1.353 0.735 0.029 0.012

149 Amiflikeia 1 I 0.485 3.690 0.184 0.093 1.836 1.674 0.032 0.040

150 Amiflikeia 2 I 0.399 2.798 0.140 0.115 2.183 1.078 0.012 0.000

151 Amiflikeia 4 I 0.358 2.465 0.123 0.174 1.903 1.234 0.008 0.017

152 Amiflikeia 3 I 0.335 2.154 0.108 0.115 1.842 1.119 0.014 0.014

153 Toumpa 3 I 0.307 2.097 0.105 0.094 1.939 1.677 0.027 0.034

154 Toumpa 5 I 0.242 3.266 0.163 0.138 1.694 1.281 0.007 0.007

155 Toumpa 6 I 0.127 2.546 0.127 0.127 1.054 0.770 0.008 0.002

156 Toumpa 4 I 0.242 2.434 0.122 0.122 2.178 1.794 0.008 0.008

157 Toumpa 4 I 0.307 2.551 0.128 0.166 1.560 1.357 0.021 0.021

158 Toumpa 3 I 0.171 2.763 0.138 0.117 1.485 0.871 0.007 0.007

159 TCV 6 III 0.063 1.872 0.094 0.006 1.716 1.789 0.030 0.030

160 TCV 5 III 0.147 1.552 0.078 0.008 1.622 1.176 0.024 0.024

161 TCV 3 III 0.161 1.777 0.089 0.084 0.987 1.077 0.008 0.008

162 Agrinion 3 III 0.168 1.372 0.069 0.099 0.974 0.938 0.017 0.017

163 TCV 4 III 0.120 2.087 0.104 0.084 1.110 0.883 0.011 0.011

164 Agrinion 6 III 0.148 2.041 0.102 0.099 1.489 1.147 0.028 0.028

165 Agrinion 5 III 0.124 1.897 0.095 0.107 1.269 0.794 0.015 0.015

166 Agrinion 4 III 0.179 1.837 0.092 0.093 0.982 0.865 0.011 0.011


Hande Karaköse 146


E. Correlation & data distribution of CQAs

Correlation table of cis and trans CQAs

cis5CQA cis45diCQa1 cis45diCQa2 cqa5 diCQA45

Spearman's

rho

cis5CQA

Correlation Coefficient 1.000 0.265** 0.183* -0.111 -0.075

Sig. (2-tailed) . 0.003 0.041 0.217 0.407

N 126 126 126 126 126

cis45diCQa1

Correlation Coefficient 0.265** 1.000 0.731** -0.012 -0.056

Sig. (2-tailed) 0.003 . 0.000 0.892 0.532

N 126 126 126 126 126

cis45diCQa2

Correlation Coefficient 0.183* 0.731** 1.000 0.035 -0.026

Sig. (2-tailed) 0.041 0.000 . 0.694 0.771

N 126 126 126 126 126

cqa5

Correlation Coefficient -0.111 -0.012 0.035 1.000 0.586**

Sig. (2-tailed) 0.217 0.892 0.694 . 0.000

N 126 126 126 126 126

diCQA45

Correlation Coefficient -0.075 -0.056 -0.026 0.586** 1.000

Sig. (2-tailed) 0.407 0.532 0.771 0.000 .

N 126 126 126 126 126

**. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.05 level (2-tailed).


Hande Karaköse 147


F. Structures of literature reported terpenes in Stevia rebaudiana

OOH

myrtenol myrtenal

OH

pinocarveol alpha pinene beta pinene

sabinene terpinene

OH

terpinen-4-ol

OH

verbenol

O

cumin aldehyde cymene limonene

OH

linalool

OH

geraniol

MeO

anethole

OH

borneol

O

camphor

OH

carvacrol

O

1,8 cineol myrcene

3-carene


Hande Karaköse 148


nerolidol

OH

selinene

copaenecubebene beta elemene

germacrene D humulene (caryophyllene)

beta-trans-farnesene

bergamotenebisabolene

gamma cadinene delta cadinene

HO

alpha cadinolcalacorene calamenene

H

H

bourbonene


Hande Karaköse 149


PUBLICATIONS

1. H. Karaköse, R. Jaiswal, N. Kuhnert “Characterisation and quantification of hydroxycinnamate

derivatives in Stevia rebaudiana leaves by LC-MSn” J. Agric. Food Chem., 2011, 59 (18),

10143–10150.

2. N. Kuhnert, F. Dairpoosh, R. Jaiswal, M. Matei, S.Deshpande, A.Golon, H. Nour, H.

Karaköse, N. Hourani “Hill coefficients of dietary polyphenolic enzyme inhibitiors: can

beneficial health effects of dietary polyphenols be explained by allosteric enzyme denaturing?” J

Chem Biol. 2011 July; 4(3): 109–116.

3. H. Karaköse, N. Kuhnert “Profiling the chlorogenic acids of Stevia rebaudiana by tandem LC-

MS” Polyphenol Commun. 2010, Vol 1, 544-546. 71.

4. N. Kuhnert, H. Karaköse, R. Jaiswal, “Analysis, characterization and pharmacokinetics of

dietary hydroxycinnamates” Invited review chapter in CRC Handbook of Food Analysis,

manuscript in press

5. G. Mikutis, H. Karaköse, R. Jaiswal, A. Le Gresley, T. Islam, M. Fernandez-Lahore, N.

Kuhnert “Phenolic promiscuity in the cell nucleus” Food and function, in Press, 2012

6. H. Karaköse, R. Jaiswal, S. Deshpande, N. Kuhnert “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” (Manuscript)

7. H. Karaköse, N. Kuhnert “Development of a LC-ESI MS method for the identification and

quantification of steviol glycosides” (Manuscript)

8. H. Karaköse, A. Golon, N. Kuhnert “Gas chromatographic analysis of lipids and volatile

terpenes in Stevia rebaudiana” (In preparation)

9. H. Karaköse, R. Shah, N. Kuhnert “Identification of proteins by MALDI-TOF MS in Stevia

rebaudiana” (In preparation)

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Kuhnert%2BN%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Dairpoosh%2BF%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Jaiswal%2BR%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Matei%2BM%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Deshpande%2BS%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Golon%2BA%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Nour%2BH%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Karak%26%23x000f6%3Bse%2BH%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Karak%26%23x000f6%3Bse%2BH%5bauth%5d

http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=search&db=PubMed&term=%20Hourani%2BN%5bauth%5d

HANDE KARAKÖSE

Address: Clamersdorfer str. 21

28757 - Bremen/Germany

Personal Details

Date of Birth 01.11.1985

Place of Birth Ankara, Turkey

Email [email protected]

Mobile +4915208494817

Education

2009 – 2012 PhD in Chemistry Jacobs University Bremen, Germany/Bremen Title: Chemical Profiling of Stevia Rebaudiana Bertoni

• Projects: Identification and profiling of all primary and secondary metabolites of stevia using HPLC-MS, MALDI-TOF and GC-MS. Method development for analysis and quantification of selected compounds and the affect of the growth conditions to the metabolite profile. Extraction of secondary metabolites. Protein extraction and isolation of stevia and identification by MALDI-TOF Lipid and volatile terpene profiling by GC-MS. Statistical analysis (e.g. PCA, ANOVA) of the dataset obtained by LC-MS. Solid phase extraction.

2007 - 2009 Master in Nanomolecular Science Jacobs University Bremen, Germany/Bremen Title: Profiling and Characterisation of Chlorogenic Acids by LC-MSn

• Projects: Optimization of the extraction technique and determination of the main chlorogenic acids in various coffee, plum and potato samples by using an HPLC-MSn method. Isolation of selected chlorogenic acids by preparative LC and identification of novel chlorogenic acids Modelling of chlorogenic acids by computational chemistry methods Comparison of experimental spectroscopic data (NMR chemical shifts, Raman spectra, IR) with the calculated spectrical data obtained by quantum mechanical calculations.

2003 - 2007 Bachelor in Chemistry University of Dokuz Eylül Faculty of Science & Arts, İzmir Title: Precontration and Solid Extraction of Uranium (VI) from various water samples using N,N-Dibutyl-N`-Benzoylthiourea

Achievements & Awards

2007 Graduation with distinction; DEÜ, Faculty of Science & Arts, Chemistry Department

2007 Full Scholarship for Master Education in Jacobs University Bremen

2009 Fellowship for PhD in Jacobs University Bremen

Articles H. Karaköse, R. Jaiswal, N. Kuhnert “Characterisation and quantification of hydroxycinnamate derivatives in Stevia Rebaudiana leaves by LC-MSn” J. Agric. Food Chem., 2011, 59 (18), 10143–10150.

N. Kuhnert, F. Dairpoosh, R. Jaiswal, M. Matei, S.Deshpande, A.Golon, H. Nour, H.

Karaköse, N. Hourani “Hill coefficients of dietary polyphenolic enzyme inhibitiors: can

beneficial health effects of dietary polyphenols be explained by allosteric enzyme

denaturing?” J Chem Biol. 2011 July; 4(3): 109–116.

H. Karaköse, N. Kuhnert “Profiling the chlorogenic acids of Stevia Rebaudiana by tandem LC-MS” Polyphenol Commun. 2010, 1, 544-546. 71

G. Mikutis, H. Karaköse, R. Jaiswal, A. Le Gresley, T. Islam, M. Fernandez-Lahore, N. Kuhnert “Phenolic promiscuity in the cell nucleus” Food and function, in Press, 2012.

H. Karaköse, N. Kuhnert “Development of a LC-ESI MS method for the identification

and quantification of steviol glycosides” (Manuscript)

Books N. Kuhnert, H. Karaköse, R. Jaiswal, “Analysis, characterization and pharmacokinetics of dietary hydroxycinnamates” Invited review chapter in CRC Handbook of Food Analysis.

Conferences & Seminars

Münster/Germany ISC 2008 – 27th International Symposium on Chromatography 21 - 25 September 2008

Münster/Germany HPLC Masterclass Advanced Method Development (LC certification) 6 - 7 Mai 2010

Montpellier/France 25. International Conference on Polyphenols 23 - 27 August 2010

Poster Presentation: Profiling of Isomers of Chlorogenic Acids by LC-MSn

H. Karaköse, N. Kuhnert

Istanbul/Turkey Terpenist 2010 26 - 29 September 2010 Poster Presentation: Sweet Terpenes in Stevia Rebaudiana; H. Karaköse, N. Kuhnert Bremen/Germany GDCh – Wissenschaftsforum Chemie 4 - 7 September 2011

Poster Presentation: Profiling, PCA Analysis and Quantification of Chlorogenic acids in Stevia Rebaudiana; H. Karaköse, N. Kuhnert.

Sitges/Spain 5th International Conference on Polyphenol and Health 17-20 October 2011

Poster Presentation: Characterization of Chlorogenic Acids in Stevia Rebaudiana

Leaves by LC-MSn ; H. Karaköse, N. Kuhnert.

Internships 03.07 - 28.07.2006 Petkim Petrokimya Holding A.Ş. (Petroleum chemicals), Izmir/Turkey

07.02 - 04.03.2006 University of Leipzig, Faculty of Chemistry and Mineralogy Institute of Analytical Chemistry. Leipzig/Germany 27.06 - 05.08.2005 Petkim Petrokimya Holding A.Ş. (Petroleum chemicals), Izmir/Turkey

Languages English (very good) German (good) Skills

Computer Skills Microsoft Windows 7& XP Microsoft Office Programs Linux / Ubuntu Open Office

Teaching Skills During my PhD and Master education, I have led seminars, supervised

undergraduates in the laboratory Regularly supervise practicals for undergraduate students and have supervised the undergraduate research projects of two final year students. I gave several seminars for undergraduates in School of Engineering and Science

Interests Swimming, travelling, reading, photography.

Reference Prof. Nikolai Kuhnert Email: [email protected]

Telephone: +49 421 200-3120

Fax: +49 421 200-3229

Published: August 02, 2011

r 2011 American Chemical Society 10143 dx.doi.org/10.1021/jf202185m | J. Agric. Food Chem. 2011, 59, 10143–10150

ARTICLE

pubs.acs.org/JAFC

Characterization andQuantification of Hydroxycinnamate Derivativesin Stevia rebaudiana Leaves by LC-MSn†

Hande Karak€ose, Rakesh Jaiswal, and Nikolai Kuhnert*

School of Engineering and Science, Chemistry, Jacobs University Bremen, 28759 Bremen, Germany

bS Supporting Information

ABSTRACT: Stevia rebaudiana leaves are used as a zero-calorie natural sweetener in a variety of food products in Asian countries,especially in Japan. In this study, the hydroxycinnamate derivatives of S. rebaudiana have been investigated qualitatively andquantitatively by LC-MSn. Twenty-four hydroxycinnamic acid derivatives of quinic and shikimic acid were detected, and 19 of themwere successfully characterized to regioisomeric levels; 23 are reported for the first time from this source. These comprise threemonocaffeoylquinic acids (Mr 354), seven dicaffeoylquinic acids (Mr 516), one p-coumaroylquinic acid (Mr 338), one feruloylquinicacid (Mr 368), two caffeoyl-feruloylquinic acids (Mr 530), three caffeoylshikimic acids (Mr 336), and two tricaffeoylquinic acids(Mr 678). Cis isomers of di- and tricaffeoylquinic acids were observed as well. Three tricaffeoylquinic acids identified in stevia leavesare reported for the first time in nature. These phenolic compounds identified in stevia might affect the organoleptic properties andadd additional beneficial health effects to stevia-based products.

KEYWORDS: Stevia rebaudiana, chlorogenic acids, hydroxycinnamic acids, caffeoylquinic acids, caffeoylshikimic acids, tandemmass spectrometry

’ INTRODUCTION

Stevia rebaudiana is a plant belonging to the Asteraceae familyof plants, which is native to Brazil and Paraguay. Due to thenatural sweetness of its leaves, S. rebaudiana has caught attentionin scientific and industrial fields to act as a natural zero-caloriesweetener in many applications in the food industry. The leavescontain ent-kaurene glycosides, comprising stevioside, rebaudio-sides A, B, C, D, E, and F, and dulcoside A. All of these diterpeneglycosides comprise a steviol backbone structure; they differ onlyin the glucose moiety at positions C13 and C19 (Figure 1).Stevioside is the main sweet-tasting glycoside in stevia and wasreported to be 250�300 times sweeter than sucrose.1 Rebaudio-side A is the second most abundant ent-kaurene and sweetestcompound in stevia; its sweetness is 400 times greater than thatof sucrose, and it has more pleasant taste and is more water-soluble than stevioside.2 The amounts of diterpene glycosidesmay vary depending on the growth conditions of stevia; however,stevioside accounts for 4�13% (w/w) and rebaudioside Aaccounts for 2�4% (w/w),3 the other glycosides being presentin lower concentrations.

The principal advantage of stevia metabolites is that they arenatural, nonsynthetic products. Stevia leaves can be used in theirnatural state (fresh or dried form), due to their high sweeteningintensity. Only small quantities are needed in comparison to whitesugar to achieve comparable sweetness. The primary use of steviais as a commercial sweetener; it is used in a wide range of productssuch as soft drinks, ice cream, chocolate, yogurt, and baked andcooked foods. Stevia products also have beneficial uses in variousconsumer care products such as toothpaste or mouthwashes.4,5

Stevia may also be used for obesity, diabetics, dental caries, andtherapeutic effects such as hypoglycemic activity.6

The majority of the annual stevia production of an estimated4000 t is produced in China and South America. The stevia crop

has been shown to be highly adaptable to cultivation in manyother parts of the world. S. rebaudiana occurs naturally on acidsoils of pH 4�5 but will also grow on soils with pH levels of6.5�7.5, making it an interesting alternative to plants cultivatedon poor soils such as tobacco.7

In addition to diterpene glycosides, a number of secondaryplant metabolites have been identified from S. rebaudina includinglabdane-type diterpenes, triterpenoids and steroids, flavonoids,and oil components. From S. rebaudiana, 10 labdane-type diter-penoids were identified, including austroinulin, isoaustroinulin,6

and sterebins (A�H).8,9 A triterpenoid, lupeol 3-palmitate, wasalso separated from stevia.10 As plant sterols, β-sitosterol, stigmas-terol, and campesterol were identified from S. rebaudiana.11

Plant phenols are a large and diverse group of compoundsincluding hydroxycinnamates, tannins, flavonoids, stilbenes, cou-marins, lignans, and lignins.12 Chlorogenic acids (CGAs) are themost commonhydroxycinnamate derivatives observed in the plantkingdom. By definition, they are a large family of esters formedbetween quinic acid and one to four residues of certain trans-hydroxycinnamic acids, most commonly caffeic, p-coumaric, andferulic; sinapic and dimethoxycinnamic acids also occur, and insome plant species various aliphatic acids may replace one or moreof the trans-cinnamic acid residues.13 CGAs are involved inbiological functions in plants such as defense against pathogensand resistance to diseases. CGAs also participate in enzyme-catalyzed browning reactions that may adversely affect the color,flavor, and nutritional quality of dietary sources.14

Several pharmacological activities of CGAs including antiox-idant activity, the ability to increase hepatic glucose utilization,15,16

Received: June 1, 2011Revised: July 15, 2011Accepted: August 2, 2011

10144 dx.doi.org/10.1021/jf202185m |J. Agric. Food Chem. 2011, 59, 10143–10150

Journal of Agricultural and Food Chemistry ARTICLE

inhibition of theHIV-1 integrase,17,18 antispasmodic activity,19 andinhibition of the mutagenicity of carcinogenic compounds20

have been revealed by in vitro, in vivo, and human intervention

studies so far. CGAs and their metabolites display additionallyhighly favorable pharmacokinetic properties.21�23 Because thepolyphenols in stevia might affect the organoleptic properties of

Figure 1. Structures and numberings of caffeoylquinic acids.

Figure 2. Base peak chromatogram of Stevia rebaudiana extract using ion trap MS in negative ion mode. For numbering, see Table 1.

http://pubs.acs.org/action/showImage?doi=10.1021/jf202185m&iName=master.img-000.png&w=399&h=456




stevia-based product and could add additional health benefits tothe product, the objective of the present study was to profile thephenolic content of S. rebaudina leaves with a particular emphasison hydroxycinnamate derivatives.

’MATERIALS AND METHODS

The chlorogenic acids, 3-caffeoylquinic acid, 4-caffeoylquinic acid,5-caffeoylquinic acid (chlorogenic acid), 3,4-dicaffeoylquinic acid, 3,5-dicaffeoylquinic acid, and 4,5-dicaffeoylquinic acid, were purchased fromPhytoLab (Vestenbergsgreuth, Germany). All other chemicals werepurchased from Sigma-Aldrich (Bremen, Germany). Stevia leaves werepurchased from a market in Bremen, Germany.Sample Preparation. Two grams of S. rebaudiana leaves was

immersed in liquid nitrogen, ground in a hammermill, and extracted firstwith 150 mL of chloroform in a Soxhlet apparatus (Buchi B-811extraction system) for 2 h and then with 150 mL of methanol foranother 2 h. Solvents were removed from the methanolic extract invacuo, and extracts were stored at �20 �C until required.

UV Irradiation. The prepared sample of stevia leaf extract (1 mL)was placed in a photoreactor (LuzchemLZC -4 V, Ottawa, Canada)under a shortwave UV lamp and irradiated at 245 nm for 40 min.LC-MSn. The LC equipment (Agilent 1100 series, Bremen,

Germany) comprised a binary pump, an autosampler with a 100 μLloop, and a diode array detector with a light-pipe flow cell (recording at320 and 254 nm and scanning from 200 to 600 nm). This was interfacedwith an ion-trap mass spectrometer fitted with an ESI source (BrukerDaltonics HCT Ultra, Bremen, Germany) operating in Auto-MSnmodeto obtain fragment ions m/z. As necessary, MS2, MS3, and MS4

fragment-targeted experiments were performed to focus only on com-pounds producing a parent ion at m/z 335.1, 337.1, 367.1, 529.2, or677.3. Tandem mass spectra were acquired in Auto-MSn mode (smartfragmentation) using a ramping of the collision energy. Maximumfragmentation amplitude was set to 1 V, starting at 30% and ending at200%. MS operating conditions (negative mode) had been optimizedusing 5-caffeoylquinic acid28 with a capillary temperature of 365 �C, adry gas flow rate of 10 L/min, and a nebulizer pressure of 50 psi.

Figure 3. Extracted ion chromatograms (EIC) of m/z 515 in negative ion mode (A) before and (B) after UV irradiation.

Figure 4. Structures and numbering of tricaffeoylquinic acids.





High-resolution LC-MS was carried out using the same HPLCequipped with a MicroTOF Focus mass spectrometer (BrukerDaltonics) fitted with an ESI source, and internal calibration wasachieved with 10 mL of 0.1 mol/L sodium formate solution injectedthrough a six-port valve prior to each chromatographic run. Calibrationwas carried out using the enhanced quadratic calibration mode.HPLC. Separation was achieved on a 150 � 3 mm i.d. column

containing diphenyl 5μmwith a 4� 3mm i.d. guard column of the samematerial (Varian, Darmstadt, Germany). Solvent A was water/formicacid (1000 + 0.05 v/v), and solvent B was methanol. Solvents weredelivered at a total flow rate of 0.5 mL/min. The gradient profile wasfrom 10 to 70% B linearly in 60 min followed by 10 min isocratic and areturn to 10% B at 80 and 10 min isocratic to re-equilibrate.Calibration Curve of Standard Compounds. Stock solutions

of the standard compounds were prepared in methanol. A series ofstandard solutions was injected (5 μL) into the LC-MS system. Theareas of the peaks of each standard from UV chromatograms were usedto make the respective standard curves.Synthesis of the Mixture of Regioisomers of Tricaffeoyl-

quinic Acids. To a solution of quinic acid (96 mg, 0.5 mmol) andDMAP (16 mg, 0.12 mmol) in CH2Cl2 (10 mL) were added triethy-lamine (4 mL) and 3,4-diacetylcaffeic acid chloride (423 mg, 1.5 mmol)

at room temperature. The reaction mixture was stirred for 6 h andacidified with 2 mol/L HCl (pH ≈1) and then stirred for an additional3 h to remove the acetyl protecting groups. The layers were separated,and the aqueous phase was re-extracted with CH2Cl2 (1� 20 mL) andEtOAc (2 � 20 mL). The combined organic layers were dried overNa2SO4 and filtered, and the solvents were removed in vacuo. Theresulting esters were analyzed by HPLC-MS.

’RESULTS AND DISCUSSION

Methanol extracts of stevia dry leaves were directly used forLC-MS analysis. Efficient separation and resolution wereachieved with diphenyl packing and acetonitrile/water as solventin the HPLC method. Negative ion mode was used for all MSmeasurements. The HPLC method used here constitutes avariation of methods employed previously,24 with variationsrequired to achieve sufficient separation of triacyl chlorogenicacids and ent-kaurene glycosides. In comparison to isolation ofCGAs from green coffee beans, no removal of proteins/peptidesby Carrez reagent was necessary.13,24

All data for chlorogenic acids and diterpene glycosides pre-sented in this paper use the IUPAC numbering system,32 and

Figure 5. Tandem mass spectra of 1,3,5-triCQA in negative ion mode.

Figure 6. Tandem mass spectra of 3,4,5-triCQA in negative ion mode.

http://pubs.acs.org/action/showImage?doi=10.1021/jf202185m&iName=master.img-004.jpg&w=365&h=202

http://pubs.acs.org/action/showImage?doi=10.1021/jf202185m&iName=master.img-005.jpg&w=360&h=162



structures are presented in Figure 1. Peak assignments of CGAshave been made on the basis of structure diagnostic hierarchicalkeys previously developed,24�26 supported by means of theirparent ion, UV spectra, and retention times relative to 5-CQAusing validated methods in our laboratory.24,27 More sensitiveand more selective fragment-targeted MSn experiments wereused for quantitatively minor components. The base peakchromatogram of stevia extract is shown in Figure 2. Abbrevia-tions and numbering are given in Figure 1. Stevia extract wasanalyzed by LC-MSn in the negative ion mode using an ESI ion-trap mass spectrometer, allowing assignments of compounds toregioisomeric level, and also by high-resolution mass spectro-metry using ESI-TOF in negative ion mode connected to LC.With the guidance of previous studies from Clifford,24�29 threeCQAs (1�3), seven di-CQAs (4�10), three FQAs (18�20),one p-CoQA (11), three CFQAs (12�14), three CSAs(15�17), and four tri-CQAs (21�24) were located in thechromatogram. For all compounds the high-resolution mass datawere in good agreement with the theoretical molecular formulas,with a mass error of below 5 ppm, confirming the elementalcompositions of all compounds investigated.

Reliable characterization of diterpene glycosides content instevia is crucial. Because the structures of single glycosides arevery similar, they have very similar retention times in LC andtherefore result in overlapping of peaks in the chromatogram.In this paper, the general profile of diterpene glycosides in

S. rebaudiana is given. Characterization of the compounds wasachieved by ion-trapmass spectrometrywith SIM, and confirmationof elemental composition was provided by ESI-TOF measure-ments (see the Supporting Information).

Table 1. Tandem Mass Spectral Data of Hydroxycinnamates in Stevia rebaudiana Leaf Extract

MS2 MS3 MS4

secondary peaks secondary peaks secondary peaks

no. compd

m/z

(neg)

base

peak m/z int m/z int m/z int

base


base


1 3-CQA 353.0 190.7 178.8 49 134.9 8 126.8 172.8 37 85.2 55 110.8 65 188.8 174.6 64 134.4 84

2 5-CQA 353.0 190.7 126.8 172.7 49 85.1 61 110.8 21 108.8

3 4-CQA 353.0 172.7 178.8 60 190.6 14 134.8 8 93.0 110.8 62 154.7 24

4 3,5-diCQA 515.1 353.0 190.8 8 190.7 178.8 49 134.9 7 126.8 93.0 98 85.2 62 172.6 48

5 3,4-diCQA 515.1 353.0 335.0 12 172.8 17 172.8 178.8 67 190.8 57 134.8 10 93.0 110.8 41 83.2 6

6 4,5-diCQA 515.1 353.0 299.0 3 254.9 7 172.8 17 172.8 178.6 52 190.8 28 135.0 7 93.0 110.8 89 83.0 18

7 a cis-3,5-diCQA 515.1 353.0 190.8 10 190.7 178.8 50 172.8 11 134.9 10 85.0 126.8 85 93.0 53 172.7 36

8 a cis-4,5-diCQA 515.1 353.0 172.8 13 172.8 178.6 66 190.6 35 134.9 11 93.0 110.9 39 83.0 10

9 cis-4,5-diCQA 515.1 353.0 172.7 7 172.7 178.8 66 190.8 59 134.9 12 93.0 110.9 23 83.0 19

10 a cis-4,5-diCQA 515.1 353.0 172.8 12 172.7 178.6 76 190.6 70 134.8 18 93.0 110.8 39 83.0 6

11 5-p-CoQA 337.1 190.7 162.8 6 126.8 172.7 42 108.8 44 92.8 32

12 3F,5CQA 529.1 367.0 353.0 12 192.7 7 178.6 2 192.7 172.6 13 133.7 14 133.7 126.6 16

13 C,FQA 529.1 367.1 349.0 7 178.7 10 178.7 160.8 73 134.8 85 134.7

14 4C,5FQA 529.1 353.0 254.8 5 172.7 17 172.7 178.6 65 190.6 24 134.7 11 93.0 110.8 32 59.4 52

15 5-CSA 335.1 178.7 172.8 11 134.8 20 134.7

16 4-CSA 335.1 178.7 160.6 82 134.8 51 134.7

17 3-CSA 335.1 178.7 160.8 4 134.8 42 134.8

18 5-FQA 367.1 190.7 172.8 3 85.0 126.8 85 172.6 28

19 FQA 367.1 178.7 190.8 33 160.8 12 134.8 72 134.7 106.8 5

20 FQA 367.2 176.8 161.6 130.8 57

21 3,4,5-triCQA 677.1 515.1 353.0 20 353.0 335.0 16 299.0 4 172.8 30 172.7 178.6 58 190.6 42 134.8 16

22 1,3,5-triCQA 677.1 515.1 353.0 16 353.0 335.0 15 254.9 5 172.7 28 190.8 178.6 72 172.6 90 136.7 10

23 triCQA 677.1 515.0 353.0 15 353.0 172.7 38 254.8 4 172.7 178.8 60 190.6 33 134.7 14

24 triCQA 677.1 515.1 353.0 20 353.0 172.7 22 254.8 4 172.8 178.6 87 190.8 90 134.7 15

Figure 7. Extracted ion chromatograms (EIC) of m/z 677 in negativeion mode (A) before and (B) after UV irradiation.




Characterization of Caffeoylquinic Acids (Mr 354) andDicaffeoylquinic acids (Mr 516). Three peaks were detectedat m/z 353.1 and assigned using the hierarchial keys previouslydeveloped24 as well-known 3-CQA, 5-CQA, and 4-CQA. Threedicaffeoylquinic acid isomers were identified by their parent ionm/z 515.2 and were assigned as 3,5-diCQA, 3,4-diCQA, and 4,5-diCQA using the hierarchial keys.24,26 Three further peakspresent as minor components showed fragmentation patternssimilar to that of 4,5-diCQA. We have recently reported on cisisomers of chlorogenic acids present in plant tissue exposed toUV light, which have formed in a photochemical trans�cisisomerization reaction.30 To confirm if the remaining three peakscorrespond to cis isomers, the extract was irradiated with UVlight at 245 nm for 40 min. After irradiation, a significant increasein the intensities of two peaks (9 and 10 in Figure 3) wasobserved, if compared to their corresponding trans isomers fromthe original plant extract. In addition, a significant increase wasobserved in the intensity of cis-3,5-diCQA (7 in Figure 3) peakaccompanied by a decrease of the 3,4-diCQA (5 in Figure 3)peak. This finding suggests that under the chromatographicconditions employed the cis isomer is coeluting with 3,4-diCQA(Figure 3).On the basis of increased intensity after UV irradiation and

fragmentation pattern, three additional cis isomers were ob-served for 4,5-diCQA. One of these isomers was assigned ascis-4,5-diCQA (9), and two of them were assigned as cis�trans(a cis) isomer, but the distinction between 4-cis,5-trans-diCQAand 4-trans-5-cis-diCQA was not possible (8 and 10).Characterization of Feruloylquinic Acid (Mr 368), p-Cou-

maroylquinic Acid (Mr 338), and Caffeoylferuloylquinic Acid(Mr 530). Only one peak was detected at m/z 337.1, which wasidentified as 5-p-CoQA according to its fragmentation pattern.Three peaks were detected at m/z 367, and one of them wasidentified as 5-FQA; the other two peaks could not be assigneddue to their uncommon fragmentation pattern.A targeted MS3 experiment at m/z 529.2 ([M � H+]�)

applied to the extract located three peaks, and two of them wereidentified as 3F,5CQA and 4C,5FQA on the basis of theircharacteristic fragmentations in MS2 and MS3 spectra. Theassigments are achieved using the hierarchial keys previouslydeveloped, and mass spectra published previously are not pre-sented here.24,31

Characterization of Caffeoylshikimic Acids (Mr 336). Caf-feoylshikimic acids (CSA) have been reported in date palms,sweet basil, and carrot,32�35 and they have been characterized toregioisomeric level in yerba mat�e leaves by tandem mass spectrapreviously.36 This class of compounds is reported here for thefirst time from the Asteraceae family of plants. A targeted MS3

experiment at m/z 335.1 ([M � H+]�) applied to the extractlocated three peaks, and they were identified by their fragmenta-tion patterns as 5-CSA, 4-CSA, and 3-CSA (15�17).36 All three

regioisomers show m/z 178 (caffeic acid fragment) in their MS2

spectra. 4-CQA shows an intense characteristic fragment ion atm/z 160, which is absent in theMS2 spectra of 3-CSA and 5-CSA.

Characterization of Tricaffeoylquinic Acid (Mr 678). Fourtriacyl CQA isomers (Figure 4) were detected in the steviaextract at 677 for tricaffeoyls in neg. mode and confirmed astricaffeoyl derivatives by targeted MS4 experiments. Assignmentof regiochemistry was assisted by an independent synthesis of amixture of all four possible regioisomers of tricaffeoylquinicacids. The chromatogram of the mixture of all theoretically pos-sible four regioisomers of tricaffeoylquinic acid obtained throughsynthesis showed two well-resolved peaks with retention timesand MS data identical to those present in the stevia extract alongwith an intense broad peak in a retention time range where thetwo remaining isomers in the stevia extract were observed (seethe Supporting Information). Detailed studies of the tandemmass spectra at various retention times within the broad peaksuggest that this broad peak must correspond to two distinctunresolved regioisomers of tricaffeoylquinic acid. Comparisonof the chromatogram of the synthetic mixture with the extractallowed unambiguous assignment of the two regioisomers in theextract by identity of the fragmentation pattern compared to thesynthetic mixture. Identification of 1,3,5-triCQA in the extractwas followed automatically due to the absence of an MS4 basepeak atm/z 173 corresponding to a dehydratedMS2 base peak ofthe quinic moiety characteristic of 4-acylated isomers. The MS4

base peak atm/z∼191 and a secondary peak atm/z 178 (72% ofbase peak) suggest the 3,5-disubstitution pattern (Figure 5).3,4,5-triCQA was identified by comparison to material describedpreviously.36,37 (Figure 6)The two remaining peaks might be cis isomers of 3,4,5-triCQA

and 1,3,5-triCQA, or they can correspond to either 1,4,5-tri-CQA and 1,3,4-tri-CQA or any of their cis isomers (see Table 1).However, current information does not allow us to discriminateunambiguously between these regioisomers at the moment. Toprobe whether cis isomers were present, the extract was againirradiated with UV light, and after chromatographic analysis, asignificant increase in the intensity of the peaks of 4-acylatedisomers was observed (Figure 7). Otherwise, the experiment wasinconclusive. It is worth noting that in theory for each tricaffeoylderivative eight stereoisomers with various trans�cis stereoche-mistries are possible, thus increasing the total number of isomerictricaffeoylquinic acids to 32. Given the identity of MS data andthe absence of characteristic shoulders in the UV spectracharacteristic for cis-caffeoyl derivatives, we tentatively assignthe two remaining isomers as 1,4,5-tri-CQA and 1,3,4-tri-CQA. Only 3,4,5-tri-CQA has been previously reported innature, whereas the remaining isomers are reported here forthe first time.Quantification of Caffeoylquinic Acids. Following the qua-

litative profiling of chlorogenic acids in S. rebaudiana, we decided

Table 2. Quantities of Mono- and Di-CQAs in S. rebaudiana Leaves

compd concn range calibration curve correl coeff calcd amount (μg/g)

3-CQA 1 μg/mL�1 mg/mL Y = 4.457x � 460.04 0.99 35.5

5-CQA 1 μg/mL�3 mg/mL Y = 17.719x � 2361.90 0.99 44.3

4-CQA 0.07 μg/mL�1 mg/mL Y = 13.288x � 1223.00 0.99 70.3

3,5-diCQA 0.09 μg/mL�1 mg/mL Y = 5.3176x � 529.76 0.99 145.6





to quantify the levels of selected compounds. Chlorogenic acidstandard solutions were analyzed by LC-MS using the samechromatographic method as used for stevia leaf extracts. For sixselected monoacyl- and diacylquinic acids, calibration curveswere obtained using six-point calibration from the UV chroma-togram recorded at 320 nm. The individual amounts calculatedfor mono- and dicaffeoylquinic acids are listed in Table 2, whichalso lists the correlation coefficient of linear regression for eachstandard sample and the concentration range.Among the monocaffeoylquinic acids, 4-CQA was found to be

the most abundant compound, and among all CQAs 3,5-diCQAwas found to be the most abundant compound. The totalchlorogenic acid amount determined here is around 370 μg/gof dry leaf.In this study we profiled the chlorogenic acids in S. rebaudiana

employing LC-MSn and LC-TOF techniques. A total of 24chlorogenic acids were detected in S. rebaudiana leaves, with23 compounds described for the first time from this source. Tri-CQAs were reported for the first time from S. rebaudiana withthree regioisomers found for the first time in nature. CSAs werecharacterized for the first time from a plant belonging to theAstareceae family by using tandem mass spectrometry. Quanti-fication of selected mono- and di-CQAs was achieved by usingthe UV chromatogram with total chlorogenic acid levels found tobe 370 μg/g of dry leaf.

’ASSOCIATED CONTENT

bS Supporting Information. Additional EIC of triacylCGAs, MS2 + MS3 data of all compounds mentioned in the text,table of high-resolution MS-TOF data for compounds identified,and structures of ent-kaurene terpenes. This material is free ofcharge via the Internet at http://pubs.acs.org

’AUTHOR INFORMATION

Corresponding Author*Phone: 49 421 200 3120. Fax: 49 421 200 3229. E-mail:[email protected].

Funding SourcesFinancial support from the European Union (project DIVAS) isgratefully acknowledged.

’ACKNOWLEDGMENT

We acknowledge the technical assistance of Anja M€uller.

’DEDICATION†This paper is dedicated to Prof.M. N. Clifford on the occasion ofhis 65th birthday.

’REFERENCES

(1) Kolb, N.; Herrera, J. L.; Ferreyra, D. J.; Uliana, R. F. Analysis ofsweet diterpene glycosides from Stevia rebaudiana: improved HPLCmethod. J. Agric. Food Chem. 2001, 49, 4538–4541.(2) Crammer, B.; Ikan, R. Sweet glycosides from the stevia plant.

Chem. Br. 1986, 22, 915.(3) Pol, J.; Hohnova, B.; Hyotylainen, T. Characterisation of Stevia

rebaudiana by comprehensive two-dimensional liquid chromatographytime-of-flight mass spectrometry. J. Chromatogr., A 2007, 1150, 85–92.

(4) Chatsudthipong, V.; Muanprasat, C. Stevioside and relatedcompounds: therapeutic benefits beyond sweetness. Pharmacol. Ther.2009, 121, 41–54.

(5) Brandle, J. E.; Rosa, N. Heritability for yield, leaf-stem ratio andstevioside content estimated from a landrace cultivar of Stevia-rebaudi-ana. Can. J. Plant Sci. 1992, 72, 1263–1266.

(6) Ibrahim, N. A.; El-Gengaihi, S.; Motawe, H.; Riad, S. A. Phyto-chemical and biological investigation of Stevia rebaudiana Bertoni;1-labdane-type diterpene. Eur. Food Res. Technol. 2007, 224, 483–488.

(7) Shock, C. C. Rebaudi’s stevia: natural noncaloric sweeteners.Calif. Agric. 1982, 4–5.

(8) Oshima, Y.; Saito, J.; Hikino, H. Sterebins A, B, C and D,bisnorditerpenoids of Stevia-rebaudiana leaves. Tetrahedron 1986, 42,6443–6446.

(9) Oshima, Y.; Saito, J.-I.; Hikino, H. Sterebins E, F, G and H,diterpenoids of Stevia rebaudiana leaves. Phytochemistry 1988, 27,624–626.

(10) Yasukawa, K.; Yamaguchi, A.; Arita, J.; Sakurai, S.; Ikeda, A.;Takido, M. Inhibitory effect of edible plant extracts on 12-O-tetradeca-noylphorbol-13-acetate-induced ear edema inmice. Phytother. Res. 1993,7, 185–189.

(11) Kinghorn, A. D. Stevia The Genus Stevia; CRC Press: BocaRaton, FL, 2001.

(12) Rice Evans, C. A.; Miller, N. J.; Paganga, G. Structure�antioxidant activity relationships of flavonoids and phenolic acids.Free Radical Biol. Med. 1996, 20, 933–956.

(13) Clifford, M. N.; Marks, S.; Knight, S.; Kuhnert, N. Character-ization by LC-MSn of four new classes of p-coumaric acid-containingdiacyl chlorogenic acids in green coffee beans. J. Agric. Food Chem. 2006,54, 4095–4101.

(14) Im, H. W.; Suh, B. S.; Lee, S. U.; Kozukue, N.; Ohnisi-Kameyama, M.; Levin, C. E.; Friedman, M. Analysis of phenoliccompounds by high-performance liquid chromatography and liquidchromatography/mass spectrometry in potato plant flowers, leaves,stems, and tubers and in home-processed potatoes. J. Agric. Food Chem.2008, 56, 3341–3349.

(15) Kono, Y.; Kobayashi, K.; Tagawa, S.; Adachi, K.; Ueda, A.;Sawa, Y.; Shibata, H. Antioxidant activity of polyphenolics in diets� rateconstants of reactions of chlorogenic acid and caffeic acid with reactivespecies of oxygen and nitrogen. Biochim. Biophys. Acta: Gen. Subj. 1997,1335, 335–342.

(16) Hemmerle, H.; Burger, H. J.; Below, P.; Schubert, G.;Rippel, R.; Schindler, P. W.; Paulus, E.; Herling, A. W. Chlorogenicacid and synthetic chlorogenic acid derivatives: novel inhibitors ofhepatic glucose-6-phosphate translocase. J. Med. Chem. 1997, 40,137–145.

(17) Robinson, W. E.; Cordeiro, M.; AbdelMalek, S.; Jia, Q.; Chow,S. A.; Reinecke, M. G.; Mitchell, W. M. Dicaffeoylquinic acid inhibitorsof human immunodeficiency virus integrase: Inhibition of the corecatalytic domain of human immunodeficiency virus integrase. Mol.Pharmacol. 1996, 50, 846–855.

(18) 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.

(19) Trute, A.; Gross, J.; Mutschler, E.; Nahrstedt, A. In vitroantispasmodic compounds of the dry extract obtained from Hederahelix. Planta Med. 1997, 63, 125–129.

(20) Stich, H. F.; Rosin, M. P.; Bryson, L. Inhibition of mutagenicityof a model nitrosation reaction by naturally-occuring phenolics, coffeeand tea. Mutat. Res. 1982, 95, 119–128.

(21) Hanlon, N.; Coldham, N.; Gielbert, A.; Kuhnert, N.; Sauer,M. J.; Kingi, L. J.; Ioannides, C. Absolute bioavailability and dose-dependent pharmacokinetic behaviour of dietary doses of the chemo-preventive isothiocyanate sulforaphane in rat. Br. J. Nutr. 2008, 99,559–564.

(22) Stalmach, A.; Mullen, W.; Barron, D.; Uchida, K.; Yokota, T.;Cavin, C.; Steiling, H.; Williamson, G.; Crozier, A. Metabolite profilingof hydroxycinnamate derivatives in plasma and urine after the ingestion



of coffee by humans: identification of biomarkers of coffee consumption.Drug Metab. Dispos. 2009, 37, 1749–1758.(23) Farah, A.; Monteiro, M.; Donangelo, C. M.; Lafay, S. Chloro-

genic acids from green coffee extract are highly bioavailable in humans.J. Nutr. 2008, 138, 2309–2315.(24) 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.(25) IUPAC Commission on the Nomenclature of Organic Chem-

istry (CNOC) and IUPAC-IUB Commission on Biochemical Nomen-clature (CBN). Nomenclature of cyclitols. Recommendations, 1973.Biochem. J. 1976, 153, 23�31.(26) 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.(27) Clifford, M. N. Chlorogenic acids and other cinnamates �

nature, occurrence, dietary burden, absorption and metabolism. J. Sci.Food Agric. 2000, 80, 1033–1043.(28) Clifford, M. N. Chlorogenic acids and other cinnamates �

nature, occurrence and dietary burden. J. Sci. Food Agric. 1999, 79, 362–372.(29) Clifford, M. N. The Analysis and Characterization of Chlorogenic

Acids and Other Cinnamates; Royal Society of Chemistry: Cambridge,U.K., 2003.(30) 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.(31) Clifford, M. N.; Wu, W. G.; Kuhnert, N. The chlorogenic acids

of Hemerocallis. Food Chem. 2006, 95, 574–578.(32) Mansouri, A.; Embarek, G.; Kokkalou, E.; Kefalas, P. Phenolic

profile and antioxidant activity of the Algerian ripe date palm fruit(Phoenix dactylifera). Food Chem. 2005, 89, 411–420.(33) Ziouti, A.; ElModafar, C.; ElMandili, A.; ElBoustani, E.;

Macheix, J. J. Identification of the caffeoylshikimic acids in the roots ofthe date palm, principle fungitoxic compounds vis-a-vis Fusarium-oxysporum f sp albedinis. J. Phytopathol. 1996, 144, 197–202.(34) Gang, D. R.; Beuerle, T.; Ullmann, P.; Werck-Reichhart, D.;

Pichersky, E. Differential production of meta hydroxylated phenylpro-panoids in sweet basil peltate glandular trichomes and leaves is con-trolled by the activities of specific acyltransferases and hydroxylases.Plant Physiol. 2002, 130, 1536–1544.(35) Kuhnl, T.; Koch, U.; Heller,W.;Wellmann, E. Chlorogenic acid

biosynthesis � characterization of a light-induced microsomal 5-O-(4-coumaroyl)-D-quinate shikimate 30-hydroxylase from carrot (Daucus-carota L) cell suspension cultures. Arch. Biochem. Biophys. 1987, 258,226–232.(36) Jaiswal, R.; Sovdat, T.; Vivan, F.; Kuhnert, N. Profiling and

characterization by LC-MSn of the chlorogenic acids and hydroxycinna-moylshikimate esters in mate (Ilex paraguariensis). J. Agric. Food Chem.2010, 58, 5471–5484.(37) Jaiswal, R.; Kuhnert, N. Hierarchical scheme for liquid chro-

matography/multi-stage spectrometric identification of 3,4,5-triacylchlorogenic acids in green Robusta coffee beans. Rapid Commun. MassSpectrom. 2010, 24, 2283–2294.

Supplemantary Information

`

TRICQAs

EIC of targeted MS for m/z 677

3,4,5‐triCQA (21)

1,3,5‐triCQA (22)

STEVIA_CGA_TAR7.D: EIC 677.0 -All MS

0.00

0.25

0.50

0.75

1.00

1.25

7x10Intens.

44 46 48 50 52 54 Time [min]

228.6293.0 338.0 448.8 515.0

579.1

754.9677.1365.1

-MS,

353.0 629.3

515.1-MS2(677.1),

172.7254.9

353.0-MS3(677.3->515.0)

136.7

190.8 -MS4(677.3->515.3->352.8

0 100

Intens. [%]

0 100

[%]

0 100

[%]

0 100

[%]

100 200 300 400 500 600 700 m/z

160.7 338.0 419.9 515.1

677.1

298.9 353.0

515.1-MS2(677.1)

172.7 254.8

353.0 -MS3(677.5->515.1),

134.8

172.7 -MS4(677.5->515.3->353.0)

0

100

Intens. [%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 700 m/z

-MS

21

glycoside


`

4‐acylated‐triCQA (23/24)

4‐acylated‐triCQA (23/24)

112.8 186.6 238.7 390.8 448.9 487.1 593.1791.3

677.1

497.0

353.0

515.1-MS2(677.1)

172.7254.8

353.0 -MS3(677.5->515.1),

134.7

172.8

-MS4(677.5->515.3->353.1)

0

100

Intens. [%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 700 800 m/z

160.7 238.6 284.9 390.8 466.1 515.1561.3

793.4677.1

658.7

-MS

353.0 467.2

515.0-MS2(677.1)

172.7 254.8

353.0 -MS3(677.2->515.1),

134.7 172.7

325.0

-MS4(677.2->515.2->353.0),

0

100 Intens.

[%]

0

100 [%]

0

100 [%]

0

100 [%]

100 200 300 400 500 600 700 800 m/z


`

Synthesis:

EIC of 677

1,3,5‐triCQA (22)

4‐acyl‐triCQA

CQAMIXTURE_1112.D: EIC 677.0 -All MS

0.00

0.25

0.50

0.75

1.00

1.25

6x10Intens.

0 10 20 30 40 50 60 Time [min]

178.7 260.7 341.1

402.8460.7

515.1544.7 602.6 642.8

677.1

-MS

353.1

515.1-MS2(677.1)

172.7

353.0 -MS3(677.4->515.1)

134.8

172.7 -MS4(677.4->515.3->352.9)

0

100

Intens. [%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 m/z

186.7 260.7 341.0 402.7 460.7

503.1

544.7

677.1

-MS

353.0

515.1 -MS2(677.1)

172.7 298.9

353.0 -MS3(677.3->514.9),

134.8

190.7 -MS4(677.3->515.2->353.0)

0

100 Intens.

[%]

0

100 [%]

0

100 [%]

0

100 [%]

100 200 300 400 500 600 m/z


`

4‐acyl‐triCQA

m/z 353

STEVIA_CGA_DP01.D: EIC 353.0 -All MS

0.0

0.5

1.0

1.5

7x10Intens.

-5 0 5 10 15 20 25 30 Time [min]

160.7 260.7 338.1

460.7515.1

582.7

677.1

353.1

515.1-MS2(677.1)

172.8 254.9 299.0

353.0 -MS3(677.4->515.1)

134.8

172.7 -MS4(677.4->515.2->353.0)

0

100

Intens.[%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 m/z


`

3‐CQA (1):

5‐CQA(2):

4‐CQA(3):

438.0 529.4

353.0

375.0

-MS

134.8

172.7 -MS2(353.0)

154.7

93.0 -MS3(353.2->172.7),

0 50

100

Intens. [%]

0 50

100

[%]

0 50

100

[%]

200 400 600 800 1000 m/z

353.0

190.8

-MS

190.7 -MS2(353.0)

85.1 172.7

126.8 -MS3(353.1->190.8)

0 50

100

Intens. [%]

0 50

100

[%]

0 50

100

[%]

200 400 600 800 1000 m/z

529.4

353.0

375.0

-MS

134.9

190.7 -MS2(353.0)

126.8

172.7

-MS3(353.2->190.8),

0

50 100

Intens.[%]

0

50 100 [%]

0

50 100 [%]

200 400 600 800 1000 m/z


`

m/z 515

3,5‐diCQA(4)

3,4‐diCQA (5)

4,5‐diCQA (6)

353.0

515.1

447.0

-MS

172.8 254.8 298.9

353.0-MS2(515.1)

134.8

172.7 -MS3(515.3->353.0)

71.3

93.0

154.7

-MS4(515.3->353.2->172.8),

0

100

Intens. [%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 700 m/z

256.9 635.0

515.1593.1

-MS

172.8

353.0 -MS2(515.1)

134.9

172.8-MS3(515.3->353.0

93.0

154.7

-MS4(515.3->353.1->172.8)

0

100

Intens. [%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 700 m/z

353.0 635.0

515.1

537.1

-MS

190.8

353.0 -MS2(515.1)

134.9

190.7 -MS3(515.3->353.0),

93.0 172.7

-MS4(515.3->353.1->190.8),

0

100 Intens.

[%]

0

100 [%]

0

100 [%]

0

100 [%]

100 200 300 400 500 600 700 m/z


`

A cis‐3,5‐diCQA (7)

cis 4,5‐diCQA (9)


515.1

353.0

-MS

172.8

353.0 -MS2(515.1)

134.9

172.8 -MS3(515.2->353.0)

59.4

93.0

154.7

-MS4(515.2->353.1->172.7),

0

100 Intens.

[%]

0

100 [%]

0

100 [%]

0

100 [%]

100 200 300 400 500 600 700 m/z

260.7 353.0 431.1 635.1 771.2

515.1

537.1

-MS

172.7

353.0 -MS2(515.1)

134.9

172.7 -MS3(515.1->353.0)

71.3

92.9

154.7

-MS4(515.1->353.0->172.8)

0

100

Intens. [%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 700 m/z

353.0 613.0

515.1

537.0

-MS

190.8

353.0 -MS2(515.1)

134.9

190.7 -MS3(515.2->353.0)

85.1 126.8 172.7

-MS4(515.2->353.1->190.7)

0

100

Intens. [%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 700 m/z


`


5‐p‐CoQA (11)

5‐FQA (18)

96.9 186.7 260.7 595.1

367.1

528.6

-MS

296.7

190.7 -MS2(367.1)

85.1

172.7

-MS3(367.2->190.9)

0

50

100

Intens. [%]

0

50

100

[%]

0

50

100

[%]

200 400 600 800 1000 m/z

92.9

260.7 380.8 725.2

337.1

529.5

-MS

190.7 -MS2(337.1)

126.8

172.7

-MS3(337.2->190.8)

0

50 100

Intens. [%]

0

50 100

[%]

0

50 100

[%]

200 400 600 800 1000 m/z

353.0

515.1

549.2

-MS

172.8 353.0

-MS2(515.1)

134.8 172.7

-MS3(515.2->353.0)

71.2

93.0

154.7

-MS4(515.2->353.0->172.8)

0

100

Intens. [%]

0

100

[%]

0

100

[%]

0

100

[%]

100 200 300 400 500 600 700 m/z


`

5‐CSA (15):

4‐CSA (16):

3‐CSA (17):

96.9260.7 380.8

733.1

335.1

528.7

-MS

134.8 260.6

178.7 -MS2(335.1)

134.8 -MS3(335.3->178.7))

0

50 100

Intens. [%]

0

50 100

[%]

0

50

100

[%]

200 400 600 800 1000 m/z

96.9260.7 529.5

353.1 335.1

-MS

290.9

178.7 -MS2(335.1)

134.7 -MS3(335.2->178.8),

0

50

100

Intens.[%]

0

50

100

[%]

0

50

100

[%]

200 400 600 800 1000 m/z

96.9 186.7 260.7 447.1 529.5 625.1 771.2

399.1 335.1

-MS

134.8 290.9

178.7 -MS2(335.1)

134.7 -MS3(335.3->178.8)

0

50 100

Intens. [%]

0

50 100

[%]

0

50 100

[%]

200 400 600 800 1000 m/z


`

m/z 529

3F,5CQA (12)

(13)

4C,5FQA (14):

176.5 238.6 312.7 396.8 447.0705.3

529.1657.3

-MS

172.7 254.8 460.7 657.2

353.0 -MS2(529.1)

134.7

172.7 -MS3(529.2->352.1)

59.4 93.0

-MS4(529.2->353.0->172.8)

0

100 Intens.

[%]

0

100 [%]

0

100 [%]

0

100 [%]

100 200 300 400 500 600 700 m/z

186.5 254.6 322.8 377.0 445.1 657.3

755.2529.1

508.7

-MS

178.7 460.7

367.1 -MS2(529.1)

134.7 178.7 -MS3(529.4->367.7),

134.7 -MS4(529.4->367.3->178.5)

0

100 Intens.

[%]

0

100 [%]

0

100 [%]

0

100 [%]

100 200 300 400 500 600 700 m/z

260.6 316.8 385.1 431.1 577.2

771.1529.1

-MS

192.7 460.7

367.0 -MS2(529.1)

133.7

192.7 -MS3(529.2->367.1)

59.4 133.7 -MS4(529.2->367.1->192.5)

0 100

Intens. [%]

0 100

[%]

0 100

[%]

0 100

[%]

100 200 300 400 500 600 700 m/z


`

TOF data

Compound No

Compound Molecular Formula

Experimentalm/z (M-H+)-

Theoretical m/z (M-H+)-

Relative Error (ppm)

1 3-CQA C16H17O9 353.0865 353.0878 3.6

2 5-CQA C16H17O9 353.0889 353.0878 3.2

3 4-CQA C16H17O9 353.0883 353.0878 1.3

4 3,5-diCQA C25H24O12 515.1199 515.1195 0.8

5 3,4-diCQA C25H24O12 515.1190 515.1195 0.9

6 4,5-diCQA C25H24O12 515.1183 515.1195 2.3

15 5-CSA C16H16O8 335.0777 335.0772 1.4

16 4-CSA C16H16O8 335.0777 335.0772 1.3

17 3-CSA C16H16O8 335.0778 335.0772 1.8

18 5-FQA C17H20O9 367.1051 367.1035 4.6

11 5-pCoQA C16H18O8 337.0940 337.0929 3.2

21 3,4,5-triCQA C34H30O15 677.1498 677.1512 2.1

- triCQA C34H30O15 677.1517 677.1512 0.8

- triCQA C34H30O15 677.1516 677.1512 0.6

22 1,3,5-triCQA C34H30O15 677.1531 677.1512 2.8


`

Compound R R1 Molecular Formula

Experimental m/z (M-H+)-

Theoretical m/z (M-H+)-

Relative Error (ppm)

Steviol H H C20H30O3 317.2093 317.2122 9.4

Steviolbioside H glc2 - 1glc C32H50O13 641.3181 641.3179 0.4

Rubusoside glc glc C32H50O13 641.3166 641.3179 2.0

Stevioside glc glc2 - 1glc C38H60O18 803.3751 803.3707 5.5

Rebaudioside A glc glc32 -1glc

1glc

C44H70O23 965.425 965.4235 1.6


1glc

C38H60O18 803.368 803.3707 2.8

Rebaudioside C (Dulcoside B)

glc glc32 -1rham

1glc

C44H70O22 949.427 949.4286 1.7


1glc

C50H80O28 1127.4726 1127.4763 3.3


Rebaudioside F glc glc32 -1xyl

1glc

C43H68O22 935.4097 935.4129 3.5

Dulcoside A glc glc2 - 1rham C38H60O17 787.3732 787.3758 3.3

OR120

18 19

31

5

CO2R

109

7

11 13

15

1617

Documents

Phytochemical Characterization of Stevia rebaudiana