ANALYTICAL STRATEGIES FOR PROFILING, ANNOTATION, AND STRUCTURE
ELUCIDATION OF SPECIALIZED TERPENOID METABOLITES
By
Ekanayaka Appuhamilage Prabodha Ekanayaka

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemistry – Doctor of Philosophy
2014

i

ABSTRACT
ANALYTICAL STRATEGIES FOR PROFILING, ANNOTATION, AND STRUCTURE
ELUCIDATION OF SPECIALIZED TERPENOID METABOLITES
By
Ekanayaka Appuhamilage Prabodha Ekanayaka
The main bottleneck in today’s plant metabolomics lies with the identification of new metabolites. A
number of plant metabolite databases that report liquid chromatography – mass spectrometry data
have been constructed. However, more than 95% of compounds reported in these databases remain
unannotated. The extensive range of unknown metabolites presents a significant challenge in
interpreting metabolome data, and therefore developing methods that accelerate annotation and
identification of previously unknown metabolites has great importance when metabolome data are
used for functional genomic research.
The challenge of metabolite annotation was addressed by using relative mass defect (RMD) filtering
of ion masses measured using liquid chromatography-mass spectrometry. Such calculated RMD
values reflect the fractional hydrogen content of each detected ion, and reflect the biosynthetic
precursors and transformations that generate metabolites in vivo. RMD filtering aids grouping of
compounds of similar relative mass defect independent of absolute mass and chromatographic
retention time. Therefore, metabolites and metabolite precursors are grouped together enabling
potential associations among related metabolites to be developed. Furthermore, a systematic variation
of RMD among the fragment/product ions observed in multiplexed collision-induced dissociation
(CID) MS or liquid chromatography-tandem mass spectrometry (LC-MS/MS) data for compounds of
interest allowed for the identification of terpene glycosides in complex matrices. However, once the
metabolites are annotated in metabolomics data sets, establishing the structure of these compounds

ii

requires the purification of the compound followed by de novo structure elucidation that relies
heavily upon 1D and 2D NMR.
Chapter 2 of this dissertation discusses the application of RMD filtering based data analysis to both
parent and fragment ions generated in LC-multiplexed CID MS metabolite profiles generated from
wild tomato species Solanum habrochaites LA1777. This resulted in the discovery of over 24 novel
sesquiterpene glycoside chemical formulas, with multiple isomers comprising a group of more than
200 sesquiterpenoid glycosides. Chapters 3 and 4 of this dissertation discuss the purification and de
novo structure elucidation of seven example compounds from wild tomato glandular trichomes using
NMR. The structures of the sesquiterpenoid cores established for these compounds are different from
the structures of known volatile sesquiterpenoid compounds found in S. habrochaites LA1777
suggesting that the synthesis of these non-volatile terpenoids involves different biosynthetic enzymes
from those involved in the synthesis of known volatile terpenoids.
Similarly, Chapter 5 of this dissertation discusses the application of these techniques to the analysis
of metabolite profiles of the medicinal plant Hoodia gordonii generated using LC-multiplexed CID
MS. This research led to the identification of 24 novel diterpene glycosides. These compounds are
believed to share the diterpenoid cores found in some of the known diterpene glycosides from Hoodia
gordonii and therefore these compounds are likely biosynthetic intermediates of the synthesis of some
of the known diterpene glycosides.

iii

Copyright by
EKANAYAKA APPUHAMILAGE PRABODHA EKANAYAKA
2014

iv

ACKNOWLEDGEMENTS
The research discussed in this dissertation would be impossible without the contribution of a number
of people. Among all these people the most important person is my advisor Dr. A. Daniel Jones to
whom I am most grateful for his guidance, influence and support. He has mentored me through both
good and bad times during my career in graduate school and I consider myself lucky to have such
a wonderful mentor who I will continue to benefit from for the rest of my life.

I also thank my collaborators Dr. Robert Last’s research group and Dr. Cornelius Barry’s research
group from Michigan State University and Dr. Eran Pichersky’s research group from the University
of Michigan. It was an honor to learn from them and to work with them. Most of the research
discussed here would have been impossible without the help of Dr. Daniel Holmes and Mr. Kermit
Johnson of Max T. Rogers NMR facility at Michigan State University and so I am thankful to both of
them for their help in generation and interpretation of NMR data presented in this dissertation.

I am also grateful to my committee members, Dr. Dana Spence, Dr. Kevin D. Walker,
and Dr. Gary Blanchard for their support and comments on my work. I would like to extend my
appreciation to many past and present members (Particularly, to Dr. Chao Li, Dr. Jiangyin Bao and
Dr. Ramin Vismeh ) of Dr. Jones’s research group for their engaging discussion and their friendship,
which bring joy and happiness to my life at Michigan State University.

Completion of this dissertation would have been difficult without the moral and financial support
from my loving wife Punsisi (Punsisi Upeka Ratnayake), my undergraduate research supervisor Dr.
R.M. Rukmal Ratnayake (currently at Open University of Sri Lanka, Polgolla) and my parents (E.A
v

Gunatilaka and Mallika Kumarihami Harasgama).I would not have pursued my PhD at Michigan
State University without the support of my undergraduate research supervisor Dr. R.M. Rukmal
Ratnayake. I am very much grateful to her for her guidance with my work and life. Furthermore, I’d
like to give a special thanks to my wife Punsisi, who always encouraged and supported me
throughout my time at MSU. Finally, I am thankful to all my friends who made my stay at MSU
enjoyable, encouraged and helped me with the completion of the research presented here.

vi

TABLE OF CONTENTS
LIST OF TABLES ............................................................................................................................... ix
LIST OF FIGURES ...............................................................................................................................x
Chapter 1 ................................................................................................................................................1
1.1 Plants as sources of drugs, food and sustainable energy ................................................................1
1.2 Modern technology for using plants to support a sustainable human life .......................................1
1.3 Growing interest in plant metabolite identification .........................................................................3
1.4 Introduction to terpenes and trichomes ............................................................................................6
1.4.1 Trichomes in the genus Solanum and their chemical composition ....................................6
1.4.2 Structures and biosynthesis of terpene metabolites in the genus Solanum ........................8
1.4.3 Biosynthesis of terpenes ....................................................................................................9
1.4.4 Tomato and its terpenoid metabolites ..............................................................................13
1.4.5 Characteristic fragment ions for identifying terpenes from plant tissues using Gas
Chromatography –Mass Spectrometry .............................................................................15
1.5 Discovery and identification of specialized metabolites ...............................................................16
1.6 Mass spectrometry-based metabolomics as a functional genomics tool .......................................18
1.6.1 Mass Spectrometry ...........................................................................................................18
1.6.2 Collision induced dissociation (CID) and MS/MS...........................................................19
1.6.3 Multiplexed CID mass spectrometry ................................................................................19
1.6.4 Mass spectrometry for discovery of novel metabolites ....................................................21
1.7 Summary of research .....................................................................................................................23
REFERENCES ....................................................................................................................................25
Chapter 2: Strategies for rapid identification of sesquiterpene glycosides from complex matrices
using relative mass defect filtering. .....................................................................................................34
2.1 Introduction ....................................................................................................................................34
2.2 Applying accurate mass and relative mass defect filtering for exploring plant metabolomes for the
identification novel terpene glycosides ........................................................................................35
2.3 Materials and Methods ...................................................................................................................41
2.3.1 LC-MS and MS/MS experiments .....................................................................................41
2.3.2 Plant material ....................................................................................................................42
2.2.3 Data processing ................................................................................................................42
2.4 Results and discussion ...................................................................................................................43
2.4.1 Recognition of sesquiterpene glycosides from ion relative mass defects ........................43
2.4.2 Discovery of sesquiterpene diol glycosides from S. habrochaites LA1777 ....................49
2.5 Conclusions ....................................................................................................................................72
REFERENCES ....................................................................................................................................74
Chapter 3: Purification and structure elucidation of campherenane diol glycosides from leaf glandular
trichomes of Solanum habrochaites LA1777 ......................................................................................77
3.1 Introduction ....................................................................................................................................77
3.2 Experimental methods ...................................................................................................................78
3.2.1 Plant material ....................................................................................................................78
vii

3.2.2 Purification of campherenane diol glycosides..................................................................79
3.2.3 NMR Experiments ............................................................................................................80
3.2.4 Acid hydrolysis, derivatization and GC-MS experiments................................................80
3.3 Results and discussion ...................................................................................................................81
3.3.1 Profiles of the campherenane diol diglycosides ...............................................................81
3.3.2 Structure elucidation of sesquiterpene diol dihexoside malonate ester and establishing the
structure of campherenane terpenoid core ........................................................................86
3.3.3 Structure elucidation of sesquiterpene monohexoside malonate ester .............................87
3.3.4 Structure elucidation of sesquiterpene diol dihexoside ....................................................89
3.3.5 Structure elucidation of sesquiterpene diol dihexoside acetate ester ...............................89
3.3.6 Using GC-MS data to support NMR based structure assignments of sugar moieties ......90
3.4 Conclusions ....................................................................................................................................91
APPENDIX ..........................................................................................................................................92
REFERENCES ..................................................................................................................................122
Chapter 4: Purification and structure elucidation of sesquiterpene I and sesquiterpene II glycosides
from leaf glandular trichomes of Solanum habrochaites LA1777 .............................................126
4.1 Introduction ..................................................................................................................................126
4.2 Plant material for metabolite purification ....................................................................................127
4.3 Purification of sesquiterpene I and sesquiterpene II glycosides ..................................................128
4.4 Structure elucidation of sesquiterpene I diol dihexoside malonate ester .....................................129
4.5 Structure elucidation of Sesquiterpene II alcohol dihexoside acetate ester .................................131
4.6 Structure elucidation of Sesquiterpene II alcohol dihexoside......................................................133
4.7 Conclusion ...................................................................................................................................135
APPENDIX ........................................................................................................................................136
REFERENCES ..................................................................................................................................163
Chapter 5: Discovering terpene glycosides from Hoodia gordonii by mining multiplexed CID
UHPLC-MS data and applying relative mass defect filtering ...........................................................165
5.1 Introduction ..................................................................................................................................165
5.2 Methods........................................................................................................................................167
5.2.1 LC-MS and LC-MS/MS experiments ............................................................................167
5.2.2 Sample preparation for metabolite profiling ..................................................................168
5.3 Results and discussion .................................................................................................................168
5.3.1 Hoodia metabolite profiling ...........................................................................................168
5.3.2 Data processing for and relative mass defect filtering for mining publically available data
sets for the discovery of novel terpenoid metabolites ....................................................169
5.3.3 Variation of RMD of terpene glycosides and other non –terpenoid compounds. ..........171
5.3.4 Putative assignments of novel steroidal glycosides from Hoodia gordonii ...................171
5.4 Conclusions ..................................................................................................................................175
APPENDIX ........................................................................................................................................176
REFERENCES ..................................................................................................................................190
Chapter 6: Concluding remarks .........................................................................................................192

viii

LIST OF TABLES
Table 2.1: Characteristic fragment ions observed in negative ion mode MS/MS experiments for
various sugar oligosaccharides and mono saccharides. The masses shown correspond to [M-H]formed by each sugar group. The MS/MS spectra of candidate terpenoid compounds were examined
for the presence of these fragment ions for identification of the presence of these oligosaccharides in
the terpenoids. ....................................................................................................................................... 39
Table 2.2: Compounds identified as terpene glycosides from S. habrochaites LA1777 based on RMD
filtering of molecular and fragment ions. ............................................................................................ 56
Table 2.3: Compounds with greatest peak areas among the list of S. habrochaites LA1777
metabolites in the RMD range 440-636 ppm ........................................................................................ 59
Table 3.1: Chemical shifts (in ppm) of 1H and 13C resonances for (a) campherenane diol diglucoside
malonate ester (peak 6a from Figure 3.1), (b) campherenane diol monoglucoside malonate ester (peak
6b), (c) campherenane diol diglucoside (peak 8c), and (d) campherenane diol diglucoside acetate ester
(peak 5d). .............................................................................................................................................. 93
Table 4.1: Chemical shift assignments to the carbons and protons of the three compounds purified
from S. habrochaites LA1777............................................................................................................. 162

ix

LIST OF FIGURES
Figure 1.1 Chemical Structure of isoprene, the simplest terpene. .......................................................... 8
Figure 1.2 Mevalonate pathway of terpene biosynthesis. ..................................................................... 11
Figure 1.3 Biosynthesis of isopentenyl pyrophosphate starting with pyruvic acid. ............................. 12
Figure 2.1 Complexity of a plant extract is evident from the number of peaks in a UHPLC-MS base
peak chromatogram generated from a leaf dip extract of S. habrochaites LA1777. Analysis was
performed using a 110 min chromatographic gradient and detected in negative ion mode. ................ 43
Figure 2.2: MarkerLynx XS data processing results in a list of markers detected from the S.
habrochaites LA1777 trichome leaf extracts. Markers detected are reported with their m/z ratio,
retention time and their abundance in each sample being considered. ................................................. 44
Figure 2.3. (A) Negative ion mode multiplexed CID mass spectrum of acylsugar S4:22, RMD = 492
ppm. (B) Negative ion mode multiplexed CID mass spectrum of acylsugar S4:23, RMD = 402 ppm.
(C) Negative ion mode multiplexed CID mass spectrum of acylsugar S4:17, RMD = 440 ppm.
Values for RMD of the major fragment ions are presented. All detected isomers displayed fragments
of the same m/z values in negative ion mode spectra. All displayed negative ion mode multiplexed
CID mass spectra were obtained using a collision potential of -60 V. ................................................. 48
Figure 2.4 Negative-ion mode product ion MS/MS spectra of metabolites extracted from S.
habrochaites LA1777. (A) MS/MS spectrum of m/z 609 from campherenane diol diglucoside; (B)
MS/MS spectrum of m/z 811 from campherenane diol triglycoside malonate ester (m/z 811.3587,
RMD = 442 ppm). Values for RMD of the major fragment ions are presented. All
chromatographically-resolved isomers (10 isomers of m/z 651 and 12 isomers for m/z 811) displayed
fragments of the same m/z values in negative ion mode spectra. All negative ion mode MS/MS data
were obtained using collision potential of -50 V. ................................................................................. 51
Figure 2.5: Negative ion mode multiplexed CID mass spectrum of tomatine from S. habrochaites
LA1777 obtained at 60 V collision potential. Note that the fragment ions all have RMD values greater
than or equal to the [M-H]- ion at m/z 1032.5. ...................................................................................... 53
Figure 2.6: Variation of RMD of fragment ions as a function of ion m/z. Fragment ions were
generated in negative ion mode MS/MS for some representative compounds (acyl sugars,
sesquiterpene diglycoside malonate esters, triterpenoid glycoside and a triglycoside malonate ester)
found in S. habrochaites LA1777 leaf dip extracts. ............................................................................. 54
Figure 2.7: Relative mass defect filtering process used for discovering conjugated terpenoids from
raw LC-MS data .................................................................................................................................... 55
Figure 2.8. (A) Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 649)
for campherenan-2,12-diol malonate ester. (B) magnified region of m/z 228-255, showing the
x

sesquiterpenoid core fragment at m/z 239, which was too small to observe in A, where peaks are
normalized to the base peak. ................................................................................................................. 60
Figure 2.9. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 487) for
campherenan-2,12-diol monoglucoside malonate ester. The lower spectrum displays the magnified
region of m/z 224-256, showing the sesquiterpenoid core fragment at m/z 239 ................................... 61
Figure 2.10. Negative ion mode product ion MS/MS spectrum of products from [M+formate]- (m/z
609) for campherenan-2,12-diol diglucoside. The lower spectrum displays the magnified region of
m/z 228-258, showing the sesquiterpenoid core fragment at m/z 239 .................................................. 62
Figure 2.11. Negative ion mode product ion MS/MS spectrum of products from [M+formate]- (m/z
651) for campherenan-2,12-diol diglucoside acetate ester. The lower spectrum displays the magnified
region of m/z 224-264, showing the sesquiterpenoid core fragment at m/z 239. .................................. 63
Figure 2.12. Negative ion mode product ion MS/MS spectrum of products from [M+formate]- (m/z
771) for a Campherenane-2,12-diol triglycoside (Compound 13 in Table 2.2) ................................... 64
Figure 2.13. Negative ion mode product ion MS/MS spectrum of products from [M+formate]- (m/z
591) for a sesquiterpene II dihexoside (Compound 1 in Table 2.2) ..................................................... 64
Figure 2.14. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 631) for
a sesquiterpene II dihexoside malonate ester (Compound 3 in Table 2.2) ........................................... 65
Figure 2.15. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 793) for
a sesquiterpene II trihexoside malonate ester (Compound 5 in Table 2.2). ......................................... 65
Figure 2.16. Negative ion mode product ion MS/MS spectrum of products from [M+HCOO]- (m/z
633) for a sesquiterpene II dihexoside acetate ester (Compound 2 in Table 2.2). ................................ 66
Figure 2.17. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 661) for
a sesquiterpene I diglycoside malonate ester (Compound 22 in Table 2.2). ........................................ 66
Figure 2.18. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 499)
sesquiterpene I monoglycoside malonate ester (Compound 20 in Table 2.2). ..................................... 67
Figure 2.19. Negative ion mode product ion MS/MS spectrum of [M-H]- (m/z 455) for Sesquiterpene
I monoglycoside acetate ester (Compound 19 in Table 2.2). .............................................................. 67
Figure 2.20. Negative ion mode product ion MS/MS spectrum of [M-H]- (m/z 413) of Sesquiterpene
I monoglycoside (Compound 18 in Table 2.2). .................................................................................... 68
Figure 2.21. Negative ion mode product ion MS/MS spectrum of [M+HCOO]- (m/z 499) for
sesquiterpene I monoglycoside malonate ester (Compound 20 in Table 2.2). ..................................... 68

xi

Figure 2.22. Negative ion mode product ion MS/MS spectrum of [M-H]- (m/z 497) of Sesquiterpene
III monoglycoside malonate ester (Compound 17 in Table 2,2). ........................................................ 69
Figure 2.23. Negative ion mode product ion MS/MS spectrum of products of [M-H]- (m/z 411) of
Sesquiterpene III monoglycoside (compound 15 in Table 2.2) ........................................................... 69
Figure 2.24. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 617) for
Sesquiterpene I diglycoside acetate ester (compound 21). .................................................................. 70
Figure 2.25. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 823) for
Sesquiterpene I triglycoside malonate ester (compound 23 in Table 2.2). ........................................... 70
Figure 2.26. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 985) for
Sesquiterpene I tetraglycoside malonate ester (compound 24 in Table 2.2) ........................................ 71
Figure 2.27. Negative ion mode product ion MS/MS spectrum of products from [M+HCOO]- (m/z
489) for Campherenane diol Monoglycoside acetate ester (compound 7 in Table 2.2). ...................... 71
Figure 2.28. Negative ion mode product ion MS/MS spectrum of products from [M+HCOO]- (m/z
753) Sesquiterpene II Triglycoside (compound 4 in Table 2.2). .......................................................... 72
Figure 3.1. Extracted ion UHPLC-MS chromatograms showing sesquiterpene glycoside metabolites
identified in extracts from Solanum habrochaites LA1777. (a) [M-H]- (m/z 649) for campherenane
diol diglucoside malonate esters, (b) [M-H]- (m/z 487) for campherenane diol monoglucoside
malonate esters, (c) [M+formate]- (m/z 609) for campherenane diol diglucosides, (d) [M+formate](m/z 651) for campherenane diol diglucoside acetate esters, (e) [M-H]- (m/z 811) for campherenane
diol triglucoside malonate esters, (f) [M+formate]- (m/z 771) for campherenane diol triglucosides, (g).
[M+formate]- (m/z 446) for campherenane diol monoglucosides and (h) [M+formate]- (m/z 489) for
campherenane diol monoglucosides acetate ester. Labels with larger font size in red designate the
four purified metabolites. ...................................................................................................................... 85
Figure 3.2: Structures of (a) campherenane-2,12-diol diglucoside malonate ester (peak 6a from Figure
1), (b) campherenane-2,12-diol monoglucoside malonate ester (peak 6b), (c) campherenane-2,12-diol
diglucoside (peak 8c), and (d) campherenane-2,12-diol diglucoside acetate ester (peak 5d) as
determined using NMR and tandem mass spectrometry. Portions of the molecule corresponding to
key fragment ions in negative ion MS/MS spectra are illustrated on each structure. Carbon atoms are
numbered in accordance with NMR assignments in Table 3.1. .......................................................... 95
Figure 3.3. GC/MS total ion chromatograms of methoxime-trimethylsilyl derivatives of (a) products
of acid hydrolysis of campherenane-2-endo-(6'-malonyl)glucoside)-12-glucoside (Compound 1; peak
6a from Figure 1), (b) glucose reference standard, (c) xylose reference standard, (d) galactose
reference standard, (e) rhamnose reference standard. ........................................................................... 96
Figure 3.4. 1H and 13C NMR spectra of the isolated campherenane-2-endo-(6'-malonyl)glucoside)12-glucoside (Compound 1; peak 6a from Figure 3.1)......................................................................... 97

xii

Figure 3.5. 1H-1H COSY NMR spectrum of the campherenane-2-endo-(6'-malonyl)glucoside)-12glucoside (Compound 1; peak 6a from Figure 3.1). ............................................................................. 98
Figure 3.6. Multiplicity edited gHSQC (1H-13C) NMR spectrum of campherenane-2-endo-(6'malonyl)glucoside)-12-glucoside (Compound 1; peak 6a in Figure 1). ............................................... 99
Figure 3.7. gHMBC NMR spectrum of campherenane-2-endo-(6'-malonyl)glucoside)-12-glucoside
(Compound 1; peak 6a from Figure 1). .............................................................................................. 100
Figure 3.8. TOCSY NMR spectrum of campherenane-2-endo-(6'-malonyl)glucoside)-12-glucoside
(Compound 1; peak 6a in Figure 1). ................................................................................................... 101
Figure 3.9. NOESY NMR spectrum of campherenane-2-endo-(6'-malonyl)glucoside)-12-glucoside
(Compound 1; peak 6a in Figure 1). ................................................................................................... 102
Figure 3.10. 1H and 13C NMR spectra of the isolated 2-endo-campherenanol-12-(6'malonyl)glucoside (Compound 2; peak 6b from Figure 1). ............................................................... 103
Figure 3.11. 1H-1H COSY NMR spectrum of the 2-endo-campherenanol-12-(6'-malonyl)glucoside
(Compound 2; peak 6b from Figure 1). .............................................................................................. 104
Figure 3.12. Multiplicity edited gHSQC (1H-13C) NMR spectrum of 2-endo-campherenanol-12-(6'malonyl)glucoside (Compound 2; peak 6b in Figure 1). .................................................................... 105
Figure 3.13. gHMBC NMR spectrum of 2-endo-campherenanol-12-(6'-malonyl)glucoside
(Compound 2; peak 6b from Figure 1). .............................................................................................. 106
Figure 3.14. TOCSY NMR spectrum of 2-endo-campherenanol-12-(6'-malonyl)glucoside (Compound
2; peak 6b in Figure 1). ....................................................................................................................... 107
Figure 3.15. NOESY NMR spectrum of 2-endo-campherenanol-12-(6'-malonyl)glucoside
(Compound 2; peak 6b in Figure 1). ................................................................................................... 108
Figure 3.16. 1H and 13C NMR spectra of the isolated campherenan-2-endo,12-diglucoside
(Compound 3; peak 8c from Figure 3.1). ........................................................................................... 109
Figure Figure 3.17. 1H-1H COSY NMR spectrum of the campherenan-2-endo,12-diglucoside
(Compound 3; peak 8c from Figure 1). .............................................................................................. 110
Figure 3.18. Multiplicity edited gHSQC (1H-13C) NMR spectrum of campherenan-2-endo,12diglucoside (Compound 3; peak 8c in Figure 1). ................................................................................ 111
Figure 3.19. gHMBC NMR spectrum of campherenan-2-endo,12-diglucoside (Compound 3; peak 8c
from Figure 1). .................................................................................................................................... 112

xiii

Figure 3.20. TOCSY NMR spectrum of campherenan-2-endo,12-diglucoside (Compound 3; peak 8c
in Figure 1) .......................................................................................................................................... 113
Figure 3.21. 2D-NOESY NMR spectrum of campherenan-2-endo,12-diglucoside (Compound 3; peak
8c in Figure 1). .................................................................................................................................... 114
Figure 3.22. Proton NMR spectrum of 2-endo-campherenanol-12-(2-(6"-acetyl)glucosyl)glucoside
(Compound 4; peak 5d in Figure 1). ................................................................................................... 115
Figure 3.23. 13C NMR spectrum of 2-endo-campherenanol-12-(2-(6"-acetyl)glucosyl)glucoside
(Compound 4; peak 5d in Figure 3.1) ................................................................................................ 116
Figure 3.24. Multiplicity edited 2D gHSQC NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). ......................................................... 117
Figure 3.25. 2D gdqCOSY NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). ......................................................... 118
Figure 3.26. 2D HMBC NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). ......................................................... 119
Figure 3.27. 2D TOCSY NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). ......................................................... 120
Figure 3.28. 2D NOESY NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). ......................................................... 121
Figure Figure 4.1: HPLC-MS extracted ion chromatogram profiles of the three compounds purified
from S. habrochaites LA1777. a.) 10 isomers of Sesquiterpene I diol dihexoside malonate ester were
separated (b) three isomers of [M+HCOO-] of Sesquiterpene II alcohol dihexoside acetate ester were
detected (c) 8 isomers of [M+HCOO-] of Sesquiterpene II alcohol dihexoside were separated ........ 137
Figure 4.2. Structures of (a) Sesquiterpene I diol dihexoside malonate ester (peak 8a from Figure 4.1),
(b) Sesquiterpene II alcohol dihexoside acetate ester (peak 3b) and (c) Sesquiterpene II alcohol
dihexoside (peak 6c), determined using NMR and tandem mass spectrometry. Portions of the
molecule corresponding to key fragment ions in negative ion MS/MS spectra are illustrated on each
structure. Carbon atoms are numbered in accordance with NMR assignments in Table 4.1 ............. 138
Figure 4.3. 1D Proton NMR spectrum of compound a (peak 8a from Figure 4.1)............................. 139
Figure 4.4. 2D COSY spectrum of compound a (peak 8a from Figure 4.1). ...................................... 140
Figure 4.5. 2D HSQC spectrum of compound a (peak 8a from Figure 4.1). ..................................... 141
Figure 4.6. 2D HMBC spectrum of compound a (peak 8a from Figure 4.1)..................................... 142

xiv

Figure 4.7. 2D TOCSY spectrum of compound a (peak 8a from Figure 4.1). .................................. 143
Figure 4.8. 2D NOESY spectrum of compound a (peak 8a from Figure 4.1). .................................. 144
Figure 4.9. 1D Proton NMR spectrum of compound b (peak 3b from Figure 4.1) ........................... 145
Figure 4.10. 2D COSY spectrum of compound b (peak 3b from Figure 4.1) ................................... 146
Figure 4.11. 2D HSQC spectrum of compound b (peak 3b from Figure 4.1) ................................... 147
Figure 4.12. 2D HMBC spectrum of compound b (peak 3b from Figure 4.1). ................................. 148
Figure 4.13. 2D TOCSY spectrum of compound b (peak 3b from Figure 4.1). ................................ 149
Figure 4.14. 1D NMR spectrum of compound c (peak 6c from Figure 4.1) ..................................... 150
Figure 4.15. 2D COSY spectrum of compound c (peak 6c from Figure 4.1). ................................... 151
Figure 4.16. 2D HSQC spectrum of compound c (peak 6c from Figure 4.1). ................................... 152
Figure 4.17. 2D HMBC spectrum of compound c (peak 6c from Figure 4.1)................................... 153
Figure 4.18. 2D TOCSY spectrum of compound c (peak 6c from Figure 4.1). ................................ 154
Figure 4.19. 2D NOESY spectrum of compound c (peak 6c from Figure 4.1). ................................ 155
Figure 4.20. GC-MS chromatogram and EI spectrum of most abundant peak from the acid hydrolysis
product of compound b. ...................................................................................................................... 156
Figure 4.21. GC-MS chromatogram and EI spectrum of most abundant peak from the acid hydrolysis
product of compound c. ...................................................................................................................... 157
Figure 4.22. HMBC correlations assigned for compound a .............................................................. 158
Figure 4.23. Important COSY correlations assigned for compound a ................................................ 158
Figure 4.24. NOE correlations assigned for compound a .................................................................. 158
Figure 4.25. HMBC correlations assigned for compound b .............................................................. 159
Figure 4.26. COSY correlations assigned for compound b ............................................................... 159
Figure 4.27. NOE correlations assigned for compound b ................................................................. 160
Figure 4.28. HMBC correlations assigned for compound c .............................................................. 160

xv

Figure 4.29. COSY correlations assigned for compound c ................................................................ 161
Figure 4.30. NOE correlations assigned for compound c ................................................................... 161
Figure 5.1: Five diterpenoid cores reported from Hoodia compounds (a-e) and the six different sugar
groups that are found attached to the terpenoid core in current literature. ......................................... 177
Figure 5.2: Some representative diterpenoids found in Hoodia. ........................................................ 178
Figure 5.3: Distribution of RMD of compounds detected in Hoodia gordonii from metabolite
profiling in different RMD range categories. ..................................................................................... 179
Figure 5.4: Abundance of known and noval metabolites annotated from Hoodia gordonii ............. 180
Figure. 5.5 Positive ion MS/MS product ion spectrum of Hoodia metabolite m/z 803 [M+H]+ ....... 181
Figure 5.6 Positive ion MS/MS product ion spectrum of Hoodia metabolite m/z 838 [M+NH4]+ .... 181
Figure 5.7 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 641.................. 182
Figure 5.8 Positive ion product ion MS/MS spectrum of Hoodia metabolite of m/z 1304. ............... 183
Figure 5.9 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 1190................. 184
Figure 5.10 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 920................. 184
Figure 5.11 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 998................. 185
Figure. 5.12 Positive ion MS/MS product ion spectrum of m/z 1316................................................. 185
Figure. 5.13 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 1206.............. 186
Figure. 5.14 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 1222.............. 186
Figure. 5.15. Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 1192............. 187
Figure. 5.16 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 916................ 187
Figure. 5.17 Positive ion MS/MS product ion spectrum of (a) m/z 983 (putative alkaloid), (b)
Variation of RMD of parent ion and product ions in the MS/MS spectrum of an alkaloid compared to
the diterpene glycoside (RMD of sugar fragments are not included). The diterpene glycoside displays
a gradual increase of RMD from the parent ion to fragment ions whereas the alkaloid displays a
constant RMD from both fragment ions and parent ions. ................................................................... 188
Figure. 5.18 Positive ion MS/MS product ion spectra of Hoodia standards. ..................................... 189

xvi

Chapter 1
1.1 Plants as sources of drugs, food and sustainable energy
Since prehistoric times, humans have depended on plants as sustainable sources of food and for
treatments of illness [1]. However this dependence has been challenged by the increasing world
population, now exceeding 7 billion, that places a greater stress on limited resources (land, water
and clean air) to generate food and other necessities for human life. In addition, advancing
standards of living in many parts of the world have changed styles of consumption, resulting in
increased demand for fruits, vegetables, meat and dairy products [2]. Furthermore, increasing
scarcity of fossil fuels has resulted in greater focus on biofuel usage and using plants as derived
sources as industrial feedstocks [3,4]. All of these factors provide a driving force to better
understand the genetic capacity of organisms to produce valuable chemicals.
Plants, in contrast to animals, accumulate a vast diversity of chemicals (phytochemicals). This
makes them an invaluable source of compounds that can be difficult or expensive to produce via
synthetic means. Plants possess this capability owing to their complicated genomes through
mutations and gene duplications that have evolved to support sessile forms of life that depend on
accumulation of chemicals for its survival (reproduction, defense, energy accumulation)[5,6].
1.2 Modern technology for using plants to support a sustainable human life
This increased dependence of human lives on plants has led to a demand for plants that are more
productive, more resistant to disease and drought, and plants that produce better-flavored
products. As a result, genetic engineering and synthetic biology approaches have been employed
to engineer plants/microbes with greater crop potential, stronger resistance to extreme conditions
and more favorable characteristics [7]. However, successful application of synthetic biology
1

approaches requires understanding relevant biosynthetic pathways and having means of
engineering microbes or plants to produce compounds of interest [7,8]. The main challenges in
the process is that biology of living systems is often a result of synergetic interactions between
multiple biosynthetic pathways, biological compartments and the external environment of the
living system [8]. Therefore, in order to properly understand biological systems it is necessary to
understand the expression of the genes (transcriptomics), expression of the proteins (proteomics)
and the metabolites synthesized (metabolomics) leading to a thorough understanding of the
biological system considered. A transcriptome represents the complete set of transcripts (mRNA,
tRNA and rRNA) in a cell, an organ or organism, including information about the quantity of
each, at a specific developmental stage or physiological condition [9]. Although transcriptomes
provide information about levels of expression of individual genes, understanding the processes
taking place within cells requires quantitative analysis of the entire complement of proteins,
known as the proteome [10,11]. While these two approaches characterize transcription of genes
and translation into proteins they represent potential cell functions, but often fail to address
realized cellular functions. The suite of small molecules, or metabolites, in cells of tissues
reflects the functions of cellular proteins, and the study of the whole spectrum of small molecules
(molecules of mass less than 1500 Da are considered small molecules) is referred to as
“metabolomics”[12]. Recent advances in biotechnology has resulted an accumulation of large
amount of transcriptomic, proteomic, and metabolomics data from various biological systems.
However, the main challenge has been understanding the relationship between these in order to
understanding the function of those biological systems. This has been the theme of a number of
investigations in recent years.

2

Such investigations begin from the central dogma of biology where one gene was believed to be
associated with a single protein. Therefore, initial interest in understanding the genomes of
organisms led to sequencing whole genomes (e.g. Arabidopsis genome, human genome project).
However, it was soon realized that understanding the functions of genes is not accomplished as
easily due to the presence of multiple gene sequences that display great sequence similarity. On
the other hand, it was also discovered that slight differences in the sequences might lead to
completely different functionalities [13]. This made assigning the functions of the genes
challenging [13].
One main approach used in assigning gene functions in plants is called functional genomics,
where correlations between mRNA expression level and accumulation levels of specific
metabolites aid identification of a list of candidate genes that may be responsible for the
synthesis of various metabolites. Reductions in cost of sequencing and quantifying expressed
mRNA (Expressed Sequence Tags –ESTs) in a global approach, combined with quantification of
metabolites from the same tissue pools to assign putative functions of the genes [14]. However,
this approach provides only a set of candidate genes, and to confirm the identified candidate
genes, further experimentation that involves “knockdown of expressed sequences” such as virus
induced gene silencing (VIGS) or RNA interference (RNAi) is required. This makes mining the
metabolome to identify the metabolites that are a crucial function to the functional genomicsbased gene discovery process.
1.3 Growing interest in plant metabolite identification
Plants display the presence of a vastly diverse set of small molecules that are classified as
primary or secondary metabolites [15]. Primary metabolites are defined as metabolites that are
common to all organisms and specialized (secondary) metabolites represent all other small
3

molecules that are found only in select groups of plants due to the specialized biosynthetic
pathways present in those groups of plants [15]. The metabolic precursors of specialized
(secondary) metabolites are often central metabolites or their activated forms (amino acids for
peptides and proteins; isoprenyl-pyrophosphates for terpenes/terpenoids; acyl-CoA thioesters for
polyketides)[16,17]. However, compared to primary metabolites, specialized metabolites display
a vast diversity that is often restricted to a narrow range of genotypes[15]. This diversity is partly
attributed to mutated enzyme functions that exhibit altered substrate selectivity. Some
compounds undergo structural changes after biosynthesis of a core molecule (e.g. glycosylations,
phosphorylations, conjugation with other secondary metabolites) that also contributes to
metabolite structure diversity [18]. In comparison, some other compounds undergo regioselective
conformational changes that makes them regiospecific toward target receptors[19]. Today a great
interest is expressed toward metabolite identification both in plant sciences and in the drug
discovery field. However, the identification of novel metabolites and establishing their structures
remains one of the great bottlenecks that limits research progress in these areas.
In plant sciences, metabolite identification is a necessary step toward complete characterization
of metabolic phenotypes. This allows for the identification of stress-induced responses in plants
that is important in agricultural applications. For instance, plants exposed to drought accumulate
different levels of compounds than those grown in normal conditions[20,21]. Similarly, exposure
to insect attacks causes changes in the chemical composition of plants, most notably in levels of
specialized metabolites[22]. Novel genome sequencing technologies have enabled the rapid
generation of genome sequence information. Applications of phenotype characterization and coexpression gene analysis allows for accurate gene function annotation [23,24]. Such function

4

annotation of genes has led to genetic engineering of plants that resulted in the development of
crops of better taste, plants with more insect resistance and drought resistance [25,26].
However, it is likely that the greatest beneficiary from natural product structure elucidation is the
pharmaceutical industry. Natural products-based drug discovery benefits from metabolite
identification for both discovering novel drugs from natural products as well as for
characterization of biotransformation of new chemical entities [27,28,29]. Nature produces a
diverse collection of molecules (both small and large molecules) that are complementarity in
structure to their targets, making some of them excellent therapeutics. This has made natural
products obtained from plants (or other natural sources including fungi, animals, bacteria) an
excellent source of candidate drugs. Low molecular weight neurotransmitter molecules (e.g.
noradrenaline, adrenaline, serotonin, melatonin, histamines) are a good example of structural
complementarity of natural products to their targets. Identifying and elucidating the structures of
these compounds allowed for development of synthetic analogs that have become “blockbuster
drugs” (e.g. antihistamines and variants of the terpene indole alkaloid camptothecin that are used
in cancer treatments).
Theories have been proposed to explain why natural products display biological activity toward
humans/animals. One proposal is that the long term co-evolution of humans/animals with the
environment (plants) has led to the accumulation of compounds in nature that displays high
structural fidelity toward enzymes/receptors in humans/animals and vice versa[1]. Another
theory proposes that plants and animals originated from a common ancestor. The theory proposes
that this common ancestor was capable of accumulating a large diversity of compounds that
aided its survival under stressed environment in early evolutionary history. Later, as the plants
and animals evolved in different paths, the animals lost their ability to synthesize these
5

compounds whereas plants maintained it. However, the animals still retained their ability to
sense these compounds in plants that enabled their survival, while depending on plants as the
carbon source. Therefore nature consists of a collection of molecules that display structural
fidelity toward receptors in animals/humans [1,30].
1.4 Introduction to terpenes and trichomes
1.4.1 Trichomes in the genus Solanum and their chemical composition
Plants display “trichomes” that consist of specialized epidermal cells found on the surfaces of
leaves, stems, and other tissues. Trichomes can be either secretory glandular or non-secretory
epidermal cells. Secreting glandular trichomes (SGT) use the space between their gland cell
walls and cuticle to accumulate large quantities of metabolic products that act as a first line of
chemical defense to protect the plant from herbivory [31]. Some of the volatile compounds (such
as terpenes) in trichomes produce unpleasant odors making the plants less attractive as a food
source for insects [32]. Some other compounds such as acyl sugars accumulated in SGTs defend
plants in a more indirect manner. The trichomes and their constituents are commonly consumed
by lepidopteran larvae. However, acyl sugars contained in the trichomes release volatile odor
compounds once they are subjected to alkaline hydrolysis in midgets of the larvae. These odor
compounds attract the predators that attack the larvae and thereby protect the plant [33].
Comparatively, non-glandular trichomes act as physical or mechanical deterrents to insect
oviposition and feeding. Some other non-glandular trichomes (hooked trichomes) are believed to
be involved in plant defense by impaling insects [34]. About 30% of vascular plants possess
secretory glandular trichomes [35]. Trichomes are found across the genus Solanum including
tomato, tobacco, and potato[36,37,38]. Cultivated tomato (Solanum lycopersicum) and its wild

6

relatives display several different types of trichomes on hypocotyls, stems, leaves, floral organs,
and immature fruit.

In 1943, Luckwill identified four distinct types of secretory glandular trichomes (SGT) among
these based on their morphologies [39,40]. In Luckwill’s classification, type I trichomes were
distinguished from others by their multicellular base, a long (∼2 mm) multicellular stalk, and a
small glandular tip. In comparison, according to Luckwill, type IV trichomes have shorter stalk
(∼0.3 mm) and have a unicellular base, a multicellular stalk and a small glandular tip. Type VI
trichomes consist of a four-celled glandular head and a short (∼0.1 mm) multicellular stalk. Type
VII trichomes consist of a yet shorter (<0.05 mm) unicellular stalk and an irregularly shaped 4to 8-celled gland [40]. Type II, III and type V trichomes found on the surfaces of cultivated and
wild species of tomato are non-glandular trichomes [40].

Among the secretory glandular trichomes in plants of the genus Solanum, not only do the
trichomes differ in their morphology and the distribution, their chemical content also varies. For
instance, type VI trichomes are selective in accumulating terpenoids metabolites whereas type I
glands are the main sites of acylsugar accumulation in S. habrochaites [41,42]. However, recent
research suggests that type I and IV identified by Luckwill, are closely similar in chemical
content and gene expression[42]. One main objective of research on trichome chemistry is to
discover the genetic basis for control of trichome chemistry.

Understanding the factors that control trichome chemistry is expected to lead to development of
pest-resistant crops that can meet tomorrow’s agricultural needs. It has been demonstrated that as

7

wild plants were domesticated, they lost their ability to synthesize some defense compounds
observed in wild plants. As a result, most domesticated plants are susceptible to attack from pests
and various diseases [42]. It is expected that reintroduction of the lost biochemical machinery for
defense compound accumulation into crop species will generate crops that are more resistant to
insect and pathogenic attack.
1.4.2 Structures and biosynthesis of terpene metabolites in the genus Solanum
Structures of terpenes

Figure 1.1 Chemical Structure of isoprene, the simplest terpene.
Terpene structure was first explained by Ruzicka and Wallach [43] based on 2-methylbutadiene
(Figure 1.1), also called “isoprene”, structure. They proposed that terpenes are synthesized by the
head-to-head, head-to-tail or tail-to-tail linkages between isoprene units. This proposal, now
known as the isoprene rule, is adhered to by some terpenes (regular terpenes) but plenty of
exceptions, known as irregular terpenes, occur in nature. The isoprene rule proposes that in
natural terpenes, 1-1 (head-to-head) or 4-4 (tail-to-tail) linkages that connect the numbered
carbon atoms (Figure 1.1) are not normal. Examples of regular terpenes include the
monoterpenes (with 10 carbons, or two isoprenoid units) myrcene and limonene, and the
diterpenoid (with 20 carbons) retinol. An example of an irregular terpene, formed by tail-to-tail
linkage, is the tetraterpene β-carotene, which has 40 carbon atoms.

8

Among natural products, terpenes and terpenoids display a vast chemical diversity, and
discovery of novel terpenoids remains an active area of research. Currently two main terpene
biosynthetic pathways are known. These are mevalonate pathway (MVA pathway) and
methylerythritol pyrophosphate pathway (MEP pathway), which are described in more detail
below.
1.4.3 Biosynthesis of terpenes
The biosynthesis of terpenes is explained using two main pathways that generate precursors of
terpenes [39]. The mevalonate pathway (MVA pathway) takes place in the cytosol of the cell,
and the non-mevalonate pathway (MEP pathway) takes place mainly in chloroplasts of higher
plants. Accumulation of experimental results from various 13C labeling experiments that could
not be explained by MVA pathway led scientists to research alternative biosynthetic pathways
for terpene precursors. In the early–mid 1990s, the methylerythritol-4-phosphate pathway (MEP)
pathway was first proposed. Further research has shown that plant plastids and most bacteria
utilize the MEP pathway to accomplish terpene biosynthesis [38]. Both MEP and MVA
pathways lead to the synthesis of the two main universal precursors of terpene synthesis, which
are isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP). It has been
discovered that the two pathways do communicate between them to produce certain terpenes that
have origins from both pathways despite them taking place in different locations of the cell [44].
Mevalonate (MVA) pathway of terpene biosynthesis
A series of enzymatic reactions in the mevalonate pathway leads to the formation of isopentenyl
pyrophosphate and dimethylallyl diphosphate from three molecules of acetyl CoA (Figure 1.2).
From the 5-carbon precursors DMAPP and IPP, other terpenes such as geranyldiphosphate (10

9

carbons, GPP), farnesyldiphosphate (15 carbons, FPP), and geranylgeranyldiphosphate (20
carbons, GGPP) are produced. The enzymes that catalyze the linking of C5 isoprenoid units are
called prenyltransferases. Linear terpenoids consisting of 15 carbons are synthesized by
connecting three isoprene groups. This is performed by either cis- or trans-isopentyl
pyrophosphate synthases (IPPS) – cis and trans-FPP synthase. The stereochemistry specific
nature of FPP synthase was established in 1959[45,46,47] and was followed up by the discovery
of the first GPP synthase in 1964 [48]. Since then a number of prenyltransferase enzymes have
been characterized from various sources ranging from microbes to higher plants and animals
[46]. The presence of stereoselective prenyltransferases that catalyze the synthesis of terpenoids
of different lengths has resulted in extensive diversity among terpenes. This is even further
complicated by the presence of terpene synthases that can cyclize the linear precursors in various
ways [49].

10

Figure 1.2. Mevalonate pathway of terpene biosynthesis.
Non-mevalonate pathway of terpene synthesis - MEP Pathway
In the non-mevalonate pathway (Figure 1.3) methylerythritol phosphate (MEP) is produced
starting with pyruvate condensation with glyceraldehyde diphosphate. MEP acts as the precursor
of IPP.

11

Figure 1.3 Biosynthesis of isopentenyl pyrophosphate starting with pyruvic acid.
Terpene synthase (TPS) enzymes catalyze the formation of hemiterpenes (C5), monoterpenes
(C10), sesquiterpenes (C15) or diterpenes (C20) from DMAPP, GPP, FPP and GGPP. Despite
understanding biological activity, mode of action, structures and biosynthetic pathways for some
terpenes, our knowledge of biological roles of these compounds within plants has remained
limited. As the functions of terpene biosynthetic genes are discovered, it becomes feasible to
engineer plants and microbes that can produce these chemicals, either as enhanced chemical
defenses for plants, or as biotechnology products with economic value. To achieve these
advances, modern analytical chemistry and molecular biology tools must be integrated to probe
the chemical diversity generated by plants and engineered microbes.
The diversity of terpene and terpenoid chemistry as well as the availability of genetic sequence
information (both genomic DNA sequences and expressed sequence information) for both wild
and domesticated tomato plants makes them an attractive model system for discovery of
12

enzymes involved in the synthesis and metabolic modification of these molecules. Among these,
the wild tomato Solanum habrochaites LA1777 possesses a rich diversity of volatile terpenoids
[50,51]. Plants of the wild tomato species Solanum habrochaites LA1777 exhibit lush production
of glandular trichomes, and a rich diversity of terpenoids in its glandular trichomes that has yet
to be fully explored [38,42]. S. habrochaites produces volatile monoterpenes and sesquiterpenes
[52] as well as oxidized sesquiterpenes in the form of sesquiterpenoid acids [53].
1.4.4 Tomato and its terpenoid metabolites
Classical GC-MS based studies of tomato leaf have identified the presence of monoterpenes and
sesquiterpenes in tomato [50,54]. However, these studies have largely been restricted to volatile
terpenes since GC-MS is not applicable for the analysis of nonvolatile compounds unless they
have been derivatized to increase their volatility. Such is not common practice in terpene
analysis. Among the tomato plants studied for volatile terpenes, wild tomato S. habrochaites
LA1777 has shown to be of particular interest due to its richness in terpenes[50].
Some of the common monoterpenes found in S. habrochaites LA1777 include pinene, terpinene,
carene, limonene, and phellandrene[50]. However compared to monoterpenes, sesquiterpenes
found in this plant display greater diversity. Two classes of volatile sesquiterpenes are known
from S. habrochaites LA1777. Class I consists of germacrenes, α-humulene and βcaryophyllene. Class II consists of α-santalene, α-bergamotene, and β-bergamotene [55]. This
classification of sesquiterpenes is based on the location of the terpene synthase genes in the
chromosomes that correspond to the synthesis of each of these compound in the plants that
commonly accumulate them – Solanum habrochaies and Solanum lycopersicum [55]. Genes
coding for the terpene synthases responsible for the synthesis of Class I terpenes are located on
chromosome 6 of both species. Similarly the genes coding for the terpene synthases that
13

synthesize class II terpenes are found on chromosome 8 of both species[55]. Despite earlier
reports of volatile terpenes from S. habrochaites, the chemistry of nonvolatile terpenoids has
remained largely unexplored. Santalenoic and bergamotenoic acids were the first sesquiterpene
acids discovered and characterized from trichomes of S. habrochaites LA1777 [53], and they are
sufficiently volatile to allow for gas chromatographic analysis. Their presence indicates the
metabolic capacity to oxidize a side chain methyl group, and this provides a potential handle for
other metabolic modifications, but intermediate alcohol or aldehyde products have yet to be
reported. These sesquiterpene acids are notable in that they display insecticidal activity [56].
However, no conjugated terpenoids (e.g. sesquiterpene glycosides that are the main theme in this
dissertation) have been reported from S. habrochaites before.
Sesquiterpene glycosides are found in other plants of the Solanaceae including potato (Solanum
tuberosum) [57] and tobacco (Nicotiana tabacum) [58]. Some sesquiterpene glycosides possess
biological activity including hypoglycemic effects[59] , inhibition of tissue factor activity by a
nerolidol glycoside from Eriobotrya japonica [60] and immunomodulatory activity by
sesquiterpenoid glycosides from Dendrobium nobile [61,62] that makes them potential
pharmacophores and precursors of medicines. Therefore, exploring terpene glycosides from S.
habrochaites was important for both investigating the chemical diversity and for discovering
functions of genes that may influence plant terpenoid composition.
However, the main challenge in such studies lies with the limitations of conventional approaches
for metabolite identification [1,63]. Identification of novel metabolites depends on first
recognizing their presence, and this is often achieved through coupling of chromatography to
mass spectrometry, since many sesquiterpenoids lack characteristic chromophores that signal
their presence. A common strategy for metabolite annotation includes performing MS/MS
14

experiments on compounds of interest and using characteristic fragment ions to identify natural
products that belong to a particular class of compounds. Flavonoid identification provides an
example of this approach [64,65]. However, in the absence of MS/MS information or when
characteristic fragment ions are not present, these methods pose limitations. The absence of a
strategy to annotate unknown compounds in plant natural product profiles owing to the limited
availability of standards has caused the pharmaceutical industry to move away from natural
products-based drug discovery to synthetic library based drug discovery over the last decade [1].
In contrast to other secondary metabolite classes including flavonoids and alkaloids, terpenoids
display a vast structural diversity owing to variation in backbone structure, stereochemistry,
double bond position, and other sites of functionalization[66]. This diversity has rendered
establishing characteristic fragment ions for the direct identification of non-volatile terpenoids
such as terpene glycosides using mass spectrometry challenging. Therefore, recognition of the
terpene glycosides in a metabolite profile has been based on the absence of fragment ions that
represent other classes of molecules such as flavonoids [67]. While GC-MS has enabled the
identification of volatile and semivolatile terpenes for many decades it is not amenable for
nonvolatile terpenoids. As a result, conjugated terpenoid characterization has been limited
mostly to the most abundant members of this class that have been purified.
1.4.5 Characteristic fragment ions for identifying terpenes from plant tissues using Gas
Chromatography –Mass Spectrometry
Volatile terpenes are characterized by the presence of some characteristic ions in their mass
spectra. For monoterpenes, these include the molecular ion (M+) at m/z 136, and fragment 121,
93 and 69. For sesquiterpenes, characteristic fragment ions are m/z 204 (M+),161, 136, 121, 93
and 69 [68]. Furthermore, monoterpenes elute earlier (at lower temperatures in the temperature
15

program) than sesquiterpenes during GC-MS [68]. These characteristic fragment ions have been
used in the detection of terpenes using GC-MS in a number of studies [68,69,70], but many
terpenes yield similar mass spectra. As a result, values of chromatographic retention indices are
often needed to provide additional evidence of terpene identities. In contrast, identification of
glycosylated terpenoids presents a challenge due to their non-volatile nature, and LC-MS is a
more appropriate strategy. However, glycosylated terpenoids often fail to yield characteristic
fragment ions in MS/MS spectra that enable their conclusive identification.
1.5 Discovery and identification of specialized metabolites
Identification of novel compounds usually requires their isolation from complex matrices.
Separating components of interest from complex matrices has been performed for thousands of
years. For instance, extraction of precious metals from ore was performed even in ancient
civilizations [71]. Classical separation methods depended on differential physical properties
including a substance’s solubility and boiling point of the compound of interest relative to the
matrix to achieve separation. “Distillation” the most commonly used method for separating
essential oils from plant leafs even today, has been used since medieval times [72]. Similarly,
liquid-liquid extraction and precipitation are used both in research and in industrial applications
widely [73]. Regardless of the application, the ultimate objective of separation methods is to
obtain a target compound/element in the maximum feasible purity.
For novel compound identifications, classical structural characterization was performed based on
elemental analysis using combustion, physical parameters (melting point, boiling point,
refractive index, optical rotation) and chemical reactivity. With the advent of ultraviolet
absorption spectroscopy, infrared spectroscopy, x-ray crystallography, mass spectrometry, and
nuclear magnetic resonance (NMR) spectroscopy, absolute structure assignment became possible
16

[74,75]. However, successful application of these methods toward the structure elucidation
requires a pure compound. This poses a challenge since specialized metabolites in plants are
present as low-abundance components of complex mixtures of compounds. Therefore,
purification of compounds is of paramount importance for their identification.
Modern analytical chemistry often employs chromatographic methods for the purification of
compounds. Earliest attempts of chromatography were performed by Mikhail Tswett [76,77].
Improvements in separation science were achieved with advancements in our understanding of
physical chemistry of compound retention on solid particles [78,79]. Since then, a number of
advances have been made in the field of chromatography, and the Nobel Prize in Chemistry
awarded in 1952 to A.J.P Martin and R.L.M Synge for “inventing partition chromatography”
stands out as a landmark [76]. Improvements in separation science were achieved with advances
in our understanding of physical chemistry of compound partition onto solid particles [78,79]
This research led to the development of column based liquid chromatographic (LC) methods
which was followed by phenomenal advances in instrumentation and column chemistries [80].
These advances led to fast high-performance LC separations, resulting in more efficient
purification of natural products. A series of hyphenated methods, involving coupling of LC
separations to a variety of detectors further accelerated use of liquid chromatography[81]. These
include gas chromatography–mass spectrometry (GC-MS), liquid chromatography – mass
spectrometry (LC-MS), liquid chromatography–NMR (LC-NMR), LC-UV, LC-UV-MS and LCMS-NMR [74,82]. These methods propelled novel compound discovery and characterization by
making structural information available at an early stage of analysis without the need to obtain a
substantial amount of compound in pure form. Such explosive growth in structural information

17

has enabled more comprehensive analytical surveys of chemical composition that is described as
the “omics” approach.
1.6 Mass spectrometry-based metabolomics as a functional genomics tool
1.6.1 Mass Spectrometry
Molecular mass is one of the single most useful measurements one can make to distinguish one
substance from another. Mass spectrometry is an experimental technique that allows for the
determination of mass-to-charge ratio (m/z) of an ion in the gas phase. Therefore, fundamental
requirements that need to be fulfilled for mass spectrometric characterization of a compound are
a) it must be taken into the gas phase and b) it must form an ion (positive or negative) in the gas
phase.
Some mass spectrometers (often called high resolution mass spectrometers) can measure the m/z
of an ion with high accuracy (within 1 ppm of the theoretical mass). In contrast, some other mass
spectrometers (low or unit resolution mass spectrometers) can measure m/z only with unit
accuracy (±1 Da). The mass reported by high-resolution mass spectrometers is termed “accurate
mass” since it is closer to the true mass of the compound measured. The mass reported by a unit
resolution instrument is termed “nominal mass” of a given compound, and is a value usually
rounded to an integer value. For instance, a compound might be detected at accurate mass of m/z
649.3035 in negative ion mode using a high-resolution mass spectrometer. The decimal
component of this reported mass (i.e. 0.3036) is termed “mass defect” of a compound. The same
compound will be reported as the nominal mass m/z 649 by a unit resolution mass spectrometer.

18

Some common high resolution mass spectrometers use time-of-flight (ToF) and Orbitrap
technologies for the detection of ions. Unit resolution mass spectrometers are commonly based
on the use of one or more quadrupoles for ion detection.
1.6.2 Collision induced dissociation (CID) and MS/MS
International union of pure and applied chemistry’s (IUPAC) has defined Collision induced
dissociation as “an ion/neutral species interaction wherein the projectile ion is dissociated as a
result of interaction with a target neutral species. This is brought about by conversion of part of
the translational energy of the ion to internal energy in the ion during collision”. Therefore, the
three key steps involved in CID are, interaction between a charged ion and a neutral species,
energy transfer and dissociation. In modern mass spectrometers nitrogen gas is used as the
neutral species and the ions are brought into collisions with the gas. The importance of CID is
that it allows for the fragmentation of ions of a compound of interest and thereby yields further
structural information about that compound. A common experiment used in this context is the
MS/MS experiment.
MS/MS experiments are mass spectrometric experiments where all ions except for the ion of
interest are filtered to isolate the ion of interest followed by colliding it with a neutral species
(such as nitrogen) to generate fragment ions. Commonly, these experiments are performed using
mass spectrometers equipped with one or more quadrupoles. Quadrupoles are used for filtering
the ion of interest from others.
1.6.3 Multiplexed CID mass spectrometry
While MS/MS experiments are useful for structural characterization of compounds usually it is
applicable to only a few compounds of interest at a time. This poses a bottleneck for discovery

19

metabolomics where obtaining mass spectral information on both parent ions and their fragment
ions for as many compounds as possible is gives the best possibility of discovering novel
compounds. Multiplexed CID mass spectrometry allows for further advancement by
multiplexing MS/MS experiments. Multiplexed CID MS is a method that enables the accurate
mass information of both parent ions and fragment ions of a compound to be accumulated
simultaneously using time-of-flight technology without isolating the ion of interest. Since no ion
isolation is performed, this method allows for fragmentation information of all parent ions that
reaches the mass spectrometer at any time to be obtained. One obvious complication here is that
multiple compounds that enter the mass spectrometer will produce an assortment of fragment
ions making spectral interpretation complicated. However, since this experiment is performed
using a Time-of-flight instrument the accurate mass of both parent ions and fragment ions will be
obtained, simplifying the spectral interpretation process. Furthermore, to simplify the data
interpretation process commonly compounds in complicated mixtures (such as plant extracts) are
separated using a chromatographic method (Ex: liquid chromatography) to minimize the number
of compounds that reach the mass spectrometer at one time to make spectral interpretation easier.
Multiplexed CID MS functions by elevating the energy applied to the ions that enter the mass
spectrometer during their movement from ion source to the ion detector. Furthermore, the
pressure of the region of the mass spectrometer where the multiplexed CID MS fragmentation
occurs is approximately 2×10-3 mbar which is about three orders of magnitude greater than
pressure in the ion detector. This allows for enough collisions between ions and collision gas to
take place enabling some of the parent ions to fragment while another proportion of them remain
unfragmented. Thus, both fragment ion mass to charge ratio and the daughter ion mass to charge
ratios can be measured simultaneously. Recent use of LC-multiplexed CID mass spectrometry
20

for rapid profiling of metabolites from wild tomato relative Solanum habrochaites accessions has
been well documented [83,84].
1.6.4 Mass spectrometry for discovery of novel metabolites
LC-MS opened gates to a world of compounds not observed before: Gas chromatography – mass
spectrometry (GC-MS) has emerged as the most powerful approach to analyze volatile
components from plant extracts [85]. However, GC-MS is restricted to non-polar to low polar
compounds that are volatile or compounds that can be derivatized to enhance their volatility.
Therefore the number of compounds observed using GC-MS from plant extracts is restricted to a
few hundred [85]. However, using liquid chromatography-mass spectrometry (LC-MS) it has
been possible to extend detection and determination of molecular masses to non-volatile
compounds that are more polar and have not been reported before [86]. This unraveled insights
into a richer plant metabolome that were not observed using GC-MS analyses. Using LC-MS,
both qualitative and quantitative assessments of the compounds in the metabolome can be made.
However, owing to the diverse structures and dynamic range of metabolites found in plants,
characterization of a given metabolome has required the use of a diverse set of tools to achieve
comprehensive coverage of the metabolites [14,87].
The vast majority of plant metabolites are unknown. Thanks to the enhancements of
instrumentation technologies the detection limits of mass spectrometers have become lower and
as a result many more compounds that were not observed in plants before are becoming apparent
(e.g. Medicinal plants consortium database - http://medicinalplantgenomics.msu.edu/index.shtml
). This enables greater amount of novel data to be accumulated, which in turn challenges the
scientists to identify them. Therefore, for annotation of these metabolites that are not found in
databases, a different approach is necessary.
21

Metabolite identification is the main challenge ahead: Identification of metabolites observed in
metabolite profiles is the main bottleneck in today’s metabolomics [88,89]. To address this
challenge, most laboratories use metabolite databases including MassBank, HMBD, LipidMaps,
METLIN and NIST databases [89]. The metabolite profiling data that are collected from a given
sample are matched with these databases based on accurate mass to find matching compounds.
Standards of metabolite annotation are used to enhance the accuracy and the quality of database
available to make compound annotations based on them more accurate [90]. A second approach
is to use authentic standards to verify the identity of compounds [91]. This involves obtaining the
authentic standards and comparing their reported mass, fragment ions formed by collision
induced dissociation (or other fragmentation method) and retention time with those found in
unknown samples to find matching compounds. A commonly employed third approach is to
apply mass defect (described under section 1.5.1) filters that would identify changes in accurate
mass to assign compounds as derivatives of known compounds [92]. However, these approaches
are useful only when searching for compounds that have already been annotated elsewhere and
have been entered into the database or in the presence of authentic standards making novel
strategies for metabolite identification a fundamental requirement.
In early days of metabolomics it was proposed that unknown metabolite annotation could be
achieved based on accurate mass and database searches [70], perhaps under the naïve assumption
that unknowns must bear similarity to known compounds. However, soon this perception was
changed and it was apparent that accurate mass alone was not adequate for compound
identification [93,94]. To enable accurate metabolite identification based on mass spectrometry,
a combination of accurate mass information, isotopic abundance information and MS/MS data
for the metabolite are used [89,93]. To obtain this information usually, first a mass spectrum is
22

obtained using a high-resolution instrument such as a ToF mass spectrometer that will allow the
establishment of the accurate mass followed by performing the MS/MS experiments on selected
masses of interest.
Despite all the information provided by such mass spectrometry experiments, these analyses
alone do not allow the structure of an unknown metabolite to be confirmed. To achieve de novo
structure elucidation, the compound of interest need to be either synthesized or purified from its
matrix, followed by performing 1D and 2D NMR based structure elucidation [12].
1.7 Summary of research
Among the secondary metabolites, glycosylated forms of terpenoids makes up a large proportion.
These are hybrid molecules that consist of a terpene group synthesized from isoprenoid
precursors and a glycoside component made up of various sugar moieties. Glycosylation is a
common form of conjugation of metabolites that enhances their water solubility, which in turn,
facilitates their intra- and inter-cellular movements within the plant. In addition, glycosylation
and other conjugations stabilize metabolites by creating forms more readily sequestered in
specific organelles (e.g. vacuoles) and allow them to be stored for longer periods of time because
they are out of contact with enzymes that may transform them further. Steroidal glycosides are a
class of compounds that is well characterized. In contrast sesquiterpene glycosides are less well
known.
Sesquiterpene glycosides consists of a terpenoid core synthesized by three isoprene groups and is
conjugated to one or more carbohydrate moieties. Sesquiterpenes could be acyclic or consist of
ring systems

23

Chapter 2, 3 and 4 of this dissertation describe the strategies for discovery, structure elucidation
and exploration of chemical diversity of various sesquiterpene glycosides from various
accessions of a wild tomato relative Solanum habrochaites. The genus Solanum is part of the
plant family Solanaceae that includes a number of agriculturally important crops including
tobacco, eggplant, pepper and petunia. This plant was selected for this research due to the
availability of the genomic DNA sequence for tomato and EST (expressed sequence tag)
sequences of wild tomato relatives. Also a number of tomato introgression lines developed are
available [95]. Both these resources enable the identification of genes involved in the synthesis
of compounds of interest. In addition, a rich history of characterization of compounds and genes
from this family exists that is useful for this research. Thirdly, tomato is a plant with a fairly
short lifecycle and a plant that is not too large in physical dimensions. This makes obtaining
offspring easy allowing for easier manipulation of the plant. All these led to selection of this as a
model plant for the characterization of metabolome in this project. Furthermore, this research has
led to the development of a novel method for rapid classification of metabolites into various
metabolite classes. Chapter 5 of this dissertation describes the application of this method for the
analysis and discovery of compounds from metabolome of Hoodia gordonii and Rosmarinus
officinalis, two plants studied in the development of the Medicinal Plants Consortium database.
H. gordonii is a plant that has been extensively studied for the presence of dietary supplements.
Similarly, R. officinalis is a medicinal plant that is used to prepare dietary supplements and is
used as a spice in most culinary practices.

24

REFERENCES

25

REFERENCES
1. H.F. Ji, X.J. Li, H.Y. Zhang, Natural products and drug discovery. Can thousands of years of
ancient medical knowledge lead us to new and powerful drug combinations in the fight
against cancer and dementia?, EMBO Rep 10 (2009) 194-200.
2. Asia-Development-Bank, Food security and poverty in asia and the pacific. Key challenges
and policy issues., (2012).
3. D. Graham-Rowe, Agriculture: Beyond food versus fuel, Nature 474 (2011) S6-8.
4. R.L. Last, A.D. Jones, Y. Shachar-Hill, Towards the plant metabolome and beyond, Nat Rev
Mol Cell Biol 8 (2007) 167-174.
5. S.A. Rensing, D. Lang, A.D. Zimmer, A. Terry, A. Salamov, H. Shapiro, T. Nishiyama, P.F.
Perroud, E.A. Lindquist, Y. Kamisugi, T. Tanahashi, K. Sakakibara, T. Fujita, K. Oishi, T.
Shin-I, Y. Kuroki, A. Toyoda, Y. Suzuki, S. Hashimoto, K. Yamaguchi, S. Sugano, Y.
Kohara, A. Fujiyama, A. Anterola, S. Aoki, N. Ashton, W.B. Barbazuk, E. Barker, J.L.
Bennetzen, R. Blankenship, S.H. Cho, S.K. Dutcher, M. Estelle, J.A. Fawcett, H. Gundlach,
K. Hanada, A. Heyl, K.A. Hicks, J. Hughes, M. Lohr, K. Mayer, A. Melkozernov, T. Murata,
D.R. Nelson, B. Pils, M. Prigge, B. Reiss, T. Renner, S. Rombauts, P.J. Rushton, A.
Sanderfoot, G. Schween, S.H. Shiu, K. Stueber, F.L. Theodoulou, H. Tu, Y. Van de Peer, P.J.
Verrier, E. Waters, A. Wood, L.X. Yang, D. Cove, A.C. Cuming, M. Hasebe, S. Lucas, B.D.
Mishler, R. Reski, I.V. Grigoriev, R.S. Quatrano, J.L. Boore, The Physcomitrella genome
reveals evolutionary insights into the conquest of land by plants, Science 319 (2008) 64-69.
6. M. Hurles, Gene duplication: The genomic trade in spare parts, Plos Biology 2 (2004) 900904.
7. J.D. Keasling, Building with biology, Nature 492 (2012) 188-188.
8. D.I. Ellis, R. Goodacre, Metabolomics-assisted synthetic biology, Curr Opin Biotechnol 23
(2012) 22-28.
9. Z. Wang, M. Gerstein, M. Snyder, RNA-Seq: a revolutionary tool for transcriptomics, Nature
Reviews Genetics 10 (2009) 57-63.
10. S.P. Gygi, B. Rist, S.A. Gerber, F. Turecek, M.H. Gelb, R. Aebersold, Quantitative analysis
of complex protein mixtures using isotope-coded affinity tags, Nature Biotechnology 17
(1999) 994-999.
11. R. Aebersold, M. Mann, Mass spectrometry-based proteomics, Nature 422 (2003) 198-207.
12. O. Fiehn, Metabolomics - the link between genotypes and phenotypes, Plant Molecular
Biology 48 (2002) 155-171.
26

13. D.G. Wang, J.B. Fan, C.J. Siao, A. Berno, P. Young, R. Sapolsky, G. Ghandour, N. Perkins,
E. Winchester, J. Spencer, L. Kruglyak, L. Stein, L. Hsie, T. Topaloglou, E. Hubbell, E.
Robinson, M. Mittmann, M.S. Morris, N. Shen, D. Kilburn, J. Rioux, C. Nusbaum, S. Rozen,
T.J. Hudson, R. Lipshutz, M. Chee, E.S. Lander, Large-scale identification, mapping, and
genotyping of single-nucleotide polymorphisms in the human genome, Science 280 (1998)
1077-1082.
14. L.W. Sumner, P. Mendes, R.A. Dixon, Plant metabolomics: large-scale phytochemistry in
the functional genomics era, Phytochemistry 62 (2003) 817-836.
15. E. Pichersky, D.R. Gang, Genetics and biochemistry of secondary metabolites in plants: an
evolutionary perspective, Trends in Plant Science 5 (2000) 439-445.
16. R. McDaniel, S. Ebertkhosla, D.A. Hopwood, C. Khosla, Engineered biosynthesis of novel
polyketides, Science 262 (1993) 1546-1550.
17. J. Bohlmann, G. Meyer-Gauen, R. Croteau, Plant terpenoid synthases: Molecular biology and
phylogenetic analysis, Proceedings of the National Academy of Sciences of the United States
of America 95 (1998) 4126-4133.
18. C.M.M. Gachon, M. Langlois-Meurinne, P. Saindrenan, Plant secondary metabolism
glycosyltransferases: the emerging functional analysis, Trends in Plant Science 10 (2005)
542-549.
19. M. Kujawa, H. Ebner, C. Leitner, B.M. Hallberg, M. Prongjit, J. Sucharitakul, R. Ludwig, U.
Rudsander, C. Peterbauer, P. Chaiyen, D. Haltrich, C. Divne, Structural basis for substrate
binding and regioselective oxidation of monosaccharides at C3 by pyranose 2-oxidase,
Journal of Biological Chemistry 281 (2006) 35104-35115.
20. B. Hu, J. Simon, H. Rennenberg, Drought and air warming affect the species-specific levels
of stress-related foliar metabolites of three oak species on acidic and calcareous soil, Tree
Physiology 33 (2013) 489-504.
21. M.A.M. Khan, C. Ulrichs, I. Mewis, Water stress alters aphid-induced glucosinolate response
in Brassica oleracea var. italica differently, Chemoecology 21 (2011) 235-242.
22. P.W. Pare, J.H. Tumlinson, Plant volatiles as a defense against insect herbivores, Plant
Physiology 121 (1999) 325-331.
23. T. Tohge, A.R. Fernie, Combining genetic diversity, informatics and metabolomics to
facilitate annotation of plant gene function, Nature Protocols 5 (2010) 1210-1227.
24. T. Tohge, A.R. Fernie, Annotation of Plant Gene Function via Combined Genomics,
Metabolomics and Informatics, (2012) e3487.

27

25. J.A. Gatehouse, Biotechnological prospects for engineering insect-resistant plants, Plant
Physiology 146 (2008) 881-887.
26. Z.G. Li, L.H. Yao, Y.W. Yang, A.D. Li, Transgenic approach to improve quality traits of
melon fruit, Scientia Horticulturae 108 (2006) 268-277.
27. C. Prakash, C.L. Shaffer, A. Nedderman, Analytical strategies for identifying drug
metabolites, Mass Spectrom Rev 26 (2007) 340-369.
28. A. Craney, S. Ahmed, J. Nodwell, Towards a new science of secondary metabolism, Journal
of Antibiotics 66 (2013) 387-400.
29. A.R. Fernie, R.N. Trethewey, A.J. Krotzky, L. Willmitzer, Metabolite profiling: from
diagnostics to systems biology, Nat Rev Mol Cell Biol 5 (2004) 763-769.
30. J. Clardy, C. Walsh, Lessons from natural molecules, Nature 432 (2004) 829-837.
31. G.J. Wagner, Secreting glandular trichomes - more than just hairs, Plant Physiology 96
(1991) 675-679.
32. J.-H. Kang, G. Liu, F. Shi, A.D. Jones, R.M. Beaudry, G.A. Howe, The Tomato odorless-2
Mutant Is Defective in Trichome-Based Production of Diverse Specialized Metabolites and
Broad-Spectrum Resistance to Insect Herbivores, Plant Physiology 154 (2010) 262-272.
33. A. Weinhold, I.T. Baldwin, Trichome-derived O-acyl sugars are a first meal for caterpillars
that tags them for predation, Proceedings of the National Academy of Sciences (2011).
34. D.A. Levin, The Role of Trichomes in Plant Defense., The quarterly review of biology 48
(1973) 3-15.
35. A. Fahn, Structure and function of secretory cells, Advances in Botanical Research
Incorporating Advances in Plant Pathology, Vol 31 2000 31 (2000) 37-75.
36. A.L. Schilmiller, R.L. Last, E. Pichersky, Harnessing plant trichome biochemistry for the
production of useful compounds, Plant Journal 54 (2008) 702-711.
37. R.W. Gibson, J.A. Pickett, Wild tomato repels aphids by release of aphid alarm pheromone.,
Nature 302 (1983) 608-609.
38. W.M. Tingey, S.A. Mehlenbacher, J.E. Laubengayer, Occurrence of glandular trichomes in
wild Solanum species, American Potato Journal 58 (1981) 81-83.
39. L. Luckwill, The genus Lycopersicon: historical, biological, and taxonomic survey of the
wild and cultivated tomatoes, Aberdeen University Press, Aberdeen, Scotland, 1943.

28

40. J.H. Kang, F. Shi, A.D. Jones, M.D. Marks, G.A. Howe, Distortion of trichome morphology
by the hairless mutation of tomato affects leaf surface chemistry, Journal of Experimental
Botany 61 (2010) 1053-1064.
41. A.L. Schilmiller, I. Schauvinhold, M. Larson, R. Xu, A.L. Charbonneau, A. Schmidt, C.
Wilkerson, R.L. Last, E. Pichersky, Monoterpenes in the glandular trichomes of tomato are
synthesized from a neryl diphosphate precursor rather than geranyl diphosphate, Proceedings
of the National Academy of Sciences of the United States of America 106 (2009) 1086510870.
42. E.T. McDowell, J. Kapteyn, A. Schmidt, C. Li, J.H. Kang, A. Descour, F. Shi, M. Larson, A.
Schilmiller, L.L. An, A.D. Jones, E. Pichersky, C.A. Soderlund, D.R. Gang, Comparative
Functional Genomic Analysis of Solanum Glandular Trichome Types, Plant Physiology 155
(2011) 524-539.
43. L. Ruzicka, The isoprene rule and the biogenesis of terpenic compounds, Experientia 9
(1953) 357-367.
44. C. Schuhr, T. Radykewicz, S. Sagner, C. Latzel, M. Zenk, D. Arigoni, A. Bacher, F.
Rohdich, W. Eisenreich, Quantitative assessment of crosstalk between the two isoprenoid
biosynthesis pathways in plants by NMR spectroscopy, Phytochemistry Reviews 2 (2003) 316.
45. G. Popjak, Cornfort.Jw, Substrate stereochemistry in squalene biosynthesis
Journal 101 (1966) 553-&.

, Biochemical

46. K.C. Wang, S. Ohnuma, Isoprenyl diphosphate synthases, Biochimica Et Biophysica ActaMolecular and Cell Biology of Lipids 1529 (2000) 33-48.
47. F. Lynen, B.W. Agranoff, H. Eggerer, U. Henning, E.M. Möslein, γ,γ-Dimethyl-allylpyrophosphat und Geranyl-pyrophosphat, biologische Vorstufen des Squalens Zur
Biosynthese der Terpene, VI1), Angewandte Chemie 71 (1959) 657-663.
48. A.A. Kandutsch, E. Levin, K. Bloch, H. Paulus, Purification of gerenylgerenyl
pyrophosphate synthase from Micrococcus lysodeiticus, Journal of Biological Chemistry 239
(1964) 2507-&.
49. B. Greenhagen, J. Chappell, Molecular scaffolds for chemical wizardry: Learning nature's
rules for terpene cyclases, Proceedings of the National Academy of Sciences of the United
States of America 98 (2001) 13479-13481.
50. E. Gonzales-Vigil, D.E. Hufnagel, J. Kim, R.L. Last, C.S. Barry, Evolution of TPS20-related
terpene synthases influences chemical diversity in the glandular trichomes of the wild tomato
relative Solanum habrochaites, Plant Journal 71 (2012) 921-935.

29

51. Y. Matsuba, T.T.H. Nguyen, K. Wiegert, V. Falara, E. Gonzales-Vigil, B. Leong, P. Schafer,
D. Kudrna, R.A. Wing, A.M. Bolger, B. Usadel, A. Tissier, A.R. Fernie, C.S. Barry, E.
Pichersky, Evolution of a Complex Locus for Terpene Biosynthesis in Solanum, Plant Cell
25 (2013) 2022-2036.
52. L. Lundgren, G. Norelius, G. Stenhagen, Leaf volatiles from some wild tomato species,
Nordic Journal of Botany 5 (1985) 315-320.
53. R.M. Coates, J.F. Denissen, J.A. Juvik, B.A. Babka, Identification of α-santalenoic and endoβ-bergamotenoic acids as moth oviposition stimulants from wild tomato leaves, Journal of
Organic Chemistry 53 (1988) 2186-2192.
54. M.A. Farag, P.W. Pare, C-6-green leaf volatiles trigger local and systemic VOC emissions in
tomato, Phytochemistry 61 (2002) 545-554.
55. R.S. van der Hoeven, A.J. Monforte, D. Breeden, S.D. Tanksley, J.C. Steffens, Genetic
control and evolution of sesquiterpene biosynthesis in Lycopersicon esculentum and Lhirsutum, Plant Cell 12 (2000) 2283-2294.
56. J.A. Juvik, M.J. Berlinger, T. Ben-David, J. Rudich, Resistance among accessions of the
generalycopersicon andsolanum to four of the main insect pests of tomato in Israel,
Phytoparasitica 10 (1982) 145-156.
57. H. Tazaki, N. Ohta, K. Nabeta, H. Okuyama, M. Okumura, Structure of sesquiterpene
glucosides from potato leaves, Phytochemistry 34 (1993) 1067-1070.
58. X. Feng, J.S. Wang, J. Luo, L.Y. Kong, Two new sesquiterpene glucosides from the leaves
of Nicotiana tabacum, Journal of Asian Natural Products Research 11 (2009) 675-680.
59. J. Chen, W.L. Li, J.L. Wu, B.R. Ren, H.Q. Zhang, Hypoglycemic effects of a sesquiterpene
glycoside isolated from leaves of loquat (Eriobotrya japonica (Thunb.) Lindl.),
Phytomedicine 15 (2008) 98-102.
60. M.H. Lee, K.S. Yeon, N.H. Yong, Tissue factor inhibitory sesquiterpene glycoside from
Eriobotrya japonica, Archives of Pharmacal Research 27 (2004) 619-623.
61. Q. Ye, G. Qin, W. Zhao, Immunomodulatory sesquiterpene glycosides from Dendrobium
nobile, Phytochemistry 61 (2002) 885-890.
62. W. Zhao, Q. Ye, X. Tan, H. Jiang, X. Li, K. Chen, A.D. Kinghorn, Three New Sesquiterpene
Glycosides from Dendrobium nobile with Immunomodulatory Activity, Journal of Natural
Products 64 (2001) 1196-1200.
63. B. Zhou, J.F. Xiao, L. Tuli, H.W. Ressom, LC-MS-based metabolomics, Molecular
Biosystems 8 (2012) 470-481.

30

64. Y.L. Ma, Q.M. Li, H. VandenHeuvel, M. Claeys, Characterization of flavone and flavonol
aglycones by collision-induced dissociation tandem mass spectrometry, Rapid
Communications in Mass Spectrometry 11 (1997) 1357-1364.
65. E.A.P. Ekanayaka, C. Li, A.D. Jones, Sesquiterpenoid glycosides from glandular trichomes
of the wild tomato relative Solanum habrochaites, Phytochemistry.
66. D.O. Kennedy, E.L. Wightman, Herbal Extracts and Phytochemicals: Plant Secondary
Metabolites and the Enhancement of Human Brain Function, Advances in Nutrition 2 (2011)
32-50.
67. J.L. Ward, J.M. Baker, A.M. Llewellyn, N.D. Hawkins, M.H. Beale, Metabolomic analysis
of Arabidopsis reveals hemiterpenoid glycosides as products of a nitrate ion-regulated,
carbon flux overflow, Proc Natl Acad Sci U S A 108 (2011) 10762-10767.
68. G. Wang, L. Tian, N. Aziz, P. Broun, X. Dai, J. He, A. King, P.X. Zhao, R.A. Dixon,
Terpene Biosynthesis in Glandular Trichomes of Hop, Plant Physiology 148 (2008) 12541266.
69. S.M. Colby, J. Crock, B. Dowdle-Rizzo, P.G. Lemaux, R. Croteau, Germacrene C synthase
from Lycopersicon esculentum cv. VFNT Cherry tomato: cDNA isolation, characterization,
and bacterial expression of the multiple product sesquiterpene cyclase, Proceedings of the
National Academy of Sciences of the United States of America 95 (1998) 2216-2221.
70. O. Fiehn, J. Kopka, R.N. Trethewey, L. Willmitzer, Identification of uncommon plant
metabolites based on calculation of elemental compositions using gas chromatography and
quadrupole mass spectrometry, Analytical Chemistry 72 (2000) 3573-3580.
71. D. Killick, King Croesus' gold: Excavations at Sardis and the history of gold refining,
American Journal of Archaeology 106 (2002) 113-114.
72. V.D. Zheljazkov, T. Astatkie, B. O'Brocki, E. Jeliazkova, Essential Oil Composition and
Yield of Anise from Different Distillation Times, Hortscience 48 (2013) 1393-1396.
73. M. Giulietti, M.M. Seckler, S. Derenzo, M.I. Re, E. Cekinski, Industrial crystallization and
precipitation from solutions: State of the technique, Brazilian Journal of Chemical
Engineering 18 (2001) 423-440.
74. D.A. Dias, S. Urban, U. Roessner, A Historical Overview of Natural Products in Drug
Discovery, Metabolites 2 (2012) 303-336.

75. G. Pataki, J. Borko, Some Recent Advances in Thin Layer Chromatography I. New
Applications in Amino Acid Peptide und Nucleic Acid Chemistry, Chromatographia 1 (1968)
406-417.

31

76. L.S. Ettre, K.I. Sakodynskii, Tswett, M.S and the discovery of chromatography 1. Early work
(1899-1903), Chromatographia 35 (1993) 223-231.
77. L.S. Ettre, K.I. Sakodynskii, Tswett, M.S and the discovery of chromatography 2.
Completion of the developemnt of Chromatography (1903-1910), Chromatographia 35
(1993) 329-338.
78. E. Heftmann, Chromatography, Analytical Chemistry 36 (1964) 14R - 35R.
79. K.R. Williams, Colored Bands: History of Chromatography, Journal of Chemical Education
79 (2002) 922-923.
80. T.L. Chester, Recent Developments in High-Performance Liquid Chromatography Stationary
Phases, Analytical Chemistry 85 (2013) 579-589.
81. G. Rajput, H. Patel, J. Patel, K. Patel, M. Patel, Introduction to hyphenated techniques and
their applications in pharmacy.
82. E.H. Liu, L.W. Qi, J. Cao, P. Li, C.Y. Li, Y.B. Peng, Advances of modern chromatographic
and electrophoretic methods in separation and analysis of flavonoids, Molecules 13 (2008)
2521-2544.
83. F. Shi, Profiling of specialized metabolites in glandualr trichomes of the genus Solanum
using liquid chromatography and mass spectrometry., Chemistry, Michigan State University,
East Lansing, 2009.
84. C. Li, Mass spectrometric profiling and localization of metabolites in biological samples.,
Chemistry, Michigan State University, East Lansing, 2011.
85. O. Fiehn, J. Kopka, P. Dormann, T. Altmann, R.N. Trethewey, L. Willmitzer, Metabolite
profiling for plant functional genomics, Nat Biotechnol 18 (2000) 1157-1161.
86. R.C. De Vos, S. Moco, A. Lommen, J.J. Keurentjes, R.J. Bino, R.D. Hall, Untargeted largescale plant metabolomics using liquid chromatography coupled to mass spectrometry, Nat
Protoc 2 (2007) 778-791.
87. K. Saito, F. Matsuda, Metabolomics for Functional Genomics, Systems Biology, and
Biotechnology, in: S. Merchant, W.R. Briggs, D. Ort (Eds.), Annual Review of Plant
Biology, Vol 61, 2010, pp. 463-489.
88. R.D. Hall, Plant metabolomics in a nutshell: potential and future challenges, Annual Plant
Reviews 43 (2011) 1–24.

32

89. R. Tautenhahn, K. Cho, W. Uritboonthai, Z. Zhu, G.J. Patti, G. Siuzdak, An accelerated
workflow for untargeted metabolomics using the METLIN database, Nature Biotechnology
30 (2012) 826–828.
90. F. Matsuda, Y. Shinbo, A. Oikawa, M.Y. Hirai, O. Fiehn, S. Kanaya, K. Saito, Assessment of
Metabolome Annotation Quality: A Method for Evaluating the False Discovery Rate of
Elemental Composition Searches, Plos One 4 (2009).
91. A.D. Hegeman, Plant metabolomics-meeting the analytical challenges of comprehensive
metabolite analysis, Briefings in Functional Genomics 9 (2010) 139-148.
92. H.Y. Zhang, D.L. Zhang, K. Ray, M.S. Zhu, Mass defect filter technique and its applications
to drug metabolite identification by high-resolution mass spectrometry, Journal of Mass
Spectrometry 44 (2009) 999-1016.
93. T. Kind, O. Fiehn, Metabolomic database annotations via query of elemental compositions:
Mass accuracy is insufficient even at less than 1 ppm, Bmc Bioinformatics 7 (2006).
94. T. Kind, O. Fiehn, Seven Golden Rules for heuristic filtering of molecular formulas obtained
by accurate mass spectrometry, Bmc Bioinformatics 8 (2007) 105.
95. D. Zamir, Improving plant breeding with exotic genetic libraries, Nat Rev Genet 2 (2001)
983-989.

33

Chapter 2: Strategies for rapid identification of sesquiterpene glycosides from complex
matrices using relative mass defect filtering.
The research findings described in this chapter have, in part, been published in the journal
article: Sesquiterpenoid glycosides from glandular trichomes of the wild tomato relative
Solanum habrochaites. Phytochemistry 98: 223-231
2.1 Introduction
Terpenes represent the largest group of plant natural products, with over 55,000 known
compound structures [1]. Their diversity stems from an assortment of isoprenoid metabolites that
vary in length and configuration, and these undergo a wide variety of metabolic cyclizations and
secondary transformations (e.g. hydroxylation, reduction, glycosylation, and acylation) [2].
However, isoprene remains the fundamental structural unit of all terpenes as discussed in
Chapter 1. As a result, despite common secondary transformations, terpenoids almost always
consist of a hydrocarbon core, but recognition of metabolites derived from terpenoid biosynthetic
pathways has remained inefficient. Despite the vast diversity of terpenoids, a systematic
approach that can distinguish metabolites that have a terpene hydrocarbon core from the rest of
the molecules in a complex matrix can allow for the distinctive identification of terpenes among
others. This information can guide research into the roles of specific enzymes in terpene
biosynthesis and degradation. To exploit this characteristic, this Chapter presents a method for
classification of compounds based on the concept of relative mass defect (described below) that
is expected to accelerate identification of terpene glycosides from complicated matrices.

34

2.2 Applying accurate mass and relative mass defect filtering for exploring plant metabolomes for
the identification novel terpene glycosides

Mass spectrometry plays important roles as a structure elucidation tool in modern metabolomics
[3]. The growing use of medium- to high-resolution mass spectrometers has provided greater
mass measurement accuracy, often with low part-per-million (ppm) mass measurement errors.
For metabolites of low molecular mass, such measurements often provide sufficient information
to assign molecular formulas without ambiguity, but the number of possible formulas within a
given error tolerance grows rapidly as a function of mass. For metabolites of higher (> 500 Da)
molecular masses, multiple elemental formulas may be consistent with measured masses within
measurement errors of about 1 ppm. Moreover, assignments of molecular formulas often fail to
yield reliable assignments of metabolites to specific biosynthetic origins. This is because a given
elemental formula can correspond to a number of different chemical structures. For instance, a
“simple search” in the ChemSpider database (www.chemspider.com) for the elemental formula
of glucose C6H12O6 results in 256 chemical structures with identical monoisotopic mass!
Among the methods proposed for associating compounds with specific metabolite classes using
mass spectrometry, relative mass defect (RMD) is a property that can be calculated for both
molecular and fragment ions in mass spectra. RMD facilitates classification of compounds to
various biosynthetic classes, since RMD is correlated with fractional hydrogen content [4].
Relative mass defect is ratio of the mass defect to accurate mass of a given compound, expressed
in parts per million. Relative mass defect (RMD) is calculated in ppm, with RMD = (mass
defect/measured monoisotopic mass) × 106. Similar exploitation of absolute mass defect
filtering has been helpful for identifying peptides [5]. The main drawback of using absolute mass
defect filtering instead of relative mass defect filtering is that the former does not scale well with
molecular mass, whereas normalizing mass defect to ion mass yields values that better reflect
35

biosynthetic origins. For example, monoterpene (C10H16), diterpene (C20H32), and triterpene
(C30H48) hydrocarbons would all have different absolute mass defects, reflecting differences in
the number of hydrogen atoms in each molecule, but the relative mass defect, obtained by
normalizing to molecular mass, would be the same, reflecting their common biosynthetic origins.
For example, the monoisotopic mass of a monoterpene of formula C10H16 is 136.1252 Da, a
diterpene of formula C20H32 has a monoisotopic mass of 272.2504 Da, and triterpene C30H48 with
monoisotopic mass of 408.3756. All three compounds have RMD of 920 ppm. These
calculations demonstrate that regardless of the size of the terpenoid, hydrocarbons with constant
proportions of hydrogen to carbon yield the same RMD value, even though their absolute mass
defects are quite different.
A positive absolute mass defect usually reflects a large number of hydrogen atoms because the
atomic mass of hydrogen (1.00783 Da) is slightly greater than the rounded-off integer value (by
7.83 mDa), and carbon (12.00000 Da) does not contribute to mass defect because its
monoisotopic mass is exactly an integer value. Oxygen (15.99491 Da) has a small negative mass
defect (-5.09 mDa) and nitrogen (14.00307 Da) has a positive defect (+3.07 mDa). Since most
metabolites have much larger numbers of hydrogen atoms than oxygen or nitrogen, the RMD for
most metabolites is largely governed by the fraction of the mass that is made up of hydrogen
atoms. Compounds rich in hydrogen such as alkanes and fatty acids often display RMD greater
than those of terpenoids because the hydrogen content of terpenoids is lower. For example, RMD
values of stearic acid (C18H36O2; 12.76 %H; 955 ppm) and palmitic acid (C16H32O2; 12.58 %H;
937 ppm) are slightly higher than isoprene (C5H8 and its oligomers (C5H8)n , all of which have
11.84 %H and RMD = 919 ppm. However, it is common that terpenoids undergo metabolic
oxidation resulting in a decrease of RMD value. For example, the RMD of farnesol (C15H26O;
36

11.79 %H) equals 893 ppm. However, it is common that terpenoids undergo extensive metabolic
oxidation or conjugation by oxygen-rich groups (e.g. carbohydrates, phosphates, and oxygen-rich
esters), and the RMD values decrease as a result of increased fractional content of oxygen and
decreased fractional hydrogen content. This difference in RMD reflects the extent of oxidation of
the compound and enables a rapid assignment of ions to broad structural classes based on
hydrogen content. Because conjugation decreases the RMD value of an intact terpenoid, this one
value alone does not allow conclusive assignment of a compound as a terpenoid based on RMD
of a molecular ion. For example, a common tetraacylsucrose found in glandular trichomes of
tomato has an elemental formula of C29H48O15, which gives an RMD value of 470 ppm.

further

support annotation of metabolites as terpenoids, the RMDs of fragment ions in mass spectra that
are generated using collision-induced dissociation need to be considered. For most conjugated
terpenoids, the RMD of fragment ions remaining after the loss of conjugate groups is greater
than the RMD of the molecular ion because the conjugate groups have lower fractional hydrogen
content than the intact molecules.
For molecules that are rich in reduced carbon such as isoprene, which is the building block of
terpenes, the RMD is 920 ppm, reflecting the 11.8 wt % hydrogen content of C5H8. This value
does not change with oligomerization, as a sesquiterpene such as -santalene (C15H24) also has
11.8% hydrogen. Oxidative transformations usually involve either increases in molecular mass
without increasing hydrogen content, or decreases in hydrogen content. In both cases, the
fraction of the molecular mass accounted for by hydrogen atoms decrease, and values of RMD
also decrease. For example, oxidative conversion of -santalene (C15H24) to α-santalenol
(C15H24O) by addition of a single oxygen atom shifts RMD downward to 830 ppm, and insertion
of a second oxygen atom (C15H24O2) yields RMD of 752 ppm. This information has particular
37

relevance when searches for terpenoids employ LC-MS for analysis because electrospray
ionization usually requires molecules to have acidic or basic groups. For terpenoids, this means
that one or more oxygen atoms must be in the molecule for it to be ionized. In many organisms,
terpenoids may be conjugated to polar groups (e.g. glycosides or phosphates) that decrease a
terpenoid’s RMD value further because they decrease the fractional hydrogen content. For
example, each glycosylation by a hexose adds C6H10O5 to a terpenoid, so the glycoside of a
sesquiterpene alcohol (C21H34O6) would have RMD = 616 ppm. Additional oxidation or
conjugation by malonate (addition of C3H2O3) would decrease the RMD value of terpenoid
metabolites further. The diterpene glycosides and their malonate esters [6] provide examples.
The RMD of nicotianoside VII (a diterpenoid glycoside of formula C50H78O24 with two malonate
esters) is 426 ppm and nicotianoside VI (one malonyl group less with elemental formulae of
C47H76O24) has RMD of 461 ppm. Since conjugated terpenoids usually consist of a terpenoid
core that is rich in reduced carbon and conjugate groups of low fractional hydrogen content, the
RMD of terpenoids conjugated to sugar molecules vary from ~400 ppm to ~600 ppm . In
contrast, polyphenolic metabolites have low hydrogen content, and their RMD is less than 300
ppm (230 ppm for the phenolic acid salicylic acid, 167 ppm for the flavonoid kaempferol).
Negative ion mode mass spectrometry aids identification of terpenoid conjugate functional
groups based on neutral losses observed, [7] but negative ion tandem mass spectra often fail to
exhibit product (fragment) ions characteristic of the unmodified terpenoid core. For instance,
negative ion mode MS/MS spectra of common hexose derivatives (e.g. mono-, di- and trisaccharides) display fragment ions (Table 2.1) characteristic of carbohydrate groups, as well as
fragments derived from neutral mass losses that indicate losses of those carbohydrates from the
pseudomolecular ions. Positive ion mode MS/MS spectra also allows for the annotation of mono38

Table 2.1: Characteristic fragment ions observed in negative ion mode MS/MS experiments for
various sugar oligosaccharides and mono saccharides. The masses shown correspond to [M-H]formed by each sugar group. The MS/MS spectra of candidate terpenoid compounds were
examined for the presence of these fragment ions for identification of the presence of these
oligosaccharides in the terpenoids. 1

Negative ion mode fragment ion
m/z (Theoretical exact mass)
503.1618

Common sugar moiety
Triglycoside(Hexose-Hexose-hexose; C18H31O16-)
Triglycoside - H2O (C18H29O15-)

485.1512
589.1622

Triglycoside malonate ester (C21H33O19-)
Triglycoside malonate ester - H2O

571.1516
341.1089

Diglycoside (hexose-hexose; C12H21O11-)
Diglycoside - H2O (C12H19O10-)

323.0984
179.0561
221.0667

Monoglycoside (Hexose; C6H11O6-)
Monoglycoside acetate ester (C8H13O7-)

161.0455

Monoglycoside - H2O (C6H9O5-)

101.0232

Fragment ion from hexoses (C4H5O3-)

113.0228

Fragment ion from hexoses

125.0244

Fragment ion from hexoses

, di- and tri-saccharides and their respective malonylated and acetylated forms based on fragment
ions characteristic of the sugar residues and neutral losses characteristic of the sugar and
malonate groups [8,9]. These results are similar to MS/MS characterization of flavonoid

39

glycosides from Arabidopsis thaliana using LC-MS/MS [9]. Based on the presence of
characteristic fragment ions in these MS/MS spectra, the presence of disaccharides rhamnose –
glucose, xylose-glucose and trisaccharides (rhamnose – glucose – glucose) was established [9].
Carbohydrate-derived fragment ions have RMD values around 300 ppm, and product ions in
MS/MS spectra with RMD values that fall in this range are suggestive of glycosides. In the case
of a terpenoid glycoside, one would expect RMD of fragments derived from carbohydrates to lie
around 300 ppm, the molecular RMD to fall in the range of 400-600 ppm, and fragments derived
from the terpenoid core (without carbohydrates) to approach 800 ppm. Such patterns of RMD
values in MS/MS spectra aid annotation of unknown metabolites as candidate terpenoid
glycosides.
Owing to the scarcity of information regarding non-volatile terpenoids in the genus Solanum, this
study has explored LC-MS metabolite profiling data generated for extracts of the wild tomato S.
habrochaites LA1777. To accelerate recognition of terpenoid metabolites, filtering of molecular
and fragment masses using RMD was performed. The study is explained in detail below.

40

2.3 Materials and Methods
2.3.1 LC-MS and MS/MS experiments
All LC-MS and LC-MS/MS experiments were performed using a Waters Xevo G2-S QToF mass
spectrometer equipped with the standard electrospray ionization ion source. This instrument was
coupled to a Waters Acquity ultra-high pressure liquid chromatography system. Separations were
performed using an Ascentis Express C-18 UHPLC column (2.1 mm × 100 mm, 2.7 µm;
Supelco Co. USA). For negative ion mode experiments, solvents were 0.15% aqueous formic
acid - pH 2.85 (A) and methanol (B). Solvent gradient (A:B) used for profiling of conjugated
terpenoids was as follows: 0-1.00 min (99:1), 1.01-4.00 (linear ramp to 55:45), hold at (55:45)
4.01-9.00 min, step to (50:50) and hold at (50:50) over 9.01-14.00 min, step to (1:99) and hold
over 14.01-17.00 min, step to (99:1) and hold over 17.01-20.00 min. Flow rate was 0.30 ml/min
and column temperature was held at 35 °C. Mass spectra were acquired using negative-ion mode
electrospray ionization and dynamic range extension over m/z 50-1500, with mass resolution
(M/M, full width-half maximum) approximately 20000. Five parallel collision energy functions
were used, with 0.1 s per function. Collision cell potentials used for negative ion mode
fragmentation for each function were 5, 15, 25, 35 and 45 V respectively. Other parameters
include capillary voltage of 3.0 kV, desolvation temperature of 350 °C, source temperature of
100 °C, Cone gas (N2) at 0 L/hr and desolvation gas (N2) at 800 L/hr. Positive ion mode MS/MS
experiments were performed using the same instrument specifications and LC method, except,
for solvent A, 10 mM ammonium formate (pH 2.65) was used instead of 0.15% formic acid
solution. In all analyses, leucine enkephalin (0.1 ng/µL) was used as a lock mass reference
infused at a flow rate of 3.0 µL/min and lock mass spectra were acquired for 100 ms at 10 sec
intervals. Capillary voltage for lock spray was 2.5 kV.
41

All MS/MS experiments were performed on the same instrument employing the same LC
methods. For negative ion mode MS/MS experiments a collision potential of -50 V was used and
for positive ion mode experiments +30 V was used.
2.3.2 Plant material
Sample preparation for metabolite profiling: Solanum habrochaites LA1777 plants were grown
in Michigan State University plant growth chambers (28 °C, 16:8 h day/night cycle, 150 µmol m2 -1

s , 96% humidity) for six weeks using seeds obtained from the C. M. Rick Tomato Genetics

Resource Center at the University of California-Davis. Ten leaflets were harvested from each
plant (six weeks post germination) and were extracted by dipping in 5 mL of methanol: water
(80:20 v/v) for about 30 sec. Three biological replicates were used for the profiling. Extracts
were concentrated by drying under a stream of N2 gas at room temperature, and the residues
were redissolved in 0.5 ml methanol: water (80:20 v/v). The extract was centrifuged (10000g for
10 min at 25 °C) to remove debris, the supernatant was transferred to an autosampler vial, and
2.5 µL was introduced to the LC column. Automated peak detection, integration, and retention
time alignment were performed using Waters MarkerLynx XS software, and lists of m/z values,
retention times, and extracted ion chromatogram peak areas were exported as text files and
processed further using Microsoft Excel software.
2.2.3 Data processing
Automated chromatographic peak detection was performed using Waters MarkerLynx XS
software. The lowest collision energy function (function 1) was used for peak detection,
integration, retention time alignment, and deisotoping. The parameters used with MarkerLynx
processing were as follows. Marker intensity threshold: 800 counts, mass window 0.05 Da,

42

retention time window: 0.25 min, m/z range: 100 – 1500, retention time range 0.5 – 20.0 min. No
peak smoothing was applied.
2.4 Results and discussion
2.4.1 Recognition of sesquiterpene glycosides from ion relative mass defects
Analysis of leaf extracts of S. habrochaites LA1777 using LC-multiplexed CID MS in negative
ion mode yielded evidence of complex mixtures of chemicals (Figure 2.1).

Figure 2.1 Complexity of a plant extract is evident from the number of peaks in a UHPLC-MS
base peak chromatogram generated from a leaf dip extract of S. habrochaites LA1777. Analysis
was performed using a 110 min chromatographic gradient and detected in negative ion mode. 4
Automated peak detection, deisotoping, integration, and retention time alignment using Waters
MarkerLynx XS software yielded a total of 3280 m/z-retention time pairs, estimated to arise from
about 1000 distinct metabolites (Fig. 2.2).

43

List of samples
processed

List of
markers

Figure 2.2: MarkerLynx XS data processing results in a list of markers detected from the S.
habrochaites LA1777 trichome leaf extracts. Markers detected are reported with their m/z ratio,
retention time and their abundance in each sample being considered. 5
Examination of the RMD values for the entire data set indicated that 3199 (98 %) of the ions
detected in automated peak picking had positive absolute mass defects, with the remainder
having negative absolute mass defects typical of inorganic salt cluster ions and instrument
contaminants (e.g. trifluoroacetate, NaHPO4-). Of the ions with positive mass defects, 1805 (55%
of total) displayed RMD in the range of 400 to 650 ppm and 1177 (36% of total) possessed RMD
from 200 to 400 ppm, the latter range being more typical of aromatic, probably polyphenolic
metabolites. Since the objective of this exercise was identification of sesquiterpene glycosides
from this data set, three theoretical boundary condition approximations were made: 1) based on
the theoretical m/z of 383.2439 for [M-H]- of farnesol monoglycoside (C21H35O6-), the maximum
RMD a sesquiterpene glycoside (with minimum of one hexose moiety) can display is 636 ppm;

44

2) based on the theoretical m/z of 869.4024 for [M-H]- farnesol tetraglycoside, minimum RMD a
sesquiterpene glycoside (with maximum of four hexose moieties) can display is 463 ppm; 3)
based on the theoretical m/z of 383.2439 for [M-H]- of farnesol monoglycoside, minimum
nominal m/z of a sesquiterpenoid monoglycoside should be 383. It is notable that some terpenoid
compounds display malonylation as evidenced by the malonylated diterpene glycosides of
Nicotiana attenuata and malonylated sesquiterpenes of Panax ginseng. [6,10,11,12]. Similarly,
acetylated terpenes and acetylated terpene glycosides have been isolated from Iphiona scabra
and Combretum imberbe [13,14]. Dozens of such malonylated and acetylated terpenoids have
been reported [15,16]. As a result, rule #2 described above is modified to account for such
possibilities, setting the lower limit of RMD to 440 ppm to account for malonylated
sesquiterpene compounds (RMD of a farnesol triglycoside malonate ester would be 441 ppm).
Similarly, to account for minor mass detection errors the maximum RMD possible for
sesquiterpene glycosides with minimum of one hexose moiety was set to 640 ppm.
Applying the first two conditions, the number of compounds with RMD 440 to 640 ppm detected
is 1280 (38% of total markers). Applying the third condition to this set of 1280 markers results a
list of 1074 markers (33% of total) that are putative terpene glycosides. Next, this list of markers
was arranged in the descending order of peak area and the 200 most abundant markers were
selected as candidate terpenoid glycosides. The first 38 compounds in this list are listed in Table
2.3.
Setting the rules based on the requirement: The boundary conditions set in rules discussed here
were set based on the objective of identifying sesquiterpene glycosides. These need to be reestablished if the objective was diterpene or triterpene glycoside identification. For instance for
diterpene glycoside identification the rules will be established based on geranylgeranyl
45

monoglycoside since geranylgeranyl acts as the precursor of diterpenes. This will result in the
following rules. 1). Based on theoretical mass of geranylgeranyl monoglycoside the minimum
m/z of [M-H]- of a diterpene glycoside would be 451.3065 and the maximum RMD a diterpene
glycoside can display is 679 ppm. 2.) Based on theoretical mass of geranylgeranyl tri glycoside
the minimum m/z of [M-H]- of a diterpene triglycoside would be 775.4122 and the minimum
RMD a diterpene glycoside can display is 531 ppm. 3) based on the theoretical m/z of 451.3065
for [M-H]- of geranylgeranyl monoglycoside, minimum nominal m/z of a diterpenoid
monoglycoside should be 451. Therefore, compounds in RMD range 500 – 700 ppm will need to
be considered if the objective was to discover diterpene glycosides.
Distinguishing terpenoid glycosides from other compounds: Applying the RMD criteria
described allows for the elimination of most, if not all, compounds that are not terpene
glycosides. However, the final list of compounds may possess a number of non-terpenoid
compounds, and to distinguish these from terpenoids, calculation of RMD of the fragment ions
generated using either MS/MS or multiplexed CID should be performed. Inclusion of nonterpenoid compounds in this list occurs because a given elemental formula (and m/z value) can
correspond to a number of different chemical structures as discussed in introduction section.
Therefore, in order to identify terpenoids, additional information Fragment ion RMD values
distinguish terpenoid glycosides from other compounds since losses of all carbohydrate moieties
will yield a fragment ion corresponding to a terpenoid core, with RMD values > 800 ppm.
Furthermore, even if all sugars are not removed, the RMD of the fragment ions formed by
terpenoid glycosides will be greater than that of the pseudomolecular ion because RMDs are
about 325-350 ppm for hexoses and hexose fragments, values much less than for terpenoid cores.
Therefore, terpenoid glycosides are characterized by fragment ions that display an increasing
46

RMD as their mass decreases due to the loss of sugar groups. This phenomenon is illustrated in
Figure 2.3, which shows the MS/MS spectra of three of the five metabolites with
pseudomolecular ion RMD falling in the 440 to 636 ppm range and with greatest peak areas.
Among the fragment ions of these compounds, only the fragments m/z 199 (RMD ~ 840 ppm;
Fig. 2.3.A and 2.3.B) display RMD close to that of isoprene (919 ppm), but, m/z 199 has a mass
too low to be a sesquiterpenoid core (C15H24 would be 204 Da). In addition, there is no
systematic increase of RMD with as fragment masses decrease among the fragment ions of in
any of these compounds, suggesting the groups being lost have hydrogen contents similar to, or
greater than, the intact molecule. These findings suggest the molecules are not terpenoid
glycosides, and in fact, they are all acylsucroses. In contrast, the MS/MS spectra observed for
glycosides of the sequiterpene alcohol campherenane diol (discussed below) shown in Figure 2.3
shows a clear systematic increase in RMD as fragment masses decrease, consistent with a
terpenoid glycoside.

47

Figure 2.3. (A) Negative ion mode multiplexed CID mass spectrum of acylsugar S4:22, RMD =
492 ppm. (B) Negative ion mode multiplexed CID mass spectrum of acylsugar S4:23, RMD =
402 ppm. (C) Negative ion mode multiplexed CID mass spectrum of acylsugar S4:17, RMD =
440 ppm. Values for RMD of the major fragment ions are presented. All detected isomers
displayed fragments of the same m/z values in negative ion mode spectra. All displayed negative
ion mode multiplexed CID mass spectra were obtained using a collision potential of -60 V. 6
48

2.4.2 Discovery of sesquiterpene diol glycosides from S. habrochaites LA1777
In the list of the 200 S. habrochaites LA1777 metabolites with greatest peak areas within the
RMD range 440-640 ppm (as discussed in Section 2.4.1), the metabolite m/z of 609 was ranked
17th in peak area. Negative-ion mode multiplexed CID mass spectra of this compound revealed a
systematic increase of RMD with decreasing fragment masses, as is characteristic of terpenoid
glycosides. In order to ensure that these fragment ions were derived from the proposed
pseudomolecular ion, MS/MS spectra were generated.
The MS/MS spectrum of this compound (Figure 2.4A) revealed m/z 563.3045 as a fragment ion.
This was 46 Da less than the parent ion mass. Molecular mass of formic acid is also 46 Da. Since
formic acid was in the mobile phase used for this experiment it was concluded m/z 609
represented a formate adduct of m/z 563. Such ionization behavior is typical of compounds that
lack acidic functional groups, but can form pseudomolecular ions by anion adduction. The next
most abundant fragment was m/z 401, which was 162 Da less than m/z 563. 162 Da neutral loss
is a common hexose sugar neutral loss and therefore, 401 represents a form that has one less
sugar moiety than 563. Note that the RMD of each of these fragments (539 ppm and 631 ppm for
m/z 563 and 401 respectively) increases as the m/z of fragment decreased (Fig. 2.4A). Since this
is a characteristic feature of terpene glycosides, the observations are consistent with this
metabolite being a member of this class. However, no prominent fragment ions with high RMD
typical of terpenoid cores were observed with m/z < 401. An elemental formula calculation for
m/z 401.2532 yielded a formula of C21H37O7-, which has 6 more than the 15 carbon atoms that
are usually found in sesquiterpenes. Therefore, potentially m/z 401 represents the terpenoid core
with other groups present in it. The low mass fragment ion of m/z 323 is a characteristic fragment
corresponding to [dihexose-H-H2O]- (C12H19O10-) suggesting the presence of a diglucoside in the
49

compound. Based on the presence of this m/z 323 it can be proposed that this compound has a
minimum of two hexose groups attached to it, and that they are linked to one another. It is
important to note that for further characterization of this compound required that it be purified
and subjected to structure determination using 1D and 2D NMR (Chapter 3), since further
fragmentation of the core was not observed in negative ion mode.

50

Figure 2.4 Negative-ion mode product ion MS/MS spectra of metabolites extracted from S.
habrochaites LA1777. (A) MS/MS spectrum of m/z 609 from campherenane diol diglucoside;
(B) MS/MS spectrum of m/z 811 from campherenane diol triglycoside malonate ester (m/z
811.3587, RMD = 442 ppm). Values for RMD of the major fragment ions are presented. All
chromatographically-resolved isomers (10 isomers of m/z 651 and 12 isomers for m/z 811)
displayed fragments of the same m/z values in negative ion mode spectra. All negative ion mode
MS/MS data were obtained using collision potential of -50 V. 7
51

Discovery of sesquiterpene diol triglycoside malonate ester from S. habrochaites LA1777
provides an example that involves the presence of additional groups other than sugar moieties.
The exact mass of this molecule was m/z 811.3587 in negative ion mode (RMD 442 ppm). In the
MS/MS spectrum (Figure 2.3B) of the compound it fragments by loss of CO2 to give m/z
767.3707 (RMD 483 ppm) followed by the loss of C2H2O to give m/z 725.3587 (RMD 495
ppm). The loss of CO2 (44 Da) followed by the loss of another 42 Da (C2H2O) is characteristic of
malonate esters. Next, the loss of the first anhydroglucose (neutral loss of 162 Da) leads to
fragment ion m/z 563.3057 (RMD 543 ppm), and loss of a second anhydroglucose group resulted
in m/z 401.2516 (RMD 627 ppm). Loss of a third anhydroglucose results in fragment ion of m/z
239.2017 (RMD 841 ppm). Fragment ion 239.2017 represents the sesquiterpenoid core
(C15H27O2-) of the molecule as it does not undergo any further fragmentation in negative ion
mode. This compound also displays the characteristic feature of conjugated terpenoids –
presence of fragments with increasing RMD value with decreasing m/z. Based on these observed
characteristics this compound can be annotated as a sesquiterpenoid triglycoside malonate ester.
The described method can also be applied to other types of terpenoids. For instance tomatine, a
triterpenoid glycoside also displays similar behavior. RMD of the [M+formate]- ion of tomatine
in negative ion mode is 520 ppm. The major fragments display a gradual increase of RMD from
520 ppm to 674 ppm, and correspond to losses of carbohydrate moieties (Fig. 2.5).

52

Figure 2.5: Negative ion mode multiplexed CID mass spectrum of tomatine from S. habrochaites
LA1777 obtained at 60 V collision potential. Note that the fragment ions all have RMD values
greater than or equal to the [M-H]- ion at m/z 1032.5. 8
Plotting RMD as a function of m/z for some example S. habrochaites metabolites provides a
clear image of this difference of RMD variation among terpene glycoside fragment ions from the
fragment ions of other compounds (Fig. 2.6).

53

900
800

RMD (ppm)

700
600
500
400
300
200
0
Tomatine

200

400

600
m/z

800

Sesquiterpene TriGlycoside-811

1000

1200
AS2-765

Figure 2.6: Variation of RMD of fragment ions as a function of ion m/z. Fragment ions were
generated in negative ion mode MS/MS for some representative compounds (acyl sugars, ,
triterpenoid glycoside and a triglycoside malonate ester) found in S. habrochaites LA1777 leaf
dip extracts. 9
The proposed process of applying RMD filtering for the discovery of novel glycosylated
terpenoids is described graphically in Figure 2.7. Applying this procedure to the most abundant
200 compounds in the list allowed for the identification of a total of 224 different terpene
glycoside (including the isomers) from S. habrochaites plant extract (Table 2.2). Three different
sesquiterpenes cores were established including the campherenane diol core. The structures of
some of these compounds were established using NMR after purifying them. Structure
elucidation of these compounds (including the sesquiterpenes I, II and III terpenoid cores listed
in Table 2.2) is described in Chapter 3 of this dissertation. The MS/MS data generated for each
of these compounds and the RMD of each fragment ion are presented in Figs. 2.8-2.28.

54

Figure 2.7: Relative mass defect filtering process used for discovering conjugated
terpenoids from raw LC-MS data 10

55

Table 2.2: Compounds identified as terpene glycosides from S. habrochaites LA1777 based on RMD filtering of molecular and
fragment ions. 2

Co
mp
oun
d#

Proposed
elemental
Experimental RMD
∆m
formula of
m/z
(ppm)
(ppm)
neutral
molecule

∆m
(ppm)

RMD of
terpenoid
core
fragment
(ppm)

# of
Iso
mer
s
obs
erve
d

C14H21O2-

-9

691

8

221.1529

C14H21O2-

-8

692

1

221.1565

C14H21O2-

8

708

2

Compound type

Measured
m/z of
terpenoid
core
fragment ion

Proposed
elemental
formula of
terpenoid
core
fragment ion

Sesquiterpene II
Diglycoside

221.1527

1

545.2932

538

C27H46O11

-1

2

587.3049

519

C29H48O12

-3

3

631.2929

464

C37H44O13

3

4

707.3112

440

C32H52O17

-3

Sesquiterpene II
Triglycoside

221.1539

C14H21O2-

-4

696

4

5

793.3474

438

C36H58O19

1

Sesquiterpene II diol
Triglycoside
malonate ester

221.1527

C14H21O2-

-9

691

11

6

401.252

628

C21H38O7

-6

Campherenane diol
Monoglycoside

239.1997

C15H27O2-

-8

836

6

Sesquiterpene II
Diglycoside acetate
ester
Sesquiterpene II
Diglycoside malonate
ester

56

Table 2.2 (cont’d)

C15H27O2-

-9

834

14

239.1999

C15H27O2-

-7

836

5

239.1994

C15H27O2-

-9

834

9

239.2048

C15H27O2-

13

857

7

239.1993

C15H27O2-

-10

834

9

239.1993

C15H27O2-

-10

834

14

Campherenane diol
triglycoside

239.2007

C15H27O2-

-4

840

6

-2

Campherenane diol
triglycoside malonate
ester

239.2006

C15H27O2-

-4

839

13

-9

Sesquiterpene III
monoglycoside

249.1481

C15H21O3-

-4

595

13

-9

Sesquiterpene III
monoglycoside
acetate ester

249.1483

C15H21O3-

-3

596

14

443.2632

594

C23H40O8

-4

8

445.2421

544

C22H38O9

-4

9

487.2536

520

C24H40O10

-3

10

563.3055

542

C27H48O12

-3

11

605.3152

521

C29H50O13

-4

12

649.3042

469

C30H50O15

-5

13

725.3607

497

C33H58O17

2

14

811.3587

442

C36H60O20

15

411.1986

483

C21H32O8

16

453.2082

459

C23H34O9

Campherenane diol
Monoglycoside
acetate ester
Campherenane diol
Monoglycoside
derivative
Campherenane diol
Monoglycoside
malonate ester

239.1994

7

Campherenane diol
diglycoside
Campherenane diol
diglycoside acetate
ester
Campherenane diol
diglycoside malonate
ester

57

Table 2.2 (cont’d)
17

497.2011

404

C24H34O11

-3

Sesquiterpene III
monoglycoside
malonate ester

18

413.2143

519

C21H34O8

-9

Sesquiterpene I
monoglycoside

19

455.227

499

C22H36O9

-4

20

499.2162

433

C24H36O12

-5

21

617.2795

453

C29H45O14

-2

22

661.2693

407

C30H46O16

-3

23

823.3245

394

C36H56O21

1

24

985.3742

380

C30H50O26

-3

Sesquiterpene I
monoglycoside
acetate ester
Sesquiterpene I
monoglycoside
malonate ester
Sesquiterpene I
diglycoside acetate
ester
Sesquiterpene I
diglycoside malonate
ester
Sesquiterpene I
triglycoside malonate
ester
Sesquiterpene I
tetraglycoside
malonate ester

58

249.1475

C15H21O3-

-6

592

11

251.1634

C15H23O3-

-8

651

17

251.1638

C15H23O3-

-6

653

21

251.1643

C15H23O3-

-4

655

11

251.1636

C15H23O3-

-7

652

4

251.1639

C15H23O3-

-6

653

8

251.164

C15H23O3-

-5

653

5

251.1641

C15H23O3-

-5

654

11

Table 2.3: Compounds with greatest peak areas among the list of S. habrochaites LA1777
metabolites in the RMD range 440-636 ppm 3

Ret.
Time
15.3
15.3
17.8
17.5
15.2
15.1
15.2
15.0
17.3
14.8
15.3
11.0
14.5
15.3
15.1
15.4
6.8
15.1
15.0
12.4
6.5
13.7
15.2
15.3
16.1
14.2
15.0
7.3
13.2
15.2
11.3
10.7
8.6
7.4
15.0

Average
m/z
Peak
area
765.3836
219.44
751.3706
172.28
707.4408
170.61
707.4412
167.58
723.3353
140.38
681.3032
122.81
737.3490
120.67
871.4180
109.13
707.4419
98.18
873.3857
86.43
925.5072
64.32
975.4882
63.70
857.3901
63.41
911.5283
60.36
709.3192
59.31
967.5318
52.59
609.3073
47.11
901.4240
43.67
887.4166
40.09
609.2946
34.21
609.3055
31.96
413.2069
30.53
695.3402
29.89
723.3752
29.74
663.3840
29.15
455.2153
28.87
873.4442
28.52
1078.5147
27.86
455.2149
27.17
1011.5815
26.43
651.3034
25.33
455.2166
24.14
651.3154
23.74
1078.5139
23.15
885.4205
22.62

RMD
(ppm)
501
493
623
624
464
445
473
480
625
442
548
500
455
580
450
550
504
470
469
484
501
501
489
519
579
473
509
477
472
575
466
476
484
476
475
59

Figure 2.8. (A) Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z
649) for campherenan-2,12-diol malonate ester. (B) magnified region of m/z 228-255, showing
the sesquiterpenoid core fragment at m/z 239 which was too small to observe in A, where peaks
are normalized to the base peak. 11

60

Figure 2.9. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z 487)
for campherenan-2,12-diol monoglucoside malonate ester. The lower spectrum displays the
magnified region of m/z 224-256, showing the sesquiterpenoid core fragment at m/z 239. 12

61

Figure 2.10. Negative ion mode product ion MS/MS spectrum of products from [M+formate](m/z 609) for campherenan-2,12-diol diglucoside. The lower spectrum displays the magnified
region of m/z 228-258, showing the sesquiterpenoid core fragment at m/z 239.13

62

Figure 2.11. Negative ion mode product ion MS/MS spectrum of products from [M+formate](m/z 651) for campherenan-2,12-diol diglucoside acetate ester. The lower spectrum displays the
magnified region of m/z 224-264, showing the sesquiterpenoid core fragment at m/z 239. 14

63

Figure 2.12. Negative ion mode product ion MS/MS spectrum of products from [M+formate](m/z 771) for a campherenane-2,12-diol triglycoside (Compound 13 in Table 2.2). 15

Figure 2.13. Negative ion mode product ion MS/MS spectrum of products from [M+formate](m/z 591) for a sesquiterpene II dihexoside (Compound 1 in Table 2.2). 16

64

Figure 2.14. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z
631) for a sesquiterpene II dihexoside malonate ester (Compound 3 in Table 2.2).17

Figure 2.15. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z
793) for a sesquiterpene II trihexoside malonate ester (Compound 5 in Table 2.2).18
65

Figure 2.16. Negative ion mode product ion MS/MS spectrum of products from [M+HCOO](m/z 633) for a sesquiterpene II dihexoside acetate ester (Compound 2 in Table 2.2). 19

Figure 2.17. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z
661) for a Sesquiterpene I diglycoside malonate ester (Compound 22 in Table 2.2). 20

66

Figure 2.18. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z
499) sesquiterpene I monoglycoside malonate ester (Compound 20 in Table 2.2). 21

Figure 2.19. Negative ion mode product ion MS/MS spectrum of [M-H]- (m/z 455) for
Sesquiterpene I monoglycoside acetate ester (Compound 19 in Table 2.2). 22

67

Figure 2.20. Negative ion mode product ion MS/MS spectrum of [M-H]- (m/z 413) of
Sesquiterpene I monoglycoside (Compound 18 in Table 2.2). 23

Figure 2.21. Negative ion mode product ion MS/MS spectrum of [M+HCOO]- (m/z 499) for
sesquiterpene I monoglycoside malonate ester (Compound 20 in Table 2.2). 24

68

Figure 2.22. Negative ion mode product ion MS/MS spectrum of [M-H]- (m/z 497) of
Sesquiterpene III monoglycoside malonate ester (Compound 17 in Table 2,2). 25

Figure 2.23. Negative ion mode product ion MS/MS spectrum of products of [M-H]- (m/z 411)
of Sesquiterpene III monoglycoside (compound 15 in Table 2.2).26

69

Figure 2.24. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z
617) for Sesquiterpene I diglycoside acetate ester (compound 21). 27

Figure 2.25. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z
823) for Sesquiterpene I triglycoside malonate ester (compound 23 in Table 2.2). 28

70

Figure 2.26. Negative ion mode product ion MS/MS spectrum of products from [M-H]- (m/z
985) for Sesquiterpene I tetraglycoside malonate ester (compound 24 in Table 2.2) 29

Figure 2.27. Negative ion mode product ion MS/MS spectrum of products from [M+HCOO](m/z 489) for Campherenane diol Monoglycoside acetate ester (compound 7 in Table 2.2). 30

71

Figure 2.28. Negative ion mode product ion MS/MS spectrum of products from [M+HCOO](m/z 753) Sesquiterpene II Triglycoside (compound 4 in Table 2.2). 31

2.5 Conclusions
Analysis of the RMD variation among precursor ions and product/fragment ions generated
information that accelerates the terpene glycoside discovery process by eliminating signals from
the numerous compounds that do not give RMD values consistent with sesquiterpenoid
glycosides. While the method may not exclusively represent terpenoids, it allows for the
elimination of many non-terpenoid compounds from the complicated list of large number of
metabolites detected in a given plant extract. The proposed method has allowed for the
annotation of over 200 novel sesquiterpene glycosides from a plant system that has been studied
for many decades. This is an indication that the method explained here will enable the rapid
identification of terpene glycosides are largely understudied due to the lack of a systematic
method for recognition of their presence.

72

Multiplexed CID MS/MS allows for the rapid examination of MS/MS spectra of compounds
eliminating the requirement for separate MS/MS experiments for all markers detected using
automates peak picking.
The research discussed here used the negative ion mode MS/MS for the annotation of terpene
glycosides from complex matrices, and provides the first evidence for an extensive and diverse
group of sesquiterpenoid glycosides in S. habrochaites LA1777. However, the proposed method
is perceived as equally applicable to positive ion mode data sets as well. This is because the
relative mass defect is an intrinsic property of the compound being considered and is
independent of ionization mode being used for the detection.

73

REFERENCES

74

REFERENCES
1. T.J. Maimone, P.S. Baran, Modern synthetic efforts toward biologically active terpenes, Nat
Chem Biol 3 (2007) 396-407.
2. D. Tholl, Terpene synthases and the regulation, diversity and biological roles of terpene
metabolism, Current Opinion in Plant Biology 9 (2006) 297-304.
3. Z. Lei, D.V. Huhman, L.W. Sumner, Mass spectrometry strategies in metabolomics, J Biol
Chem 286 (2011) 25435-25442.
4. M.C. Stagliano, J.G. DeKeyser, C.J. Omiecinski, A.D. Jones, Bioassay-directed fractionation
for discovery of bioactive neutral lipids guided by relative mass defect filtering and
multiplexed collision-induced dissociation, Rapid Commun Mass Spectrom 24 (2010) 35783584.
5. M.L. Toumi, H. Desaire, Improving mass defect filters for human proteins, J Proteome Res 9
(2010) 5492-5495.
6. S. Heiling, M.C. Schuman, M. Schoettner, P. Mukerjee, B. Berger, B. Schneider, A.R. Jassbi,
I.T. Baldwin, Jasmonate and ppHsystemin regulate key Malonylation steps in the biosynthesis
of 17-Hydroxygeranyllinalool Diterpene Glycosides, an abundant and effective direct defense
against herbivores in Nicotiana attenuata, Plant Cell 22 (2010) 273-292.
7. K. Levsen, H.M. Schiebel, J.K. Terlouw, K.J. Jobst, M. Elend, A. Preiss, H. Thiele, A.
Ingendoh, Even-electron ions: a systematic study of the neutral species lost in the dissociation
of quasi-molecular ions, J Mass Spectrom 42 (2007) 1024-1044.
8. V.A. Marinos, M.E. Tate, P.J. Williams, Protocol for FAB MS/MS Characterization of
Terpene Disaccharides of Wine, Journal of Agricultural and Food Chemistry 42 (1994) 24862492.
9. P. Kachlicki, J. Einhorn, D. Muth, L. Kerhoas, M. Stobiecki, Evaluation of glycosylation and
malonylation patterns in flavonoid glycosides during LC/MS/MS metabolite profiling, J Mass
Spectrom 43 (2008) 572-586.
10.
C.-C. Ruan, Z. Liu, X. Li, X. Liu, L.-J. Wang, H.-Y. Pan, Y.-N. Zheng, G.-Z. Sun, Y.-S.
Zhang, L.-X. Zhang, Isolation and Characterization of a New Ginsenoside from the Fresh
Root of Panax Ginseng, Molecules 15 (2010) 2319-2325.
11.
S. Guangzhi, L. Zhi, L. Xianggao, Z. Yinan, W. Jiyan, Isolation and Identification of
Two Malonyl-ginsenosides from the Fresh Root of Panax Ginseng, Chinese Journal of
Analytical Chemistry 33 (2005) 1783-1786.

75

12.
L.R. Sun, J. Yan, L. Zhou, Z.R. Li, M.H. Qiu, Two New Triterpene Glycosides with
Monomethyl Malonate Groups from the Rhizome of Cimifuga foetida L, Molecules 16 (2011)
5701-5708
.
13.
C.B. Rogers, Pentacyclic triterpenoid rhamnosides from Combretum imberbe leaves,
Phytochemistry 27 (1988) 3217-3220
.
14.
M.G. El-Ghazouly, N.A. El-Sebakhy, A.A.S. El-Din, C. Zdero, F. Bohlmann,
Sesquiterpene xylosides from Iphiona scabra, Phytochemistry 26 (1987) 439-443.
15.

H. Pfander, H. Stoll, Terpenoid glycosides, Nat Prod Rep 8 (1991) 69-95.

16.

B.M. Fraga, Natural sesquiterpenoids, Natural Product Reports 30 (2013) 1226-1264.

76

Chapter 3: Purification and structure elucidation of campherenane diol glycosides from
leaf glandular trichomes of Solanum habrochaites LA1777

3.1 Introduction
A systematic investigation of methanol leaf dip extracts of Solanum habrochaites LA1777 was
performed in order to purify and elucidate the structures of sesquiterpenoid specialized
metabolites that were annotated using relative mass defect filtering. This chapter discusses the
purification and structure elucidation of the campherenane diol derivatives introduced in chapter
two. The research described in this Chapter has already been published in the journal
Phytochemistry [1]
Among the sesquiterpenes reported in the literature, campherenanes are rare[2], but bear some
structural similarity to santalenes(Fig. 3.1 a. and b.) [3]. Campherenane derivatives have been
isolated from Santalum album [4], Santalum austrocaledonicum [5] and Illicium tsangii [6] but
not from any plants of the genus Solanum (tomato, potato). The biosynthetic pathway of
santalenes in Solanum habrochaites LA1777 leaf trichomes have been proposed by Sallaud et al
[7]. Furthermore, Sallaud et al. has proposed that depending on the stereochemistry of the
precursor and the position of ring closure different types of sesquiterpenes can result. For
instance, another sesquiterpene -bergamotene – also is synthesized starting with the same Z,Zfarnesyl diphosphate. Therefore, it is likely that the precursors of campherenane also are either
Z,Z-FPP or E,E-FPP however, a group of enzymes that convert these precursors specifically into
campherenane compounds must be present in these plants.
Glycosylated terpenes are seldom reported from genus Solanum. Some acyclic diterpene
glycosides have been reported from tobacco, a member of genus Solanum [8]. Yet, glycosylated
77

sesquiterpenes are not known from this genus. Glycosylation results in changes of physical
properties of terpenoids such as volatility[9]. Therefore, much interest has been placed on
developing terpene glycosides as slow releasing aroma and flavor compounds[10,9]. This has
resulted in development of methods to glycosylate terpenoids by chemical or enzymatic means
[11]
The abundance of genetic information about tomato (S. lycopersicum) [12] combined with the
availability of introgression lines derived from S. habrochaites and S. lycopersicum [13]
facilitates discovery of sesquiterpenoid metabolism genes from metabolomic and transcriptomic
data. An introgression line is an offspring that is genetically closely similar to parent plant with
only a few genetic differences. Therefore, by comparing the chemical constituents of parent plant
and the offspring plant it is possible to assign chemical traits to genes.
3.2 Experimental methods
3.2.1 Plant material
Five Solanum habrochaites LA1777 plants were grown using seeds obtained from Tomato
Genetics Resource Center of University California, Davis, USA, in a plant growth chamber (28
°C, 16 hour day time, 96% humidity; 15 µmol m-2 s-1 irradiance), and leaf and stem tissues were
harvested and extracted in six-week intervals. Final mass of fresh plant material extracted was
about 5 kg. Extraction was performed by dipping each branch in 1 L of 100% methanol for about
30 s. This approach aims for selective extraction of glandular trichomes. Extracts were
subsequently combined, and solvent was evaporated under vacuum on a rotary evaporator
without heating, and the residue was redissolved in about 10 ml of 100% methanol. This extract

78

was divided into approximately three equal portions, and a single compound was isolated from
each portion.
3.2.2 Purification of campherenane diol glycosides
Semipreparative HPLC was performed on an Acclaim 120 C18 column (4.6 × 150 mm, 5 µm
particles; Dionex Co. USA) using a Waters semipreparative HPLC system consisting of two
Waters 510 pumps controlled using a Waters 610 gradient controller. The system was equipped
with a Waters 717 plus autosampler and fractions were collected using a LKB BROMMA 221
Superrac fraction collector. For the semipreparative HPLC, flow rate was 2.0 ml/min and the
injection volume for each injection was 150 µL. The linear solvent gradient used for the
purification employed water (solvent A) and methanol (solvent B) was as follows (A:B): 0-1.0
min (99%:1%), 4.00 min (60%:40%), 4.50 min (55%:45%), and held at this composition until
9.00 min, 9.50 min (50%:50%), and held until 14.00 min, 14.50 min (0%:100%), and held until
17.00 min, then returned to initial conditions at 17.01min and held until 25.00 min (99%:1%).
Forty fractions were collected (30 s intervals for the first 20 min), and each fraction was tested
for the presence of compound of interest using a 5.50 min UHPLC-MS method in negative ion
mode ESI using the same instrument configuration used for profiling. This gradient was (A =
0.15% aqueous formic acid; B = methanol (A:B): 0.0-2.0 min (95%:5%), step at 2.01 min
(40%:60%), linear increase to 3.00 min (0%: 100%), and held at this condition until 5.0 min,
then returned to initial condition at 5.01 min and held until 5.5 min. From these analyses, the
fractions that were over 90% pure based on total ion chromatogram peak areas were identified
and combined, and solvent was removed by evaporating to dryness at room temperature under a
N2 gas flow.

79

3.2.3 NMR Experiments
All NMR experiments were performed in a Bruker Avance 900 spectrometer, equipped with TCI
triple-resonance inverse detection CryoProbe, at 900 MHz (1H frequency) at 25˚C. NMR
experiments were performed using CD3OD matched NMR tubes (Shigemi Inc., Allison Park, PA
USA). Once dried, each purified compound was redissolved in 99.8% CD3OD (Sigma-Aldrich,
St. Louis, MO, USA) and analyzed using NMR. All spectra were calibrated using residual nondeuterated solvent signal as δ1H (CD3OD) = 3.34 ppm and δ13C (CD3OD) = 49.08 ppm.
Proton,13C, gHSQC, gHMBC, DQF-COSY, TOCSY and NOESY spectra were acquired.
3.2.4 Acid hydrolysis, derivatization and GC-MS experiments
Acid hydrolysis was performed to identify the sugars attached to the terpenoid core. A 200 µL
aliquot of a 30-ml semipreparative HPLC fraction containing the purified compound (peak 6a in
Fig. 3.1) was mixed with 200 µL of 2.0 M sulfuric acid and incubated in a sealed vessel at 25 ˚C
for 6 hrs. Next, 1.0 M BaCl2 was added dropwise until no further precipitation of BaSO4 was
observed. The mixture was centrifuged (10,000xg) for 5 min, and the supernatant was
evaporated to dryness on a SpeedVac. To the dried residue, 150 µL of 20 mg/mL methoxyamine
hydrochloride/pyridine was added and incubated at 25 ˚C for 2 hrs. Next, 150 µL of MSTFA (Nmethyl-N-trimethylsilyltrifluoroacetamide) was added to the mixture and incubated for another
hour. Sugar standards (glucose, xylose, galactose, rhamnose purchased from Sigma-Aldrich, St.
Louis, MO, USA) were derivatized using the same procedure, and these were analyzed using an
Agilent 5975 mass spectrometer with 6890N GC and autosampler. The column was an Agilent
DB-5 (10 m x 0.18 mm, 0.34 µm film) using a temperature program as follows; Initial
temperature 40°C, hold for 1.00 min; increase at 30 ˚C/min to 90 ˚C, then 5 ˚C/min to 110 ˚C, 40
˚C/min to 165 ˚C, 5 ˚C/min up to 180 ˚C, and 40 ˚/min to 320 ˚C, and maintained at 320 ˚C for
80

one minute. Ionization employed 70 eV electron ionization. The same procedure was used to
verify the identity of carbohydrates in fractions 6b and 8c.
3.3 Results and discussion
3.3.1 Profiles of the campherenane diol diglycosides
The metabolite profiling performed in the experiments discussed in Chapter 2 allowed for the
initial annotation of campherenane diol diglycosides from Solanum habrochaites LA1777 leaf
extract. Negative ion mode UHPLC-MS profiling of leaf dip metabolites from wild tomato
species S. habrochaites LA1777 showed more than 50 chromatographic peaks not evident in
analyses of extracts from other Solanum species. Multiplexed non-selective collision-induced
dissociation (CID) time of flight (TOF) mass spectrometry enabled the rapid annotation of these
molecules based on accurate pseudomolecular and co-eluting fragment ion masses, as well as the
collision energy-dependence of fragment ion abundances. Assignments of fragment ions in
multiplexed nonselective CID spectra may be subject to artifactual false positives when
metabolites co-elute, but in this study the origins of characteristic fragment ions were confirmed
using MS/MS experiments that employed selection of specific precursor ions before CID.
Chapter 2 described the annotation of campherenane diol diglycoside detected in negative-ion
mode LC/MS as m/z 609 and campherenane diol triglycoside malonate ester detected as m/z 811.
At least 10 different isomers were identified for m/z 609 from the leaf dip extract of LA1777
(Fig. 3.1c). Peak 8c (Fig. 3.1) was purified using HPLC and the structure was elucidated using
NMR. Similarly, at least 8 isomers were identified for m/z 811 from the same extract (Fig. 3.1e).
The annotation of the other campherenane diol derivatives is described below.

81

One of the metabolite peaks (Rt = 9.0 min; peak 6a in Fig. 3.1) showed [M-H]- at m/z 649
(experimental m/z 649.3089; theoretical m/z for C30H49O15- = 649.3077; m = +1.8 ppm), and
MS/MS revealed fragment ions corresponding to facile loss of CO2 (44 Da) and loss of an
additional 42 Da (C2H2O). These neutral losses suggested presence of a malonate ester.
Additional neutral loss of 162.05 Da suggested the presence of a hexose group, and accurate
mass measurements of the remaining fragment at m/z 401 (experimental m/z 401.2536;
theoretical m/z for C21H37O7- = 401.2545; m = -2.2 ppm) yielded a relative mass defect of 640
ppm which, when combined with the observed resistance to further fragmentation, were
consistent with fractional hydrogen content of a glycosylated sesquiterpenoid core. Though
initial inspection of nonselective CID mass spectra suggested loss of only a single neutral
carbohydrate fragment, product ion MS/MS spectra for m/z 649 at higher collision energies
yielded a fragment at m/z 239 (experimental m/z 239.2006; theoretical for C15H27O2- = 239.2017;
∆m = -4.6 ppm) consistent with an additional loss of 162 Da, and led to annotation as a malonate
ester of a sesquiterpene diol dihexoside. At least 9 additional metabolites were observed in the
UHPLC-MS/MS profile that shared the same molecular mass and common fragment ions in the
CID mass spectra, suggesting a diverse group of isomeric sesquiterpenoid dihexoside malonate
esters.
Another peak in the total ion chromatogram (peak 6b in the extracted ion chromatogram in
Figure 3.1b) exhibited a base peak ion at m/z 487 (experimental m/z 487.2539; theoretical m/z for
C24H39O10- = 487.2549; ∆m = -2.0 ppm) that was 162 Da less than 649. This molecule also
underwent collision-induced dissociation to give m/z 401 and 239 fragments. Since the molecular
mass was 162 Da less than the malonate ester of sesquiterpene diol dihexoside, it was concluded

82

to be a malonate ester of a sesquiterpene diol monohexoside. At least six isomers of this
molecule were detected in UHPLC-MS/MS profiles from S. habrochaites LA1777 (Fig. 3.1b).
Another metabolite (peak 5d in Fig. 3.1d) was observed in the UHPLC-MS analyses at m/z 651
(experimental m/z 651.3233; theoretical m/z for C30H51O15- = 651.3210; ∆m = +3.5 ppm). The
MS/MS spectrum yielded [M-H]- at m/z 605 (experimental m/z of [M-H]- = 605.3164; theoretical
m/z 605.3179; ∆m = -2.5 ppm), which was 44 Da lower in mass than the sesquiterpene diol
dihexoside malonate (peak 6a in Fig. 3.1), and 42 Da greater than the sesquiterpene diol
dihexoside, consistent with an acetate ester in place of the malonate ester. Therefore, this
metabolite was annotated as an acetate ester of a sesquiterpene diol dihexoside. At least 10
chromatographically-resolved isomers of this could be annotated from UHPLC-MS/MS profiles
of S. habrochaites LA1777 (Fig. 3.1d) The MS/MS spectra of all isomers showed the loss of the
acetyl group (neutral loss of 42 Da) followed by the sequential losses of two anhydrohexoses
(162 Da) to yield m/z 401 and aglycone core of m/z 239.
A non-malonylated form of sesquiterpene diol trihexoside described in Chapter 2 also was
detected (Fig. 3.1f, peak 6f) as m/z 771 (experimental m/z 771.3621; theoretical m/z for
C34H59O19- = 771.3656; m = -4.5 ppm; annotated as [M+formate]-). The MS/MS product ion
spectrum exhibited [M-H]- at m/z 725 and fragments at m/z 563, 401 and 239 that corresponded
to three sequential neutral losses of 162 Da indicating the presence of the three hexoses. Eight
isomers of this sesquiterpene diol trihexoside were evident in UHPLC-MS profiles (Fig. 3.1f)
and yielded the same fragment ion masses.
Further exploration of the UHPLC-MS profile yielded recognition of a set of metabolites
yielding base peaks at m/z 447 (experimental m/z 447.2585 for peak 2g in Figure 3.1g;

83

theoretical m/z for C22H37O9- = 447.2600; ∆m = -3.3 ppm), consistent with [M+formate]- of the
sesquiterpenoid monoglycoside described above (peak 6b) but without the malonate group.
Formate adducts exhibited a characteristic neutral loss of formic acid (46 Da) under collisioninduced dissociation conditions. The [M-H]- ion co-eluting with the above formate adduct is
observed in MS/MS spectrum of products from m/z 447 at m/z 401 (experimental m/z 401.2545;
theoretical m/z for C21H37O7- = 401.2520; ∆m = -6.2 ppm), and 239, corresponding to the loss of
anhydrohexose (C6H10O5). Therefore, this metabolite was annotated as a sesquiterpene diol
monohexoside. Eight isomers with matching pseudomolecular and fragment masses were
observed in extracts from S. habrochaites LA1777 (Fig. 3.1g).
Similarly, another compound detected at m/z 489 (experimental m/z 489.2687 for peak 4h in
Figure 3.1h; theoretical m/z for C24H41O10- = 489.2705; ∆m = -3.6 ppm), consistent with
[M+formate]- of a sesquiterpenoid monoglycoside acetate ester. Formate adducts exhibited a
characteristic neutral loss of formic acid (46 Da) under collision-induced dissociation conditions.
Like the formate adduct of sesquiterpenoid monoglycoside described above, the [M-H]- ion coeluting with the above formate adduct is observed in MS/MS spectrum of products from m/z 489
at m/z 443 (experimental m/z 443.2632; theoretical m/z for C23H40O8- = 443.2650; ∆m = -4.0
ppm), 401 and 239, corresponding to the sequential loss of the acetate group (42 Da) followed by
the loss of anhydrohexose (C6H10O5). Therefore, this metabolite was annotated as a
sesquiterpene diol monohexoside acetate ester. Four isomers with matching pseudomolecular
and fragment masses were observed in extracts from S. habrochaites LA1777 (Fig. 3.1h).

84

Retention time (min)
Figure 3.1. Extracted ion UHPLC-MS chromatograms showing sesquiterpene glycoside
metabolites identified in extracts from Solanum habrochaites LA1777. (a) [M-H]- (m/z 649) for
campherenane diol diglucoside malonate esters, (b) [M-H]- (m/z 487) for campherenane diol
monoglucoside malonate esters, (c) [M+formate]- (m/z 609) for campherenane diol diglucosides,
(d) [M+formate]- (m/z 651) for campherenane diol diglucoside acetate esters, (e) [M-H]- (m/z
811) for campherenane diol triglucoside malonate esters, (f) [M+formate]- (m/z 771) for
campherenane diol triglucosides, (g). [M+formate]- (m/z 447) for campherenane diol
monoglucosides and (h) [M+formate]- (m/z 489) for campherenane diol monoglucosides acetate
ester. Labels with larger font size in red designate the four purified metabolites. 32
85

3.3.2 Structure elucidation of sesquiterpene diol dihexoside malonate ester and establishing
the structure of campherenane terpenoid core
One of the major isomers of sesquiterpene diol dihexoside malonate ester (Fig. 1a, peak 6a. Rt =
9.0 min) was purified from extracts of S. habrochaites LA1777, and its structure was elucidated
using 1D and 2D NMR. The structure shown in Fig. 3.2a was assigned to this molecule based
upon NMR and mass spectrometry information. In the proton NMR spectrum (Fig. 3.4 in
appendix) the anomeric protons corresponding to the two glucose moieties could be identified at
proton resonances δH 4.63 ppm (J = 8.52 Hz), δH = 4.28 ppm (J = 7.97 Hz) as two doublets. This
supported and anchored the annotation of two hexose moieties. Coupling constants of each
anomeric proton suggested that the glycosides are bound by beta-substitution. Three methyl
group resonances could be identified. Two of these were singlets and one was a doublet (δH 0.99
ppm (J = 7.42 Hz)). The doublet was indicative that one of the three methyl groups was
connected to a tertiary carbon atom with a single proton. Other protons were not well resolved in
the proton spectrum, and therefore to assign these, the multiplicity edited gHSQC spectrum was
used (Fig. 3.6) , and all protons were readily identifiable along with their corresponding carbon
connectivity. To identify neighboring protons, COSY, gHMBC and TOCSY spectra (Figs. 3.5,
3.7, 3.8 in appendix) were used. Starting from the anomeric protons, 1H and 13C resonances of
the two hexose moieties were assigned. Next, the structure of terpenoid core was assigned using
the same approach. In the terpenoid core, protons located on primary and secondary carbons with
glycoside attachments were distinctive, with methylene protons at δH 3.36 ppm and δH 3.43 ppm
for the primary carbon position (C12) and δH 4.00 ppm at the secondary carbon position (C2).
The correlations observed in gHMBC spectra between these protons and anomeric carbons of the
hexose moieties were important in identifying the position of each hexose attachment. Similarly,

86

protons on C2 carbon atom correlated in gHMBC spectra with the C22 carbon resonance and
vice versa. The gHMBC spectrum was important in assigning the quaternary carbon atoms also.
The two quaternary carbon atoms (δC 51.14 ppm and δC 52.06 ppm) displayed correlations in
gHMBC spectra with the neighboring protons (Fig. 3a) and this enables their correct
assignments. Next, the methylene protons in malonate were weakly visible in the gHSQC
spectrum (δH 4.43 ppm and δH 4.33 ppm). The position of the malonate ester attachment was
assigned using correlations observed in both gHMBC and NOESY (Fig. 3.9) spectra. Positioning
of malonate could be inferred based on HMBC correlations between the hexose moiety protons
(δH 3.71 ppm) and the carbonyl carbon (δC 172.45 ppm) (Fig. 3.4). In the NOESY spectrum,
malonate protons showed correlations with protons in the hexose attached to the secondary
position of the terpenoid core. (δH 4.32 ppm and δH 4.00 ppm of terpenoid core, and δH 4.32 and
δH 3.18 ppm of hexose). Furthermore, in the NOESY spectrum, the proton at the secondary
alcohol position (C2) shows correlation with protons on methyl attached to C7 position of
terpenoid core (δH 4.00 ppm and δH 0.91 ppm terpenoid core) confirming the position of the
secondary alcohol group. This information helps distinguish this terpenoid core structure from
the previously reported campherenane type cores [4]. Based on these observations the final
structure shown in Fig. 3.2a was assigned and the carbon and proton resonances for each carbon
and proton in the molecule are shown in Table 3.1.
3.3.3 Structure elucidation of sesquiterpene monohexoside malonate ester
Three main isomers of sesquiterpene monohexoside malonate molecule are observed in extracts
of LA1777 (Fig. 1b m/z 487). The major isomer (Fig. 3.1b peak 6b, Rt = 15.3 min) of this
compound was purified and analyzed using 1D and 2D NMR spectroscopy. Based on NMR and
mass spectrometry experiments the structure presented in Fig. 3.2b was assigned to this
87

compound. Proton and gHSQC NMR spectra of the compound (Fig. 3.10 and 3.12) were
suggestive of the presence of three methyl groups corresponding to the sesquiterpene core and a
single anomeric proton from the hexose moiety on the molecule (δH 4.25, d (J=8.04 Hz)). Based
on coupling constant of the anomeric proton the glycosidic bond that connected the hexose
moiety to the terpenoid core was identified to be a beta bond. One of the resonances
corresponding to a methyl was a doublet (δH 0.99 ppm (J=7.09 Hz) suggesting its located
attached to a tertiary carbon atom. Proton – carbon connectivity of each CH, CH2 and CH3 was
assigned using the multiplicity edited gHSQC data (Fig. 3.12). To assign neighboring carbons
and protons, COSY and gHMBC spectra were used. To ascertain the positions of the quaternary
carbons the gHMBC spectrum was used (Fig. 3.13). Starting from the anomeric proton, the
resonances corresponding to the hexose moiety were assigned using gHSQC, COSY and
gHMBC. Next, the structure of the terpenoid core was assembled using the same approach. In
the terpenoid core, the protons on the primary alcohol and secondary carbon where the hexose
group is attached (C2 and C12 respectively) could be identified (δH 3.72 ppm and δH 3.42 ppm
for C2 and δH 3.98 ppm for secondary carbon (C12). To confirm the position of the primary
alcohol group, the NOESY spectrum (Fig. 3.15) also was important. In the NOESY spectrum,
the proton at C2 position correlates with protons on methyl group attached to C7 position (δH
3.98 ppm and δH 0.90 ppm of terpenoid core). The position of hexose was inferred based on the
correlations observed in gHMBC spectra between these protons and the anomeric carbon atoms.
The anomeric proton correlated with C12 carbon atom and the protons of C12 carbon atom
correlated with anomeric carbon (C16) atom. Using this, the position of hexose moiety was
determined. Position of the malonate group was inferred based on NOESY and gHMBC
experiments. One of the methylene protons of malonate (δH 4.40 ppm) displayed NOEs with

88

hexose moiety protons (δH 3.36 ppm) indicating it is located proximal to hexose in space. The
exact position of malonate was assigned using correlations observed in gHMBC between the
hexose group proton (δH 3.87 ppm) and carbonyl carbon. The resonances assigned to each proton
and carbon atom are shown in Table 3.1.
3.3.4 Structure elucidation of sesquiterpene diol dihexoside
A third novel molecule isolated from S. habrochaites LA1777 was the sesquiterpene dihexoside.
At least ten isomers of this molecule were observed from UHPLC-MS profiling (Fig. 3.1c). This
molecule displayed the presence of two hexose moieties but no sign of malonate was detected by
NMR or mass spectrometry, and the observation of [M+formate]- rather than [M-H]- supports the
notion that this metabolite lacks acidic groups. Two major isomers of this molecule were
detected along with a number of less abundant isomers from LA1777, and the more abundant
isomer (Fig. 1c Isomer 8c at Rt = 12.8 min) was purified and analyzed using NMR (Fig. 3.2c and
Table 3.1). The 1H NMR spectrum (Fig. 3.16) shows the presence of two anomeric protons (δH
4.60 ppm (J=7.84 Hz) and δH 4.37 ppm (J=7.82 Hz)). Coupling constants of the anomeric
protons suggest beta glycosidic bonds that connect the hexoses. Positions of attachment for the
hexose moieties to the terpenoid core were inferred based on HMBC couplings of anomeric
protons and anomeric carbons with the terpenoid core (Fig. 3.19). Spectra suggest that all three
compounds share a common sesquiterpene diol (campherenane diol) core. (Fig. 3.16 to Fig. 3.21)
3.3.5 Structure elucidation of sesquiterpene diol dihexoside acetate ester
Sesquiterpene dihexoside acetate ester was detected from S. habrochaites LA1777 at m/z 651 as
[M+formate]- which yielded a [M-H]- at m/z 605 upon collisional activation. The presence of two
hexose moieties and an acetate group in this molecule was judged based on neutral losses
observed during collisional activation of [M+formate]- ion. One of the most abundant isomers,
89

(Fig. 3.1. 5d) was purified using preparative HPLC and NMR spectral analysis was performed.
The NMR spectra showed the presence of two anomeric protons (δH 4.66 ppm (J=8.2 Hz) and δH
4.43 ppm (J=7.68 Hz)) confirming the presence of two hexoses (Fig. 3.22). The coupling
constants of the two anomeric protons suggest beta glycosidic bonds that connect the hexoses.
HMBC spectra confirmed that the two hexose moieties are connected to one another. Also the
connectivity of the hexoses to the terpenoid core was confirmed using HMBC. HSQC and
proton displayed the presence of four methyl groups in the molecule. Three of these were
attributed to the terpenoid core and the last one was inferred to be present in the acetate group.
Spectra suggest that all four isolated compounds share a common sesquiterpene diol (2,12campherenane diol) core with identical relative stereochemical configuration at C7 (Fig. 3.22 to
Fig. 3.28).
3.3.6 Using GC-MS data to support NMR based structure assignments of sugar moieties
Despite the usefulness of NMR in de novo structure assignments, assigning NMR chemical shifts
for the accurate assignment of sugar structures is challenging since different sugars (ex: glucose,
galactose) can show similar chemical shifts. This can be particularly challenging in the presence
of multiple sugars in a given molecule due to overlapping chemical shifts. However, the gas
chromatographic retention times of MSTFA derivatives of different sugars are characteristic, and
therefore offer a reliable alternative to NMR for assigning the sugars.
The sugar composition of these glycoconjugates was examined using acid hydrolysis followed
by derivatization and GC-MS, which distinguishes sugars based on their retention times and
mass spectra. The GC-MS retention times and spectra were compared with results from reference
monosaccharide standards, and the only sugar detected in any of these analyses was glucose.
Results of these analyses are presented in Fig. 3.3.
90

3.4 Conclusions
Compared to the santalenes, campherenanes identifying in this study do not contain any double
bonds or the three membered ring structures. Additionally, two hydroxyl groups are found in the
campherenane diol core. These structural differences indicate that campherenanes potentially
undergo more extensive redox reactions during the biosynthesis compared to santalenes.
Furthermore, the additional glycosylation, acetylation and malonylation are indicative of the
involvement of a number of additional secondary reactions that transforms the basic terpenoid
core into a non-volatile glycosylated terpene form. Such extensive transformation of
sesquiterpene precursors into glycosylated sesquiterpenoid form is an indication of the
involvement of a large number of different enzymes in the biosynthetic pathway that are still to
be explored. Thus, these findings open the doors to a novel class of compounds that have been
hidden from researchers studying Solanum habrochaites LA1777 for many decades.

91

APPENDIX

92

Table 3.1: Chemical shifts (in ppm) of 1H and 13C resonances for (a) campherenane diol
diglucoside malonate ester (peak 6a from Figure 3.1), (b) campherenane diol monoglucoside
malonate ester (peak 6b), (c) campherenane diol diglucoside (peak 8c), and (d) campherenane
diol diglucoside acetate ester (peak 5d). 4
a

c
δH

δC

c
δH

δC

d

Position

δC

1

51.1

2

78.7

4.00

3

39.6

0.99

4

43.9

5

44.9

1.41

6

27.6

2.00

7

52.1

8

35.0

1.48

1.01

34.8

1.48

1.02

34.5

1.45

0.96

9

43.1

1.42

1.13

31.1

1.32

1.32

31.4

1.24

10

36.2

1.50

1.14

36.3

1.51

1.14

35.7

1.45

11

35.2

1.81

35.1

1.80

34.8

1.77

12

77.0

3.36

76.9

3.72

76.9

3.70

13

18.3

0.99
(3H,d)

17.8

0.99
(3H,d)

17.8

14

16.0

0.91
(3H,s)

16.1

0.9
(3H,s)

15

14.0

0.86
(3H,s)

14.2

16

105.6

4.28

17

75.9

18
19

51.2

δH

δC

51.2

78.8

3.98

39.7

0.99

43.7

1.78

1.77

29.4

1.25

1.29

27.6

2.00

δH

52.0
78.5

4.00

39.7

0.97

43.5

1.77

1.65

29.6

1.27

1.67

1.28

27.5

2.01

1.30

34.9

1.48

1.00

1.31

22.8

1.48

1.23

1.1

36.0

1.49

1.13

35.2

1.80

76.8

3.75

0.94
(3H,d)

17.7

0.99
(3H,d)

16.1

0.87
(3H,s)

16.5

0.91
(3H,s)

0.86
(3H,s)

13.8

0.82
(3H,s)

14.5

0.89
(3H,s)

105.8

4.25

103.5

4.37

103.9

4.43

3.21

76.2

3.23

83.4

3.48

83.2

3.50

78.6

3.39

71.9

3.40

78.6

3.58

78.6

3.60

72.7

3.31

72.3

3.31

78.8

3.38

71.95

3.34

2.22

78.0

3.95

39.0

0.95

43.0

1.77

1.66

29.0

1.21

1.29

27.0

1.95

2.22

51.6

3.43

2.19

51.6

93

3.42

2.21

51.9

3.45

3.48

Table 3.1 (cont’d)
20

75.8

3.52

78.73

3.31

21

63.6

3.88

63.0

3.89

22

106.3

4.63

170.9

4.63

106.00

4.66

23

83.7

3.43

65.0

76.1

3.24

76.6

3.28

24

82.7

3.52

170.9

78.9

3.41

78.6

3.44

25

79.0

3.58

72.6

3.84

72.3

3.32

26

72.8

3.18

72.3

3.34

78.7

3.31

27

63.6

3.71

63.6

3.90

63.0

3.89

28

172.5

29

66.3

30

170.4

3.69

78.8

3.35

63.2

3.67

4.4

3.88

3.87

4.35

76.9

3.27

63.7

3.68

105.8

3.88

3.68

172.6
4.43

4.33

32.3

94

1.37

3.72

3.72

Figure 3.2: Structures of (a) campherenane-2,12-diol diglucoside malonate ester (peak 6a from
Figure 1), (b) campherenane-2,12-diol monoglucoside malonate ester (peak 6b), (c)
campherenane-2,12-diol diglucoside (peak 8c), and (d) campherenane-2,12-diol diglucoside
acetate ester (peak 5d) as determined using NMR and tandem mass spectrometry. Portions of the
molecule corresponding to key fragment ions in negative ion MS/MS spectra are illustrated on
each structure. Carbon atoms are numbered in accordance with NMR assignments in Table 3.1.

95

33

Figure 3.3. GC/MS total ion chromatograms of methoxime-trimethylsilyl derivatives of (a)
products of acid hydrolysis of campherenane-2-endo-(6'-malonyl)glucoside)-12-glucoside
(Compound 1; peak 6a from Figure 1), (b) glucose reference standard, (c) xylose reference
standard, (d) galactose reference standard, (e) rhamnose reference standard.34

96

Figure 3.4. 1H and 13C NMR spectra of the isolated campherenane-2-endo-(6'malonyl)glucoside)-12-glucoside (Compound 1; peak 6a from Figure 3.1). 35
97

Figure 3.5. 1H-1H COSY NMR spectrum of the campherenane-2-endo-(6'-malonyl)glucoside)12-glucoside (Compound 1; peak 6a from Figure 3.1). 36

98

Figure 3.6. Multiplicity edited gHSQC (1H-13C) NMR spectrum of campherenane-2-endo-(6'malonyl)glucoside)-12-glucoside (Compound 1; peak 6a in Figure 1).37

99

Figure 3.7. gHMBC NMR spectrum of campherenane-2-endo-(6'-malonyl)glucoside)-12glucoside (Compound 1; peak 6a from Figure 1). 38

100

Figure 3.8. TOCSY NMR spectrum of campherenane-2-endo-(6'-malonyl)glucoside)-12glucoside (Compound 1; peak 6a in Figure 1). 39

101

Figure 3.9. NOESY NMR spectrum of campherenane-2-endo-(6'-malonyl)glucoside)-12glucoside (Compound 1; peak 6a in Figure 1). 40

102

Figure 3.10. 1H and 13C NMR spectra of the isolated 2-endo-campherenanol-12-(6'malonyl)glucoside (Compound 2; peak 6b from Figure 1). 41
103

Figure 3.11. 1H-1H COSY NMR spectrum of the 2-endo-campherenanol-12-(6'malonyl)glucoside (Compound 2; peak 6b from Figure 1). 42

104

Figure 3.12. Multiplicity edited gHSQC (1H-13C) NMR spectrum of 2-endo-campherenanol-12(6'-malonyl)glucoside (Compound 2; peak 6b in Figure 1). 43

105

Figure 3.13. gHMBC NMR spectrum of 2-endo-campherenanol-12-(6'-malonyl)glucoside
(Compound 2; peak 6b from Figure 1). 44

106

Figure 3.14. TOCSY NMR spectrum of 2-endo-campherenanol-12-(6'-malonyl)glucoside
(Compound 2; peak 6b in Figure 1). 45

107

Figure 3.15. NOESY NMR spectrum of 2-endo-campherenanol-12-(6'-malonyl)glucoside
(Compound 2; peak 6b in Figure 1). 46

108

Figure 3.16. 1H and 13C NMR spectra of the isolated campherenan-2-endo,12-diglucoside
(Compound 3; peak 8c from Figure 3.1). 47

109

Figure Figure 3.17. 1H-1H COSY NMR spectrum of the campherenan-2-endo,12-diglucoside
(Compound 3; peak 8c from Figure 1). 48

110

Figure 3.18. Multiplicity edited gHSQC (1H-13C) NMR spectrum of campherenan-2-endo,12diglucoside (Compound 3; peak 8c in Figure 1). 49

111

Figure 3.19. gHMBC NMR spectrum of campherenan-2-endo,12-diglucoside (Compound 3;
peak 8c from Figure 1). 50

112

Figure 3.20. TOCSY NMR spectrum of campherenan-2-endo,12-diglucoside (Compound 3; peak
8c in Figure 1). 51

113

Figure 3.21. 2D-NOESY NMR spectrum of campherenan-2-endo,12-diglucoside (Compound 3;

peak 8c in Figure 1). 52

114

Figure 3.22. Proton NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). 53

115

Figure 3.23. 13C NMR spectrum of 2-endo-campherenanol-12-(2-(6"-acetyl)glucosyl)glucoside
(Compound 4; peak 5d in Figure 3.1) 54

116

Figure 3.24. Multiplicity edited 2D gHSQC NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). 55

117

Figure 3.25. 2D gdqCOSY NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). 56

118

Figure 3.26. 2D HMBC NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). 57

119

Figure 3.27. 2D TOCSY NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). 58

120

Figure 3.28. 2D NOESY NMR spectrum of 2-endo-campherenanol-12-(2-(6"acetyl)glucosyl)glucoside (Compound 4; peak 5d in Figure 1). 59

121

REFERENCES

122

REFERENCES
1. Ekanayaka EAP, Li C, Jones AD (2014) Sesquiterpenoid glycosides from glandular trichomes
of the wild tomato relative Solanum habrochaites. Phytochemistry 98:223-231.
doi:10.1016/j.phytochem.2013.11.011
2. Liu Y-N, Su X-H, Huo C-H, Zhang X-P, Shi Q-W, Gu Y-C (2009) Chemical Constituents of
Plants from the Genus Illicium. Chemistry & Biodiversity 6 (7):963-989.
doi:10.1002/cbdv.200700433
3. Coates RM, Denissen JF, Juvik JA, Babka BA (1988) Identification of alpha-santalenoic acid
and endo-beta-bergamotenoic acids as moth oviposition stimulants from wild tomato leaves.
Journal of Organic Chemistry 53 (10):2186-2192. doi:10.1021/jo00245a012
4. Kim TH, Ito H, Hatano T, Takayasu J, Tokuda H, Nishino H, Machiguchi T, Yoshida T
(2006) New antitumor sesquiterpenoids from Santalum album of Indian origin. Tetrahedron
62 (29):6981-6989. doi:http://dx.doi.org/10.1016/j.tet.2006.04.072
5. Alpha T, Raharivelomanana P, Bianchini J-P, Faure R, Cambon A (1997) A sesquiterpenoid
from Santalum austrocaledonicum var. austrocaledonicum. Phytochemistry 46 (7):1237-1239.
doi:http://dx.doi.org/10.1016/S0031-9422(97)80018-0
6. Ngo K-S, Brown GD (1999) Santalane and isocampherenane sesquiterpenoids from Illicium
tsangii. Phytochemistry 50 (7):1213-1218. doi:http://dx.doi.org/10.1016/S00319422(98)00652-9
7. Sallaud C, Rontein D, Onillon S, Jabes F, Duffe P, Giacalone C, Thoraval S, Escoffier C,
Herbette G, Leonhardt N, Causse M, Tissier A (2009) A Novel Pathway for Sesquiterpene
Biosynthesis from Z,Z-Farnesyl Pyrophosphate in the Wild Tomato Solanum habrochaites.
Plant Cell 21 (1):301-317. doi:10.1105/tpc.107.057885
8. Jassbi AR, Zamanizadehnajari S, Kessler D, Baldwin IT (2006) A new acyclic diterpene
glycoside from Nicotiana attenuata with a mild deterrent effect on feeding Manduca sexta
larvae. Zeitschrift Fur Naturforschung Section B-a Journal of Chemical Sciences 61 (9):11381142
9. Herrmann A (2007) Controlled release of volatiles under mild reaction conditions: From
nature to everyday products. Angewandte Chemie-International Edition 46 (31):5836-5863.
doi:10.1002/anie.200700264
10. Goldstein H, Murakami AA, Navarro AL, Ryder DS (2004) Use of glycosides extracted from
hop plant parts to flavor malt beverages. Google Patents,

123

11. Caputi L, Lim EK, Bowles DJ (2008) Discovery of new biocatalysts for the glycosylation of
terpenoid scaffolds. Chemistry-A European Journal 14 (22):6656-6662.
doi:10.1002/chem.200800548
12. Sato S, Tabata S, Hirakawa H, Asamizu E, Shirasawa K, Isobe S, Kaneko T, Nakamura Y,
Shibata D, Aoki K, Egholm M, Knight J, Bogden R, Li CB, Shuang Y, Xu X, Pan SK, Cheng
SF, Liu X, Ren YY, Wang J, Albiero A, Dal Pero F, Todesco S, Van Eck J, Buels RM,
Bombarely A, Gosselin JR, Huang MY, Leto JA, Menda N, Strickler S, Mao LY, Gao S,
Tecle IY, York T, Zheng Y, Vrebalov JT, Lee J, Zhong SL, Mueller LA, Stiekema WJ,
Ribeca P, Alioto T, Yang WC, Huang SW, Du YC, Zhang ZH, Gao JC, Guo YM, Wang XX,
Li Y, He J, Li CY, Cheng ZK, Zuo JR, Ren JF, Zhao JH, Yan LH, Jiang HL, Wang B, Li HS,
Li ZJ, Fu FY, Chen BT, Han B, Feng Q, Fan DL, Wang Y, Ling HQ, Xue YBA, Ware D,
McCombie WR, Lippman ZB, Chia JM, Jiang K, Pasternak S, Gelley L, Kramer M, Anderson
LK, Chang SB, Royer SM, Shearer LA, Stack SM, Rose JKC, Xu YM, Eannetta N, Matas AJ,
McQuinn R, Tanksley SD, Camara F, Guigo R, Rombauts S, Fawcett J, Van de Peer Y, Zamir
D, Liang CB, Spannagl M, Gundlach H, Bruggmann R, Mayer K, Jia ZQ, Zhang JH, Ye ZBA,
Bishop GJ, Butcher S, Lopez-Cobollo R, Buchan D, Filippis I, Abbott J, Dixit R, Singh M,
Singh A, Pal JK, Pandit A, Singh PK, Mahato AK, Dogra V, Gaikwad K, Sharma TR,
Mohapatra T, Singh NK, Causse M, Rothan C, Schiex T, Noirot C, Bellec A, Klopp C,
Delalande C, Berges H, Mariette J, Frasse P, Vautrin S, Zouine M, Latche A, Rousseau C,
Regad F, Pech JC, Philippot M, Bouzayen M, Pericard P, Osorio S, del Carmen AF, Monforte
A, Granell A, Fernandez-Munoz R, Conte M, Lichtenstein G, Carrari F, De Bellis G, Fuligni
F, Peano C, Grandillo S, Termolino P, Pietrella M, Fantini E, Falcone G, Fiore A, Giuliano G,
Lopez L, Facella P, Perrotta G, Daddiego L, Bryan G, Orozco M, Pastor X, Torrents D, van
Schriek K, Feron RMC, van Oeveren J, de Heer P, daPonte L, Jacobs-Oomen S, Cariaso M,
Prins M, van Eijk MJT, Janssen A, van Haaren MJJ, Jo SH, Kim J, Kwon SY, Kim S, Koo
DH, Lee S, Hur CG, Clouser C, Rico A, Hallab A, Gebhardt C, Klee K, Jocker A, Warfsmann
J, Gobel U, Kawamura S, Yano K, Sherman JD, Fukuoka H, Negoro S, Bhutty S, Chowdhury
P, Chattopadhyay D, Datema E, Smit S, Schijlen EWM, van de Belt J, van Haarst JC, Peters
SA, van Staveren MJ, Henkens MHC, Mooyman PJW, Hesselink T, van Ham R, Jiang GY,
Droege M, Choi D, Kang BC, Kim BD, Park M, Kim S, Yeom SI, Lee YH, Choi YD, Li GC,
Gao JW, Liu YS, Huang SX, Fernandez-Pedrosa V, Collado C, Zuniga S, Wang GP, Cade R,
Dietrich RA, Rogers J, Knapp S, Fei ZJ, White RA, Thannhauser TW, Giovannoni JJ, Botella
MA, Gilbert L, Gonzalez R, Goicoechea JL, Yu Y, Kudrna D, Collura K, Wissotski M, Wing
R, Schoof H, Meyers BC, Gurazada AB, Green PJ, Mathur S, Vyas S, Solanke AU, Kumar R,
Gupta V, Sharma AK, Khurana P, Khurana JP, Tyagi AK, Dalmay T, Mohorianu I, Walts B,
Chamala S, Barbazuk WB, Li JP, Guo H, Lee TH, Wang YP, Zhang D, Paterson AH, Wang
XY, Tang HB, Barone A, Chiusano ML, Ercolano MR, D'Agostino N, Di Filippo M, Traini
A, Sanseverino W, Frusciante L, Seymour GB, Elharam M, Fu Y, Hua A, Kenton S, Lewis J,
Lin SP, Najar F, Lai HS, Qin BF, Qu CM, Shi RH, White D, White J, Xing YB, Yang KQ, Yi
J, Yao ZY, Zhou LP, Roe BA, Vezzi A, D'Angelo M, Zimbello R, Schiavon R, Caniato E,
Rigobello C, Campagna D, Vitulo N, Valle G, Nelson DR, De Paoli E, Szinay D, de Jong HH,
Bai YL, Visser RGF, Lankhorst RMK, Beasley H, McLaren K, Nicholson C, Riddle C,
Gianese G, Tomato Genome C (2012) The tomato genome sequence provides insights into
fleshy fruit evolution. Nature 485 (7400):635-641. doi:10.1038/nature11119

124

13. Monforte AJ, Tanksley SD (2000) Development of a set of near isogenic and backcross
recombinant inbred lines containing most of the Lycopersicon hirsutum genome in a L.
esculentum genetic background: A tool for gene mapping and gene discovery. Genome 43
(5):803-813. doi:10.1139/g00-043

125

Chapter 4: Purification and structure elucidation of sesquiterpene I and sesquiterpene II
glycosides from leaf glandular trichomes of Solanum habrochaites LA1777
4.1 Introduction
A systematic investigation of methanol leaf dip extracts of Solanum habrochaites LA1777 plant
was performed in order to purify and elucidate structures of the sesquiterpene I and II glycosides
that were annotated using relative mass defect filtering in Chapter 2. Assigning the structure of
these compounds is an important first step in the road to the discovery of, biosynthetic
precursors, enzymes involved in biosynthesis and the genes regulating the expression of the
enzymes. This becomes an even more important endeavor for compounds that display structural
modifications such as cyclizations, shifts in methyl group positions, glycosylations, and
acylations. This is because such modifications in structure are suggestive that synthesis of these
compounds involves a number of additional reaction steps involving additional enzymes, and
thus exploring these compounds can lead to the discovery of novel biosynthetic pathways or
chemical precursors. In establishing the structures of compounds, identifying the stereochemistry
(relative or absolute) of the chiral centers as much as possible is important for establishing the
identity of the precursors. As discussed in chapter 3 of this dissertation, it has been established
that depending on the stereochemistry of the precursor, the end product also varies [1].
Furthermore, the biological activity of two enantiomers may not be the same due to enzyme
specificity toward one enantiomer than the other[2].
Establishing the structure of these compounds requires the purification and obtaining the NMR
data for establishing the structure. The purification of these compounds involved the use of
different chromatographic conditions than the campherenane derivative purification. Two

126

purification cycles had to be performed at two different pH conditions until pure compound
could be obtained.
Linear diterpene glycosides with mild deterrent effect on feeding of Manduca sexta larvae have
been reported from Nicotiana attenuata [3]. Similarly, an acyclic sesquiterpene glycoside is
reported from Sapindus trifolzatus [4] and a diglycosylated farnesol is reported from Guioa
crenulata [5]. The compound reported from Guioa crenulata has acyl groups on the sugar
moieties. Recently, four acyclic sesquiterpene glycosides with inhibitory effects on tumor
necrosis factor-α-induced cytotoxicity were discovered from pericarps of Sapindus rarak [6].
Three bicyclic sesquiterpene glycosides are reported from Elsholtzia bodinieri [7]. Two of the
three compounds reported herewith are bicyclic sesquiterpenes and the other is an acyclic
sesquiterpene glycoside. All compounds carry two sugar groups attached and in the acyclic
compound a malonate ester is identified. In contrast with the compounds with malonylation
observed so far, this compound has the malonylation on a hydroxyl group attached to the
terpenoid core of the molecule. In previously reported compounds, the malonylation or
acetylation is observed on sugar group.
4.2 Plant material for metabolite purification
Five Solanum habrochaites LA1777 plants were grown in a plant growth chamber (28 °C, 16
hour day time, 96% humidity, lighting 150 μmol-m-2-s-1 ), and leaf and stem tissues were
harvested and extracted in six week intervals. Leaf trichomes were extracted by dipping each
branch in 1 L of 100% methanol for about 30 s. After extraction, plant material was dried at 60
°C and dry mass was determined to be 183 g. Extracts were subsequently combined, and solvent
was evaporated under vacuum on a rotary evaporator without heating, and the residue was

127

redissolved in about 10 ml of 100% methanol. Next, the three compounds were purified from
this concentrated extract.
4.3 Purification of sesquiterpene I and sesquiterpene II glycosides
Semipreparative HPLC was performed on an Acclaim 120 C18 column (4.6 × 150 mm, 5 µm
particles; Dionex Co. USA) using a Waters 2795 semipreparative HPLC system. Fractions eluted
from the column were collected using a LKB BROMMA 221 Superrac fraction collector. For the
semipreparative HPLC, flow rate was 2.0 ml/min and the injection volume for each injection was
150 µL. Purification of these three compounds required employing two different gradients in
two chromatographic cycles. The first linear solvent gradient used for the purification employed
water with 0.15% formic acid of pH 2.90 (solvent A) and acetonitrile (solvent B) was as follows
(A:B): initial (99%:1%), 1.00 min (70%:30%), 31.00 min (53%:47%), 31.01min (0%:100%),
33.00 min (0%:100%), 33.01 min (99%:1%), 36.00 min (99%:1%). Twenty fractions were
collected (1-min intervals), and each fraction was tested for the presence of compound of interest
using a 5.50 min UHPLC-MS method in negative ion mode ESI using a Waters LCT Premier
TOF mass spectrometer. Instrument parameters were: capillary voltage of 3.0 kV, desolvation
temperature of 350 °C, source temperature of 100 °C, Cone gas (N2) at 40 L/hr and desolvation
gas (N2) at 350 L/hr. This gradient was (A = 0.15% aqueous formic acid; B = methanol (%A:
%B): 0.0-2.0 min (95:5), step at 2.01 min (40:60), linear increase to 3.00 min (0: 100), and held
at this condition until 5.0 min, then returned to initial condition at 5.01 min and held until 5.5
min. From these analyses, the fractions that contained the compound of interest were combined
and concentrated to about 5 ml volume using Speedvac without heating. The concentrated
sample was again chromatographed using the same gradient, however now water with 10 mM
ammonium formate pH 7.50 (solvent A) and acetonitrile (solvent B) were used as solvents.
128

Again 20 fractions were collected and the fractions were analyzed using the 5.5 min LC-MS
procedure above described on TOF mass spectrometer. From these analyses, the fractions that
were over 90% pure based on total ion chromatogram peak areas were identified and combined,
and solvent was removed by evaporating to dryness under a N2 gas flow. All NMR experiments
were performed as explained before [8]. Also the acid hydrolysis experiments and the chemical
derivatizations of the sugars obtained from the hydrolysis was performed on compounds purified
from peaks 8a, 3b and 6c of Fig. 4.1 as explained before [8].
4.4 Structure elucidation of sesquiterpene I diol dihexoside malonate ester
This compound is observed at m/z 661 (observed m/z 661.2693; theoretical m/z for [C30H45O16-]
661.2713; ∆m = -3.0 ppm ) as [M-H]-. The MS/MS product ion spectrum of this compound
displays a neutral loss of 44 Da (corresponding to CO2) followed by a neutral loss of 42 Da
(corresponding to C2H2O) giving fragment ions m/z 617 (observed m/z 617.2791; theoretical m/z
for [C29H45O14-] 661.2815; ∆m = -3.9 ppm) and 575 (observed m/z 575.2739; theoretical m/z for
[C27H43O13-] 575.2709; ∆m = -5.2 ppm ). This is followed by the two sequential neutral losses of
162 Da (corresponding to two anhydroglucoses) to give fragment ions m/z 413 (observed m/z
413.2166; theoretical m/z for [C21H33O8-]413.2181; ∆m = -3.6 ppm) and 251 (observed m/z
251.1650; theoretical m/z [C15H23O3-] 251.1653; ∆m = -1.2 ppm ). Based on accurate mass of the
putative terpenoid core elemental formula of C15H23O3- was assigned. To further characterize the
structure of this compound it was purified using semi-preparative HPLC followed by 1D and 2D
NMR spectroscopic analysis (Fig. 4.3 to 4.8). Structure shown in Fig. 4.2a was assigned.

About 8 isomers of this compound were separated from S. habrochaites LA1777 and one of the
major isomers (Fig. 4.1, peak 8a. Rt = 7.1 min) was purified and structure was elucidated using
129

1D and 2D NMR. Final structure shown in Fig. 4.2a was assigned to this molecule using NMR
and mass spectrometry information. In the proton NMR spectrum the anomeric protons
corresponding to the two glucose moieties could be identified at proton resonances δH 5.56 ppm
(J = 7.81 Hz), δH = 4.31 ppm (J = 7.80 Hz) as two doublets. This confirmed the presence of two
hexose moieties and the coupling constants of each anomeric proton suggest that the glycosides
are bound by beta-substitutions. Based on the correlations observed in the HMBC spectrum, it
was concluded that the two sugar groups are connected by a 1-4 glycosidic bond. Three methyl
groups were identified as singlets from proton NMR spectra and were confirmed by HSQC
spectra. This indicated that the methyl groups were attached to quaternary carbons. COSY and
HMBC spectra were used to identify the neighboring protons. Next, starting from the anomeric
protons, the two hexose moieties were assigned followed by the assignment of the structure of
terpenoid core. In the terpenoid core, the protons located on alkene groups and the two primary
alcohol groups were distinctive. The alkene protons were detected at δH 5.41 ppm and δH 6.93
ppm. The two primary alcohol positions hosts two protons each. (δH 4.22 and 4.33) at C12 and
(δH 3.99 and 3.99) ppm at C1. The correlations observed in the HMBC spectrum between the
C12 protons and the anomeric carbon (C16) of the sugar moiety aided assignment of the position
of the sugar moiety. Similarly the correlations of protons on C2 and carbonyl carbon at C28
enabled the identification of the position of the malonyl group. Furthermore, the correlations
observed between the protons of methyl groups and the alkene carbons enabled the correct
assignment of the positions of the methyl groups.

Based on the correlations observed in the NOE spectrum, the relative stereochemistry of methyl
groups was established. NOE spectrum showed correlations between the proton at C11 and

130

methyl group at C15 (attached to C10) indicating these two groups are located in close proximity
in space. Therefore these are located in a cis configuration at the C-C double bond between C10
and C11. Similar correlations were observed in the NOE spectrum between the proton at C8 and
the methyl group attached to C7 indicating these are also located in a cis configuration.
Additionally the two methyl groups (methyl group attached to C7 and C5 respectively) also
displayed correlations in the NOE spectrum. However, the proton at C4 and the methyl group at
C5 showed no NOE correlation, indicating that these two are located in a trans configuration at
the double bond between C4 and C5.
4.5 Structure elucidation of Sesquiterpene II alcohol dihexoside acetate ester
This compound is observed at m/z 633 (observed m/z 633.3115; theoretical m/z for [C29H48O12 +
HCOO-] 633.3128; ∆m = 2.0 ppm) as [M+ HCOO]-. The compound displays a neutral loss of 42
Da (corresponding to C2H2O) giving fragment ion m/z 545 (observed m/z 545.2923; theoretical
m/z for C27H45O11 545.2967; ∆m = 8.0 ppm) which is followed a neutral loss of 162 Da
(corresponding to an anhydroglucose) to give fragment ion m/z 383 (observed m/z 383.2429;
theoretical m/z for [C21H35O6-] 383.2439; ∆m = 2.6 ppm). To further characterize the structure of
this compound it was purified using semi-preparative HPLC followed by 1D and 2D NMR
spectroscopic analysis (Fig. 4.9 to 4.13). Proposed structure is shown in Fig. 4.2b.

Four isomers of this compound were separated from S. habrochaites LA1777 and one of the
major isomers (Fig. 4.1, peak 3b. Rt = 14.9 min) was purified and structure was elucidated using
1D and 2D NMR. Final structure shown in Fig. 4.2b was assigned to this molecule using NMR
and mass spectrometry information. In the proton NMR spectrum two anomeric protons
corresponding to the two glucose moieties could be identified at proton resonances δH 4.65 ppm
131

(J = 7.96 Hz), δH = 4.67 ppm (J = 7.69 Hz) as two doublets. This supported the presence of two
hexose moieties (in contrast with the LC-MS/MS spectra that showed one neutral loss of 162 Da
corresponding to only one anhydroglucose). In contrast to Sesquiterpene I diol dihexoside
malonate ester discussed above, correlations observed in the HMBC spectrum of Sesquiterpene
II alcohol dihexoside acetate ester indicated that the two sugar groups are attached through a 1-6
glycosidic bond. The coupling constants of each anomeric proton suggest that the glycosides are
bound by beta-substitutions. Four methyl groups were identified from the proton NMR spectrum
and were confirmed by HSQC. In the terpenoid core, the proton located on alkene group
(detected at δH 6.76 ppm) was distinctive. COSY and HMBC spectra were used to identify the
neighboring protons. The HMBC correlations between the alkene carbon and the tertiary alcohol
carbon were instrumental in establishing the structure of the terpenoid core. Next, starting from
the anomeric protons, the two hexose moieties were assigned. The structure of the terpenoid core
was assigned starting from the alkene proton at C2. Using the H-H correlations observed in the
COSY spectrum most of the terpenoid core could be established. (Fig. 4.2b) This assignment
was confirmed by the presence of distinctly identifiable proton spin systems in the TOCSY
spectrum, which revealed longer range H-H correlations. Therefore, this serves as supporting
evidence to confirm structural assignments made using COSY. In addition, the correlations
observed in HMBC supported the assigned terpenoid core structure.

Relative stereochemistry was assigned using NOE spectra. The methyl group and the proton
attached to C1 and C2 positions display correlations in NOE spectra indicating they are at a cis
configuration. One of the two protons at C6 also shows NOE correlations with methyl at C1 and
the proton at C5 show correlations with proton at C6. Indicating that these groups (protons at C6,

132

C5, C2 and methyl at C1) are in same plane. However, no NOE correlations were observed
between the protons at C5 and C4 and it was concluded that these are located on opposite planes.

To confirm the m/z of the terpenoid core, acid hydrolysis was performed on the purified
compound followed by GC-MS analysis. The GC-MS indicated a peak appearing to be a [MH2O]+•. The m/z of this fragment ion was 204 (observed m/z 204.1862; theoretical m/z for
[C14H20O-] 204.1878; ∆m = 7.8 ppm) (Fig. 4.20)
4.6 Structure elucidation of Sesquiterpene II alcohol dihexoside
This compound is observed at m/z 591 (observed m/z 591.3013; theoretical m/z for [C27H46O11 +
HCOO-] 591.3022; ∆m = 1.5 ppm) as [M+HCOO]-. The MS/MS spectra of the compound
displays [M-H]- ion at m/z 545 (observed m/z 545.2947; theoretical m/z 591.2967; ∆m = 3.7
ppm), which is followed by a neutral losses of 162 Da (corresponding to two anhydroglucoses)
to give fragment ions m/z 383. To further characterize the structure of this compound it was
purified using semi-preparative HPLC followed by 1D and 2D NMR spectroscopic analysis and
the structure was established (Fig. 4.14 to 4.19 and 4.2c).

Eight isomers of this compound were separated from S. habrochaites LA1777 and one of the
major isomers (Fig. 4.1, peak 6c. Rt = 10.7 min) was purified and structure was elucidated using
1D and 2D NMR. Final structure shown in Fig. 4.2c was assigned to this molecule using NMR
and mass spectrometry information. In the proton NMR spectra two anomeric protons
corresponding to the two glucose moieties could be identified at δH 4.66 ppm (doublet with J =
8.10 Hz) and δH = 4.63 ppm (triplet of J = 7.92 Hz). This supported the presence of two hexose
moieties. Four methyl groups were identified from proton NMR spectra based on integrals and
133

were confirmed by HSQC spectra. The proton at alkene group was detected at δH = 5.89 ppm
(triplet of J = 3.13 Hz). COSY and HMBC spectra were used to identify the protons and carbon
atoms neighboring the methyl groups and the alkene proton in the structure. Also the HMBC
correlations between the tertiary carbon hosting the –OH and the neighboring protons were
important in elucidating the structure. Starting from the anomeric protons, the two hexose
moieties were assigned. Similar to the sesquiterpene alcohol dihexoside acetate ester the
structure of the terpenoid core was assigned using COSY, TOCSY, HMBC,HSQC and NOE
spectra. Based on this NMR evidence, the structure shown in Fig. 4.2c was assigned to this
compound.

Similar to the sesquiterpene alcohol dihexoside acetate ester, to confirm the m/z of the terpenoid
core, acid hydrolysis was performed on the purified compound followed by GC-MS analysis.
The GC-MS indicated a peak at m/z 204 (observed m/z 204.1862; theoretical m/z for [C14H20O-]
204.1878; ∆m = 7.8 ppm) corresponding to [M-H2O]+• (Fig. 4.21). This observation further
supported the structure established using NMR.

134

4.7 Conclusion
Our efforts have resulted in the discovery of another two novel terpenoid cores that were not
reported from Solanum habrochaites before. While these terpenoid compounds possess structural
similarity to farnesol and naphthalenol, their presence in S. habrochaites LA1777 presents the
opportunity for discovering the genes involved in their synthesis. Furthermore, the GC-MS
profiling of volatile terpenoids from S. habrochaites reported in literature does not report of any
compound that similar in structure to these. This is an important implication in that the nonvolatile terpenoids in these plants presents very different structural characteristics from the
volatile terpenoids. The proposed structures indicate that these compounds possess a number of
chiral centers. Therefore, it can be expected that a number of stereoisomers of these compounds
would be present. Our metabolite profiling work reported here (Fig. 4.1) and chapter 2 of this
dissertation indicated that for all of these compounds multiple isomeric forms are detected. It can
be perceived that some of the isomers could be the result of stereochemical differences in the
terpenoid core.

135

APPENDIX

136

Figure 4.1: HPLC-MS extracted ion chromatogram profiles of the three compounds purified
from S. habrochaites LA1777. a.) 10 isomers of Sesquiterpene I diol dihexoside malonate ester
were separated (b) three isomers of [M+formate]- of Sesquiterpene II alcohol dihexoside acetate
ester were detected (c) 8 isomers of [M+formate]- of Sesquiterpene II alcohol dihexoside were
separated. 60

137

Figure 4.2. Structures of (a) Sesquiterpene I diol dihexoside malonate ester (peak 8a from Figure
4.1), (b) Sesquiterpene II alcohol dihexoside acetate ester (peak 3b) and (c) Sesquiterpene II
alcohol dihexoside (peak 6c), determined using NMR and tandem mass spectrometry. Portions of
the molecule corresponding to key fragment ions in negative ion MS/MS spectra are illustrated
on each structure. Carbon atoms are numbered in accordance with NMR assignments in Table
4.1. 61
138

Figure 4.3. 1D Proton NMR spectra of compound a (peak 8a from Figure 4.1). 62

139

Figure 4.4. 2D COSY spectrum of compound a (peak 8a from Figure 4.1).

Figure 4.4. 2D COSY spectrum of compound a (peak 8a from Figure 4.1). 63

140

Figure 4.5. 2D HSQC spectrum of compound a (peak 8a from Figure 4.1). 64

141

Figure 4.6. 2D HMBC spectrum of compound a (peak 8a from Figure 4.1).65

142

Figure 4.7. 2D TOCSY spectrum of compound a (peak 8a from Figure 4.1). 66

143

Figure 4.8. 2D NOESY spectrum of compound a (peak 8a from Figure 4.1).67

144

Figure 4.9. 1D Proton NMR spectrum of compound b (peak 3b from Figure 4.1). 68

145

Figure 4.10. 2D COSY spectrum of compound b (peak 3b from Figure 4.1). 69

146

Figure 4.11. 2D HSQC spectrum of compound b (peak 3b from Figure 4.1). 70

147

Figure 4.12. 2D HMBC spectrum of compound b (peak 3b from Figure 4.1). 71

148

Figure 4.13. 2D TOCSY spectrum of compound b (peak 3b from Figure 4.1). 72

149

Figure 4.14. 1D 1H NMR spectrum of compound c (peak 6c from Figure 4.1). 73

150

Figure 4.15. 2D 1H COSY spectrum of compound c (peak 6c from Figure 4.1). 74

151

Figure 4.16. 2D HSQC spectrum of compound c (peak 6c from Figure 4.1). 75

152

Figure 4.17. 2D HMBC spectrum of compound c (peak 6c from Figure 4.1). 76

153

Figure 4.18. 2D TOCSY spectrum of compound c (peak 6c from Figure 4.1). 77

154

Figure 4.19. 2D NOESY spectrum of compound c (peak 6c from Figure 4.1). 78
.

155

Figure 4.20. GC-MS chromatogram and EI spectrum of most abundant peak from the acid
hydrolysis product of compound b. 79

156

Figure 4.21. GC-MS chromatogram and EI spectrum of most abundant peak from the acid
hydrolysis product of compound c. 80

157

Figure 4.22. HMBC correlations assigned for compound a 81

Figure 4.23. Important COSY correlations assigned for compound a 82

Figure 4.24. NOE correlations assigned for compound a 83

158

Figure 4.25. HMBC correlations assigned for compound b 84

Figure 4.26. COSY correlations assigned for compound b 85

159

Figure 4.27. NOE correlations assigned for compound b 86

Figure 4.28. HMBC correlations assigned for compound c 87
160

Figure 4.29. COSY correlations assigned for compound c 88

Figure 4.30. NOE correlations assigned for compound c 89

161

Table 4.1: Chemical shift assignments to the carbons and protons of the three compounds
purified from S. habrochaites LA1777.

Compound a
Position
δC
δH
66.30 3.99 3.99
1
21.70 2.22 2.22
2
28.70 2.36 2.36
3
146.20 6.93
4
129.00
.
5
169.10
6
136.00
7
127.70 5.20
8
33.74 2.10 2.18
9
142.80
10
123.80 5.41
11
66.96 4.33 4.22
12
12.70 1.86
13
24.00 1.71
14
23.95 1.78
15
103.70 4.31
16
75.80 3.18
17
78.80 3.31
18
83.15 3.49
19
78.70 3.61
20
63.00 3.87 3.69
21
96.70 5.56
22
74.70 3.41
23
64.50 3.72
24
71.60 3.25
25
74.50 3.45
26
65.58 3.67 3.57
27
165.35
28
25.40 2.17 2.01
29
183.55
30

5

Compound b
δC
δH
173.30
144.20
6.76
36.10
1.82
51.7
1.80
48.60
1.90
24.90
1.69
82.90
27.50
1.72
28.65
2.41
19.91
1.84
50.30
1.55
20.25
1.30
25.90
1.28
19.64
1.63
23.50
1.25
98.10
4.67
83.10
3.54
78.90
3.60
72.50
3.26
78.50
3.38
63.50
3.88
106.40
4.65
76.90
3.28
78.70
3.42
88.30
3.26
68.00
4.44
65.50
3.86
178.40
23.60
1.11

162

2.34

1.77
2.77
1.74

3.66

3.69

Compound c
δC
δH
151.30
124.20
5.89
24.36
2.06
49.10
2.06
51.60
1.80
37.20
1.67
81.70
23.14
1.60
27.90
1.71
50.50
1.54
19.96
1.64
18.50
1.24
29.20
1.30
25.40
1.31
23.40
1.27
106.20
4.66
77.00
3.23
79.00
3.65
72.40
3.32
82.00
3.57
63.50
3.86
98.00
4.63
83.40
3.50
79.00
3.60
72.50
3.26
78.50
3.38
63.50
3.86

2.36

1.77
1.77
2.78

3.69

3.69

REFERENCES

163

REFERENCES
1. Jassbi AR, Zamanizadehnajari S, Kessler D, Baldwin IT (2006) A new acyclic diterpene glycoside
from Nicotiana attenuata with a mild deterrent effect on feeding Manduca sexta larvae. Zeitschrift Fur
Naturforschung Section B-a Journal of Chemical Sciences 61 (9):1138-1142
2. Kasai R, Nishi M, Mizutani K, Miyahara I, Moriya T, Miyahara K, Tanaka O (1988) Trifolioside-II ,
an acyclic sesquiterpene oligoglycoside from pericarps of Sapindus-trifoliatus. Phytochemistry 27
(7):2209-2211
3. Magid AA, Voutquenne-Nazabadioko L, Litaudon M, Lavaud C (2005) Acylated farnesyl diglycosides
from Guioa crenulata. Phytochemistry 66 (23):2714-2718. doi:10.1016/j.phytochem.2005.09.009
4. Morikawa T, Xie YY, Ninomiya K, Okamoto M, Muraoka O, Yuan D, Yoshikawa M, Hayakawa T
(2010) Inhibitory Effects of Acylated Acyclic Sesquiterpene Oligoglycosides from the Pericarps of
Sapindus rarak on Tumor Necrosis Factor-alpha-Induced Cytotoxicity. Chemical & Pharmaceutical
Bulletin 58 (9):1276-1280
5. Hu HB, Jian YF, Zheng XD, Cao H (2007) Three sesquiterpene glycosides from Elsholtzia bodinieri.
Bulletin of the Korean Chemical Society 28 (3):467-470
6. Ekanayaka EAP, Li C, Jones AD (2013) Sesquiterpenoid Glycosides from Glandular Trichomes of the
Wild Tomato Relative Solanum habrochaites. Phytochemistry

164

Chapter 5: Discovering terpene glycosides from Hoodia gordonii by mining multiplexed
CID UHPLC-MS data and applying relative mass defect filtering
5.1 Introduction
Hoodia gordonii is a plant considered to be rich in appetite suppressants and therefore is used in
inducing weight loss [1], though scientific evidence of its effectiveness has been lacking.
Extracts of Hoodia contain an assortment of diterpenoid glycosides known as Hoodigosides, and
it has been proposed that one of these, termed P57AS, is the main active ingredient [2]. The
appetite suppressant activity of P57AS has been attributed to its being able to increase the ATP
content in the hypothalamus neurons that regulate food intake of humans [2]. Similarly, another
appetite suppressant compound was isolated from Hoodia recently and was tested for its appetite
suppressing activity using mouse models [3]. In addition to the known Hoodigosides with
appetite suppressant activity, eleven biologically active (antioxidant and cytotoxic) oxypregnane
glycosides have been recently reported from Hoodia gordonii [4]. Furthermore, H. gordonii has
also been patented for its antidiabetic activity [5] and as a protectant against acetaminopheninduced gastrointestinal damage based on in vitro bioactivity assay [6].
Hoodia is rich in various diterpene glycoside compounds [7-9]. Five main terpenoid cores are
observed in Hoodia [7]. These are Hoodigogenin A/Gordonoside A, Calogenin, Isoramanone,
Hoodistanal and Dehydrohoodistanal (Fig. 5.1a-e) [7]. Diterpene glycosides found in Hoodia
consist of sugars that are not commonly found in glycosides. Six main types of sugar groups
found in diterpenoid glycosides from Hoodia are cymarose, oleandrose, digitoxose, thevetose, 3O-methyl-6-deoxyallose and glucose (Fig. 5.1f-k) [2]. These sugar groups get attached to the
terpenoid core to produce some of the most “decorated” terpene glycosides in nature (Fig. 5.2).

165

These findings suggest that Hoodia gordonii is a plant that possesses an assortment of
biologically important compounds, but the mechanisms responsible for their formation have not
been established. Therefore, investigating the chemical composition of this plant using a
systematic study that involves deep profiling of specialized metabolites of different tissues of
this plant is an important endeavor. Moreover, discovering the genes involved in regulation of
these compounds will aid the development of means for engineering plants or microbes for the
synthesis of these compounds at larger scale[10].
A systematic study was performed to analyze Hoodia stem tissues (primary step pith and spine
tissues) using UHPLC-multiplexed CID MS. Relative mass defect (RMD) filtering was applied
to the data generated from this analysis to aid recognition of terpenoid metabolites, including
low-abundance compounds that are potential metabolic intermediates. The same RMD filtering
method can be extended to analyzing metabolite data from Hoodia gordonii available online at
the Medicinal Plants Consortium database (available at http://metnetdb.org/PMR/). The RMD
filtering method correctly recognizes all 29 compounds that are annotated in database as terpene
glycosides. Among the 100 most abundant metabolites in this category, only 18 have been
annotated as known compounds. Using RMD filtering along with MS/MS analysis led to the
annotation of 24 additional novel steroidal glycoside compounds among these unknown
molecules, as described below.

166

5.2 Methods
5.2.1 LC-MS and LC-MS/MS experiments
All LC-MS experiments were performed using a Waters LCT Premier mass spectrometer using
the standard electrospray ion source. This instrument was coupled to a Shimadzu Prominence
Liquid Chromatography system (LC-20AD pumps with high pressure mixing). Separations were
performed using an Ascentis Express C-18 UHPLC column (2.1 mm × 100 mm; 2.7 µm fused
core particles, Supelco Co. USA). Aqueous 10 mM ammonium formate adjusted to pH 2.55 was
used as solvent A and 100% acetonitrile was used as solvent B for the separation of the
compounds. The LC gradient started with 1% B/ 99% A, held for one minute, increased to 30%
B at three minutes, then 70% B at 31 minutes, then 100% B at 31.01 minutes, held at 100 % B
until 33 minutes, decreased to 1% B at 33.01 minutes, held at 1% B. Total analysis time was 35
minutes. Total flow rate was 0.30 ml/min and column temperature was held at 50 °C. Mass
spectra were acquired using positive-ion mode electrospray ionization, centroid peak detection
and dynamic range extension over m/z 50-1400, with mass resolution (M/M, full width-half
maximum) approximately 10000. Five parallel collision energy functions were used, with 0.1 s
per function. Collision cell potentials used for negative ion mode fragmentation for each function
were 5, 20, 40, 60 and 80 V respectively.
All LC-MS/MS analyses were performed using a Waters Xevo G2-S QToF mass spectrometer
coupled to a Waters Acquity ultra-high performance liquid chromatography system. The LC
gradient used here was the same as the one used for LC-MS profiling (see above). Other
parameters include capillary voltage of 3.0 kV, desolvation temperature of 350 °C, source
temperature of 100 °C, cone gas (N2) at 0 L/hr and desolvation gas (N2) at 800 L/hr. In all
analyses, leucine enkephalin (0.1 ng/µL) was used as a lock mass reference infused at a flow rate
167

of 3.0 µL/min, and lock mass spectra were acquired for 100 ms at 10 sec intervals. Capillary
voltage for lock spray was 2.5 kV.
All MS/MS experiments were performed on the same instrument employing the same LC
method, and argon as collision gas. For positive ion mode MS/MS experiments a collision
potential +30 V was used.
5.2.2 Sample preparation for metabolite profiling
Aliquots (100 µL) of Hoodia gordonii extracts (chloroform: methanol 1:1 v/v) delivered to
Michigan State University from the laboratory of Professor Joe Chappell of the University of
Kentucky were transferred into 200 µL glass inserts, and the inserts were placed into
microcentrifuge tubes. Solvent was evaporated under vacuum at room temperature in a
SpeedVac, and residues were reconstituted in 100 µL of 20/80 v/v methanol/water. The inserts
were placed in 2 mL HPLC autosampler vials and briefly sonicated. The samples were spiked
with 5 µL of a 10 µM solution of telmisartan (achieving 0.5 µM final concentration). The HPLC
autosampler vials were labeled with MSU barcodes. The order of analysis was randomized using
Microsoft Excel.
5.3 Results and discussion
5.3.1 Hoodia metabolite profiling
Hoodia gordonii metabolite profiling was performed using the samples prepared above (5.2.2)
and using the LC-MS method described in 5.2.1. The data generated were subjected to
automated peak picking using Waters MarkerLynx peak picking algorithm. Peak picking
parameters used were as follows. Peak Width at 5% Height = 18 s, masses per retention time 15,
minimum intensity as a percentage of the base peak 0.02%, mass window 0.05 Da, retention time
168

window 0.4 min. Peak picking using these parameters resulted in a list of 1535 signals organized
as m/z – retention time pairs.
5.3.2 Data processing for and relative mass defect filtering for mining publically available
data sets for the discovery of novel terpenoid metabolites
The objective of this exercise was to mine the Hoodia gordonii LC-MS data for steroidal
glycosides (which may be diterpene glycosides) and therefore the three boundary conditions
discussed in Chapter 2 were re-established. As discussed in Chapter 1, the simplest form of
diterpene is geranylgeranyl. Therefore, rules can be established based on theoretical derivatives
of geranylgeranyl. 1). Based on theoretical mass of geranylgeranyl mono glycoside, the
minimum m/z of [M+H]+ of a diterpene monoglycoside would be 453.3211 and the maximum
RMD a diterpene glycoside can display is 708 ppm. 2). Similarly, the molecular mass of
Hoodigoside S molecule which reported from H. gordonii, has a diterpenoid core that is
extensively “decorated” with sugars and sugar derivatives was established as the maximum mass
a diterpene glycoside from this plant can exhibit. Therefore 1479.7730 was considered as the
upper limit of m/z for [M+H]+ and 522 ppm was established as the lower limit of relative mass
defect of diterpene glycosides from this plant. The distribution of RMD among compounds is
shown in Fig. 5.3. From this analysis it is clear that a majority of the compounds detected from
Hoodia consists of RMD in the range 500 – 700 ppm.

Therefore, for exploring diterpene glycosides from Hoodia gordonii, compounds of the RMD
range 500 – 710 ppm were selected as candidate compounds for diterpene discovery. Next, this
list of markers was organized in the descending order of peak area reported in the database and
higher collision energy fragments observed in the multiplexed CID MS data was examined for
169

each marker in this list to establish their terpenoid identities.

Examining higher energy functions of multiplexed CID MS data enabled annotation of those
compounds that generated fragment ions with increasing RMD compared to the precursor ion,
and these were selected as candidate terpenoids. Next, the MS/MS spectra were generated for
these compounds to confirm the fragment ions observed in multiplexed CID MS. Furthermore,
since a number of diterpene glycosides are known from H. gordonii, MS/MS spectra were
generated for some of these standards and the MS/MS spectra from the candidate terpenoid
compounds selected in the previous step were compared with these data.

The Medicinal Plants Consortium database annotates 29 compounds as steroidal glycosides all of
which display a RMD between 500 - 710 ppm. Among the 100 most abundant compounds with
RMD values in this range, only 16 compounds are known. RMD filtering and comparing the
MS/MS data for unknown compounds with those from standards led to the discovery of 21 novel
steroidal glycoside compounds that are putative derivatives of known compounds (based on the
presence of common fragment ions assigned based on the accurate mass) and five alkaloid
compounds among the 100 most abundant markers with RMD range 500 – 710 ppm. The
abundance of these novel compounds and the compounds annotated in the database are shown in
Fig. 5.4.

170

5.3.3 Variation of RMD of terpene glycosides and other non –terpenoid compounds.
In addition to all the terpene glycosides that appeared in the RMD range of 500 – 710 ppm, the
second most abundant compound in this list (m/z 983.5523) appeared to be an alkaloid. The
MS/MS spectrum (Fig. 5.17 a) of this compound revealed a number of even mass fragment ions
(m/z 520.3010, 492.3056, 466.2877, 310.1608, 282.1655) that suggested this compound
contained an odd number of nitrogen atoms, suggesting a putative alkaloid.

Alkaloids can be distinguished from terpenoids based on the presence of even mass fragment
ions (indicative of nitrogen) in the MS/MS spectra of alkaloids in positive ion mode. In addition,
product ions generated from alkaloids also display a different variation in RMD compared to
those generated from terpenoids (Fig. 5.17 b). For diterpene glycosides the characteristic gradual
increase of RMD with decreasing mass of fragment ions can be observed. The decreasing RMD
of the fragment ions was observable among fragment ions of this compound. Minimum of two
isomeric forms were separated and detected for this compound. Based on accurate mass
information, a glycosylated form of this compound also was assigned from the metabolite profile
data.
5.3.4 Putative assignments of novel steroidal glycosides from Hoodia gordonii
A number of terpenoid glycosides that are potentially modified structures of the known
compounds were assigned based on accurate mass information and the observation of product
ions in MS/MS spectra that are common to standards. These compounds were wither Calogenin
or Hoodigogenin A derivatives.
Anhydrocalogenin triglycoside: This compound was detected as [M+H]+ at m/z 803.4136
(theoretical m/z 803.4060 with elemental formula C39H63O17 for [M+H]- ; ∆m = 9.5 ppm). The
171

MS/MS spectrum of this compound (Fig. 5.5) showed the presence of m/z 487.1651 representing
anhydro triglucose (theoretical m/z 487.1657 with elemental formula C18H31O15 for [M-H2O+H];∆m = 1.2 ppm). m/z 641.3565, 479.3042 and 299.2383 that represent diglycosylated, mono
glycosylated and aglycone form of anhydrocalogenin core is formed as fragment ions after the
sequential loss of the sugar moieties. (theoretical m/z 299.2369 for calogenin core; ∆m = 4.6
ppm, 479.3042 and 641.3532 for calogenin mono and diglycoside ; ∆m = 8.1 and 5.1 ppm
respectively). Six isomeric forms of this compound were separated chromatographically.
Fragment ions of similar m/z was observed in MS/MS spectra of Hoodigoside L, Hoodigoside M,
Hoodigoside R, Hoodigoside P, Hoodigoside Q, and Hoodigoside H, all of which are calogenin
derivatives, further supporting that this compound represents anhydrocalogenin triglycoside
(Figs 5.18 and 5.6). A related compound (a hydrated form representing of calogenin triglycoside)
was observed as [M+NH4]+ at m/z 838 (experimental m/z 838.4514; theoretical 838.4431 for
elemental formula C39H64O18 + NH4+; ∆m = 9.8 ppm) that corresponded to calogenin
triglycoside. This compound also gave m/z 479, 461, 325 and 299 as fragment ions in MS/MS
spectra (Fig. 5.5) that were also observed in anhydrocalogenin triglycoside. Similarly a
compound that was 162 Da less in mass than anhydrocalogenin triglycoside was observed at m/z
641 [M+H]+ (experimental m/z 641.3590; theoretical 641.3532 for elemental formula C33H53O12
+ H+; ∆m = 9.0 ppm). The MS/MS spectrum of this compound (Fig. 5.6) also displayed the
presence of m/z 299. However, no m/z 325 was observed in this spectrum indicating that the two
glucose moieties exist separately in the molecule instead of as a disaccharide.
A glycosylated form of Hoodigoside M was detected as a [M+NH4]+ at m/z 1304 (experimental
m/z 1304.6605; theoretical 1304.6481 for elemental formula C59H98O30 + NH4+; ∆m = 9.5 ppm).
This compound displayed the presence of two isomeric forms that were chromatographically

172

resolved. The MS/MS spectra of this compound and Hoodigoside M standard (Fig. 5.8 a,b, and c
respectively) displayed fragments of m/z 641.3552 corresponding to protonated form of
diglycosylated anhhydrocalogenin (theoretical m/z for elemental formula of C33H53O12+ is
641.3532; ∆m = 3.1 ppm). However, in isomeric form 1, a fragment ion of m/z 485
(experimental m/z 485.2275;theoretical m/z for C20H37O13+ 485.2229) is observed. This
corresponds to the trihexose group formed by the attachment of an additional glucose group to
the dihexoside formed by D-thevetose and D-oleandrose of Hoodigoside M. In isomeric form 2,
m/z 485 is not observed and instead m/z 649 (experimental m/z 649.2223, theoretical m/z
649.2186; ∆m = 5.7 ppm) occurs. This fragment corresponds to a tetraglycoside group indicating
that the additional sugar group is attached to the triglucoside of Hoodigoside M. Additionally,
the fragments corresponding to triglycoside (m/z 487) and diglycoside (m/z 325) groups are also
observed indicating the presence of the sugar groups in both isomeric forms. Fragments
corresponding to the calogenin terpenoid core m/z 317 and 299 are observed in both the
Hoodigoside M standard and in both isomeric forms.

Another calogenin derivative was observed as [M+NH4]+ at m/z 1190 (experimental m/z
1190.6805; theoretical 1190.6681 for elemental formula C59H96O23 + NH4+; ∆m = 10.4 ppm).
The MS/MS spectrum of this compound is shown in Figure 5.7. This compound yields a product
ion at m/z 479, which corresponds to the glycosylated calogenin as a prominent fragment ion. In
addition it displays m/z 317 and 299 (theoretical m/z 317.2475; ∆m = 2.2 ppm and m/z
299.2369; ∆m = 6.3 ppm) as fragment ions that represent the presence of the calogenin core. The
base peak observed in MS/MS of this compound m/z 227 (experimental m/z 227.1296;
Theoretical m/z 227.1278; ∆m = 7.9 ppm) corresponds to an anhydrocymarose with the aliphatic

173

2-methylbutenoate chain attached to it.

Calogenin forms another abundant compound detected as [M+NH4]+ at m/z 920 (experimental
m/z 920.4928; theoretical 920.4850 for elemental formula C44H70O19 + NH4+; ∆m = 8.4 ppm).
The MS/MS spectrum of this compound (Figure 5.10) indicates the presence of a disaccharide
group (detected as fragment ion with m/z 325.1157 (Theoretical m/z 325.1129; ∆m = 7.4 ppm)
and the glycosylated calogenin core detected as m/z 461. Similar compound is observed at m/z
998 as a [M+NH4]+ (experimental m/z 998.5272). The MS/MS of this compound (Figure 5.11)
also displays fragment ions corresponding to a disaccharide group (detected as fragment ion with
m/z 325.1105 (Theoretical m/z 325.1129; ∆m = -7.0 ppm) and glycosylated calogenin core
(detected as fragment ion with m/z 461.2925 (theoretical m/z 461.2898; ∆m = 5.5 ppm)
indicating it is a calogenin derivative. Other common calogenin fragment ions m/z 299 and 281
(Experimental m/z 299.2365 and 281.2285 respectively; theoretical m/z 299.2369 and 281.2264;
∆m = 1.3 ppm and 7.4 ppm respectively) are observed further supporting this conclusion.

Hoodigogenin A derivatives: Aula et al proposed a fragmentation pathway for P57 that explains
the formation fragment ions m/z 457, 313 and 295 [1] . These same fragment ions are observed
in the MS/MS fragmentation of 1316 (Fig. 5.12), 1206 (Fig. 5.13), 1222 (Fig. 5.14), 1192 (Fig.
5.15) and 916 (Fig. 5.16), indicating these compounds share a similar terpenoid core as P57
(Hoodigogenin A) and glycosylation. Based on accurate mass it can be inferred that m/z 1316
represents a [M+NH4]+ form of a P57 modified with one additional oleandrose and two
digitoxose moieties. All these compounds form fragment ions m/z 457 (experimental m/z
457.2986; theoretical 457.2949 for elemental formula C28H41O5+; ∆m = 8.0 ppm), 313

174

(experimental m/z 313.2177; theoretical 313.2162 for elemental formula C21H29O2+; ∆m = 4.8
ppm) and 295 (experimental m/z 295.2074; theoretical 295.2056 for elemental formula
C21H27O+; ∆m = 6.1 ppm).
5.4 Conclusions
Our investigation has disclosed of the presence of an assortment of novel diterpene glycosidic
derivatives of known compounds, their isomers and novel alkaloids. This is the first report to our
knowledge that reports of alkaloids from Hoodia gordonii. This is an important finding since a
number of alkaloids are associated with manipulation of neuronal activity [7]. Therefore,
potentially these alkaloids also might be responsible for the biological effects of Hoodia extracts.
Further investigation is required for establishing the structure of these compounds and their
biological activity.

Applying RMD filtering to metabolite profiles of Hoodia has resulted in the discovery of a
number of novel terpene glycosides. It is likely that some of these are biosynthetic intermediates
of the diterpene glycosides observed in Hoodia. For instance the calogenin derivatives such as
anhydrocalogenin tri glycoside and monoglycoside can be expected to be biosynthetic
intermediates involved in the synthesis of larger calogenin compounds. Therefore, these
compounds may hold the key to understanding the diterpene biosynthesis in Hoodia.

The discovery of a number of novel diterpene glycosides is an indication that the chemical
diversity of terpenoid glycosides in Hoodia is even greater than currently reported. The multiple
types of glycosylations and the presence of a number of isomeric forms of a given compound add
yet further complexity to the chemical diversity of Hoodia.
175

APPENDIX

176

Figure 5.1: Five diterpenoid cores reported from Hoodia compounds (a-e) and the six different
sugar groups that are found attached to the terpenoid core in current literature. 90

177

Figure 5.2: Some representative diterpenoids found in Hoodia. 91

178

Number of markers detected

1200
1000
800

600
400
200
0
0-100

100-200

200-400

400-500
500-700
RMD (ppm)

700-900

900 +

Figure 5.3: Distribution of RMD of ion detected in Hoodia gordonii from metabolite profiling in
different RMD range categories. 92

179

Hoodigoside H and I Isomers_1186_19.31
Hoodistanaloside B_914_10.53
Hoodigoside T isomer 2_1334_22.11
Hoodigoside T isomer 1_1334_21.02
Hoodigoside T derivative
Hoodigoside R isomer 2_1352_15.52
Hoodigoside R isomer 1_1352_14.37
Hoodigoside Q_1368_10.97
Hoodigoside Q derivative
Hoodigoside P isomer 2_1208_13.44
Hoodigoside P isomer 1_1208_12.53
Hoodigoside O_1062_11.2
Hoodigoside O derivative
Hoodigoside M_1142_4.51
Hoodigoside M + Glu_1304_4.19
Hoodigoside L isomer 2_1224_10.08
Hoodigoside L isomer 1_1224_9.28
Hoodigoside J_1172_17.34
Hoodigoside H and Hoodigoside I_1186_19.38
Hoodigoside H and Hoodigoside I_1186_19.39
Hoodigoside F and Hoodigoside G_1202_15.72
Hoodigoside C or W, Gordonside B or…
Hoodigogenin A derivaive_1316_20.42
Hoodigogenin A derivaive_1316_19.7
Hoodigogenin A derivaive_1222_15.12
Hoodigogenin A derivaive_1206_18.64
Hoodigogenin A derivaive_1192_16.59
Gordonoside H_1058_13.52
Calogenin triglycoside + NH4+_838_4.5
Calogenin triglycoside + NH4+_838_9.28
Calogenin triglycoside – H2O_803_18.09
Calogenin triglycoside – H2O_803_15.49
Calogenin triglycoside – H2O_803_13.43
Calogenin triglycoside – H2O_803_9.28
Calogenin triglycoside – H2O_803_4.51
Calogenin diglycoside_641_11.2
Calogenin diglycoside_641_4.98
Calogenin diglycoside + Digitoxose_982_9.28
Calogenin diglycoside + Cymarose_982_4.41
Calogenin derivative_920_9.38
Calogenin derivative_1190_18.11
Alkaloid 1 + Glu_1145
Alkaloid 1_492
Alkaloid 1_983
Alkaloid 1_983_4.07
Alkaloid 1_492_4.05
Alkaloid 1 + Glu Isomer 2_1145_4.35
0

Novel Alkaloid metabolites
Novel diterpene glycosides
Known Metabolites annotated

Peak area
50000

100000

150000

Figure 5.4: Abundance of known and novel metabolites annotated from Hoodia gordonii 93
180

Figure. 5.5 Positive ion MS/MS product ion spectrum of Hoodia metabolite m/z 803 [M+H]+ 94
351 ppm

634 ppm

Figure 5.6 Positive ion MS/MS product ion spectrum of Hoodia metabolite m/z 838 [M+NH4]+
95

181

794 ppm

643 ppm

384 ppm

Figure 5.7 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 641. 96

182

360 ppm

333 ppm
798 ppm
342 ppm

354 ppm

787 ppm
634 ppm

561 ppm
577 ppm

Figure 5.8 Positive ion product ion MS/MS spectrum of Hoodia metabolite of m/z 1304. 97
183

567 ppm
545 ppm

811 ppm
638 ppm

544 ppm

Figure 5.9 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 1190. 98

RMD = 799 ppm
RMD = 805 ppm
RMD = 371
ppm

RMD = 356 ppm

RMD = 323
ppm
RMD = 634 ppm

Figure 5.10 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 920 99

184

340 ppm
790 ppm

812 ppm

634 ppm

Figure 5.11 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 998 100

498 ppm

407 ppm

713 ppm 649 ppm
699 ppm
644 ppm

530 ppm

Figure. 5.12 Positive ion MS/MS product ion spectrum of m/z 1316. 101

185

698 ppm

686 ppm

586 ppm
628 ppm

582 ppm

661 ppm
586 ppm

Figure. 5.13 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 1206 102

697 ppm
690 ppm
585 ppm

621 ppm
645 ppm

575 ppm
590 ppm

Figure. 5.14 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 1222 103

186

697 ppm
691 ppm
587 ppm
411 ppm

650 ppm
618 ppm

595 ppm

601 ppm

595 ppm

Figure. 5.15. Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 1192 104

723 ppm

703 ppm
442 ppm

653 ppm

Figure. 5.16 Positive ion MS/MS product ion spectrum of Hoodia metabolite of m/z 916 105

187

a.
585 ppm

669 ppm

620 ppm

517 ppm
577 ppm

577 ppm

Relative mass defect (PPM)

b.
900
800
700
600
500
400
300

100
Hoodigoside L

600

1100
m/z
Hoodigoside Q

1600
Alkaloid

Figure. 5.17 Positive ion MS/MS product ion spectrum of (a) m/z 983 (putative alkaloid), (b)
Variation of RMD of parent ion and product ions in the MS/MS spectrum of an alkaloid
compared to the diterpene glycoside (RMD of sugar fragments are not included). The diterpene
glycoside displays a gradual increase of RMD from the parent ion to fragment ions whereas the
alkaloid displays a constant RMD from both fragment ions and parent ions.106

188

347 ppm

492 ppm
792 ppm

626
ppm

626 ppm
545 ppm

527 ppm
779 ppm
529 ppm

345 ppm

791 ppm
347 ppm
625 ppm

347 ppm
792 ppm

492 ppm
626 ppm

547 ppm 500 ppm

440 ppm
726 ppm
672 ppm

551 ppm

Figure. 5.18 Positive ion MS/MS product ion spectra of Hoodia standards. 107
189

REFERENCES

190

REFERENCES
1. Avula, B., et al., Identification and structural characterization of steroidal glycosides in
Hoodia gordonii by ion-trap tandem mass spectrometry and liquid chromatography coupled
with electrospray ionization time-of-flight mass spectrometry. Rapid Commun Mass
Spectrom, 2008. 22(16): p. 2587-96.
2. van Heerden, F.R., Hoodia gordonii: A natural appetite suppressant. Journal of
Ethnopharmacology, 2008. 119(3): p. 434-437.
3. van Heerden, F.R., et al., An appetite suppressant from Hoodia species. Phytochemistry, 2007.
68(20): p. 2545-2553.
4. Pawar, R.S., et al., New oxypregnane glycosides from appetite suppressant herbal supplement
Hoodia gordonii. Steroids, 2007. 72(6–7): p. 524-534.
5. Bindra, J., M. Cawthorne, and I. Rubin, Extracts, compounds & pharmaceutical compositions
having anti-diabetic activity and their use, 2002, Google Patents.
6. Hakkinen, J., R.M. Horak, and V. Maharaj, Steroidal glycosides or plant extracts for treatment
of gastric acid secretion damage. 2001: patent; EP1099444.
7. Vermaak, I., J.H. Hamman, and A.M. Viljoen, Hoodia gordonii: An Up-to-Date Review of a
Commercially Important Anti-Obesity Plant. Planta Medica, 2011. 77(11): p. 1149-1160.
8. Shukla, Y.J., et al., Pregnane glycosides from Hoodia gordonii. Phytochemistry, 2009. 70(5):
p. 675-683.
9. Dall’Acqua, S. and G. Innocenti, Steroidal glycosides from Hoodia gordonii. Steroids, 2007.
72(6–7): p. 559-568.
10. Schilmiller, A.L., R.L. Last, and E. Pichersky, Harnessing plant trichome biochemistry for
the production of useful compounds. Plant Journal, 2008. 54(4): p. 702-711.

191

Chapter 6: Concluding remarks
Increasing world population, climate change, depletion of non-renewable energy sources such as
fossil fuels has made discovering novel sources of chemical feed stocks an important endeavor
for the survival of mankind. This requirement has resulted in increased dependence on plants as a
food and energy source of humans. The field of plant sciences has approached this challenge by
genetically modifying plants to enhance production of food and renewable substances. This has
prompted the requirement for discovering the genes involved in plant specialized metabolism in
order to understand their functionality and to manipulate their behavior. Plant metabolomics has
contributed to this by discovering the compounds that are important for humans from plants and
providing information of their structure, quantity and location.
Research discussed in Chapter 2 of this dissertation resulted in the development of a novel
strategy for the discovery of glycosylated terpenoids, a class of compounds for which annotation
has posed a significant challenge for many years owing to their low levels of abundance and
limitations of methods for their annotation in mixtures. We addressed this challenge by applying
Relative Mass Defect (RMD) filtering which allowed for separating a substantial amount of
unrelated compounds from glycosylated terpenoids. RMD filtering is applicable for the
discovery of any kind of conjugated terpenoid as demonstrated in the application of the method
for the discovery of sesquiterpene glycosides and diterpene glycosides (Chapter 5).
However, the discovery of these compounds has opened up a number of novel research
challenges. First, what are the structural features that yield multiple isomers of each group of
metabolites? RMD filtering allows for putative annotation of glycosylated terpenoids, but does
not provide the final word regarding metabolite structure. The presence of numerous isomers of
so many terpenoid glycosides presents a substantial challenge in purifying sufficient quantities
192

for NMR analysis. This issue has been addressed to some extent by purifying and elucidating the
structures of several terpenoid metabolites (Chapter 3 and 4), but absolute configurations of
some chiral centers remain to be determined. Owing to the lack of plant material, the compounds
discovered from Hoodia gordonii (Chapter 5) are annotated only based on mass spectrometric
evidence, and elucidating their structures using NMR still remains to be performed.
All the profiling and RMD filtering work reported here has resulted in the identification of a
number of isomeric forms for each compound. Therefore an important question that needs
answering is what are the structural differences among these isomers? Some of the isomerism
may arise from the differences in the carbohydrate linkages, and these are often not apparent
from mass spectrometry data alone. Also, the positions of substitution of additional groups such
as malonyl or acetate esters may be different from one isomer to another. To establish these
differences, more isomeric forms of each compound need to be purified and their structures
elucidated using NMR and x-ray crystallography.
A third challenging questions is, do different accessions of the same plant species accumulate the
same compounds or different compounds? For instance in the work presented in this dissertation,
only Solanum habrochaites LA1777 leaf trichomes have been discussed. Preliminary data (not
presented here) in metabolite profiling from other plant accessions of tomato in addition to
LA1777 (These were LA1624, LA2975, LA1362, LA2107, LA1392, LA2098, LA2409,
LA2101 LA2144, LA1223, LA1927 LA2106 and LA2204) has provided evidence that the
isomer profile of a given compound (e.g. campherenane-2,12-diol diglycoside malonate ester) is
different between different accessions. These data need further exploration to develop
comparisons between different accessions and to explore these nonvolatile conjugated terpenoids
further.
193

In summary our research presented here has opened up a number of opportunities for future
scientists to explore plant terpenoid chemistry that has not been known for many decades and
this can potentially result in the discovery of future drugs, chemical feed stock for industries,
methods to develop better crops, food additives, and flavor compounds .

194