GENETIC AND GENOMICS APPROACHES TO UNDERSTANDING THE
BIOSYNTHESIS OF SPECIALIZED METABOLITES IN TRICHOMES OF THE
CULTIVATED TOMATO AND ITS WILD RELATIVES
By

Jeongwoon Kim

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Plant Biology
2011

ABSTRACT

GENETIC AND GENOMICS APPROACHES TO UNDERSTANDING THE
BIOSYNTHESIS OF SPECIALIZED METABOLITES IN TRICHOMES OF THE
CULTIVATED TOMATO AND ITS WILD RELATIVES
By

Jeongwoon Kim

Trichomes are specialized epidermal appendages that cover the surface of plant tissues.
Glandular secreting trichomes produce a variety of plant specialized metabolites.
Trichome metabolites across the kingdom Plantae are extremely diverse and include
those important for plant defense and therapeutic purposes. Due to these benefits,
biosynthetic pathways leading to the production of trichome metabolites haven been
intensively studied. Recent advances and development of analytical chemistry and
genetic and genomic tools facilitate the study of trichome biochemistry. In this study,
analytical chemistry, genetic and genomics approaches were used to understand
biosynthetic pathways for the production of non-volatile metabolites in Solanum
trichomes. Specifically, this study included three individual projects: (i) identification of
tomato EMS mutants altered in biosynthetically diverse trichome non-volatile
metabolites and phenotypic characterization of glycosylated flavonoid mutant (ii)
identification and analysis of Solanum lycopersicum genetic variants altered in trichome
methylated myricetin biochemistry, and (iii) chemical analysis of diverse trichome
acylsugar profiles in 80 accessions from the wild species of cultivated tomato S.

habrochaites.

ACKNOWLEDGEMENTS

First of all I would like to thank Dr. Robert Last for his consistent support,
encouragement and patience, which helped me explore and enjoy working on science
despite great challenges from tough course work, exams, unexpected experimental results
and endless learning during my graduate school years.

I would also like to thank my committee members Drs. Dan Jones, Gregg Howe and
Dean DellaPenna for their constructive and supportive advice. Especially, I appreciate
Dr. Dan Jones for teaching me mass spectrometry performance and data analysis.

In addition, it was wonderful to work with Last lab members and tomato trichome project
members. Scientific discussion and everyday conversation were among the most pleasant
parts of time in graduate school. I especially thank Dr. Tony Schilmiller and Ms. Amanda
Charbonneau for their suggestion and help. I would also like to thank my friends at MSU
for being great companies sharing the sense of achievement and concerns.

Last, I would like to thank my family, my parents and a brother, for their love and
supports and being there. I would have not been able to make any achievement in my life
without them. Thank you and love you.

iv

TABLE OF CONTENTS

List of Tables ·······································································vii
List of Figures ·····································································viii
Chapter 1 Introduction: understanding plant specialized metabolites using
the Solanum trichome as a model system ········································ 1
Plant specialized (secondary) metabolites ········································1
Examples of the biosynthesis of plant specialized metabolites and their values ··2
Examination of metabolite profiles - metabolite profiling and metabolomics ···16
Three major steps to understand plant specialized metabolite profiles ··········19
Solanum trichomes as a model system to study plant specialized metabolites ·· 24
References ······································································28
Chapter 2 Chemical screening of tomato EMS mutants to identify altered
trichome non-volatile metabolites ················································37
Abstract ········································································37
Introduction ····································································38
Results ·········································································41
Discussion ·····································································68
Materials and methods ·························································72
References ·····································································76
Chapter 3 Analysis of genetic variants of Solanum lycopersicum with
altered trichome methylated flavonoids ·········································82
Abstract ········································································82
Introduction ····································································83
Results ·········································································86
Discussion ····································································119
Materials and methods ························································123
Acknowledgement ····························································129
v

References ····································································130

Chapter 4 Mass spectrometry reveals strong diversity in glandular trichome
acylsugars of the wild tomato Solanum habrochaites························135
Abstract ·······································································135
Introduction ···································································136
Results ········································································142
Discussion ····································································175
Materials and methods ························································183
References ····································································189

vi

LIST OF TABLES

Table 2.1. The list of 35 analytical signals tested for targeted data analysis ············48

Table 2.2. Summary of mutant screen results ··········································51

Table 2.3. Chemical phenotypes of the 24 mutants identified from mutant screening ··52

Table 3.1. The list of genetic markers generated for JP117 genetic mapping ···········98

Table 3.2. The list of co-dominant amplified polymorphic sequence (CAPS) markers
used to map MOMTs on chromosome 6 ·······································128
Table 4.1. The location of original collection site for S. habrochaites accessions ·····143

Table 4.2. The list of analytical signals characterized as acylsugars in LC-TOF-MS
analysis ·······································································153

vii

LIST OF FIGURES

Figure 1.1. Structural similarity between the primary metabolite proline and specialized
metabolite pipecolic acid ··························································3
Figure 1.2. Biosynthetic pathways for flavonoid production ····························6

Figure 1.3. Examples of flavonoid and anthocyanin compounds ························7

Figure 1.4. Chemical structures of acylsugars ··········································11

Figure 1.5. Overview of terpene biosynthesis ··········································13

Figure 1.6. Examples of terpene compounds ···········································15

Figure 1.7. Biosynthetic pathways for alkaloid production ·····························17

Figure 1.8. Examples of alkaloid compounds ··········································18

Figure 1.9. Seven types of Solanum trichomes ·········································25

viii

Figure 2.1. Structure of the mutant population used in this study ······················42

Figure 2.2. EMS mutant screening workflow ··········································45

Figure 2.3. Targeted data analysis identified an M2 putative mutant JP314 as having
increased kaempferol 3-O-rutinoside ··········································54
Figure 2.4. Non-targeted data analysis could not identify M2 putative mutant not even
JP314 ·········································································55
Figure 2.5. Targeted data analysis of JP314 M3 progeny identified chemical phenotypes
not seen in M2 screening ······················································58
Figure 2.6. Non-targeted data analysis of JP314 M3 progeny validated the flavonol
glycoside (m/z 949.3) phenotype and identified unannotated analytical signals
highly correlated with the m/z 949.3 ··········································60
Figure 2.7. LC-ToF-MS with MUX-CID analysis suggested that the flavonol glycoside
with m/z 949.3 is myricetin derivative and revealed adduct ions of this myricetin
glycoside ······································································63
Figure 2.8. Pale leaf color reveals reduced total anthocyanin accumulation in JP314 ··66

Figure 2.9. Overview of flavonoid and anthocyanin biosynthetic pathways ···········71

Figure 3.1. Increased mono-Me-M and decreased di- and tri-Me-M accumulation in
JP117 EMS mutant identified from LC-ToF-MS analysis of single leaf dip
extracts ········································································87
Figure 3.2. The chemical structure of unmodified myricetin ··························89
ix

Figure 3.3. The Me-M levels measured in isolated glands of types 1 and 6 trichomes
from JP117 mutant and M82 plants ···········································90
Figure 3.4. Evidence that a single recessive mutation is responsible for the JP117 mutant
phenotype ·····································································92
Figure 3.5. The JP117 mutation maps to the bottom of chromosome 6 ················95

Figure 3.6. Higher accumulation of 3Me-M in IL6-3/4 plants and a genetic map of
chromosome 6 ILs ·····························································99
Figure 3.7. The Me-M levels measured in isolated glands of types 1 and 6 trichomes from
IL6-3/4 and M82 plants ·······················································101
Figure 3.8. Chemical profiles of IL6-4 genetic cross plants suggest that the S. pennellii
allele is dominant for IL6-4 chemical phenotypes and that the phenotypes
segregate genetically ··························································103
Figure 3.9. Summary of the chemical phenotypes and chromosomal locations of the
JP117 mutant and the IL6-4 region ···········································105
Figure 3.10. Search for candidate O-methyltransferases (OMTs) based on sequence
homology ····································································107
Figure 3.11. Predicted gene structures of slMOMT1a and 1b and the corresponding S.
pennellii homologs ··························································109
Figure 3.12. Evolutionary relationships among various plant OMTs ················110

x

Figure 3.13. Field experiments revealed altered Manduca sexta feeding on S.
lycopersicum variants ······················································113
Figure 3.14. Greenhouse analysis of M. sexta feeding ······························115

Figure 3.15. Field experiments reveal genotype-dependent differential Epitrix cucumeris
damage ·····································································118
Figure 4.1. Light microscope images for leaf trichomes and whole leaf from three S.
habrochaites accessions ····················································138
Figure 4.2. Structural classes of acylsugars in Solanum species ····················140

Figure 4.3. Acylsugar analysis workflow ···········································149

Figure 4.4. An example of the use of LC-TOF with multiplexed collision-induced
dissociation to annotate acylsugar S4:26 (5,5,5,11) ·························150
Figure 4.5. Extracted ion chromatograms of three accessions illustrating chemical
diversity in acylsugar profiles ···············································156
Figure 4.6. Hierarchical clustering analysis (HCA) of acylsugar chemistry in Solanum
accessions ··································································160
Figure 4.7. Types of glucose and sucrose esters in Solanum accessions ············165

Figure 4.8. The composition of fatty acids in trichome acylsugars ·················167

xi

Figure 4.9. Test of reproducibility in acylsugar quantity and composition in eight M82
replicate plants ·····························································169
Figure 4.10. Normalized total acylsugar levels in S. habrochaites accessions ······173

Figure 4.11. Total acylsugar levels in S. habrochaites accessions without additional
normalization for batch-to-batch comparison ·······························174
Figure 4.12. Correlation between geographic distribution and acylsugar chemistry in S.
habrochaites accessions ····················································177
Figure 4.13. Hypothetical biosynthetic pathways to produce a subset of major acylsugars
in LA1777 ··································································181

xii

Chapter 1. Introduction: understanding plant specialized metabolites using the
Solanum trichome as a model system

Plant specialized (secondary) metabolites
A variety of low-weight molecules are produced as intermediates or final products from
chemical reactions involved in plant metabolism. These compounds are commonly
categorized to two groups, primary and secondary metabolites. Primary metabolites are
defined as chemical compounds that are directly involved in plant growth and
development (Aharoni and Galili 2011, Pichersky and Gang 2000, Schwab 2003). In
contrast to primary metabolites, the term secondary metabolite refers to natural products
that are traditionally thought to indirectly affect biological processes in plants (Aharoni
and Galili 2011, Schwab 2003). To date, more than 200,000 plant secondary metabolites
have been reported (Aharoni and Galili 2011). This term 'secondary metabolite' came
from the traditional concept that these small molecules are not essential for plant survival.
However, increasing numbers of studies revealed that these chemical compounds have
critical impacts on fitness, for example, protecting plants against diverse environmental
stresses (Baldwin et al. 2006, Kang et al. 2010a, Kang et al. 2010b) and promoting
symbiotic interaction with beneficial organisms (Djordjevic et al. 1987, Zhang et al.
2009).

Despite the traditional view that plant metabolites can be divided to the categories of
'primary and secondary', there is a no solid rule to distinguish two groups of metabolites.
One relatively simple and clear distinction is whether their biosynthesis is universal for

	
  

1

all plants or specific to different plant species. While the production of primary
metabolites such as amino acids, sugars and lipids is common to all plant species
(Aharoni and Galili 2011), secondary metabolite biosynthesis is taxon-specific (Pichersky
and Gang 2000). For example, plants in the family Brassicaceae and genus Catharanthus
produce glucosinolates (Wittstock and Halkier 2002) and indole alkaloids (Ziegler and
Facchini 2008), respectively. Based on the taxon-specificity of secondary metabolite
biosynthesis and discovery of increasing numbers of biological functions, the term
'specialized metabolite' has gained popularity in this field (Aharoni and Galili 2011,
Pichersky and Gang 2000). This term will be used throughout the remainder of this
dissertation.

The category of primary and specialized metabolites is determined based on the function
of chemical compounds rather than on their structures. In fact, a number of specialized
metabolites are derived from identical biosynthetic pathways using the same precursors
as primary metabolites (Aharoni and Galili 2011, Frey et al. 1997). Therefore, some
specialized metabolites are structurally similar to primary metabolites. For example,
despite structural similarity (Buchana et al. 2002) (Figure 1.1), proline is categorized as a
primary metabolite and the flowering-inducing molecule pipecolic acid as a specialized
metabolite (Fujioka and Sakurai 1997).

Examples of the biosynthesis of plant specialized metabolites and their values
Among the various groups of plant specialized metabolites, this section will focus on two
classes of compounds, flavonoids and acylsugars, because this dissertation work focuses

	
  

2

on these classes of compounds (See 'Chapters 2 and 3' for flavonoids and Chapter 4 for
acylsugars, respectively). The biosynthetic pathways for terpenoids and alkaloids are also
introduced briefly, because these are, along with flavonoids, two of the three major
classes of plant specialized metabolites.

Figure 1.1. Structural similarity between the primary metabolite proline and specialized
metabolite pipecolic acid.

Flavonoids
Flavonoids such as chalcones, flavonones, flavones, flavonols, isoflavones and
anthocyanins are well-studied and widely distributed plant specialized metabolites
(Boudet 2007, Buer et al. 2010, Dixon and Steele 1999). The biological roles of
	
  

3

flavonoids have been intensively investigated. Beside the impact on flower color (Mol et
al. 1998), pollen viability (Taylor and Jorgensen 1992) and protecting plants against UV
light (Li et al. 1993) and reactive oxygen species stress (Taylor and Grotewold 2005),
flavonoids facilitate the interaction between plants and microorganisms such as in
legume-bacterium symbiosis (Djordjevic et al. 1987) and nodule organogenesis in plant
roots (Wasson et al. 2006, Zhang et al. 2009). In addition to the biological roles in plants,
flavonoids were shown to have health benefits as anti-oxidants (Boudet 2007) and antimicrobials (Nowack and Schmitt 2008).

The biosynthesis of flavonoids begins with chalcone production by chalcone synthase
using malonyl-CoA and coumaryl-CoA as substrates (Buchana et al. 2002) (Figure 1.2).
Chalcone is then converted to structurally diverse flavonoid compounds such as
flavanones, flavones, flavonols and isoflavonoids by reactions of oxidation, reduction and
isomerization (Boudet 2007, Buchana et al. 2002, Buer et al. 2010, Dixon and Steele
1999) (Figure 1.3). The flavonoids are also modified by glycosylation and/or methylation,
resulting in a wide diversity of chemical structures (Jones et al. 2003, Jorgensen et al.
2005, Matsuda et al. 2009, Schmidt et al. 2011).

A number of glycosylated flavonols have been identified from various plant tissues such
as leaf tissues (Braca et al. 2001, Schmidt et al. 2011), stems and bark (Min et al. 2003),
fruits (Reynertson et al. 2008), flowers (Modolo et al. 2009, Wu et al. 2008) and isolated
glandular trichomes (Kang et al. 2010a). The structure of glycosylated flavonols vary
depending on the types of flavonol aglycone and the types, numbers and positions of

	
  

4

sugars attached to flavonol ring structure (See Figure 1.2 for flavonoid ring structure and
numbering system). The unmodified flavonol aglycone can be one of three flavonols,
kaempferol, quercetin and myricetin, which differ by the number of hydroxyl groups on
B ring (See Figure 1.2 for flavonoid ring structure and numbering system). While
kaempferol has a hydroxyl group at C4', quercetin and myricetin have two and three
hydroxyl groups at C3' and C4', and C3', C4' and C5' on the B ring, respectively. Many
sugars, for instance, glucose, galactose, rhamnose, xylose or arabiose, are found in
varying combinations and positions on glycosylated flavonols (Li et al. 2008).

The number of sugars in glycosylated kaempferol and - quercetin compounds is reported
to range from one to five (Williams and Grayer 2004, Wu et al. 2008). When the flavonol
is glycosylated with multiple numbers of sugars, individual sugars can be attached at the
different positions or multiple sugars can be bonded to each other and then attached to a
single position of flavonol. Glycosylation has been reported at the positions of C3, C7,
C3' and C'4 in kaempferol or quercetin, among which C3 is the most often glycosylated
position followed by C7 (Min et al. 2003, Modolo et al. 2009, Williams and Grayer
2004). The glycosyaltion of myricetin is also most commonly found at C3 (Braca et al.
2001, Williams and Grayer 2004). However, unlike the myriad of glycosylated derivates
identified for kaempferol and quercetin, relatively few glycosylated myricetin compounds
are known. In addition, while kaempferol and quercetin are often found glycosylated with

	
  

5

4-coumaryl CoA + malonyl CoA

Chalcones
Aurones
Flavanones

HO 7

8
A

6

5
OH

2’

1
O

3’
B

2 1’ 6’

C
4
O

4’ OH
5’

3
Flavones

Naringenin
(a flavanone)

Isoflavonoids

Catechins
Dihydroflavonols

Anthocyanidins
Proanthocyanidins

Flavonols
Leucoanthocyanidins

Anthocyanins

Figure 1.2. Biosynthetic pathways for flavonoid production.
Biosynthesis of flavonoid compounds starts with the production of chalcones. Chalcones
are then converted to structurally diverse flavonoid compounds through a serious of
chemical reactions. The chemical structure of naringenin is shown with numbering
system for the ring structure. This biosynthetic pathway is modified from Buer et al.
2010.

	
  

6

Figure 1.3. Examples of flavonoid and anthocyanin compounds.

	
  

7

polysaccharides, myricetin is mostly glycosylated with one or two sugars and highly
glycosylated forms of myricetin are rare (Williams and Grayer 2004). Identification and
characterization of a myricetin tetraglycoside compound is described in Chapter 2.

In addition, O-methylation of hydroxyl groups also contributes to the structural diversity
of flavonol compounds. Structurally diverse methylated flavonols are described from
various plant tissues including leaf (Kumari et al. 1984, Stevens et al. 1995), fruits
(Brunet and Ibrahim 1986), flowers (Jay et al. 1982), roots (Ibrahim 2005) and isolated
glandular trichomes (Roda et al. 2003, Schmidt et al. 2011, Stevens et al. 1995). Similar
to glycosylated flavonols, the structures of methylated flavonols also vary by the types of
flavonol aglycones and the numbers of methyl groups attached to flavonol backbone and
their positions: C3, C5, C7 and available hydroxyl groups such as C3', C4' and C5' on B
ring for the given flavonols (Ibrahim 2005, Kumari et al. 1984, Schmidt et al. 2011,
Stevens et al. 1995) (Figure 1.2). Various plants accumulate polymethylated flavonols, of
which Chrysosplenium americanum (Golden Saxifrage) is a well-studied example. The
identification and characterization of O-methyltransferases involved in the production of
polymethylated quercetins in C. americanum revealed that O-methylation at C3 position
is the first committed step in the sequential methylation reactions (Collins et al. 1981, De
Luca and Ibrahim 1982, De Luca and Ibrahim 1985, Ibrahim 2005).

In contrast to few glycosylated myricetin compounds known, a variety of methylated
myricetin compounds were identified (Kang et al. 2010a, Schmidt et al. 2011, Stevens et
al. 1995). For instance, previous study revealed that leaf trichomes from the wild relative

	
  

8

of cultivated tomato, S. habrochaites LA1777, accumulate polymethylated myricetins
such as tri-, tetra- and pentamethylated myricetins, but not glycosylated myricetins
(Schmidt et al. 2011). The lack of glycosylated myricetins raises a question whether
myricetin tends to be methylated rather than glycosylated, whereas lack of methylation
allows kaempferol and quercetin to be glycosylated. Especially, in that C3 of the flavonol
is most likely the first position to be glycosylated or methylated (De Luca and Ibrahim
1982, De Luca and Ibrahim 1985, Ibrahim 2005), it is suspected that this position is
critical to determine the modification type of myricetin in plants. In this dissertation study,
various S. lycopersicum variants including an EMS mutant and S. pennellii introgression
lines were identified as having altered trichome myricetin methylation phenotypes.
Characterization of the altered methylated myricetin chemistry in these genetic variants is
discussed in Chapter 3.

Acylsugars
Acylsugars are sugars esterified with acyl chains of various lengths at one or more
hydroxyl groups and are found in plants of the family Solanaceae (Figure 1.4). Plants in
the genus Solanum, including the cultivated tomato (Solanum lycopersicum) and its wild
relatives, are known to accumulate acylsugars (Ghangas and Steffens 1993). Of particular
note, the wild tomato plant S. pennellii LA0716 accumulates acylsugar exudates up to 20%
of leaf dry weight, making the leaf surface sticky (Fobes et al. 1985, Goffreda et al.
1989). These sticky chemical compounds improve plant defense against herbivore insects
by deterring insect feeding and oviposition (Hawthorne et al. 1992, Mutschler et al.
1996). The previous report that volatile branched chain aliphatic acids produced from the

	
  

9

metabolism of Nicotiana attenuata trichomes acylsugars by Manduca sexta caterpillars
attract carnivorous ants is a fascinating example of indirect plant defense against the
herbivore (Weinhold and Baldwin 2011). The diverse documented direct and indirect
defensive roles of acylsugars suggest that these compounds are of widespread ecological
importance.

Acylsugar compounds are structurally diverse, with differences in the sugar backbone
(typically glucose or sucrose), and the number and length of individual acyl chains. In
particular, chemical analysis of S. pennellii accessions revealed that differences in
acylsugar chemistry are associated with the geographic distribution of accessions. For
example, S. pennellii accessions from the southern part of the Andes mountains,
including LA0716, predominately accumulate acylglucoses, whereas accessions from the
northern range have higher ratios of acylsucroses to acylglucoses (Shapiro et al. 1994).
This observation suggests that the diverse acylsugar profiles in S. pennellii accessions
result from evolutionary adaptation. However, this question remains to be answered
because relatively little is known about the biosynthetic and regulatory pathways leading
to the production of chemically diverse acylsugar metabolites. Chemical diversity of
acylsugar compounds and its correlation with the geography for 80 S. habrochaites
accessions will be discussed in Chapter 4.

	
  

10

(A)

R

R

R

(B)

R

R

R
R

Figure 1.4. Chemical structures of acylsugars.
(a) Schematic structure of an acylglucose. This acylglucose consists of glucose esterified
with three acyl chains. ‘R’ indicates various numbers of carbons in the individual acyl
chains. Typical acyl chains from S. pennellii LA0716 range from C4 to C12.
(b) Schematic structure of an acylsucrose. The structure shown here depicts sucrose
tetraester, a sucrose esterified with four acyl chains. Typical acyl chains from S.
lycopersicum M82 range from C2 to C12.
	
  

11

Terpenes
Terpenoids are structurally diverse compounds that include volatiles of commercial
importance (Pichersky and Gershenzon 2002). Terpenes are defined based on the number
of carbons in the terpene skeleton; for example, C10, C15, C20, C30 and C40 for mono-,
sesqui-, di-, tri and tetraterpenes, respectively (Buchana et al. 2002, Tholl 2006, Tholl
and Lee 2011). Once terpene skeletons of various lengths are generated through the
condensation of C5 isoprene units, they are converted to structurally diverse terpene
compounds through cyclization and modification such as glycosylation and oxidation
(Pichersky and Gershenzon 2002, Tholl 2006) (Figures 1.5). Beside improvement of
plant defense against herbivore insects (Baldwin et al. 2006), terpene compounds
contribute to plant flavor and aroma (Pichersky and Gershenzon 2002, Schilmiller et al.
2008). Some terpenes are used as important drugs; for examples the antimalarial
arteminisin (Liu et al. 2011) and anti-cancer agent taxol (Roberts 2007) (Figure 1.6).
Collaborative work with Dr. Cornelius Barry's group in the Michigan State University
Department of Horticulture revealed chemical diversity of terpene profiles in 80 S.
habrochaites accessions and its correlation with geographic distribution of the individual
accessions. This work is described in a manuscript entitled 'Evolution of TPS20-related
terpene synthases influences chemical diversity in the glandular trichomes of the wild
tomato relative Solanum habrochaites' submitted for publication in November, 2011.

	
  

12

Acetyl-CoA (x3)
MVP pathway
In cytosol

Pyruvate + GAP
MEP pathway
In plastid

IPP

C5

DMAPP

OPP
OPP

OPP
OPP

C10

Geranyl diphosphate

Monoterpenes

C15

Farnesyl diphosphate

Sesquiterpenes

x2
C20

Geranylgeranyl diphosphate

Diterpenes

x2
C30

Squalene

Triterpenes

C40

Phytoene

Tetraterpenes

Figure 1.5. Overview of terpene biosynthesis.

	
  

13

Figure 1.5. Continued
The biosynthesis of terpene compounds begins with condensation of five carbon isoprene
units, isopentenyl diphosphate (IPP) and its isomer dimethylallyl diphosphate (DMAPP).
IPP is synthesized by the mevalonate pathway (MVP) and methylerythritol phosphate
pathway (MEP) operating in the cytosol and plastid respectively. ‘GAP’ indicates
glyceraldehyde 3-phosphate. Rendering of the biosynthetic pathway is modified from
Buchana et al. 2002. Previously trans-prenyldiphosphates such as E-geranyl diphosphate
and E,E-farnesyl diphosphate were considered the universal intermediates for
monoterpene and sesquiterpene biosynthesis, respectively. However, recent studies
revealed that S. lycopersicum trichomes synthesize monoterpenes from the cisprenyldiphosphate intermediate neryldiphosphate (Schilmiller et al. 2009), whereas the
all cis Z,Z-farnesyl diphosphate intermediate is used to produce sesquiterpenes in S.
habrochaites trichomes (Sallaud et al. 2009).

	
  

14

$-caryophyllene
$-caryophyllene
!-humulene
!-humulene

sene
arnesene

rmacrene B
germacrene B

germacrene C C
germacrene
$-caryophyllene
$-caryophyllene

2-carene
2-carene
2-carene
2-Carene
"-elemene
2-carene
"-elemene

germacrene CC
germacrene

!-pinene
!-pinene
!-pinene

limonene
"-phellandrene
!-phellandrene
limonene
"-phellandrene
limonene
!-phellandrene
$-caryophyllene
"-phellandrene Limonene
!-phellandrene
!-Phellandrene "-Phellandrene
#-elemene!-humulene
limonene
#-elemene
"-phellandrene
!-phellandrene
"-elemene
"-elemene
#-elemene
#-elemene

!-pinene
germacrene C
!-Pinene

$-farnesene
$-farnesene

"-elemene

#-elemene

!-santalene
!-santalene

$-bergamotene
$-bergamotene
"-Bergamotene

!-santalene
$-farnesene!-santalene
!-humulene
!-humulene
$-bergamotene "-Farnesene
$-bergamotene
!-Santalene
!-Humulene
!-humulene
!-humulene

germacren
germacre

$-caryophyllene
$-caryophyllene
"-Caryophyllene

germacrene C C
germacrene

$-caryophyllene
$-caryophyllene

!-santalene

$-bergamotene

"-elemene
"-elemene
germacrene D
germacrene D
germacrene D
germacrene D
germacrene C
germacrene BB
germacrene
!-humulene
$-caryophyllene
Germacrene B
Germacrene D Germacrene C
$-farnesene
$-farnesene

H
H
O
O

H3C
H3C

O
$-farnesene O
H
H

CH3
CH3

$-farnesene
"-elemene $-farnesene #-elemene
"-elemene
#-elemene
#-Elemene
$-Elemene

H
H

O
O

CH3
CH3

OH
OH

O
O
Artemisinin

OH
OH

!-santalene
Taxol

$-bergamotene
germacrene B B
germacrene

germacrene D D
germacrene

Figure 1.6. Examples of terpene compounds.
germacrene B

	
  

$-bergamotene
$-bergamotene

O OH
OOCCH
OOCCH3 O OH
3 !-santalene
#-elemene
!-santalene
$-bergamotene
NH $-bergamotene
O
O NH
O
O

"-elemene

O
O

O
O

germacrene D

germacrene B

germacrene D

15

H
O
O

H

O
O
OOCCH3
OOCCH

germacrene BB
germacrene
O
O

3

germacrene D
germacrene D

Alkaloids
Nitrogen-containing alkaloids include biosynthetically and structurally diverse chemical
compounds (Facchini and St-Pierre 2005). In particular, biosynthetic pathways for the
production of benzylisoquinoline- and monoterpenoid indole alkaloids have been
intensively studied (Facchini and St-Pierre 2005, Ziegler and Facchini 2008) (Figure 1.7).
Alkaloid compounds are also important pharmaceuticals; for example vinblastine is an
anti-cancer drug and ajmaline is an efficient drug for antiarrythmic heart disorders
(Figure 1.8).

Examination of metabolite profiles - metabolite profiling and metabolomics
Cataloguing the types of metabolites and their abundance is a prerequisite to
understanding plant specialized metabolism. Metabolite cataloguing can be divided into
two main categories: metabolite profiling and metabolomics (Fiehn 2002, Last et al.
2007). The main distinction between the two approaches is determined mainly by the
number of metabolites in an analysis (Sweetlove et al. 2003). Metabolite profiling aims
to detect and quantify relatively small numbers of metabolites, typically of similar
chemical properties (Fiehn 2002, Last et al. 2007). As a result, metabolite profiling
approaches typically are highly sensitive and accurate (Sweetlove et al. 2003). While
metabolite profiling has strong advantages for studying compounds of known structure
and physical properties, it is not well suited for more open-ended investigation of a large
numbers of metabolites of diverse physical properties, including uncharacterized
compounds. Metabolomics is an alternative approach where large numbers of structurally
diverse compounds are analyzed (Fiehn 2002, Sweetlove et al. 2003). Although there are

	
  

16

(a)

Dopamine + p-hydroxyphenylacetaldehyde

H3CO
H3CO
S-Reticuline

N CH3 3
CH

HO

H

HO

H3CO
H3CO
S-Reticuline

Benzylisoquinoline alkaloids

(b)

Tryptamine + Secologanin

NH

NH

Strictosidine

H

H

OGlc
H

O

H33CO2C
H CO2C
Monoterpenoid indole alkaloids

Strictosidine

Figure 1.7. Biosynthetic pathways for alkaloid production.
(a) Benzylisoquinoline alkaloid production. S-reticuline is used to produce structurally
diverse benzylisoquinoline alkaloids. The synthesis of S-reticuline is initiated with
condensation of dopamine and p-hydroxyphenylacetaldehyde.
(b) Monoterpenoid indole alkaloid production. Strictosidine is produced by the
condensation of tryptamine and secologanin and it serves as an intermediate for synthesis
of diverse monoterpenoid indole alkaloid compounds.
The biosynthetic pathways (a-b) are modified from Facchini and St-Pierre 2005.
	
  

17

(a)

H CO
H33CO

HO

O

O
N

N

CH3
CH3

CH3
CH3

H

H
HO

HO

Codeine

Morphine

HO
C2H5
C2H5
(b)

Vinblastine

N
H

N
H
H3CO2C
H3CO2C

N

C2H5
C2H5
N

CH3O
CH3 O

OCOCH 3
OCOCH3
CO2CH3
CO CH

HO
CH3
CH3

2

3

OH
H
N

N
H

H

CH
CH3

Figure 1.8. Examples of alkaloid compounds.
(a) Examples of benzylisoquinoline alkaloids
(b) Examples of monoterpenoid indole alkaloids
	
  

18

OH

Ajamaline

generally tradeoffs between the number of metabolites analyzed in a given method and
the accuracy, reproducibility and selectivity of the assay, metabolomics is beneficial to
understand a large number of metabolites even including poorly or uncharacterized
metabolites in an open-ended analysis (Sumner et al. 2003).

Three major steps to understand plant specialized metabolite profiles
Investigation of the types of plant specialized metabolites and their abundance requires
three major steps: 1. chemical extraction of the metabolites; 2. use of an analytical
chemistry platform; 3. data analysis. Details of these individual steps are discussed
below.

Chemical extraction of plant specialized metabolites
Because plants accumulate a myriad of specialized metabolites with varying chemical
and physical properties, the choice of extraction system determines the types of
metabolites harvested from plant materials (Kim et al. 2011, Last et al. 2007, Shin et al.
2010). Therefore, selecting an extraction system requires consideration of the chemical
and physical characteristics of metabolites such as solubility and concentration range
(Sweetlove et al. 2003). Furthermore, the accumulation of plant specialized metabolites is
highly influenced by the environment. In addition, some classes of metabolites (for
example, glucosinolates) are so highly reactive that special care must be taken to prevent
their degradation or conversion during extraction or storage (Kim et al. 2011, Shin et al.
2010). Therefore, various methods for immediate quenching or cessation of all

	
  

19

metabolism are commonly used to prevent unwanted alteration of metabolite profiles
during tissue harvest and chemical extraction (Kim et al. 2011).

Analytical chemistry platforms
Different analytical chemistry techniques have varying characteristics including
sensitivity, selectivity and speed. Thus, understanding the benefits and technical
limitations for each technique is essential in selecting suitable analytical instruments for
specific experimental goals (Lei et al. 2011). Because liquid chromatography-mass
spectrometry (LC-MS) was employed in this dissertation study, MS-based approaches are
mainly discussed in this section.

MS is commonly used to investigate metabolite profiles due to three benefits. First, the
high sensitivity of MS allows detection of low abundance metabolites down to the
nanomolar concentration range (Last et al. 2007). Second, MS enables identification of a
large number of metabolites based on characteristic mass to charge ratio (m/z) and
quantification of specific metabolites by measuring signal intensity for the individual ion.
Third, use of multidimensional or high mass accuracy MS can reveal information about
the class of molecule being analyzed and guide structural elucidation of some chemical
compounds (Lei et al. 2011, Sumner et al. 2003). The MS method is even more powerful
when combined with chromatographic separation such as gas- or liquid chromatography.
The availability of authentic standards enables unambiguous identification of compounds
and heavy isotope-labeled standards permit accurate quantification of target metabolites
(Gu et al. 2007).

	
  

20

Gas chromatography (GC)-MS is suitable for profiling volatile metabolites such as
terpene compounds (Sallaud et al. 2009, Schilmiller et al. 2009). GC-MS is also used to
analyze non-volatile metabolites after derivatization steps that increase the volatility of
chemical compounds (Sumner et al. 2003). GC-MS has three major advantages: 1.
complete ionization, minimizing ion suppression effects; 2. reproducible retention times,
allowing comparisons of samples across time, instruments and laboratories; 3. availability
of large mass spectral databases (Aprea et al. 2011, Kopka et al. 2005, Sawada et al.
2009, Sweetlove et al. 2003). These characteristics promote metabolite identification,
data interpretation and cross comparisons of results from different studies and
laboratories, making GC-MS an excellent tool to study plant specialized metabolites.

LC-MS techniques are suitable for analyzing non-volatile metabolites without
derivatization and can be faster than GC-MS methods, makingLC-MS efficient for highthroughput analysis. These benefits make LC-MS one of the most popular analytical
platforms despite existence of technical challenges such as matrix effects (i.e. ion
suppression), lack of comparability of retention times and ionization characteristics and
paucity of comprehensive mass spectral databases (Sumner et al. 2003, Sweetlove et al.
2003).

The combination of approaches such as GC- and LC-MS in the study of complex
mixtures or elucidation of metabolite structures has advantages compared with the more
rapid direct-infusion MS. For example, chromatographic separation provides additional

	
  

21

information such as retention time. Therefore, GC-MS or LC-MS can generate twodimensional information indexes for the characteristic ions with m/z and retention time
(referred to as ‘analytical signals’ in this dissertation). Furthermore, chromatographic
separation prior to MS decreases the number of chemical compounds delivered to the MS
at a given time, reducing ion suppression and allowing more accurate quantification and
data analysis (Last et al. 2007, Lei et al. 2011).

Although MS-based approaches, especially combined with GC or LC, are efficient tools
to study plant metabolites, MS analysis is not sufficient to unambiguously elucidate
completely novel structures of chemical compounds and generally does not provide
information about stereochemistry. This limitation can often be compensated by using
nuclear magnetic resonance (NMR) methods (Kim et al. 2011, Sweetlove et al. 2003).
NMR enables the identification and quantification of metabolites based on chemical shift,
intensity, fine structure and magnetic relaxation properties. Because these signals are
directly related to the chemical environment of nuclei in atoms in the metabolites, the
NMR technique is suitable for elucidating the chemical structure of novel compounds,
sometimes including stereochemistry (Krishnan et al. 2005). Disadvantages of these
powerful methodologies are that they typically require relatively large amounts of a
metabolite in a mixture for identification and quantification of known compounds and of
relatively pure preparations of metabolites for structural elucidation of compounds for
which standards are unavailable.

	
  

22

Targeted and non-targeted data analysis of large MS datasets
Because MS-based approaches generate large amounts of data, analysis and interpretation
challenging. Analysis methods can be divided to two categories: targeted and nontargeted. The category is determined based on whether the identity and properties of
metabolites to be analyzed is already known or not. Targeted data analysis aims to assess
relatively small numbers of ‘known’ metabolites, whereas non-targeted method
investigates datasets in an open-ended manner without previous knowledge or prediction
of metabolite class or specific compound structures. Therefore, non-targeted data analysis
is beneficial when studying poorly or uncharacterized metabolites.

Advances in bioinformatics facilitate the analysis of large MS datasets. For example,
many software packages have been developed for noise reduction, analytical signal
detection and normalization (Aprea et al. 2011, Tolstikov et al. 2003). In addition,
computer-based statistical tests help to explore enormous datasets reliably and extract
useful information. Multivariate analyses are often used to examine and interpret large
datasets, especially for non-targeted data analysis (Kim et al. 2011, Sumner et al. 2003).
For example, the application of the unsupervised principal component analysis (PCA)
and hierarchical clustering analysis (HCA) methods assists data interpretation by
reducing the dimensionality of datasets and efficiently visualizing them (Schilmiller et al.
2010a, Sumner et al. 2003). Supervised method such as partial least squares discriminant
analysis (PLS-DA) is also useful to identify characteristic compounds that separate
different sample groups, including metabolites of low abundance (Kim et al. 2011). The
application of multivariate tests for non-targeted data analysis is discussed in Chapter 2.

	
  

23

Solanum trichomes as a model system to study plant specialized metabolites
Because the biosynthesis of specific types of plant specialized metabolites is often
confined to taxa and or tissue types, selecting a proper model system is critical for their
study (Pichersky and Gang 2000, Ziegler and Facchini 2008). Trichomes are specialized
epidermal structures populating the surfaces of plant tissues. In particular, glandular
trichomes or secreting glandular trichomes (SGTs) are 'chemical factories' where a
number of specialized metabolites are produced and thus are useful as model systems to
elucidate their biochemistry (Schilmiller et al. 2008, Wagner 1991). The genus Solanum
includes commercially valuable crop species such as the cultivated tomato and their
SGTs accumulate or exude high-levels of various specialized metabolites (Besser et al.
2009, Kandra et al. 1990, Slocombe et al. 2008).

In addition, morphological and chemical diversity in Solanum trichomes makes them
especially appealing models for specialized metabolism studies (McDowell et al. 2011,
Schilmiller et al. 2008). Seven different types of trichomes are documented in the
cultivated and wild tomato species, four of which are known or proposed to be glandular
(types I, IV, VI and VII in Figure 1.9). The types and density of trichomes vary across
Solanum species and also according to their location on the plant (Luckwill 1943,
McDowell et al. 2011). In addition to morphological diversity, SGTs have characteristic
compositions of chemicals they produce, accumulate and/or secrete. Some specific types
of SGTs in Solanum species have shown to accumulate certain classes of chemicals, for
example, acylsugars are secreted by type 4 of S. pennellii (Fobes et al. 1985, McDowell
et al. 2011) and methylketones are produced in type 6 of S. habrochaites (Fridman et al.

	
  

24

2005). The accumulation of gland-specific metabolites suggests that distinct biosynthetic
pathways are present in particular types of SGTs. The great diversity in morphology and
chemistry of trichomes in different Solanum species provides a good opportunity to
identify trichome-synthesized metabolites and the major biosynthetic pathways operating
in each type of trichomes.

Another advantage of this system compared with tissues and organs that produce other
specialized metabolites is that purification of SGTs is relatively straightforward.
Preparation of samples enriched with specific glands enables detailed chemical profiling,

Figure 1.9. Seven types of Solanum trichomes.
Types I, IV, and VI are documented to be glandular trichomes, type VII was proposed to
be glandular and the rest are non-glandular trichomes. Trichome types are drawn with
modification of Luckwill 1943.

	
  

25

EST library construction and proteomics for individual types of SGTs (Schilmiller et al.
2008, Schilmiller et al. 2010b, Schmidt et al. 2011). The enriched sample preparation is
beneficial for discovering novel genes and metabolites, especially when the accumulation
of target transcripts and metabolites is at low levels (Schmidt et al. 2011).

The availability of various genetic and genomic resources also facilitates study of
Solanum trichome metabolism. For example, an EMS (ethyl methane sulfonate) tomato
mutant population is available (Menda et al. 2004) for forward genetics. Chromosomal
substitution lines such as the S. lycopersicum M82 X S. pennellii introgression lines (ILs)
(Eshed and Zamir 1994, Eshed and Zamir 1995) have proven to be highly useful
resources for studies of trichome metabolism (Schilmiller et al. 2010a, Schilmiller et al.
2009). Details about the use of the tomato EMS mutant population and M82 X S.
pennellii ILs are discussed in Chapters 2 and 3. In addition, accessions of wild species
from different areas of the Andes Mountains and other areas are also available
(http://tgrc.ucdavis.edu/). The use of natural variation in studies of trichome metabolites
is discussed in Chapter 4.

Finally, Solanum trichomes are an increasingly powerful model system due to recent
advances in genomic information and biological resources for the cultivated tomato and
its wild relatives; these include whole genome sequence, ESTs and BAC clones and
sequence (http://solgenomics.net/). The availability of genomic tools is beneficial in that
they efficiently assist gene discovery, functional elucidation of genes and understanding
evolutionary relationships of the genes (Buell and Last 2010, DellaPenna and Last 2008,

	
  

26

Falara et al. 2011). The application of tomato genomic resources will be discussed in
Chapter 3.

As a part of this project aimed at elucidating biosynthetic pathways of trichome
metabolites in the cultivated tomato and its wild relatives, three individual studies were
performed for this dissertation work: (i) identification of tomato EMS mutants altered in
trichome non-volatile metabolites and characterization of the glycosylated flavonoid
phenotype in the JP314 EMS mutant (ii) identification and analysis of S. lycopersicum
genetic variants such as the JP117 EMS mutant and S. pennellii IL6-3/4 for trichome
methylated myricetin biochemistry, and (iii) chemical analysis of diverse trichome
acylsugar profiles in 80 accessions from the wild tomato S. habrochaites species.

	
  

27

References

	
  

28

References
Aharoni, A. and Galili, G. (2011) Metabolic engineering of the plant primary-secondary
metabolism interface. Curr. Opin. Biotechnol., 22, 239-244.
Aprea, E., Gika, H., Carlin, S., Theodoridis, G., Vrhovsek, U. and Mattivi, F. (2011)
Metabolite profiling on apple volatile content based on solid phase
microextraction and gas-chromatography time of flight mass spectrometry. J.
Chromatogr. A., 1218, 4517-4524.
Baldwin, I.T., Halitschke, R., Paschold, A., von Dahl, C.C. and Preston, C.A. (2006)
Volatile signaling in plant-plant interactions: "talking trees" in the genomics era.
Science, 311, 812-815.
Besser, K., Harper, A., Welsby, N., Schauvinhold, I., Slocombe, S., Li, Y., Dixon,
R.A. and Broun, P. (2009) Divergent regulation of terpenoid metabolism in the
trichomes of wild and cultivated tomato species. Plant Physiol., 149, 499-514.
Boudet, A.M. (2007) Evolution and current status of research in phenolic compounds.
Phytochemistry, 68, 2722-2735.
Braca, A., Bilia, A., Mendez, J. and Morelli, I. (2001) Myricetin glycosides from
Licania densiflora. Fitoterapia, 72, 182-185.
Brunet, G. and Ibrahim, R.K. (1986) O-methylation of flavonoids by cell-free extracts
of calamondin orange. Phytochemistry, 19, 741-746.
Buchana, B., Gruissem, W. and Jones, R. (2002) Biochemistry & Molecular Biology of
Plants: Wiley.
Buell, C.R. and Last, R.L. (2010) Twenty-first century plant biology: impacts of the
Arabidopsis genome on plant biology and agriculture. Plant Physiol., 154, 497500.
Buer, C.S., Imin, N. and Djordjevic, M.A. (2010) Flavonoids: new roles for old
molecules. J. Integr. Plant Biol., 52, 98-111.
Collins, F.W., De Luca, V., Ibrahim, R.K., Voirin, B. and Jay, M. (1981)
Polymethylated flavonol synthesis in Chrysosplenium americanum: identification
and enzymatic synthesis. Z. Naturforsch, 36c, 730-736.
De Luca, V. and Ibrahim, R.K. (1982) Characterization of three distinct Omethyltransferases from Chrysosplenium americanum. Phytochemistry, 21, 15371540.

	
  

29

De Luca, V. and Ibrahim, R.K. (1985) Enzymatic synthesis of polymethylated
flavonols in Chrysosplenium americanum. I. Partial purification and some
properties of S-adenosyl-L-methionine:flavonol 3-, 6-, 7- and 4'-Omethyltransferases. Arch. Biochem. Biophys., 238, 595-605.
DellaPenna, D. and Last, R.L. (2008) Genome-enabled approaches shed new light on
plant metabolism. Science, 320, 479-481.
Dixon, R.A. and Steele, C.L. (1999) Flavonoids and isoflavonoids - a gold mine for
metabolic engineering. Trends Plant Sci., 4, 394-400.
Djordjevic, M.A., Redmond, J.W., Batley, M. and Rolfe, B.G. (1987) Clovers secrete
specific phenolic compounds which either stimulate or repress nod gene
expression in Rhizobium trifolii. EMBO J., 6, 1173-1179.
Eshed, Y. and Zamir, D. (1994) A genomic library of Lycopersicon pennellii in L.
esculentum: A tool for fine mapping of genes. Euphytica, 79, 175-179.
Eshed, Y. and Zamir, D. (1995) An introgression line population of Lycopersicon
pennellii in the cultivated tomato enables the identification and fine mapping of
yield-associated QTL. Genetics, 141, 1147-1162.
Facchini, P.J. and St-Pierre, B. (2005) Synthesis and trafficking of alkaloid
biosynthetic enzymes. Curr. Opin. Plant Biol., 8, 657-666.
Falara, V., Akhtar, T.A., Nguyen, T.T., Spyropoulou, E.A., Bleeker, P.M.,
Schauvinhold, I., Matsuba, Y., Bonini, M.E., Schilmiller, A.L., Last, R.L.,
Schuurink, R.C. and Pichersky, E. (2011) The tomato terpene synthase gene
family. Plant Physiol., 157, 770-789.
Fiehn, O. (2002) Metabolomics--the link between genotypes and phenotypes. Plant Mol.
Biol., 48, 155-171.
Fobes, J.F., Mudd, J.B. and Marsden, M.P.F. (1985) Epicuticular lipid accumulation
on the leaves of Lycopersicon pennellii (Corr.) D'Arcy and Lycopersicon
esculentum Mill. Plant Physiol., 77, 567-570.
Frey, M., Chomet, P., Glawischnig, E., Stettner, C., Grun, S., Winklmair, A.,
Eisenreich, W., Bacher, A., Meeley, R.B., Briggs, S.P., Simcox, K. and Gierl,
A. (1997) Analysis of a chemical plant defense mechanism in grasses. Science,
277, 696-699.
Fridman, E., Wang, J., Iijima, Y., Froehlich, J.E., Gang, D.R., Ohlrogge, J. and
Pichersky, E. (2005) Metabolic, genomic, and biochemical analyses of glandular

	
  

30

trichomes from the wild tomato species Lycopersicon hirsutum identify a key
enzyme in the biosynthesis of methylketones. Plant Cell, 17, 1252-1267.
Fujioka, S. and Sakurai, A. (1997) Conversion of lysine to L-pipecolic acid induces
flowering in Lemna paucicostata 151. Plant Cell Physiol., 38, 1278-1280.
Ghangas, G.S. and Steffens, J.C. (1993) UDPglucose: fatty acid transglucosylation and
transacylation in triacylglucose biosynthesis. Proc. Nat. Acad. Sci. USA, 90,
9911-9915.
Goffreda, J.C., Ave, D.A., Mutschler, M.A., Steffens, J.C. and Tingey, W.M. (1989)
Aphid deterrence by glucose esters in the glandular trichome exudate of the wild
tomato, L. pennellii. J. Chem. Ecol., 15, 2135-2147.
Gu, L., Jones, A.D. and Last, R.L. (2007) LC-MS/MS assay for protein amino acids
and metabolically related compounds for large-scale screening of metabolic
phenotypes. Anal Chem, 79, 8067-8075.
Hawthorne, D.J., Shapiro, J.A., Tingey, W.M. and Mutschler, M.A. (1992)
Trichome-borne and artificially applied acylsugars of wild tomato deter feeding
and oviposition of the leafminer Liriomyza trifolii. Entomologia Experimentalis et
Applicata, 65, 65-73.
Ibrahim, R. (2005) A forty-year journey in plant research: original contributions to
flavonoid biochemistry. Can. J. Bot., 83, 433-450.
Jay, M., De Luca, V. and Ibrahim, R.K. (1982) Meta-methylation of rings A (8-) and B
(3'-) is catalyzed by two distinct O-methyltransferases in Lotus corniculatus. Z.
Naturforsch, 38, 413-417.
Jones, P., Messner, B., Nakajima, J., Schaffner, A.R. and Saito, K. (2003) UGT73C6
and UGT78D1, glycosyltransferases involved in flavonol glycoside biosynthesis
in Arabidopsis thaliana. J. Biol. Chem., 278, 43910-43918.
Jorgensen, K., Rasmussen, A.V., Morant, M., Nielsen, A.H., Bjarnholt, N.,
Zagrobelny, M., Bak, S. and Moller, B.L. (2005) Metabolon formation and
metabolic channeling in the biosynthesis of plant natural products. Curr. Opin.
Plant Biol., 8, 280-291.
Kandra, G., Severson, R. and Wagner, G.J. (1990) Modified branched-chain amino
acid pathways give rise to acyl acids of sucrose esters exuded from tobacco leaf
trichomes. Eur. J. Biochem., 188, 385-391.
Kang, J.H., Liu, G., Shi, F., Jones, A.D., Beaudry, R.M. and Howe, G.A. (2010a) The
tomato odorless-2 mutant is defective in trichome-based production of diverse

	
  

31

specialized metabolites and broad-spectrum resistance to insect herbivores. Plant
Physiol., 154, 262-272.
Kang, J.H., Shi, F., Jones, A.D., Marks, M.D. and Howe, G.A. (2010b) Distortion of
trichome morphology by the hairless mutation of tomato affects leaf surface
chemistry. J. Exp. Bot., 61, 1053-1064.
Kim, H.K., Choi, Y.H. and Verpoorte, R. (2011) NMR-based plant metabolomics:
where do we stand, where do we go? Trends Biotechnol., 29, 267-275.
Kopka, J., Schauer, N., Krueger, S., Birkemeyer, C., Usadel, B. and Bergmuller, E.
(2005) GMD@CSB.DB: the Golm Metabolome Database. Bioinformatics, 21,
1635-1638.
Krishnan, P., Kruger, N.J. and Ratcliffe, R.G. (2005) Metabolite fingerprinting and
profiling in plants using NMR. J. Exp. Bot., 56, 255-265.
Kumari, G., Rao, L. and Rao, N. (1984) Myricetin methyl ethers from Solanum
pubescens. Phytochemistry, 23, 2701-2702.
Last, R.L., Jones, A.D. and Shachar-Hill, Y. (2007) Towards the plant metabolome and
beyond. Nat. Rev. Mol. Cell. Biol., 8, 167-174.
Lei, Z., Huhman, D.V. and Sumner, L.W. (2011) Mass spectrometry strategies in
metabolomics. J. Biol. Chem., 286, 25435-25442.
Li, J., Ou-Lee, T.M., Raba, R., Amundson, R.G. and Last, R.L. (1993) Arabidopsis
flavonoid mutants are hypersensitive to UV-B irradiation. Plant Cell, 5, 171-179.
Li, Z., Ngojeh, G., DeWitt, P., Zheng, Z., Chen, M., Lainhart, B., Li, V. and Felpo,
P. (2008) Synthesis of a library of glycosylated flavonols. Tetrahedron Letters,
49, 7243-7245.
Liu, B., Wang, H., Du, Z., Li, G. and Ye, H. (2011) Metabolic engineering of
artemisinin biosynthesis in Artemisia annua L. Plant. Cell. Rep., 30, 689-694.
Luckwill, L. (1943) The genus Lycopersicon: A historical, biological, and taxonomic
survey of the wild and cultivated tomatoes Aberdeen, Scotland: Aberdeen
University Press.
Matsuda, F., Yonekura-Sakakibara, K., Niida, R., Kuromori, T., Shinozaki, K. and
Saito, K. (2009) MS/MS spectral tag-based annotation of non-targeted profile of
plant secondary metabolites. Plant J., 57, 555-577.
McDowell, E.T., Kapteyn, J., Schmidt, A., Li, C., Kang, J.H., Descour, A., Shi, F.,
Larson, M., Schilmiller, A., An, L., Jones, A.D., Pichersky, E., Soderlund,

	
  

32

C.A. and Gang, D.R. (2011) Comparative functional genomic analysis of
Solanum glandular trichome types. Plant Physiol., 155, 524-539.
Menda, N., Semel, Y., Peled, D., Eshed, Y. and Zamir, D. (2004) In silico screening of
a saturated mutation library of tomato. Plant J., 38, 861-872.
Min, B., Lee, S., Kim, J., Lee, J., Kim, T., Kim, D., Kim, Y., Joung, H., Lee, H. and
Nakamura, N. (2003) Anti-complement activity of constituents from the stembark of Juglans mandshurica. Biol. Pharm. Bull., 26.1042-1044.
Modolo, L.V., Li, L., Pan, H., Blount, J.W., Dixon, R.A. and Wang, X. (2009) Crystal
structures of glycosyltransferase UGT78G1 reveal the molecular basis for
glycosylation and deglycosylation of (iso)flavonoids. J Mol Biol, 392, 1292-1302.
Mol, J., Grotewold, E. and Koes, R. (1998) How genes paint flowers and seeds. Trends
Plant Sci., 3, 212-217.
Mutschler, M.A., Doerge, R.W., Liu, S.C., Kuai, J.P., Liedl, B.E. and Shapiro, J.A.
(1996) QTL analysis of pest resistance in the wild tomato Lycopersicon pennellii:
QTLs controlling acylsugar level and composition Theor. Appl. Genet. , 92, 709718.
Nowack, R. and Schmitt, W. (2008) Cranberry juice for prophylaxis of urinary tract
infections – Conclusions from clinical experience and research. Phytomedicine,
15, 653-667.
Pichersky, E. and Gang, D.R. (2000) Genetics and biochemistry of secondary
metabolites in plants: an evolutionary perspective. Trends Plant Sci., 5, 439-445.
Pichersky, E. and Gershenzon, J. (2002) The formation and function of plant volatiles:
perfumes for pollinator attraction and defense. Curr. Opin. Plant Biol., 5, 237243.
Reynertson, K., Yang, H., Jiang, B., Basile, M. and Kennelly, E. (2008) Quantitative
analysis of antiradical phenolic constituents from fourteen edible Myrtaceae
fruits. Food Chem. , 109, 883-890.
Roberts, S.C. (2007) Production and engineering of terpenoids in plant cell culture. Nat.
Chem. Biol., 3, 387-395.
Roda, A.L., Oldham, N.J., Svatos, A. and Baldwin, I.T. (2003) Allometric analysis of
the induced flavonols on the leaf surface of wild tobacco (Nicotiana attenuata).
Phytochemistry, 62, 527-536.
Sallaud, C., Rontein, D., Onillon, S., Jabes, F., Duffe, P., Giacalone, C., Thoraval, S.,
Escoffier, C., Herbette, G., Leonhardt, N., Causse, M. and Tissier, A. (2009)

	
  

33

A novel pathway for sesquiterpene biosynthesis from Z,Z-farnesyl pyrophosphate
in the wild tomato Solanum habrochaites. Plant Cell, 21, 301-317.
Sawada, Y., Akiyama, K., Sakata, A., Kuwahara, A., Otsuki, H., Sakurai, T., Saito,
K. and Hirai, M.Y. (2009) Widely targeted metabolomics based on large-scale
MS/MS data for elucidating metabolite accumulation patterns in plants. Plant Cell
Physiol., 50, 37-47.
Schilmiller, A., Shi, F., Kim, J., Charbonneau, A.L., Holmes, D., Daniel Jones, A.
and Last, R.L. (2010a) Mass spectrometry screening reveals widespread
diversity in trichome specialized metabolites of tomato chromosomal substitution
lines. Plant J., 62, 391-403.
Schilmiller, A.L., Last, R.L. and Pichersky, E. (2008) Harnessing plant trichome
biochemistry for the production of useful compounds. Plant J., 54, 702-711.
Schilmiller, A.L., Miner, D.P., Larson, M., McDowell, E., Gang, D.R., Wilkerson, C.
and Last, R.L. (2010b) Studies of a biochemical factory: tomato trichome deep
expressed sequence tag sequencing and proteomics. Plant Physiol., 153, 12121223.
Schilmiller, A.L., Schauvinhold, I., Larson, M., Xu, R., Charbonneau, A.L.,
Schmidt, A., Wilkerson, C., Last, R.L. and Pichersky, E. (2009) Monoterpenes
in the glandular trichomes of tomato are synthesized from a neryl diphosphate
precursor rather than geranyl diphosphate. Proc. Natl. Acad. Sci. U S A, 106,
10865-10870.
Schmidt, A., Li, C., Shi, F., Jones, A.D. and Pichersky, E. (2011) Polymethylated
myricetin in trichomes of the wild tomato species Solanum habrochaites and
characterization of trichome-specific 3'/5'- and 7/4'-myricetin Omethyltransferases. Plant Physiol., 155, 1999-2009.
Schwab, W. (2003) Metabolome diversity: too few genes, too many metabolites?
Phytochemistry, 62, 837-849.
Shapiro, J.A., Steffens, J.C. and Mutschler, M.A. (1994) Acylsugars of the wild
tomato Lycopersicon pennellii in relation to geographic distribution of the
species. Biochem. Syst. Ecol., 22, 545-561.
Shin, M.H., Lee do, Y., Liu, K.H., Fiehn, O. and Kim, K.H. (2010) Evaluation of
sampling and extraction methodologies for the global metabolic profiling of
Saccharophagus degradans. Anal. Chem., 82, 6660-6666.
Slocombe, S.P., Schauvinhold, I., McQuinn, R.P., Besser, K., Welsby, N.A., Harper,
A., Aziz, N., Li, Y., Larson, T.R., Giovannoni, J., Dixon, R.A. and Broun, P.
(2008) Transcriptomic and reverse genetic analyses of branched-chain fatty acid

	
  

34

and acyl sugar production in Solanum pennellii and Nicotiana benthamiana. Plant
Physiol., 148, 1830-1846.
Stevens, J., Hart, H. and Wollenweber, E. (1995) The systematic and evolutionary
significance of exudate flavonoids in Aeonium Phytochemistry 39, 805-813.
Sumner, L.W., Mendes, P. and Dixon, R.A. (2003) Plant metabolomics: large-scale
phytochemistry in the functional genomics era. Phytochemistry, 62, 817-836.
Sweetlove, L.J., Last, R.L. and Fernie, A.R. (2003) Predictive metabolic engineering: a
goal for systems biology. Plant Physiol., 132, 420-425.
Taylor, L. and Jorgensen, R. (1992) Conditional male-fertility in chalcone synthasedeficient Petunia. J. Hered., 83, 11-17.
Taylor, L.P. and Grotewold, E. (2005) Flavonoids as developmental regulators. Curr.
Opin. Plant Biol., 8, 317-323.
Tholl, D. (2006) Terpene synthases and the regulation, diversity and biological roles of
terpene metabolism. Curr. Opin. Plant Biol., 9, 297-304.
Tholl, D. and Lee, S. (2011) Terpene Specialized Metabolism in Arabidopsis thaliana.
Tolstikov, V.V., Lommen, A., Nakanishi, K., Tanaka, N. and Fiehn, O. (2003)
Monolithic silica-based capillary reversed-phase liquid
chromatography/electrospray mass spectrometry for plant metabolomics. Anal.
Chem., 75, 6737-6740.
Wagner, G.J. (1991) Secreting glandular trichomes: more than just hairs. Plant Physiol.,
96, 675-679.
Wasson, A.P., Pellerone, F.I. and Mathesius, U. (2006) Silencing the flavonoid
pathway in Medicago truncatula inhibits root nodule formation and prevents
auxin transport regulation by rhizobia. Plant Cell, 18, 1617-1629.
Weinhold, A. and Baldwin, I.T. (2011) From the Cover: Trichome-derived O-acyl
sugars are a first meal for caterpillars that tags them for predation. Proc Natl Acad
Sci U S A, 108, 7855-7859.
Williams, C. and Grayer, R. (2004) Anthocyanins and other flavonoids. Nat. Prod.
Rep., 21, 539-573.
Wittstock, U. and Halkier, B.A. (2002) Glucosinolate research in the Arabidopsis era.
Trends Plant Sci., 7, 263-270.

	
  

35

Wu, J.H., Huang, C.Y., Tung, Y.T. and Chang, S.T. (2008) Online RP-HPLC-DPPH
screening method for detection of radical-scavenging phytochemicals from
flowers of Acacia confusa. J Agric Food Chem, 56, 328-332.
Zhang, J., Subramanian, S., Stacey, G. and Yu, O. (2009) Flavones and flavonols play
distinct critical roles during nodulation of Medicago truncatula by Sinorhizobium
meliloti. Plant J., 57, 171-183.
Ziegler, J. and Facchini, P.J. (2008) Alkaloid biosynthesis: metabolism and trafficking.
Annu. Rev. Plant Biol., 59, 735-769.

	
  

36

Chapter 2. Chemical screening of tomato EMS mutants to identify altered trichome
non-volatile metabolites

Abstract
EMS M2 family mutants of the cultivated M82 tomato were analyzed to identify altered
trichome non-volatile metabolites using high-throughput LC-ToF-MS and targeted and
non-targeted data analyses. Self-cross M3 progeny plants from M2 putative mutants were
analyzed to test for inheritance of the phenotypes. Of 1,387 M2 mutant plants from 58
M2 families, 24 mutants were confirmed for altered trichome non-volatile chemistry. The
mutant JP314 was identified as having decreased accumulation of a flavonol
tetraglycoside with m/z 949.3 from targeted analysis. Non-targeted analysis validated the
flavonol tetraglycoside mutant phenotype and revealed unannotated analytical signals
highly correlated with the m/z 949.3. To further characterize these metabolites, tomato
leaf extracts were chemically profiled in more detail. The results suggested that the
flavonol tetraglycoside with m/z 949.3 is a myricetin derivative and the other
unannotated analytical signals are adduct ions of this flavonol tetraglycoside. JP314
mutant also have decreased anthocyanins in leaf tissue. The phenotypes of reduced
myricetin tetraglycoside and total anthocyanins lead to the hypothesis that flavonoid
accumulation is impaired in this mutant. The JP314 mutant case study demonstrates the
value of 'brute force' M2 mutant metabolite screening in uncovering mutations that affect
specialized metabolism, improving metabolite annotation and generating testable
hypothesis.

	
  

37

Introduction
Trichomes are specialized epidermal 'hairs' covering the surface of plant tissues
(Schilmiller et al. 2008). Trichomes are categorized into non-glandular or glandular types
based on the presence of one or more glandular cells on the top of a stalk (Luckwill 1943,
Wagner 1991). In particular, secretory and glandular trichomes (SGTs) serve as 'chemical
factories' where plant specialized metabolites are produced and accumulated. SGTs are
often used as models to study plant specialized (also known as 'secondary') metabolism
(Besser et al. 2009, McConkey et al. 2000, Schilmiller et al. 2008, Turner et al. 2000).
For example, SGTs from plants in the genus Solanum including the cultivated tomato
Solanum lycopersicum and its wild relatives have been studied to understand volatile
metabolites such as terpenoids (Falara et al. 2011, Sallaud et al. 2009, Schilmiller et al.
2010b, Schilmiller et al. 2009) and methylketones (Fridman et al. 2005, Yu 2010).
Previous studies also reported that Solanum SGTs accumulate a variety of specialized
non-volatile metabolites such as flavonoids, alkaloids and acylsugars (Fobes et al. 1985,
McDowell et al. 2011, Schilmiller et al. 2010a).

Forward and reverse genetics are efficient approaches to understand gene functions.
Reverse genetics looks for phenotypes when the function of a specific gene or set of
genes is deliberately altered. In contrast, forward genetics is used to screen large
collections of genetic variants for altered phenotypes and then identify the allele
responsible for a trait that differs from the wild type or reference organism. Mutagenesis
with ethylmethane sulphonate (EMS) is commonly used to create genetic diversity for
forward genetics in plants, microbes and animals (Jander et al. 2002, Peters et al. 2003).

	
  

38

EMS mutagenesis causes high densities of G:C to A:T point mutations at random
chromosomal locations (Jander et al. 2003), allowing efficient sampling of changes in
many genes in the genome by screening a relatively small number of mutants (Buttner et
al. 2010, Marti et al. 2010). This makes it an excellent mutagen for generating genetic
variants for mutant screens. Performing forward genetic screens on EMS mutant
populations allows the discovery of genes that influence a phenotype without prior
knowledge of gene function. This is beneficial for the investigation of plant specialized
metabolism, especially when studying unknown or poorly understood biosynthetic
pathways.

Careful choice of phenotypic assay is also important for forward genetic screens. These
can range from a simple test looking for differences in plant morphology to technically
complex chemical profiling to identify altered metabolites. For the identification of
mutants altered in specialized metabolites, selecting the optimal analytical approach is an
important prerequisite. Special challenges arise because plants accumulate complex
mixtures of specialized metabolites varying in concentration range and chemical
properties. To maximize the value of the screen, plant metabolite profiling requires
analytical approaches with high sensitivity, accuracy and reproducibility. There are
several reasons why mass spectrometry (MS) is a powerful tool to profile plant
specialized metabolites. First, MS can detect metabolites down to the nanomolar
concentration range (Last et al. 2007). The high sensitivity and selectivity of MS allows
simplifying extraction methods (for instance, eliminating the need for two phase
extraction), thus shortening sample preparation time while still allowing detection of low

	
  

39

abundance specialized metabolites. MS can be more powerful when combined with
chromatographic separation such as liquid chromatography (LC) or gas chromatography
(GC) (Last et al. 2007). The efficacy of MS combined with GC was revealed from
chemical profiling of volatile terpenes in different tissue types of M82 cultivated tomato
(Falara et al. 2011, Schilmiller et al. 2010b). LC-MS was used to investigate non-volatile
metabolite profiles in S. pennellii introgression lines (Schilmiller et al. 2010a) and S.
habrochaites SGTs (Schmidt et al. 2011).

Despite advances in detection and quantification of specialized metabolites, metabolite
annotation still remains limited (Sumner et al. 2003, Wink 1999). One reason why
structural characterization remains difficult is that there is enormous structural diversity
and complexity of metabolites in plants. In addition, structural annotation is even harder
due to low abundance of metabolites, lack of authentic standards and the formation of
various types of adducts for a single metabolites in MS (Last et al. 2007). When
metabolite annotation is not known, non-targeted data analysis (i.e. the open-ended data
analysis) is useful. Unlike targeted analysis, non-targeted data analysis does not require
previous knowledge of target metabolites. For instance, non-targeted analysis of LC-MS
datasets revealed spatial distribution of metabolites in strawberry flowers without prior
knowledge of metabolite composition (Hanhineva et al. 2008). Non-targeted data
analysis also helped identify volatile metabolites in ripe tomato fruits (Tikunov et al.
2005). Non-targeted data analysis is advantageous to understand metabolite profiles, even
including incompletely characterized metabolites (See 'Chapter 1. Introduction' for more
details about targeted and non-targeted data analyses).

	
  

40

In this study, tomato EMS M2 family mutants were screened by high-throughput LCTime of Flight (ToF)-MS analysis to identify altered non-volatile trichome chemistry.
The workflow of mutant screening is described with the JP314 mutant as a case study.
Another class of mutants from this screen is described in Chapter 3.

Results
The structure of the EMS M2 family mutant population and its use for mutant
screening
A cultivated tomato (Solanum lycopersicum, cultivar M82) M2 mutant population was
previously generated by EMS mutagenesis (Menda et al. 2004)Sol Genomics Network,
http://solgenomics.net/) and provided by Dr. Dani Zamir, Hebrew University, Rehovot,
Israel. This population was chosen to screen for trichome metabolites because it has a
high mutation rate, and was previously shown to have mutants of many morphological
phenotypes (Menda et al. 2004)The structure of this mutant population, with M2 families
of seed from individual M1 plants (Figure 2.1), also helps identify siblings and
independent alleles, supporting the discovery of heritable mutations (Krieger et al. 2010).
Chemical extraction was performed by a single leaf dip method with a short incubation;
this method was chosen to maximize trichome metabolite extraction, while minimizing
the extraction of compounds from epidermal pavement cells (Schilmiller et al. 2010a). A
negative mode LC-ToF-MS assay was employed to profile non-volatile metabolites from
plant extracts enriched with trichome metabolites (Figure 2.2).

	
  

41

M0
(Wild type)

M1
A

B

x

x

Self-pollination

M2 families

A1

A2

A3

A4

B1

B2

B3

B4

A5

A6

A7

A8

B5

B6

B7

B8

x

x

x

x

Self-pollination

M3
progeny
plants

Progeny of A7

Progeny of A8

Progeny of B5

Progeny of B6

Figure 2.1. Structure of the mutant population used in this study.
Tomato EMS M2 family mutant seeds were previously generated (Menda et al. 2004)
and kindly supplied by Dani Zamir, Hebrew University, Rehovot, Israel. The M2 plants
were screened to identify altered trichome non-volatile metabolites. Once an M2 putative
mutant was identified, the self-pollinated M3 progeny were analyzed to test for
inheritance of the chemical phenotype. M2 plants derived from a single M1 plant are
siblings and any common phenotype is presumed to result from the same original
mutation.
	
  

42

M2 mutant plants were considered putative mutants when their chemical phenotypes
varied from wild type controls by two serial statistical cutoffs. These putative mutants
were self-pollinated to obtain M3 progeny. Because individual M2 mutant plants were
assayed, putative mutants were confirmed by testing chemical phenotypes in M3 progeny
plants. These progeny plants were analyzed with at least four biological replicates to test
whether the M2 putative mutant phenotype was consistently observed in M3 progeny
(Figures 2.1 and 2.2). I also took advantage of the mutant population structure in two
ways (Menda et al. 2004). First, if two different M2 siblings from the same family
showed identical phenotypes (for example, A7 and A8, or B5 and B6 in Figure 2.1), it is
likely that they have shared mutations. Second, if similar phenotypes were found in M2
lines from two independent families (for example, A8 and B5, or A7 and B6 in Figure
2.1), it suggests a high chance that they are independent alleles for the same mutant
genes.

The heritability of M2 putative mutant phenotypes was validated by chemical analysis of
the self-pollinated M3 progeny plants. In addition to confirmation of the M2 phenotype,
rescreening of four or more M3 plants increases the statistical power to detect other
phenotypes (Figure 2.1).

Identification and characterization of trichome non-volatile metabolite mutants by
high-throughput LC-ToF-MS with targeted and non-targeted data analysis
The chemical profiles of EMS M2 mutant plants were analyzed and compared with
control M82 wild type plants and a set of control S. pennellii introgression lines (ILs)

	
  

43

previously found to have quantitative changes in trichome chemistry (Schilmiller et al.
2010a) (Figure 2.2, See 'Materials and Methods' for details about plant growth). The
rapid LC-ToF-MS chemical analysis generated a list of thousands of 'analytical signals',
which is defined as the combination of retention time in LC and mass-to-charge ratio
(m/z) with high mass accuracy for corresponding ions detected in MS (Gu et al. 2010,
Last et al. 2007, Schilmiller et al. 2010a). The analytical signals were evaluated by
targeted and non-targeted data analysis (Figure 2.2).

Thirty-five analytical signals were tested for targeted data analysis (Table 2.1). Target
metabolites were chosen based on previously characterized annotation (Schilmiller et al.
2010a, Table 2.1 and Figure 2.2). Unannotated analytical signals were also included
when reproducibly detected (Schilmiller et al. 2010a). Two serial cutoffs were applied to
identify M2 putative mutants in the targeted data analysis: 1. median +/- 5 median
absolute deviation (MAD); 2. > 3-fold change compared to the median (Figure 2.2).
When the alteration of trichome metabolites in M2 mutant plants met two serial cutoffs,
they were considered putative mutants in M2 screening (Figure 2.2). Non-targeted data
analysis such as principal component analysis (PCA) was performed to identify M2
putative mutants that targeted data analysis might have failed to reveal (Figure 2.2).

The same LC-ToF-MS assay was used to analyze M3 progeny of putative mutants.
Targeted data analysis was performed to identify M3 progeny altered in the 35 analytical
signals using two serial cutoffs: 1. > 3-fold change in the mean between mutant and M82

	
  

44

Single leaf dip
extraction of EMS M2
mutant plants

Rapid LC-ToF-MS analysis

Targeted data analysis
•  Quantification of target
metabolites
•  Application of two serial
cutoffs
1. median +/- 5 median
absolute deviation (MAD)
2. > 3 fold change

Non-targeted data
analysis
•  Principal Component
Analysis (PCA)

Putative hits in M2

x

Self-pollination

Validation of chemical
phenotypes in M3 progeny
plants

Targeted data analysis
•  Quantification of target
metabolites
•  Application of two serial
cutoffs to test weather
altered chemistry from M2
putative hits was inherited
1. > 3 fold change
2. p < 0.05 in student t-test

Non-targeted data analysis
•  PCA
•  Partial least squaresdiscriminant analysis
(PLS-DA)

Identification of mutant
phenotypes

Figure 2.2. EMS mutant screening workflow.

	
  

45

Figure 2.2. Continued
LC-ToF-MS (negative ion mode) analysis of single leaf dip extracts was used to profile
trichome non-volatile metabolites. Targeted and non-targeted data analyses were
performed to investigate LC-MS datasets. Two serial cutoffs were used in targeted data
analysis to identify M2 putative mutants and validate the phenotypes in M3. Multivariate
analyses such as PCA and PLS-DA was performed for non-targeted data analysis.

	
  

46

wild type; 2. statistically significant alteration corresponding to p < 0.05 from Student's ttest (Figure 2.2). In addition to PCA, the partial least squares discriminant analysis (PLSDA) supervised method was also employed for non-targeted data analysis in M3 progeny.
PLS-DA allows identification of minor factors (i.e. analytical signals) contributing to the
separation between mutant and wild-type plants and searching for the correlation among
these analytical signals (Westerhuis et al. 2007).

Twenty-four mutants with altered trichome non-volatile metabolites
This mutant screening tested 1,387 M2 mutants from 58 families. Of 1,387 plants, 223
(16%) were identified as M2 putative mutants and 175 yielded M3 seed for rescreening.
In total, 24 mutants had consistent phenotypes in M3 progeny (Table 2.2). The 24
identified mutants revealed altered chemistry in diverse classes of trichome non-volatile
metabolites ranging from methylated flavonoids to acylsugars (Table 2.3). Notably,
multiple mutants were discovered for five types of trichome chemistry, methylated
myricetins, flavonol tetraglycoside with m/z 949.3, kaempferol 3-O-rutinoside, acylsugar
S4:15 and an unannotated analytical signal with m/z 493.3. Based on the M2 mutant
population structure, they are likely to be either siblings (from the same M2 family, for
example, JP1913 and JP1938 for flavonol glycoside phenotype) or independently derived
alleles (from different M2 families, for example, JP857 and JP2883 for acylsugar S4:15
(2,4,4,5) phenotype) (Table 2.3). Seven mutants from three M2 families were identified
with methylated myricetin phenotypes; thus this collection contains at least three
independent alleles for this phenotype (see Chapter 3 for detailed analysis of one such

	
  

47

Table 2.1. The list of 35 analytical signals tested for targeted data analysis.
Thirty-five analytical signals tested for targeted data analysis are listed. The metabolite
annotation, LC retention time, m/z and ion type is from Schilmiller et al. 2010a.
Unannotated or unknown analytical signals were also included when reproducibly
detected from M82 plants. Acylsugar nomenclature includes type of sugar, number of
acyl chains, total number of carbons in acyl chains and expected carbon number for each
acyl chain; for example, S4:17 (2,5,5,5) indicates sucrose esterified with 4 acyl chains,
one acyl chain with 2 carbons and three with 5 carbons respectively, for a total of 17
carbons. The abundance of metabolites was estimated based on peak areas for the
corresponding analytical signals. Relative abundance was obtained by comparing the
peak areas of the individual analytical signals with that of S4:17 (2,5,5,5), the most
abundant acylsugar in M82 cultivated tomato (Schilmiller et al, 2010) and indicated as
follows; < 1% (---), 1~3 (--), 3~5 (-), 5~10 (+), 10~20 (++), 20~35 (+++), 35~50 (++++)
and 50~100 (+++++).

	
  

48

Table 2.1. Continued

Retention
time (min)
Glucose
0.4
Sucrose
0.4
Malic acid
0.5
Quinic acid
0.5
Chlorogenic acid
1.7
Flavonol tetraglycoside
2.2
Quercetin 3-O-glucosylglucoside
2.3
Flavonol glycoside
2.4
Quercetin 3-O-rutinoside
2.5
Kaempferol 3-O-rutinoside
2.6
Mono-O-methylmyricetin
2.7
Dehydrotomatine
2.8
Tomatine
2.9
Di-O-methylmyricetin
2.9
Tri-O-methylmyricetin
3.2
Acylsugar S4:15 (2,4,4,5)
3.3
Acylsugar S3:15 (5,5,5)
3.3
Acylsugar S4:16 (2,4,5,5)
3.3
Acylsugar S4:17 (2,5,5,5)
3.4
Acylsugar S4:18 (4,4,5,5)
3.5
Acylsugar S4:20
3.5
(2,4,4,10 or 5,5,5,5)
Acylsugar S3:20
3.7
(5,5,10 or 4,5,11)
Acylsugar S4:22
3.8
(2,4,4,12 or 2,5,5,10)
Acylsugar S3:21
3.8
(5,5,11 or 4,5,12)
Acylsugar S4:23
3.8
(2,4,5,12 or 2,5,5,11)
Acylsugar S3:22 (5,5,12)
3.8
Acylsugar S3:23 (5,6,12)
3.9
Acylsugar S4:24 (2,5,5,12)
3.9
Metabolite Annotation

	
  

49

m/z (Da)

Ion Type

215.0
377.1
133.0
191.0
353.1
949.3
625.2
771.2
609.2
593.2
331.1
1076.5
1078.5
345.1
359.1
653.3
639.3
667.3
681.3
695.3

[M+Cl][M+Cl][M-H][M-H][M-H][M-H][M-H][M-H][M-H][M-H][M-H][M+formate][M+formate][M-H][M-H][M+formate][M+formate][M+formate][M+formate][M+formate]-

Relative
Abundance
-+
------+++
--+
-----++
++
+++++
--

723.4 [M+formate]-

---

709.4 [M+formate]-

--

751.4 [M+formate]-

-

723.4 [M+formate]-

---

765.4 [M+formate]-

---

737.4 [M+formate]751.4 [M+formate]779.4 [M+formate]-

+++
--++++

Table 2.1. Continued

Metabolite Annotation
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
n-Propyl-4-hydroxybenzoate
(Internal standard)

	
  

Retention
time (min)
0.5
1.4
2.2
2.2
2.2
3.0
3.1
3.0

50

m/z (Da)

Ion Type

371.1
369.1
693.4
963.3
1009.3
493.3
485.3
179.1

[M-H]-

Relative
Abundance
------------+

Table 2.2. Summary of mutant screen results
1,387 M2 mutant plants from 58 families were screened by LC-ToF-MS for altered
trichome metabolite levels (See Figures 2.1 and 2.2 for details). 223 of 1,387 plants were
identified as M2 putative mutants. Self-pollination yielded viable progeny for 175 of the
223 putative mutants, and 24 of these lines revealed consistent phenotypes upon
rescreening (See Table 2.3 for details of chemical phenotypes in the 24 identified
mutants).

The stage of screening

Description

Screening M2 individual mutant plants
Identification of putative mutants in M2

1387
223

From 58 M2 families

Validation of chemical phenotypes in M3
progeny

175

Identification of mutants

	
  

No. of plants

24

51

No seed (39),
No germination (9)
for M3 progeny
From 14 M2 families

Table 2.3. Chemical phenotypes of the 24 mutants identified from this screening.
Twenty-four mutants were identified for altered trichome non-volatile metabolites. SGN
M2 family ID (http://zamir.sgn.cornell.edu/mutants/) is listed with mutant names
assigned in this study.

Mutant chemical phenotypes

Mutant Name

Increased mono-O-methylmyricetin and
decreased di-O-methylmyricetin and tri-Omethylmyricetin

JP117, JP130,
JP334
JP256, JP446
JP3045, JP3060

Decreased flavonol tetraglycoside (m/z
949.3, RT 2.2) and unknown (m/z1009.3,
JP314
RT 2.2)
Increased flavonol glycoside (m/z 949.3,
JP1913, JP1938
RT 2.2) and unknown (m/z1009.3, RT 2.2)
JP761
JP416
Increased kaempferol 3-O-rutinoside
JP3427
JP3123
Decreased kaempferol 3-O-rutinoside
JP1229
Decreased tomatine
JP1885
Increased total acylsugars
JP269
Decreased acylsugars S4:20 (2,4,4,10 or
JP389
5,5,5,5) and S4:18 (4,4,5,5)
Increased acylsugar S4:23 (2,4,5,12 or
JP757
2,5,5,11)
JP857
Increased acylsugar S4:15 (2,4,4,5)
JP2883
Increased unknowns (m/z 693.4, RT 2.2;
JP316
m/z 371.1, RT 0.5)
Increased unknown (m/z 493.3, RT 3.0)
JP691, JP612

	
  

52

SGN M2
Family ID
e0004b
e0007b
e0139b
e0004b
e0124b
e0019b
e0014b
e0142b
e0139b
e0116b
e0117b
e0005b
e0009b
e0019b
e0016b
e0131b
e0005b
e0118b

mutant). In summary, this screen identified 24 mutants for trichome non-volatile
chemistry, providing valuable genetic variants for further experiments.

A Case Study: Identification of JP314 as a mutant altered in flavonoid glycosides
Identification of JP314 as having increased kaempferol 3-O-rutinoside
Targeted analysis of LC-ToF-MS assay data revealed higher accumulation of kaempferol
3-O-rutinoside in the M2 mutant JP314. This putative mutant was identified using two
serial cutoffs, 5 MAD and > 3 fold-change compared to the median of cohort plants
(Figure 2.2). In this screen, thirteen cohort M2 plants were grown in the same flat and
screened together with JP314 and one plant for each M82 wild type and IL11-3 (See
'Materials and Methods' for details of plant growth). As previously reported IL11-3
revealed decreased total acylsugar accumulation (Schilmiller et al. 2010a), with no
alteration in the remaining 22 analytical signals compared to M82 (Table 2.1), supporting
the reproducibility of chemical assay and data analysis in this mutant screening (Figure
2.3).

Non-targeted data analysis with PCA did not identify any putative mutants in M2
screening (Figure 2.4), which might be because PCA is often insensitive to subtle
differences, especially when the metabolites causing the phenotypes are low abundance
compounds. PCA with all 13 cohort plants distinguished IL11-3 control plant from others
(Figure 2.4ab). The strong chemical phenotypes in IL11-3 (Schilmiller et al. 2010a) led
to the hypothesis that the presence of IL11-3 prevents PCA from distinguishing potential
mutants with subtle phenotypes, especially for low abundance compounds. Therefore,

	
  

53

Peak area IS-1 gdw-1

250
200
150
100
50
0
-50
-100
-150

MED + 5 MAD

MED

MED - 5 MAD

JP314

M82

IL11-3

Individual M2 plants

Figure 2.3. Targeted data analysis identified an M2 putative mutant JP314 as having
increased kaempferol 3-O-rutinoside.
The altered accumulation of kaempferol 3-O-rutinoside in JP314 met the two serial
cutoffs used for the targeted data analysis in M2 screening: median + 5 MAD and > 3
fold change compared to the median.

	
  

54

(b)

-60

-20

20

PC 2 (1 %)

0.0e6

-1.0e7
0.0e6
PC 1 (88.1 %)

0.0e6

30
10
-220

PC 1 (97.9 %)

JP314
IL11-3

-180

-30

-1.0e7

-200

-10

PC 2 (0.3 %)

(d)

-2.0e7

-220

PC 1 (97.8 %)

0e6
-4e6

PC 2 (1.2 %)

(c)

-2.0e7

-4e6

-1.0e7

PC 2 (10.3 %)

1.0e7

60

(a)

-200

-180

PC 2 (98.7 %)

M82
M2 plants

Figure 2.4. Non-targeted data analysis could not identify M2 putative mutant, not even
JP314.

	
  

55

Figure 2.4. Continued
Non-targeted data analysis with PCA was not a sensitive approach to identify M2
putative mutants. Scores plot is shown with PC1 and PC2. The 95% confidence region is
marked with a circle. Although all 13 plants (12 plants for c and d) were included in this
analysis, some plants are almost superimposed with others due to their close position in
the scores plot. Filtering was performed by median intensity and scaling by log
transformation using MetaboAnalyst webtools (Xia et al. 2009). See 'Materials and
Methods' for details.
(a) PCA with IL11-3, with neither filtering nor scaling
(b) PCA with IL11-3, with filtering and scaling
(c) PCA without IL11-3, with neither filtering nor scaling
(d) PCA without IL11-3, with filtering and scaling

	
  

56

PCA was performed without IL11-3, but did not identify any outlier M2 plants in this
cohort, despite the fact that the targeted data analysis identified JP314 (Figure 2.4cd). The
failure could be due to two reasons: 1. the lack of replicate plants in M2 screening; 2.
subtle differences in low abundance metabolites.

Identification of novel metabolite changes in M3 progeny of JP314
Eighteen JP314 M3 progeny plants were analyzed by LC-ToF-MS to test the heritability
of the phenotype identified from M2 screening. Surprisingly, the altered kaempferol 3-Orutinoside phentoype was not seen in the M3 progeny, indicating that this trait was not
inherited (Figure 2.5). In addition, the M3 progeny showed a phenotype not noted in the
M2 sample: reduction in a flavonol tetraglycoside with m/z 949.3 and two previously
unannotated analytical signals with m/z 963.3 and 1009.3 (Figure 2.5). These are low
abundance ions whose signal intensity is less than 1% of S4:17 (2,5,5,5), the most
abundant acylsugar in the cultivated tomato (Table 2.1). These chemical alterations met
two serial cutoffs, > 3-fold change and p-value < 0.05 in Student's t-test (Figure 2.2). To
test whether these chemical phenotypes were consistently seen in the JP314 M3 progeny,
another set of nine M3 plants were grown and assayed at different time. The phenotype
was consistent in the second set of M3 progeny test (data not shown). These results
indicated that the JP314 M3 progeny are true breeding for a mutation that caused the
reduction in the flavonol tetraglycoside with m/z 949.3 and unannotated signals with m/z
963.3 and 1009.3.

	
  

57

Non-targeted data analysis to search for additional phenotypes in JP314 M3 progeny
Non-targeted data analysis with PCA and PLS-DA was performed to identify additional
analytical signals associated with JP314 M3 plants, and to search for potential correlation

JP314 (n=18)
M82 (n=6)

20000

5000

***

10000

***

15000

***

Peak area IS-1 gdw-1

25000

0
flavonoid glycoside unknown m/z 963!unknown m/z 1009!
K-H!
Flavonol
Unknown
Unknown Kaempferol 3m/z 949!
glycoside

(m/z 949.3)

(m/z 963.3) (m/z 1009.3)

O-rutinoside

Figure 2.5. Targeted data analysis of JP314 M3 progeny identified chemical phenotypes
not seen in M2 screening.
JP314 M3 progeny testing did not confirm the kaempferol 3-O-rutinoside phenotype
identified from M2 screening (Figure 2.3). Surprisingly, targeted data analysis of M3
progeny identified novel phenotypes, reduced accumulation of three analytical signals, a
flavonol tetraglycoside with m/z 949.3 and two unannotated ions (m/z 963.3, 1009.3).
*** indicates p-value < 0.001 from Student's t-test. Pearson and Spearman correlation test
revealed that the three analytical signals are highly correlated to each other (r > 0.99 from
both tests).

	
  

58

among the signals (Figure 2.6). PCA did not distinguish JP314 M3 progeny from M82
wild type (Figure 2.6a). This is presumably because PCA works by identifying major
components driving the separation of groups; it is not uncommon that this approach to be
insensitive to differences in low abundance metabolites (Schilmiller et al. 2010a). To
identify the influence of low abundance metabolites being altered in the mutant, the
supervised method PLS-DA was used (Figure 2.6b-c). The PLS-DA separated the JP314
M3 progeny from M82 (Figure 2.6b) and revealed a list of 30 ‘Variables Important for
Projections’ (VIPs), which are analytical signals that contribute to the separation between
JP314 M3 progeny and M82 (Figure 2.6c). The VIPs included the three analytical signals
previously identified by targeted data analysis (m/z 949.3, 963.3 and 1009. 3, See 'Figure
2.5').

Of 30, 5 VIPs were shown to be highly correlated with the flavonol tetraglycoside with
m/z 949.3 (r > 0.7 from Pearson and Spearman correlation tests respectively), including
m/z 949.3 itself, 963.3, 1009.3 and two other signals: m/z 995.3 and 981.3 and (Figure
2.6c). These five analytical signals had the same LC retention time when run using the 5
min method (Figure 2.6c; Schilmiller et al. 2010a), leading to the hypothesis that they are
structurally related to the flavonoid glycoside with m/z 949.3. In fact, m/z 995.3 was
reported to be a formate adduct of the m/z 949.3 flavonol tetraglycoside (Figure 2.6c,
Table 2.1, Schilmiller et al. 2010a).

	
  

59

(b)

0
-20
-40

JP314

-40

PC 2 (3.9 %)

20

Component 2 (14.4%)
0
20
-20

40

(a)

-60

M82

-120 -100 -80
-60
-40
PC 1 (82.1 %)

-40 -30 -20 -10

-20

0

10

20 30

Component 1 (24.9%)

(c)
Rank

Metabolite (m/z_RT)

1
2
3

7
8

837.3_ 3.6
1023.3_ 2.5
583.3_ 3.8
653.3_3.3
(acylsugar S4:15*)
995.3_ 2.2
779.4_3.9
(acylsugar S4:24*)
819.4_3.6
1289.6_3.3

9
10

4

VIP
Rank
mean
2.0979 16
2.0106 17
1.9896 18

793.4_3.9
657.2_3.3
753.2_3.3

VIP
mean
1.2956
1.2866
1.2824

Metabolite (m/z_RT)

1.9619

19

387.1_0.3

1.2714

1.9375

20

977.3_2.3

1.2674

1.9337

21

741.2_2.3

1.2345

1.9140
1.7945

22
23

1.2247
1.2222

1073.3_2.5

1.7904

24

1261.6_3.3
949.3_2.2
(flavonol glycoside*)

1.6268

25

1.4579

26

12

963.3_2.2 (unknown*)

1.4309

27

13

981.3_2.2

1.4197

28

865.4_3.7
725.2_2.4
681.3_3.4
(acylsugar S4:17*)
707.2_1.7
639.3_3.3
(acylsugar S3:15*)
1076.5_2.8
(dehydrotomatine*)
1009.3_2.2
(unknown*)

1.3081

29

469.0_0.8

1.1512

1.3032

30

851.3_3.6

1.1365

5
6

11

14
15

751.4_3.9
(acylsugar S3:23*)
1497.8_3.8

Figure 2.6.

	
  

60

1.2211
1.2163
1.1932
1.1837
1.1528

Figure 2.6. Non-targeted data analysis of JP314 M3 progeny validated the flavonol
tetraglycoside (m/z 949.3) phenotype and identified unannotated analytical signals highly
correlated with the m/z 949.3.
(a) PCA did not distinguish JP314 M3 progeny from M82 wild type plants. Scores plot is
shown with PC1 and PC2. The 95% confidence region is marked with a circle.
(b) PLS-DA distinguished JP314 M3 progeny from M82 wild type plants. Scores plot is
shown with component 1 and component 2. The 95% confidence region is marked with a
circle.
(c) Top 30 Variables Important for Projection (VIPs) from PLS-DA are listed with rank,
m/z, LC retention time (RT) and VIP mean. This list includes the flavonol tetraglycoside
with m/z 949.3 and four unannotated analytical signals highly correlated with m/z 949.3
(shown in bold). Analytical signals tested with targeted data analysis are labeled with *
(See 'Table 2.1' for the details).

	
  

61

Structural characterization of the flavonol tetraglycoside with m/z 949.3
Previous study structurally characterized m/z 949.3 as a flavonol tetraglycoside but did
not provide insight into the nature of the flavonol backbone or glycosylation (Schilmiller
et al. 2010a). Detailed structural characterization of the flavonol tetraglycoside with m/z
949.3 and the other four signals correlated with the flavonol tetraglycoside is challenging
because of their low abundance (Table 2.1) and LC co-elution (Figure 2.6c). Several
steps were taken to attempt to overcome these obstacles including the analysis of
concentrated plant extracts and longer LC (20 minutes) for better separation (See
'Materials and Methods' for the details).

Twenty-minute LC-ToF-MS with multiplexed collision-induced dissociation (MUXCID) analysis provided the evidence that the flavonol tetraglycoside with m/z 949.3 is a
myricetin derivative. Under LC-ToF-MS with MUX-CID analysis, mass spectra are
collected at multiple collision energies and saved for each energy condition. MS spectra
obtained at low collision energies provide molecular mass information and those
generated at higher energies yield successively more fragment ions ((Gu et al. 2010,
McDowell et al. 2011). To test whether the fragmentation of the flavonol tetraglycoside
with m/z 949.3 yields the ion of m/z for a flavonol aglycone, MS spectra were searched
under 10eV and 55eV (Figure 2.7). Molecular ion for the flavonol tetraglycoside with
m/z 949.3 was detected from extracted ion chromatogram (XIC) at low energy conditions
(Figure 2.7a) and fragment ions with m/z 317.1, 301.1 and 285.1 corresponding to
flavonol aglycones myricetin, quercetin and kaempferol respectively at high energy

	
  

62

(a)

100
%

(e) 100
Flavonol glycoside (m/z 949.3)
at low energy condition
(10eV) in MS

%
0

949.3

995.3

981.3
963.3
940

980

1009.3
1020 (m/z)

0
5
(b) 100
%

10

15

20 (min)

10

15

20 (min)

10

15

20 (min)

10

15

20 (min)

Myricetin (m/z 317.1)
at high energy condition
(55eV) in MS

0
5
(c)

100
%

0
(d)

100
%

Quercetin (m/z 301.1)
at high energy condition
(55eV) in MS

5
Kaempferol (m/z 285.1)
at high energy condition
(55eV) in MS

0
5

Figure 2.7.

	
  

63

Figure 2.7. LC-ToF-MS with MUX-CID analysis suggested that the flavonol
tetraglycoside with m/z 949.3 is a myricetin derivative and revealed adduct ions of this
myricetin tetraglycoside.
(a)-(b). Myricetin was shown to result from fragmentation of the flavonol tetraglycoside
with m/z 949.3.
(c)-(d). Fragment ions of the flavonol tetraglycoside (m/z 949.3) did not yield ions of m/z
301.1 or 285.1 corresponding to quercetin (c) or kaempferol (d) respectively.
(e) All four ions m/z 963.3, 981.3, 995.3 and 1009.3, shown to be highly correlated with
m/z 949.3 in Figure 2.6c, were detected under the peak of m/z 949.3 at 10eV.

	
  

64

condition (Figure 2.7bcd). The LC retention time (RT) for the major peak of m/z 317.1
was identical with that of m/z 949.3. However, neither peak of m/z 301.1 nor 285.1
revealed an identical RT with m/z 949.3. These results suggest that the flavonol
tetraglycoside has a myricetin aglycone. The broad chromatographic peak of m/z 949.3
was presumably due to overloading of analytes from the highly concentrated plant extract
on the column. The XICs for all three flavonols revealed several chromatographically
separable peaks, possibly due to different types of modified flavonols such as methylated
and/or glycosylated conjugates (Figure 2.7bcd). All four analytical signals m/z 963.3,
981.3, 995.3 and 1009.3, shown to be highly correlated with the flavonoid glycoside of
m/z 949.3 (Figure 2.6c), were detected under the peak of m/z 949.3, suggesting that these
four signals are adduct ions of m/z 949.3 (Figure 2.7e). The four individual signals, m/z
963.3, 981.3, 995.3 and 1009.3, are proposed as methylated, oxidized, formate and
acetate adduct ion of m/z 949.3 respectively.

Pale leaf color due to reduced total anthocyanins in JP314 mutant
In addition to the trichome myricetin tetraglycoside phenotype, the JP314 mutant has a
brighter leaf color than M82 due to the decreased accumulation of total anthocyanins
(Figure 2.8a-d). The reduction of total anthocyanins in JP314, only 40% of wild type
levels, was estimated by measuring light absorbance of crude leaf extracts (Figure 2.8e).

	
  

65

(a)

(c)

M82
(b)

JP314

(d)

JP314

0.06
0.05
0.04
0.03

*

(e)

Absorbance (A530 –0.25A657) gdw-1

M82

0.02
0.01
0
JP314

M82

Figure 2.8. Pale leaf color reveals reduced total anthocyanin accumulation in JP314.

	
  

66

Figure 2.8. Continued
(a)-(b) Differences in the color of stems and leaves of M82 wild type (a) and EMS mutant
JP314 (b) are visible in three different 3-week-old plants. Reduced purple pigment is very
noticeable in the abaxial (c) and adaxial (d) surfaces of JP314 leaves.
(e) Light absorbance, indicative of total anthocyanin amounts, was significantly lower in
leaf extracts of JP314 than M82 (p < 0.05 in Student's t-test). The bar graphs are
presented with mean and standard error (n=4 for each genotype).
For interpretation of the references to color in this and all other figures, the reader
is referred to the electronic version of this dissertation.

	
  

67

Discussion
This M2 screen identified 24 heritable mutants altered in various trichome non-volatile
metabolites (Table 2.3). Surprisingly, qualitative chemical phenotypes such as the
presence or absence of specific metabolites in mutants were not found. In addition, no
mutant was identified for relatively high abundance compounds, for example, acylsugars
S4:17, S4:24, S3:22 or quercetin 3-O-rutinoside (Table 2.1). The lack of these
phenotypes could be due to three reasons. First, a relatively small size of mutant
population was screened. While 6,000 mutants were generated for a saturated EMS M2
mutant population (Menda et al. 2004), this screening tested 1,387 mutants. Second,
despite efforts to 'cast a wide net', cutoffs in the M2 screening might have been so
stringent that they missed putative mutants. Third, altered trichome metabolites might
have been masked by genetic or
metabolic redundancy (Liljegren et al. 2000, Sasaki et al. 2002, Vision et al. 2000) or
promiscuous enzymatic activity (Khersonsky et al. 2006, Schmidt et al. 2011). Due to
these factors, identification of mutant phenotypes might have been limited.

Furthermore, technical challenges in analytical chemistry make the identification of
altered trichome metabolites even harder. For instance, structural annotations are
available only for a limited number of plant specialized metabolites. Metabolite
identification and quantification is also challenging due to the lack of authentic standards
and LC-MS database and the inherent issues in LC-MS including matrix effects and

	
  

68

retention time shifts (Last et al. 2007, Niessen et al. 2006, Wong et al. 2008) (See
'Chapter 1. Introduction' for the details).

Despite the challenges in analytical chemistry, LC-ToF-MS was efficient to identify
altered accumulation of myricetin tetraglycoside, a low abundance metabolite, whose ion
peak is less than 1% of S4:17, the most abundant acylsugar in the cultivated tomato
(Table 2.1, Schilmiller et al. 2010a). Targeted and non-targeted data analyses identified
myricetin tetraglycoside mutant phenotypes and unannotated analytical signals highly
correlated with the metabolite. Detailed structural analysis improved metabolite
annotation of a flavonol tetraglycoside with m/z 949.3 as myricetin tetraglycoside and
identified chemical relationships between this analytical signal and adduct ions.

Although a variety of methylated myricetin compounds have been identified from various
plants, relatively few glycosylated myricetins are known. In addition, unlike quercetin
and kaempferol flavonols, which are often found as glycosylated with oligosaccharides
up to pentasaccharides, examples of myricetins decorated with multiple sugar residues
are rare (Kang et al. 2010, McDowell et al. 2011, Schilmiller et al. 2010a, Schmidt et al.
2011, Stevens et al. 1995, Williams and Grayer 2004). Therefore, it is interesting that this
study identified differences in myricetin tetraglycoside accumulation in the trichome
exudates from the cultivated tomato and EMS mutants. Specifically, the JP314 mutant
has reduced accumulation of the myricetin tetraglycoside. Furthermore, in depth
structural characterization of the myricetin tetraglycoside, including sugar types and their

	
  

69

attached positions on the myricetin aglycone backbone, will help us understand trichome
myricetin tetraglycoside biochemistry.

Beside the decreased trichome myricetin tetraglycoside, the JP314 mutant has reduced
total anthocyanins in crude leaf extracts. Given that the flavonoid and anthocyanin
biosynthetic pathways share common precursors such as dihydroflavonols, the mutant
phenotypes led to the hypothesis that the JP314 mutation alters the level of
dihydromyricetin, a precursor of myricetin and its modified conjugates and anthocyanins
structurally related to myricetin (Figure 2.9). Previous study reported that the white color
mutant of the Bilberry fruit (Vaccinium myrtillus) lacks myricetin and myricetin-derived
anthocyanin accumulation, whereas the wild type fruit highly accumulates anthocyanins
and myricetin flavonol (Jaakola et al. 2002). Jaakola et al. 2002 also described that
flavonoid 3',5'-hydroxylase, a gene converting dihydrokaempferol to dihydromyricetin, is
involved in the production of myricetin and myricetin-derived anthocyanins in Bilberry
fruits. Given the JP314 mutant phenotypes, it will be useful to analyze anthocyanin
profiles to test whether the reduced total anthocyanin accumulation in the JP314 mutant
is due to the lack of myricetin-derived anthocyanins. If this hypothesis is supported,
examining expression and sequence of the flavonoid 3',5'-hydroxylase gene will be
helpful to test the hypothesis that this gene is altered in JP314 mutant.

In summary, the combination of analytical chemistry, forward genetics and statistical
tests identified tomato EMS mutants altered in trichome non-volatile metabolites.

	
  

70

Dihydrokaempferol

Kaempferol

Kaempferol
conjugates

Dihydroquercetin

Quercetin

Quercetin
conjugates

Dihydromyricetin

Myricetin

Kaempferol

Quercetin

Anthocyanins

Myricetin
conjugates

Myricetin

Figure 2.9. Overview of flavonoid and anthocyanin biosynthetic pathways.
Schematic flavonoid and anthocyanin biosynthetic pathways are shown with grey boxes
indicating reduced amounts in JP314 mutant. Structures are presented for the flavonols
kamepferol, quercetin and myricetin.

	
  

71

Furthermore, this chemical analysis of mutant plants improved structural annotation of
incompletely characterized metabolites and led to the generation of testable hypotheses.

Materials and methods
Plant growth conditions
M82 cultivated tomato seeds were obtained from the Tomato Genetic Resource Center
(http://tgrc.ucdavis.edu/). Tomato EMS M2 mutant population was generously provided
from Dr. Dani Zamir group (The Hebrew Univeristy of Jerusalem, Faculty of
Agriculture, Rehobot, Israel). Thirty-two plants including one or two of each M82 wild
type and IL control plants were grown in a single flat for 3 weeks. The plants were
randomly located within the flat. Then, chemical assays were conducted on leaves from
the plants except ones having poor growth. Identical growth conditions and chemical
assay methods were applied as described in (Schilmiller et al. 2010a)

Chemical extraction method and LC-ToF-MS analysis
A single leaf dip extraction method and rapid LC-ToF-MS analysis with MUX-CID was
performed for mutant screening as described in (Schilmiller et al. 2010a).

Highly concentrated trichome extracts were analyzed by LC-ToF-MS with 20 min LC
performance for structural characterization of the flavonol tetraglycoside (m/z 949.3).
Because the accumulation of the metabolite was higher in M82 than JP314 mutant, M82
leaves were used for chemical extraction. Ten of the youngest leaves from a single M82
plant were collected and they were placed into 10 ml of extraction solvent used for a

	
  

72

single leaf dip method. The incubation was performed for 1 min with a gentle rocking.
The same procedure was repeated in 10 M82 plant batches, for a total of 100 leaves in
100 ml of solvent. The solvent from plant extracts was brought to 10 ml by rotary
evaporation at room temperature. The 10 ml of concentrated plant extract was transferred
to a glass test tube, evaporated to dryness under a liquid nitrogen stream and re-dissolved
in 100 µl of extraction solvent. This highly concentrated plant extract was analyzed by 20
min LC method. The gradient of mobile phase was also optimized for better separation of
flavonoids. Two solvents (solvent A – 0.15% formic acid in MilliQ water, solvent B –
methanol) were applied as mobile phases in LC with 0.4 ml/min flow rate. Gradient
profile of mobile phase was as follows: 1% B for the first 0.5 min; linear gradient to 60%
B to 12 min; 85% B to 15 min; 100% B maintained to 18 min; 1% B to 20 min.

Measurement of total anthocyanin levels
Crude leaf extraction was performed with 0.3% hydrochloric acid in methanol to extract
total anthocyanins (Kong et al. 2003). A single leaflet was collected from the identical
position where plant tissues were obtained for trichome non-volatile metabolite profiling.
The leaf was placed into 1 ml of anthocyanin extraction solvent and incubated for 24 hrs
with gentle rocking. The amount of total anthocyanins was estimated by measuring light
absorbance of the plant extracts at the specific wavelength. The light absorbance was
measured at 530 nm (A530) and 657 nm (A657) to evaluate the accumulation of total
anthocyanins and chlorophyll degradation products, respectively. Finally, A530 - 0.25 x
A657 (Spitzer-Rimon et al. 2010) was calculated as the indicative of total anthocyanin
levels.

	
  

73

Targeted data analysis
Peak areas were integrated for the individual 35 analytical signals using QuanLynx
software (Waters, http://www.waters.com/) and normalized to internal standard and leaf
dry weight. The normalized dataset was used for targeted data analysis.

Median and median absolute deviation (MAD) were calculated as described in (Lu et al.
2008), to look for outliers altered in the 35 analytical signals in M2 screening. Every
comparative analysis and statistical test was performed with plants grown in a same flat
to control for flat difference. The median and MAD were used instead of mean and
standard deviation, respectively, because the median is less likely to be affected by
extreme values, for example, outliers, compared to mean. In M2 screening, two serial
cutoffs were applied to identify putative mutants. First, median +/- 5 MAD were used to
identify outliers. Once outliers were selected, it was tested whether the amount of specific
analytical signal in the outliers has > 3-fold change compared to the median. When the
outliers met two serial cutoffs, median +/- 5 MAD and > 3-fold change, they were
considered M2 putative mutants.

Non-targeted data analysis
The signal intensity of all MS spectra ranging from 50 to 1,500 Da was obtained using
Markerlynx software (Waters, http://www.waters.com/) and normalized to internal
standard and leaf dry weight. The normalized dataset was used to perform statistical tests
including PCA, PLS-DA and correlation analysis, using MetaboAnalyst webtools (Xia et
al. 2009). PCA and PLS-DA were performed with various conditions such as

	
  

74

with/without filtering and/or scaling. Median intensity and log transformation was used
for filtering and scaling respectively when applied. Variable Importance in Projection
(VIP) was ranked based on VIP mean, the indicative of the significance of the variables
for the separation in PLS-DA (Xia et al. 2009). Pearson and Spearman correlation
analyses were also performed to search for analytical signals whose behavior was
associated with the flavonol tetraglycoside with m/z 949.3.

Chemical analysis of M3 progeny plants
Seeds were collected from the self-pollinated M3 progeny of the individual M2 putative
mutants. Identical growth conditions and chemical assay methods were applied to M3
progeny test as M2 mutant screening.

	
  

75

References

	
  

76

References
Besser, K., Harper, A., Welsby, N., Schauvinhold, I., Slocombe, S., Li, Y., Dixon,
R.A. and Broun, P. (2009) Divergent regulation of terpenoid metabolism in the
trichomes of wild and cultivated tomato species. Plant Physiol, 149, 499-514.
Buttner, B., Abou-Elwafa, S.F., Zhang, W., Jung, C. and Muller, A.E. (2010) A
survey of EMS-induced biennial Beta vulgaris mutants reveals a novel bolting
locus which is unlinked to the bolting gene B. Theor. Appl. Genet., 121, 11171131.
Falara, V., Akhtar, T.A., Nguyen, T.T., Spyropoulou, E.A., Bleeker, P.M.,
Schauvinhold, I., Matsuba, Y., Bonini, M.E., Schilmiller, A.L., Last, R.L.,
Schuurink, R.C. and Pichersky, E. (2011) The tomato terpene synthase gene
family. Plant Physiol, 157, 770-789.
Fobes, J.F., Mudd, J.B. and Marsden, M.P.F. (1985) Epicuticular lipid accumulation
on the leaves of Lycopersicon pennellii (Corr.) D'Arcy and Lycopersicon
esculentum Mill. Plant Physiol., 77, 567-570.
Fridman, E., Wang, J., Iijima, Y., Froehlich, J.E., Gang, D.R., Ohlrogge, J. and
Pichersky, E. (2005) Metabolic, genomic, and biochemical analyses of glandular
trichomes from the wild tomato species Lycopersicon hirsutum identify a key
enzyme in the biosynthesis of methylketones. Plant Cell, 17, 1252-1267.
Gu, L., Jones, A.D. and Last, R.L. (2010) Broad connections in the Arabidopsis seed
metabolic network revealed by metabolite profiling of an amino acid catabolism
mutant. Plant J., 61, 579-590.
Hanhineva, K., Rogachev, I., Kokko, H., Mintz-Oron, S., Venger, I., Karenlampi, S.
and Aharoni, A. (2008) Non-targeted analysis of spatial metabolite composition
in strawberry (Fragaria x ananassa) flowers. Phytochemistry, 69, 2463-2481.
Jaakola, L., Maatta, K., Pirttila, A.M., Torronen, R., Karenlampi, S. and Hohtola,
A. (2002) Expression of genes involved in anthocyanin biosynthesis in relation to
anthocyanin, proanthocyanidin, and flavonol levels during bilberry fruit
development. Plant Physiol, 130, 729-739.
Jander, G., Baerson, S.R., Hudak, J.A., Gonzalez, K.A., Gruys, K.J. and Last, R.L.
(2003) Ethylmethanesulfonate saturation mutagenesis in Arabidopsis to determine
frequency of herbicide resistance. Plant Physiol, 131, 139-146.
Jander, G., Norris, S.R., Rounsley, S.D., Bush, D.F., Levin, I.M. and Last, R.L.
(2002) Arabidopsis map-based cloning in the post-genome era. Plant Physiol.,
129, 440-450.

	
  

77

Kang, J.H., Liu, G., Shi, F., Jones, A.D., Beaudry, R.M. and Howe, G.A. (2010) The
tomato odorless-2 mutant is defective in trichome-based production of diverse
specialized metabolites and broad-spectrum resistance to insect herbivores. Plant
Physiol., 154, 262-272.
Khersonsky, O., Roodveldt, C. and Tawfik, D.S. (2006) Enzyme promiscuity:
evolutionary and mechanistic aspects. Curr. Opin. Chem. Biol., 10, 498-508.
Kong, J.M., Chia, L.S., Goh, N.K., Chia, T.F. and Brouillard, R. (2003) Analysis and
biological activities of anthocyanins. Phytochemistry, 64, 923-933.
Krieger, U., Lippman, Z.B. and Zamir, D. (2010) The flowering gene SINGLE
FLOWER TRUSS drives heterosis for yield in tomato. Nat. Genet., 42, 459-463.
Last, R.L., Jones, A.D. and Shachar-Hill, Y. (2007) Towards the plant metabolome and
beyond. Nat. Rev. Mol. Cell. Biol., 8, 167-174.
Liljegren, S.J., Ditta, G.S., Eshed, Y., Savidge, B., Bowman, J.L. and Yanofsky,
M.F. (2000) SHATTERPROOF MADS-box genes control seed dispersal in
Arabidopsis. Nature, 404, 766-770.
Lu, Y., Savage, L.J., Ajjawi, I., Imre, K.M., Yoder, D.W., Benning, C., Dellapenna,
D., Ohlrogge, J.B., Osteryoung, K.W., Weber, A.P., Wilkerson, C.G. and
Last, R.L. (2008) New connections across pathways and cellular processes:
industrialized mutant screening reveals novel associations between diverse
phenotypes in Arabidopsis. Plant Physiol., 146, 1482-1500.
Luckwill, L. (1943) The genus Lycopersicon: A historical, biological, and taxonomic
survey of the wild and cultivated tomatoes Aberdeen, Scotland: Aberdeen
University Press.
Marti, L., Stefano, G., Tamura, K., Hawes, C., Renna, L., Held, M.A. and Brandizzi,
F. (2010) A missense mutation in the vacuolar protein GOLD36 causes
organizational defects in the ER and aberrant protein trafficking in the plant
secretory pathway. Plant J., 63, 901-913.
McConkey, M.E., Gershenzon, J. and Croteau, R.B. (2000) Developmental regulation
of monoterpene biosynthesis in the glandular trichomes of peppermint. Plant
Physiol, 122, 215-224.
McDowell, E.T., Kapteyn, J., Schmidt, A., Li, C., Kang, J.H., Descour, A., Shi, F.,
Larson, M., Schilmiller, A., An, L., Jones, A.D., Pichersky, E., Soderlund,
C.A. and Gang, D.R. (2011) Comparative functional genomic analysis of
Solanum glandular trichome types. Plant Physiol., 155, 524-539.

	
  

78

Menda, N., Semel, Y., Peled, D., Eshed, Y. and Zamir, D. (2004) In silico screening of
a saturated mutation library of tomato. Plant J., 38, 861-872.
Niessen, W.M., Manini, P. and Andreoli, R. (2006) Matrix effects in quantitative
pesticide analysis using liquid chromatography-mass spectrometry. Mass.
Spectrom. Rev., 25, 881-899.
Peters, J.L., Cnudde, F. and Gerats, T. (2003) Forward genetics and map-based
cloning approaches. Trends. Plant. Sci., 8, 484-491.
Sallaud, C., Rontein, D., Onillon, S., Jabes, F., Duffe, P., Giacalone, C., Thoraval, S.,
Escoffier, C., Herbette, G., Leonhardt, N., Causse, M. and Tissier, A. (2009)
A novel pathway for sesquiterpene biosynthesis from Z,Z-farnesyl pyrophosphate
in the wild tomato Solanum habrochaites. Plant Cell, 21, 301-317.
Sasaki, T., Matsumoto, T., Yamamoto, K., Sakata, K., Baba, T., Katayose, Y., Wu,
J., Niimura, Y., Cheng, Z., Nagamura, Y., Antonio, B.A., Kanamori, H.,
Hosokawa, S., Masukawa, M., Arikawa, K., Chiden, Y., Hayashi, M.,
Okamoto, M., Ando, T., Aoki, H., Arita, K., Hamada, M., Harada, C.,
Hijishita, S., Honda, M., Ichikawa, Y., Idonuma, A., Iijima, M., Ikeda, M.,
Ikeno, M., Ito, S., Ito, T., Ito, Y., Iwabuchi, A., Kamiya, K., Karasawa, W.,
Katagiri, S., Kikuta, A., Kobayashi, N., Kono, I., Machita, K., Maehara, T.,
Mizuno, H., Mizubayashi, T., Mukai, Y., Nagasaki, H., Nakashima, M.,
Nakama, Y., Nakamichi, Y., Nakamura, M., Namiki, N., Negishi, M., Ohta,
I., Ono, N., Saji, S., Sakai, K., Shibata, M., Shimokawa, T., Shomura, A.,
Song, J., Takazaki, Y., Terasawa, K., Tsuji, K., Waki, K., Yamagata, H.,
Yamane, H., Yoshiki, S., Yoshihara, R., Yukawa, K., Zhong, H., Iwama, H.,
Endo, T., Ito, H., Hahn, J.H., Kim, H.I., Eun, M.Y., Yano, M., Jiang, J. and
Gojobori, T. (2002) The genome sequence and structure of rice chromosome 1.
Nature, 420, 312-316.
Schilmiller, A., Shi, F., Kim, J., Charbonneau, A.L., Holmes, D., Daniel Jones, A.
and Last, R.L. (2010a) Mass spectrometry screening reveals widespread
diversity in trichome specialized metabolites of tomato chromosomal substitution
lines. Plant J., 62, 391-403.
Schilmiller, A.L., Last, R.L. and Pichersky, E. (2008) Harnessing plant trichome
biochemistry for the production of useful compounds. Plant J., 54, 702-711.
Schilmiller, A.L., Miner, D.P., Larson, M., McDowell, E., Gang, D.R., Wilkerson, C.
and Last, R.L. (2010b) Studies of a biochemical factory: tomato trichome deep
expressed sequence tag sequencing and proteomics. Plant Physiol., 153, 12121223.
Schilmiller, A.L., Schauvinhold, I., Larson, M., Xu, R., Charbonneau, A.L.,
Schmidt, A., Wilkerson, C., Last, R.L. and Pichersky, E. (2009) Monoterpenes

	
  

79

in the glandular trichomes of tomato are synthesized from a neryl diphosphate
precursor rather than geranyl diphosphate. Proc. Natl. Acad. Sci. U S A, 106,
10865-10870.
Schmidt, A., Li, C., Shi, F., Jones, A.D. and Pichersky, E. (2011) Polymethylated
myricetin in trichomes of the wild tomato species Solanum habrochaites and
characterization of trichome-specific 3'/5'- and 7/4'-myricetin Omethyltransferases. Plant Physiol., 155, 1999-2009.
Spitzer-Rimon, B., Marhevka, E., Barkai, O., Marton, I., Edelbaum, O., Masci, T.,
Prathapani, N.K., Shklarman, E., Ovadis, M. and Vainstein, A. (2010) EOBII,
a gene encoding a flower-specific regulator of phenylpropanoid volatiles'
biosynthesis in Petunia. Plant Cell, 22, 1961-1976.
Stevens, J., Hart, H. and Wollenweber, E. (1995) The systematic and evolutionary
significance of exudate flavonoids in Aeonium. Phytochemistry, 39, 805-813.
Sumner, L.W., Mendes, P. and Dixon, R.A. (2003) Plant metabolomics: large-scale
phytochemistry in the functional genomics era. Phytochemistry, 62, 817-836.
Tikunov, Y., Lommen, A., de Vos, C.H., Verhoeven, H.A., Bino, R.J., Hall, R.D. and
Bovy, A.G. (2005) A novel approach for nontargeted data analysis for
metabolomics. Large-scale profiling of tomato fruit volatiles. Plant Physiol., 139,
1125-1137.
Turner, G.W., Gershenzon, J. and Croteau, R.B. (2000) Development of peltate
glandular trichomes of peppermint. Plant Physiol., 124, 665-680.
Vision, T.J., Brown, D.G. and Tanksley, S.D. (2000) The origins of genomic
duplications in Arabidopsis. Science, 290, 2114-2117.
Wagner, G.J. (1991) Secreting glandular trichomes: more than just hairs. Plant Physiol.,
96, 675-679.
Westerhuis, J.A., Hoefsloot, H.C.J., Smit, S., Vis, D.J., Smilde, A.K., van Velzen,
E.J.J., van Duijnhoven, J.P.M. and van Dorsten, F.A. (2007) Assessment of
PLSDA cross validation. Metabolomics, 4, 81-89.
Williams, C. and Grayer, R. (2004) Anthocyanins and other flavonoids. Nat. Prod.
Rep., 21, 539-573.
Wink, M. (1999) Biochemistry of plant secondary metabolism edn. Sheffield, England:
Sheffield Academic Press.

	
  

80

Wong, M.C., Lee, W.T., Wong, J.S., Frost, G. and Lodge, J. (2008) An approach
towards method development for untargeted urinary metabolite profiling in
metabonomic research using UPLC/QToF MS. J. Chromatogr. B., 871, 341-348.
Xia, J., Psychogios, N., Young, N. and Wishart, D.S. (2009) MetaboAnalyst: a web
server for metabolomic data analysis and interpretation. Nucleic. Acids. Res., 37,
W652-660.
Yu, G., Nguyen, T.T., Guo, Y., Schauvinhold, I., Auldridge, M.E., Bhuiyan, N., BenIsrael, I., Iijima, Y., Fridman, E., Noel, J.P. and Pichersky, E. (2010)
Enzymatic functions of wild tomato methylketone synthases 1 and 2. Plant
Physiol., 154, 67-77.

	
  

81

Chapter 3. Analysis of genetic variants of Solanum lycopersicum with altered
trichome methylated flavonoids

Abstract
Flavonoids are ubiquitous plant aromatic specialized metabolites found in a variety of
cell types and organs. Glycosylated and methylated flavonoids are found in various
Solanum plants. For example, Solanum lycopersicum M82 secreting glandular trichomes
(SGTs) accumulate both glycosylated and methylated myricetin flavonol compounds. As
part of a genetic dissection of SGT specialized metabolites in M82, random EMS mutants
and a collection of S. lycopersicum X S. pennellii (LA0716) introgression lines (ILs)
were screened for changes in flavonoid types. The JP117 EMS mutant was identified as
having reduced amounts of di- and tri-methylated myricetin, and increased monomethylated myricetin. The recessive mutation in JP117 was mapped to the bottom of
tomato chromosome 6. In contrast, introduction of the region of S. pennellii chromosome
6 defined by introgression lines (ILs) 6-3 and 6-4 caused a shift toward more highly
methylated myricetin. The overlapping introgressed region of the S. pennellii
chromosome in 6-3 and 6-4 is completely contained within the chromosomal region for
the JP117 mutant gene. Within the targeted chromosomal region, two S. lycopersicum
homologous genes for myricetin O-methyltransferases were identified, slMOMT1a and
1b, and are candidate genes for the production of Me-Ms. From insect feeding assays in
the field and green house, JP117 mutant showed higher susceptibility and IL6-4 more
resistance than M82 plants against Manduca sexta, tobacco hornworm. This trend was
consistently discovered against the flea beetle Epitrix cucumeris from an independent

	
  

82

field experiment, suggesting an impact of altered Me-M profiles on plant defense against
insect herbivores.

Introduction
Secretory glandular trichomes accumulate a wide variety of plant specialized metabolites
(Schilmiller et al. 2008, Wagner 1991). These chemical compounds protect plants against
UV-B radiation, water loss, heat stress (Wagner et al. 2004) and attack from herbivorous
insects (Kang et al. 2010a, Kang et al. 2010b, Weinhold and Baldwin 2011).
Furthermore, trichome metabolites contribute to food flavor and aroma (Gang 2005,
Wang et al. 2008) and are used as therapeutic drugs (Liu et al. 2010, van Der Heijden et
al. 2004). Due to these benefits, trichome metabolites have been intensively studied in
various plants (Schilmiller et al. 2008).

Flavonoids and structurally related compounds are examples of plant non-volatile
specialized metabolites, collectively including chalcones, flavonones, flavones, flavonols
and anthocyanins (Buer et al. 2010). The first step in flavonoid biosynthesis is chalcone
production by chalcone synthase using acetate-derived malonyl-CoA and p-coumarylCoA as substrates (Figure 1.2 from Chapter 1. Introduction). Chalcone is used as a
precursor to produce diverse flavonoids by a series of reactions such as oxidation,
reduction and isomerization (Buchana et al. 2002).

Among various flavonoids, flavonols such as kaempferol, quercetin and myricetin are
relatively abundant in plants (Heim et al. 2002, Stevens et al. 1995). These flavonols are

	
  

83

typically found as modified compounds rather than unmodified flavonol molecules in
plants (Stevens et al. 1995). For example, glycosylated and/or methylated flavonols were
identified from various plant cell types, including trichomes of the cultivated tomato and
its wild relatives (Braca et al. 2001, McDowell et al. 2011, Schilmiller et al. 2010a,
Schmidt et al. 2011, Stevens et al. 1995). Chemical modifications are known to affect the
chemical properties of flavonoids. For instance, O-methylation at hydroxyl group of
flavonoids makes the compounds more hydrophobic and changes their reactivity and
accessibility (Heim et al. 2002).

Previous studies identified plant flavonoid O-methyltransferases and characterized their
enzymatic activities (Kim et al. 2005, Muzac et al. 2000). These results commonly
showed regio-specificity of the enzymes (Kim et al. 2005, Muzac et al. 2000). Two
myricetin O-methyltransferases recently characterized from the wild species of tomato S.
habrochaites (LA1777) also exhibited regio-specificity (Schmidt et al. 2011). The regiospecificity of flavonoid O-methyltransferase provides a potential explanation of how
methylated flavonoids are produced with methyl groups at different positions. This shows
that the identification and characterization of flavonoid O-methyltransferases enzymatic
activities facilitates understanding of methylated flavonoid production.

The availability of various genetic resources makes tomato plants appealing model
systems for studies of many biological processes. For example, with EMS mutant
population (See 'Chapter 2' for the details), and chromosomal substitution lines such as
introgression lines (ILs) are highly useful plant resources. The commonly used S.

	
  

84

pennellii LA0716 ILs replace single regions of the S. lycopersicum M82 genome with a
homologous region from S. pennellii LA0716. When phenotypes are found in specific
ILs, locus identification can be efficient because individual S. pennellii ILs carry a
genetically characterized fragment of the S. pennellii chromosome. The identification of a
novel trichome monoterpene biosynthetic pathway is a recent example of the utility of
these ILs in gene discovery (Schilmiller et al. 2009).

Physiological and ecological studies revealed that plant specialized metabolites are
important for plant defense against herbivore insects. For instance, volatile trichome
terpenes were shown to improve plant defense against insects in local plants (Weinhold
and Baldwin 2011) and alarm neighbor plants in the community (Baldwin et al. 2006).
Flavonoids were also discovered to affect insect feeding (Onyilagha et al. 2004).
Specifically, flavonols inhibit insect cytochrome P450 dependent steroid hydroxylase
activity in a concentration-dependent and species-specific manner (Mitchell et al. 1993).

In this study, EMS mutants and S. lycopersicum X S. pennellii 0716 ILs with altered
degree of myricetin methylation were identified and characterized. While di-methylated
myricetin (di-Me-M) was the most abundant in the M82 cultivated tomato, the JP117
mutant and IL6-3/4 accumulated predominantly mono-Me-M and tri-Me-M, respectively.
Genetic analysis revealed that JP117 behaves as a single recessive mutation and the S.
pennellii allele is dominant in IL6-3/4. Genetic mapping placed the JP117 locus on the
bottom of chromosome 6, in the same region of the S. pennellii chromosome in IL6-3/4.

	
  

85

Insect feeding suggested an impact of altered Me-Ms on plant resistance against the
insects Manduca sexta and Epitrix cucumeris.

Results
Phenotypic characterization of altered Me-M profiles in the JP117 EMS mutant
Seven mutants with increased ratios of mono- to di-Me-M (JP117, 130, 256, 334, 446,
3045 and 3060) were identified from the tomato EMS mutant screen described in Chapter
2 (See Table 2.3 from Chapter 2). Three types of Me-Ms, mono-, di- and tri-Me-M, were
detected in M82 cultivated tomato analysis of single leaf dip trichome extracts using
liquid chromatography-time of flight-mass spectrometry (LC-ToF-MS) in negative ion
mode (Figure 3.1). While the normalized ion peak area for di-Me-M was the most
abundant in M82, mono- and tri-Me-M were also detected. In contrast, the relative
proportions were shifted in the JP117 mutant such that mono-Me-M was predominant,
with lower amount of di- and tri-Me-Ms (Figure 3.1a). The total accumulation of Me-Ms
was also significantly decreased in the JP117 mutant (approximately 64% of M82, pvalue < 0.0005 from Student's t-test) (Figure 3.1b).

Negative ion mode LC-ToF-MS analysis of single leaf dip extracts permits rapid analysis
of mono-, di- and tri-Me-Ms. However, this method provides limited molecular structure
information and might extract non-trichome epidermal cell compounds. Chao Li, Daniel
Jones and I used the complementary LC-MS/MS positive ion mode analysis of individual
trichome types. This approach revealed several similarities to the LC-ToF-MS method

	
  

86

Peak area/ IS/ leaf dry weight

(a)

80
70
60
50
1Me-M

40

2Me-M

30

3Me-M

20
10
0
JP117

(b)

M82

Genotypes

Total Me-Ms

JP117

73.3 ± 4.8

M82

p-value

115.4 ± 9.1

0.000213

Figure 3.1. Increased mono-Me-M and decreased di- and tri-Me-M accumulation in the
JP117 EMS mutant identified from LC-ToF-MS analysis of single leaf dip extracts.

	
  

87

Figure 3.1. Continued
(a) Compared to the M82 wild type plant, the JP117 mutant (M3 plant) accumulated an
increased amount of mono-Me-M with decreased di- and tri-Me-M levels. Average of
peak areas normalized to internal standard and leaf dry weight are presented with
standard error of the mean (n=24 for JP117, 21 for M82). 1Me-M, 2Me-M and 3Me-M
indicates mono-, di- and tri-Me-M respectively. An identical system is used in subsequent
figures describing LC-ToF results, unless specified otherwise.
(b) The total amount of Me-Ms is decreased in the JP117 mutant to approximately 64%
that of M82 (p-value < 0.0005 from Student's t-test). The total Me-M level was calculated
by summing normalized peak areas of all three Me-Ms from samples used in (a).

	
  

88

and to published LC-MS/MS data for S. habrochaites LA1777 (Figure 3.3; Schmidt et al.
2011). The JP117 mutant trichomes accumulated a higher amount of mono-Me-M having
a methyl group at position 3 [mono-Me-M (3)], than was detected in M82 (See Figure 3.2
for the flavonoid numbering system). In contrast, di-Me-M (3,3') levels were
approximately 14-fold higher in M82 trichomes than in JP117 (Figure 3.3). Thus, LCMS/MS positive ion mode analysis of isolated trichomes revealed the same increased
ratio of mono- to di-Me-M in JP117 trichomes compared with M82 as originally
identified by negative ion mode LC-ToF-MS analysis of single leaf dip extracts (Figure
3.1 and 3.3). Furthermore, higher accumulation of Me-Ms was found in type 1 glands
than type 6 for the JP117 mutant and M82 (Figure 3.3), consistent with recently
published data that Me-Ms are enriched in type 1 glands of S. habrochaites LA1777
(Schmidt et al. 2011).

Figure 3.2. The chemical structure of unmodified myricetin.
The structure of the flavonol myricetin is shown with the numbering system indicated.
The hydroxyl groups of myricetin are typically modified in vivo, for example, by
methylation or glycosylation.

	
  

89

Methylated myricetin (µM)

6
5
4
1Me-M (3)

3

2Me-M (3,3')

2

3Me-M (3,3',7)

1
0
JP117 T1

JP117 T6

M82 T1

M82 T6

Figure 3.3. The Me-M levels measured in isolated glands of types 1 and 6 trichomes from
JP117 mutant and M82 plants.
LC-MS/MS analysis of trichome extracts revealed higher accumulation of Me-Ms in type
1 (T1) glands than type 6 (T6) from both the JP117 mutant and M82 wild type plants.
While 1Me-M levels were higher in JP117 mutant trichomes than M82, 2Me-M was
higher in M82 trichomes than JP117. LC-MS/MS also revealed the position of methyl
groups for individual Me-Ms. For instance, 3Me-M (3,3’,7) has methyl groups on at the
3, 3’ and 7 positions (See Figure 3.2 for the numbering system). The amount of
methylated myricetin (µM) is presented with the average and standard error (See
'Materials and Methods' for metabolite detection and identification). Fifty glands of
specific types of trichomes were collected from each of three biological replicate plants.

	
  

90

In addition to validating previously identified chemical profiles, the LC-MS/MS analysis
expands our understanding about Me-M phenotypes in two ways. First, it provides
information, regarding the position of methyl groups on Me-Ms. Second, the use of
commercially available methylated myricetins as a flavonoid calibration standard (See
'Materials and Methods' for the details) provided corroboration for the relative
concentrations of Me-Ms in M82 and JP117 compared with quantititave Me-M profiles
obtained by normalized peak areas under LC-ToF-MS method.

Genetic analysis of JP117
To test whether JP117 behaves as a recessive monogenic mutation, it was backcrossed to
the parental M82 wild type. Backcross F1 progeny plants (BC F1) were analyzed by LCToF-MS and their Me-M profiles compared with those of JP117 and M82. The Me-M
profiles in BC F1 were similar to those of M82 in that di-Me-M was most abundant in
every F1 plant (Figure 3.4a). These results are consistent with the hypothesis that the
JP117 mutation is recessive to wild type. However, while di-Me-M is restored to wild
type levels, the overall F1 phenotype differs from M82 in that it accumulated mono-MeM intermediate between that of JP117 and M82 (Figure 3.4a). This gene dosage effect
suggests a defect in a structure gene(s) rather than a regulatory gene(s) in JP117.

BC F2 progeny were analyzed to test whether the JP117 mutation segregates as a
monogenic Mendelian trait. The ratio of di- to mono-Me-M (di-Me-M/ mono-Me-M) was
used as a diagnostic phenotype. Of 23 BC F2 plants, 19 (82.6%) showed similar profiles

	
  

91

(b)

Peak area/ IS/ leaf dry weight
Peak area/ IS/ leaf dry weight

(a)

(c)

140
120
100
80
60
40
20
0

1Me-M
2Me-M
3Me-M
BC F1
BC1 F1

JP117

M82

200
150
1Me-M

100

2Me-M

50

3Me-M

0
JP117

BC1F2 (1)
BC F2 (1)

BC1 F2 (2)
BC F2 (2)

M82

2Me-M / 1Me-M

5
4
3
2
1
0
1

3

5

7

9

JP117
JP117 BC1 BC (1)
F2

11 13 15 17 19 21 23 25 27 29 31
BC F2F2 (2)
M82
BC1 (2)
M82

F2 (1)

Figure 3.4. Evidence that a single recessive mutation is responsible for the JP117 mutant
phenotype.

	
  

92

Figure 3.4. Contined
(a) Me-M profiles in trichomes of M82 X JP117 BC F1 plants are generally similar to
those of M82, with levels of 1Me-M intermediate between M82 and JP117 mutant. Four
replicate plants were assayed for each genotype.
(b) Averages of classes of F2 BC plants. Of 23 BC F2 plants, four labeled as BC F2 (1)
predominately accumulated 1Me-M, while 19 labeled as BC F2 (2) 2Me-M. The
grouping of BC F2 plants was based on the individual chemical profiles from (c). Four
replicate plants were tested for JP117 and M82 respectively.
(c) The ratio of di- to mono-Me-M (2Me-M/1Me-M) is shown for individual plants tested
in (b). Although BC F2 plants showed continuous values for the ratio, the 2Me-M/1MeM ratios of the 4 plants from BC F2 (1) were consistently as low as the JP117 mutant and
the ratios of 19 plants from BC F2 (2) were as high as M82.

	
  

93

as M82, while four samples (17.4%) resembled JP117 mutant (Figure 3.4bc). Although
the total number of progeny tested is low, Chi-square test revealed that these numbers fit
the segregating ratio expected for monogenic Mendelian trait at p-value of 0.05. These
results suggested that the JP117 mutation is recessive and segregates as a monogenic
recessive trait.

To test whether JP117 behaves as a single locus, genetic mapping was performed (Jander
et al. 2002). Because the genetic diversity of S. lycopersicum is relatively low
(Rodriguez-Saona et al. 2011), a polymorphic mapping population was obtained by
crossing JP117 to the wild species S. pennellii LA0716. S. pennellii LA0716 was chosen
because it is sexually compatible with the cultivated tomato and there are large numbers
of DNA polymorphisms between the two species. Chemical profiles were analyzed for
289 outcross (OC) F2 plants by LC-ToF-MS. Because no Me-Ms were detected in
LA0716 (data not shown), Me-M profiles of OC F2 were compared with those of JP117
and M82. The ratio of di- to mono-Me-M was highly variable in the OC F2 mapping
population, ranging from 0.09 to 203.97 (corresponding to -1.05 and 2.31 respectively on
a log10 scale, as presented in Figure 3.5a). The ratio was noticeably lower in the JP117
mutant than M82 despite the value varied among replicate plants for each genotype. The
ratio of di-Me-M/mono-Me-M ranged from 0.15 to 0.60 in JP117 mutant and 3.45 to 5.01
in M82 (-0.82 to -0.22 and 0.54 to 0.70 respectively with log10 scale in Figure 3.5a),
revealing that even the highest ratio from JP117 was lower than the lowest from M82.
The highest ratio among the JP117 mutant replicates 0.60 (-0.22 with log10 scale in

	
  

94

(a)
Log10 (2Me-M / 1Me-M)

2.5

***

2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5

JP117

(b)

OC F2

M82

Chr 6

1Marker

1

2Genetic

3Physical 4No. of plants with

Location
(cM)

Location
S. lycopersicum
(Mbp)
homozygous allele

1

46.0

-

8

B

74.5

-

19

4

At1g29900

80.7

-

21

5

sc06019

-

44.24

21

sc03622-1

-

44.30

21

sc03622-2

-

44.52

27

8

sc03622-3

-

44.52

28

9

sc03622-4

-

44.58

28

10

sc03622-5

-

44.83

30

11

sc03622-6

-

44.83

30

12

sc03622-7

-

45.00

30

13

10
12

At1g22850

7

8

2

6

7
9

-

3

3
4
5
6

12.7

2

2

U231369

U146140

97.2

45.08

30

11
13

Figure 3.5. The JP117 mutation maps to the bottom of chromosome 6.

	
  

95

Figure 3.5. Contined
JP117 was crossed to S. pennellii LA0716 and the F2 progeny phenotypically and
genotypically analyzed.
(a) The ratio of 2Me-M/1Me-M is shown for 289 individual F2 plants, as well as the
JP117 mutant and M82 controls. Due to the great variation seen in the interspecific cross,
the ratio is presented as a log scale. Using a cutoff of -0.22, the highest value among the
replicates for the JP117 mutant, identified 30 OC F2 with the JP117 mutant phenotype:
these are labeled with *** and used for genetic mapping in (b).
(b) Results of genotyping of the 30 OC F2 plants selected from (a) localizes the JP117
mutation within the 1.5Mb intervals from the bottom of chromosome 6. Thirteen genetic
markers for chromosome 6 used for genetic mapping

1

(See 'Materials and Methods' for

2

the details). Genetic locations of the markers were obtained from the Sol Genomics
3

Network (SGN; http://solgenomics.net). Physical locations were determined based on
tomato whole genome sequence (http://solgenomics.net/, SL2.31 chromosome sequence).
4

'-' indicates no information was available at SGN. The number of plants with S.
lycopersicum homolozygous alleles (LL) was also shown for the individual markers. All
30 plants showed LL genotype when tested with markers 10 - 13.

	
  

96

Figure 3.5a) was used as a cutoff to identify OC F2 plants having the mutant phenotypes
and 30 F2 plants were selected.

Thirty OC F2 plants that met the criteria for having the JP117 mutant phenotype were
used for genetic mapping (Figure 3.5). First-pass mapping results with genetic markers
throughout the Solanum genome (the SGN, http://solgenomics.net/) suggested that the
JP117 mutation is located on chromosome 6 (data not shown). Fine mapping was
performed with polymorphic genetic markers throughout chromosome 6 from the SGN
and others generated in this study (Figure 3.5b). Genotyping with markers closely linked
to the mutation are expected to show a preponderance of S. lycopersicum homozygous
alleles (LL), rather than heterozygous (LP) or S. penellii homozygous alleles (PP). In
fact, the number of plants with LL alleles increased as genotyping was performed with
genetic markers progressing toward the bottom of chromosome 6. All 30 F2 plants
revealed LL genotype with markers 10 ~ 13, indicating that the JP117 mutant gene is
closely linked to these markers (Figure 3.5b, Table 3.1). The physical locations of
markers 8 and 9 and the end of chromosome 6 are known to be 44.52, 44.58 and 46.04
Mbp, respectively (http://solgenomics.net, Figure 3.5b). Taken together, these genetic
mapping results indicate that the JP117 gene is located within the last 1.5 Mbp at the
bottom of chromosome 6.

Phenotypic characterization of altered Me-M profiles in IL6-3/4
As a parallel project aimed at identification of genes involved in SGT metabolism, LCToF-MS analysis of S. pennellii introgressions (ILs) was performed to identify regions of

	
  

97

Table 3.1. The list of genetic markers generated in this project for JP117 genetic
mapping. Primer sequences and restriction enzymes are shown for individual markers
used for JP117 mutant genetic mapping. See 'Materials and Methods' for details about
genotyping procedures.
Marker
Name

Primer
Type

Primer Sequence
(5'-3')

Restriction
Enzyme

sc06019

F

AAAGGAAAAGGAGC
AGAAGTTAAAGC
TTGGGTCAGATTTAA
CATTTTTCTCACG
TCCGAGTGAATGGA
ACGC
TTAGTAGAAAAGTAC
AGTGTAAAATAGTTG
ATGAGCTTAGTTGGA
GCATTATGG
GAAGGAAGAAGACG
ACCAAAGAAG
TATAATTTGATTTGA
CATTTGTGTTGCTT
CTATGGTCCATTTAA
ACATAAACGTTC
ATGACCGTGGTGTC
CAGG
CCAAAATAAGCAGA
GATGGCTTG
ACCTCACTTGTCATT
TGCCAC
TTCTGATTGTCGCCG
GAAAC
GAAACTGAACATCTT
GATCCAAAAGATT
CACCTATACCTTCTC
CTCCAGCC
ATGGATAATGACTGC
TTCTCTAATGG

EcoRV

R
sc03622-1

F
R

sc03622-2

F
R

sc03622-3

F
R

sc03622-4

F
R

sc03622-5

F
R

sc03622-6

F
R

sc03622-7

F
R

	
  

ATCATCTTCATTGCCC
TTTTGTTC

98

Expected product
size
(bp)
L
P
320,
706
386

PsiI

131,
203

334

MspI

1079

538,
541

ApoI

196,
266

102,
164,
196

ApoI

161,
398

559

HaeIII

297,
878

1175

HaeIII

282,
783

1065

PvuII

428,
445

873

(c)
70

1Me-M
2Me-M

50

3Me-M

6-1

60

40
30
6-2-2

20
10
0
IL6-3

IL6-4

M82

6-2

Peak area/ IS/ leaf dry weight

(a)

(b)
Genotypes 1Me-M (%)

2Me-M (%)

71 cM

3Me-M (%)

23.51

66.87

IL6-4

10.33

24.51

65.16

M82

28.10

53.84

18.05

6-3

9.62

6-4

IL6-3

96 cM
101 cM

Figure 3.6. Higher accumulation of 3Me-M in IL6-3/4 plants and a genetic map of
chromosome 6 ILs.
(a) IL6-3 and 6-4 consistently revealed increased accumulation of 3Me-M. Four replicate
plants were tested for each genotype.
(b) The percentage composition of individual Me-Ms to total Me-Ms was calculated
based on LC-ToF peak area. While 2Me-M was the most abundant signal, representing
54% of total Me-Ms in M82, IL6-3/4 predominately accumulated 3Me-M with up to
65~67% of total Me-M peak area.
(c) The genetic map of IL6-3/4 (Eshed and Zamir 1994, Eshed and Zamir 1995) reveals
that IL6-4 is contained within the IL6-3 region.

	
  

99

the S. pennellii genome that alter S. lycopersicum M82 metabolite profiles (Schilmiller et
al. 2010a). ILs 6-3 and 6-4 caused strong increases in tri-Me-M accumulation compared
with M82 (Figure 3.6a). These two ILs have chromosomal substitutions at the bottom of
chromosome 6 (Figure 3.6c), with IL6-4 completely contained within IL6-3 (Eshed and
Zamir 1994, Eshed and Zamir 1995). In contrast to M82, which accumulated only up to
18% of total Me-Ms as tri-Me-M, IL6-3/4 trichomes accumulated tri-Me-M up to
65~67% of total (Figure 3.6b). The consistent phenotypes in the two overlapping ILs
indicates that gene(s) responsible for differences in trichome myricetin methylation are
located between 96-101 cM at the bottom of chromosome 6 (Figure 3.6c).

To complement the negative ion mode LC-ToF-MS analysis of single leaf dip extracts,
positive ion mode LC-MS/MS analysis of individual trichome types was employed. The
LC-MS/MS approach also revealed higher accumulation of tri-Me-Ms in IL6-3/4
trichomes than M82 confirming IL6-3/4 phenotypes identified by the LC-ToF-MS
method (Figures 3.6 and 3.7). Furthermore, the LC-MS/MS approach expands our
understanding of IL6-3/4 Me-M profiles beyond what was learned from the LC-ToF-MS
analysis. For example, this analysis distinguished two isomers of tri-Me-M: the more
abundant species (3,3',7) is seen in all genotypes while a minor one (3,3',5') is found in
IL6-3/4 samples. Furthermore, the tetra-Me-Ms (3,3',5',7) was detected in the IL6-3/4
samples but not in M82 trichomes, or in any sample under negative ion mode LC-ToFMS analysis of single leaf dip extracts.

	
  

100

Methylated myricetin (µM)

0.7

1Me-M (3)

0.6

2Me-M (3,3')
3Me-M (3,3',7)

0.5

3Me-M (3,3',5')

0.4

4Me-M (3,3',5',7')

0.3
0.2
0.1
0
IL6-3 T1

IL6-3 T6

IL6-4 T1

IL6-4 T6

M82 T1

M82 T6

Figure 3.7. The Me-M levels measured in isolated glands of types 1 and 6 trichomes from
IL6-3/4 and M82 plants.
LC-MS/MS analysis with trichome extracts revealed Me-M profiles including the
position of methyl groups for individual Me-Ms. In contrast to M82, which shows the
enrichment of Me-Ms in type 1 glands, IL6-3/4 reveals relatively high accumulation of
Me-Ms in type 6 glands as well as in type 1. While three Me-Ms were detected from M82
including 1Me-M, 2Me-M and 3Me-M (3,3',7), two additional Me-Ms, 3Me-M (3,3',5')
and 4Me-M (3,3'7,5') were also found in IL6-3/4 glands. See Figure 3.2 for the
numbering system. Compared to M82, IL6-3/4 accumulate higher amount of
polymethylated Me-Ms such as 3Me-Ms and 4Me-M. The amount of methylated
myricetin (µM) is presented with the average and standard error. A sample of fifty glands
of specific types of trichomes was assayed from three independent plants.

	
  

101

Genetic analysis of IL6-4
To ask whether the S. pennellii IL6-4 locus conditions increased myricetin methylation,
M82 X IL6-4 F1 and F2 progeny plants were analyzed. The three F1 progeny exhibited
increased tri-Me-M reminiscent of the IL6-4 parent, consistent with the hypothesis that
the S. pennellii allele is dominant to M82 (Figure 3.8a). F2 progeny analysis gave results
consistent with the hypothesis that the S. pennelli IL6-4 high tri-MeM is caused by a
single dominant gene or multiple linked genes; nine of 12 plants showed the same Me-M
profiles as IL6-4, and the other three F2 plants resembled M82 (Figure 3.8b and c).

The search for candidate genes influencing myricetin methylation in JP117 and IL64
Analysis of the JP117 EMS mutant and IL6-3/4 ILs indicated the presence of genes that
influence myricetin methylation at the bottom of chromosome 6 in S. lycopersicum M82
and S. pennellii LA0716 (Figure 3.9ab). A simple two-part hypothesis is that: 1. JP117 is
mutated in an O-methyltransferase (OMT) that normally functions in conversion of
mono-Me-M to di-Me-M; 2. the S. pennellii IL6-4 region contains an OMT that
efficiently converts di-Me-M to tri-Me-M and even tetra-Me-M, which is absent from or
non-functional in M82 (Figure 3.9).

The map positions of JP117 and IL6-4 are consistent with the hypothesis that there are
two or more OMTs near the bottom of chromosome 6. During the time that this project
was being conducted, progress was made in sequencing and assembly of the S.
lycopersicum Heinz 1706 variety genome (http://solgenomics.net, SL2.31 chromosome

	
  

102

(b)

Peak area/ IS/ leaf dry weight
Peak area/ IS/ leaf dry weight

(a)

(c)

200
1Me-M

150

2Me-M

100

3Me-M

50
0
IL6-4

BC1 F1
F1

M82

200
150

1Me-M
2Me-M
3Me-M

100
50
0
IL6-4

F2 (1)
BC1 F2 (1)

F2 (2)
BC1 F2 (2)

M82

3Me-M / 2Me-M

5
4
3
2
1
0
IL6-4

F2 (1)

Figure 3.8.

	
  

103

F2 (2)

M82

Figure 3.8. Chemical profiles of IL6-4 genetic cross plants suggest that the S. pennellii
allele is dominant for IL6-4 chemical phentoypes and that the phenotypes segregate
genetically.
(a) Me-M profiles in M82 X IL6-4 F1 plant leaf dips are similar to those of IL6-4. Three
replicate plants were tested for each genotype.
(b) The IL6-4 chemical phenotypes segregate in F2 plants. Of 12 F2 plants, 9 labeled as
F2 (1) revealed the IL6-4 phenotype, while 3 labeled as F2 (2) show the M82 phenotype.
The grouping of BC F2 plants is based on the individual chemical profiles from (c).
Three replicate plants were tested for JP117 and M82.
(c) The ratio of tri- to di-Me-M (3Me-M/2Me-M) is shown for individual plants tested in
(b). The 3Me-M/2Me-M ratios of the nine plants from F2 (1) were similarly high as IL64 and those of the three plants from F2 (2) are as low as M82.

	
  

104

(a)

(b)
Chr 6

120
100
80

1Me-M
2Me-M
3Me-M

60
40
20
0
JP117

M82

IL6-4

IL6-4

JP117

OCH3

(c)
OCH3
OCH

?

3
OCH3

?

H3CO

OCH

3
OCH3

OCH3
OCH3

?

Figure 3.9. Summary of the chemical phenotypes and chromosomal locations of the
JP117 mutant and the IL6-4 region.
(a) While the JP117 mutant revealed higher accumulation of 1Me-M and reduced total
Me-M amounts, IL6-4 plants showed increased tri-Me-M accumulation.
(b) The regions associated with JP117 mutant and IL6-4 genes overlap at the bottom of
chromosome 6.
(c) Proposed pathways for the production of Me-Ms in S. lycopersicum variants. The
indicated positions of methyl groups on Me-Ms are based on the LC-MS/MS analysis
performed by Chao Li. The structure of 3Me-M (3,3',7) is shown as it is the most
abundant isomer of 3Me-M (See Figure 3.7 for the details).

	
  

105

sequence) and shotgun sequence data became available for S. pennellii LA0716. These
assemblies were searched for sequences homologous to two trichome-specific myricetin
OMTs from S. habrochaites LA1777 (shMOMT1 and 2) (Pearson and Lipman 1988,
Schmidt et al. 2011, Vingron and Waterman 1994). Five S. lycopercisum MOMTs were
discovered to have high sequence identity to shMOMTs from chromosome 6, including
three genes for shMOMT1 and two for shMOMT2; these were named slMOMT1a, 1b
and 1c and slMOMT2a and 2b, respectively (Figure 3.10). S. pennellii OMT genes were
also identified by BLAST analysis from shotgun DNA sequence using shMOMTs and
slMOMTs sequences. Despite the relatively short genomic sequence assemblies for this
species, six homologous genes were identified: spMOMT1a, 1b, 1c, 2a, 2b and an
additional gene spMOMT1aa (Figures 3.11 and 3.12). All sl- and spMOMTs contained
characteristic plant OMT domains (Ibrahim et al. 1998).

The specific position of each slMOMT genes on chromosome 6 was identified using
genome assembly and confirmed by genetic mapping. The S. lycopersicum genome
assembly

(http://solgenomics.net/,

SL2.31

chromosome

sequences)

contained

slMOMT1a and 1b at the bottom of chromosome 6, in the region containing the JP117
mutation and IL6-4 (Figure 3.10). In contrast, slMOMT1c, slMOMT2a and 2b were
located on the top and middle of chromosome 6 respectively. Because the tomato genome
sequence was found to have mistakes for the terpene synthase gene family, including
genes annotated on incorrect chromosomal regions and hybrid assemblies (Falara et al.
2011), each slMOMT gene was genetically mapped. This was done by identifying a
polymorphic sequence for each gene and looking for the presence of the S. pennellii

	
  

106

6-2-2

6-1

SlMOMT1c

SlMOMT2a
SlMOMT2b
6-2

44,524,611 bp

Cytochrome b5 reductase

SlMOMT1b
SlMOMT1a

WD-40 repeat family

46,041,039 bp
Gene ID

Chromosome Location (bp)

SlMOMT1c

1,816,555 - 1,820,094

SlMOMT2a

36,549,479 - 36,551,598

SlMOMT2b

36,559,878 - 36,561,067

Cytochrome b5 reductase

45,134,673 - 45,136,744

SlMOMT1b

45,137,707 - 45,140,902

SlMOMT1a

45,144,339 - 45,145,525

WD-40 repeat family

45,159,823 - 45,158,037

Figure 3.10.

	
  

107

JP117

SlMOMT1a

6-4

6-3

SlMOMT1b

Figure 3.10. Search for candidate O-methyltransferases (OMTs) based on sequence
homology.
Five myricetin OMT (MOMT) homologous genes were identified on chromosome 6 of S.
lycopersicum and named as slMOMT1a, 1b, 1c, 2a and 2b. Two slMOMTs, slMOMT1a
and 1b, were identified within the target region for JP117 and IL6-4. The chromosomal
locations of five slMOMTs were obtained from the tomato whole genome sequence
(http://solgenomics.net/, SL2.31 chromosome sequence) and confirmed by genetic
mapping. Sequence truncation was found at 5' end of slMOMTa (See Figure 3.11 for
details).

	
  

108

5’ ! 3’

slMOMT1a

spMOMT1a

spMOMT1aa

slMOMT1b

spMOMT1b

Exon

Start codon

Intron

Stop codon

Transposon-like inverted repeat sequence
Lack of sequence due to incomplete whole genome sequence

Figure 3.11. Predicted gene structures of slMOMT1a and 1b and the corresponding S.
pennellii homologs.
Gene models were predicted for slMOMT1a, 1b and S. pennellii homologous genes using
the Softberry computer program (www.softberry.com). slMOMT1a, spMOMT1a and
spMOMT1aa are predicted to have similar gene structures, except for a truncation at the
5' end of slMOMT1a. BLAST analysis with the upstream sequence of this truncated
region revealed transposon-like sequences at the 5' end.

	
  

109

Medicago Sativa (chalcone L10211)
100

81

99

Populus tremuloides (lignin X62096)
Arabidopsis thaliana (flavonol AAB96879)
Chrysosplenium americanum (flavonol U16794)
95
shMOMT1 (flavonol ADZ76433)
78
spMOMT1aa
87
slMOMT1a
100
spMOMT1a
slMOMT1b
spMOMT1b
slMOMT1c
100 spMOMT1c
100 shMOMT2 (flavonol ADZ76434)
100
slMOMT2a
spMOMT2a
slMOMT2b
100
spMOMT2b

100

94

100
86

Cantharanthus roseus (flavonol AY127568)
Cantharanthus roseus (flavonol AAR02419)
Pisum sativum (pterocarpan U69554)
Medicago sativa (isoflavone AF023481)

100

100
92
0.05

Figure 3.12. Evolutionary relationships among various plant OMTs.
Protein sequence was obtained for previously characterized plant OMTs. The detailed
information is presented for individual OMTs including the name of plant species from
which the OMT was characterized, their preferred substrate identified from biochemical
experiments and accession number of the protein sequence. The phylogeny tree was
generated by the neighbor-joining method (Saitou and Nei 1987). Bootstrap testing
(Felsenstein 1985) was performed to validate the reliability of the tree with n=500 and the
result is shown next to each branch. MEGA5 was used to generate trees and for statistical
tests, (Tamura et al. 2011). As expected, slMOMTs and spMOMTs cluster together with
shMOMTs. They were also closely related with other OMTs that use flavonols as
preferred substrates.
	
  

110

LA0716 polymorphism on the four ILs covering chromosome 6 (IL6-1, 2, 3 and 4 in
Figure 3.10). In each case the S. pennellii polymorphism was found associated with the
expected IL (Table 3.2, Figure 3.10).

Gene structures were predicted for slMOMT1a, 1b and the corresponding to S. pennellii
homologs using genomic DNA sequence obtained from the BLAST analysis (Figure
3.11). slMOMT1a, spMOMT1a and spMOMT1aa are predicted to have high similarity
with the exception that slMOMT1a has a 5' end deletion. This truncation would result in
the complete loss of the first exon, while spMOMT1a and 1aa are predicted to encode
functional proteins. BLAST analysis was performed with 10 kb upstream sequence from
the beginning of the second exon of slMOMT1a and a transposon-like repeated inverted
sequence was found at the 5' end of this gene (Figure 3.11).

To test the relationship of slMOMTs with previously characterized OMTs, a phylogenetic
tree was generated with plant OMT protein sequences (Figure 3.12). As expected,
slMOMTs were closely related to shMOMTs and spMOMTs. In addition, slMOMT
genes showed closer relationships with OMTs preferentially using flavonol substrates.

Testing the roles of Me-Ms in plant-herbivorous insect interactions
Trichome metabolites are known to improve plant defense against herbivorous insects by
affecting insect feeding or attracting predators of herbivore insects (Kang et al. 2010a,
Weinhold and Baldwin 2011). To test the hypothesis that altered trichome Me-M profiles
affect plant-insect interactions, an insect feeding assay was performed with JP117

	
  

111

mutant, M82 and IL6-4. The odorless-2 (od-2) mutant was used as a control because it
lacks the major trichome metabolites, including mono-Me-M, and is more susceptible to
insects compared to its isogenic wild type CastleMart (CM) (Kang et al. 2010a).

Because we did not have predictions about which, if any, insects are affected by Me-Ms,
Field experiments were chosen to ask whether herbivores found in mid-Michigan would
respond differentially to the genotypes tested. A Solanaceous specialist Manduca sexta,
tobacco hornworm, was found on plants grown in the field in early September 2010
(Figure 3.13a). The JP117 mutant showed strong damage by the tobacco hornworm
Manduca sexta, whereas IL6-4 was least attacked by M. sexta with M82 damaged at a
level intermediate between JP117 and IL6-4 (Figure 3.13b). The number of insects on
plants was measured at a single time point. The largest number of M. sexta individuals
was found on the JP117 mutant and the lowest on IL6-4 (Figure 3.13c). This result
suggests that the degree of myricetin methylation is inversely correlated with hornworm
herbivory. The plant damage by M. sexta was higher in the od-2 mutant compared to the
CM wild type, consistent with published results (Kang et al. 2010a). No insects were
found on od-2 or CM possibly due to the small population (n=2 for both genotypes).

Given that M. sexta feeding varied on JP117, M82 and IL6-4 from the field experiment,
the feeding assay was repeated under green house conditions. The weight of M. sexta
larvae fed on different genotypes was measured. Of 60 larvae fed on JP117, M82 and
IL6-4, 17, 11 and 9 larvae survived, respectively (Figure 3.14a). Also, insect weight was
approximately 2 fold higher in JP117 than M82. The larvae weight was slightly higher in

	
  

112

(a)

A

*

AB

od-2
ordorless-2

CM

ordorless-2
od-2

CM

IL6-4

B

M82

1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0

JP117

Damage degree

(b)

A

AB
B

113

IL6-4

Serious damage (3)

Figure 3.13

	
  

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

M82

Intermediate
damage (2)

Light damage (1)

JP117

No damage (0)

No. of insects found

(c)

Figure 3.13. Field experiments revealed altered Manduca sexta feeding on S.
lycopersicum variants.
(a) Representative photographs showing that M. sexta caused various degrees of plant
damage to JP117 mutant, M82 and IL6-4 plants. Plant damage was scored from values of
0 to 3, depending on how much plant material was lost to insect herbivory. See 'Materials
and Methods' for details.
(b) M. sexta feeding caused the most serious damage on the JP117 mutant and the least
on IL6-4. One-way ANOVA revealed that the variation of plant damage among JP117,
M82 and IL6-4 is statistically significant (p-value = 0.013). Different lower case letters in
the bar graph indicate statistically different groups. Because the odorless-2 (od-2) mutant
and its wild type Castlemart (CM) control plants are not in the M82 genetic background,
they were not included in the ANOVA test. The identical system is used in subsequent
figures describing the ANOVA test results, unless specified otherwise. As previously
reported (Kang et al. 2010a), od-2 mutant plants were significantly more susceptible than
CM (p-value from student's t-test < 1 x 10-4). Plant damage is presented with average
and standard error (n = 28, 25, 26, 2 and 2 for JP117, M82, IL6-4, ordorless-2 and CM,
respectively).
(c) A single timepoint describing the number of M. sexta visiting plants. This analysis
was performed with the same plants monitored in (b). The largest number of insects was
observed on the JP117 mutant and the lowest on IL6-4 (p-value = 0.024 from one-way
ANOVA test). No M. sexta was found on the od-2 mutant or CM plants, possibly due to
the small population number of plants and single assay time.

	
  

114

(a)

(b)

*

Insect weight (g)

1.00
0.80
0.60
0.40

A
BD

0.20

CD

M82

IL6-4

0.00
JP117

Figure 3.14.

	
  

115

od-2
ordorless-2

CM

Figure 3.14. Greenhouse analysis of M. sexta feeding.
(a) Sixty M. sexta larvae were placed on each genotype and living larvae collected at the
end of 2 weeks from JP117, M82 and IL6-4 respectively. Larvae from JP117 mutant were
bigger than those from M82 and IL6-4 plants. In addition, M. sexta from od-2 (= terp)
mutant were bigger than ones from CM as previously reported (Kang et al. 2010a).
(b) Insect weight for M. sexta larvae shown in (a). Larvae that fed on JP117 mutant were
heaviest and on IL6-4 the lightest (p-value = 0.0005 from one-way ANOVA test). M.
sexta from od-2 mutant was significantly heavier than larvae from CM (p-value = 0.03
from Student's t-test), consistent with previously published data (Kang et al. 2010a).
Larval weight is presented with average and standard error.

	
  

116

M82 compared to IL6-4 (Figure 3.14b). Larvae from the control plant od-2 were
approximately 3-fold heavier than those from CM, as previously reported (Kang et al.
2010a) (Figure 3.14b). The green house assay results were consistent with the
observation from the field test and reinforce the idea that hornworm feeding is affected
by altered Me-M profiles.

Hydroperiod and seasonality affect insect community structure by varying insect types
and their density (Fontanarrosa et al. 2009, Stamp and Yang 1996) and the impact of
plant specialized metabolites on herbivore insects is species-specific (Mitchell et al.
1993). To look for further evidence that insects native to mid-Michigan respond to
differences in trichome myricetin methylation patterns, another independent field
experiment was performed on plants transplanted to the field in early August 2011. In
this field test, the flea beetle Epitrix cucumeris, was found to attack S. lycopersicum
variants. As the flea beetle makes holes on plant leaves, the number of holes was counted
to estimate plant damage (Figure 3.15a). The number of holes was approximately 2-fold
higher in JP117 than M82, and M82 had a slightly larger number of damage sites than
IL6-4 (Figure 3.15b). As expected (Kang et al. 2010a), a large difference in damage site
numbers was seen between the control plants od-2 and CM (Figure 3.15b). E. cucumeris
feeding followed the similar trend as M. sexta, supporting the hypothesis that trichome
Me-Ms composition has an impact on tomato defense against herbivore insects.

	
  

117

(a)

(b)

*

No. of holes

800
600
400

A
BD

200

CD

CM

ordorless-2
od-2

IL6-4

M82

JP117

0

Figure 3.15. Field experiments reveal genotype-dependent differential Epitrix cucumeris
damage.
(a) E. cucumeris, flea beetle, feeding left holes on leaf tissues.
(b) The number of holes was highest in JP117, followed by M82 and IL6-4 (p-value <
0.0001 from one-way ANOVA test). Control plants od-2 mutant and CM showed
dramatic difference (p-value from Student's t-test < 1 x 10-4), consistent with published
data (Kang et al. 2010a). The number of holes is presented with the average and standard
error (n = 28, 23, 18, 6 and 6 for JP117, M82, IL6-4, od-2 and CM, respectively).

	
  

118

Discussion
The identification of altered Me-M phenotypes in the JP117 mutant and IL6-3/4 benefited
from two factors: 1. high sensitivity of the LC-ToF-MS method, 2. the use of the ratios of
differentially

methylated

myricetins

(for

examples,

tri:dimethylated

or

di:monomethylated) as robust diagnostic phenotypes. Despite the low abundance of MeMs - < 1% of the most dominant acylsugar in M82 cultivated tomato (Schilmiller et al.
2010a) (Table 2.1 in Chapter 2) - the LC-ToF-MS method reproducibly measures Me-M
levels.

The combination of negative mode LC-ToF-MS analysis and the single leaf dip
extraction method was efficient for a fast chemical screen to identify genetic variants
altered in trichome metabolites. However, positive mode LC-MS/MS analysis of
individual trichomes was a useful complementary method for three reasons. First, the
different ionization modes (i.e, positive and negative mode) provide distinct information.
For glycosylated flavonols as an example, negative ion mode can identify the flavonol
aglycone and the types of sugars attached to this flavonol core while positive ion mode
MS can reveal information about the position of sugars on the flavonol ring structure
(Tian et al. 2002). Second, LC-MS/MS analysis allows distinguishing positional isomers
with identical m/z. Third, the analysis of isolated trichomes confirms that Me-Ms are
trichome metabolites. As preparations of both type 1 and 6 trichomes were previously
shown to contain Me-Ms (Schmidt et al. 2011), these two trichome types were tested in
this study. Consequently, positive mode LC-MS/MS analysis of isolated trichomes
characterized Me-Ms as trichome metabolites and generally validated JP117 and IL6-3/4

	
  

119

phenotypes identified from negative LC-ToF-MS analysis of single leaf extracts (Figures
3.1, 3.3, 3.6 and 3.7). Furthermore, LC-MS/MS analysis extended the LC-ToF-MS
analysis by distinguishing two isomers of tri-Me-M and detecting a tetra-Me-M. LC-ToFMS analysis of single leaf extracts is effective for fast screening, whereas LC-MS/MS
analysis of the individual trichomes provides more in depth structural information such as
the confirmation of isomers. Therefore, the combination of two complementary
approaches was advantageous to reveal the Me-M phenotypes.

The comprehensive understanding of JP117 and IL6-3/4 Me-M phenotypes using two
complementary analytical approaches leads to a hypothesis on the causes of altered MeM profiles. In contrast to the JP117 mutant, which has decreased total Me-M levels as
well as a higher ratio of mono- to di-Me-M compared to M82 (Figures 3.1 and 3.3), IL63/4 shows a strong shift toward higher accumulation of polymethylated myricetins such
as tri-Me-Ms and tetra-Me-Ms without altering total Me-M level (Figures 3.6 and 3.7). In
LC-MS/MS analysis of individual trichomes, total Me-M levels were only approximately
1.5-fold higher in IL6-3/4 than M82 and this was not statistically significant (p-value >
0.05 from Student's t-test) (Figure 3.7). LC-ToF-MS analysis of single leaf dip extracts
also identified no significant alteration of total Me-Ms between IL6-3/4 and M82. In the
first analysis (Figure 3.6), total Me-M levels were 2~3 fold higher in IL6-3/4 than M82
and the alteration was shown to be statistically significant only in IL6-3 (p-value = 0.005
and 0.09 from Student's t-test for IL6-3 and IL6-4 respectively) (Figure 3.6). Therefore,
other independent sets of IL6-3/4 were analyzed and the alteration was not consistent
(Data not shown). Higher accumulation of tri-Me-Ms in IL6-3/4 could result from the

	
  

120

activity of S. pennellii OMT allele(s) introgressed into M82 genetic background, which
shifted Me-M profiles toward polymethylated myricetins without altering total Me-M
levels. The appearance of tetra-Me-M (3,3',7,5') in LC-MS/MS analysis of IL6-3/4
samples could also be caused by conversion of the abundant tri-Me-Ms by an activity
encoded on chromosome 6 or elsewhere in the genome. Furthermore, while M82 and
JP117 revealed enrichment of Me-Ms in type 1 glands, IL6-3/4 accumulated high amount
of Me-Ms in isolated type 6 glands as well as in type 1 (Figure 3.7). Although this could
be from cross-contamination during harvesting individual types of trichomes due to
rupture or leakage of glands, an exciting hypothesis is that it might represent a real shift
in synthesis site for Me-M production in IL6-3/4.

Recent advance of tomato genomics assisted genetic approaches, allowing identification
of two candidate genes, slMOMT1a and 1b, for the production of Me-Ms in tomato
trichomes. Searching for the genes responsible for the phenotypes in JP117 and IL6-3/4
relying on only genetic mapping is challenging for two reasons. First, phenotypic and
genetic analyses of altered trichome metabolites are extremely time- and labor-intensive.
In particular, only a small portion of the mapping population revealed JP117 mutant
phenotypes and could be used for further genetic analysis (For example, 30 of 289 OC F2
plants as shown in Figure 3.5). This low yield might be because of the interspecific cross
used for mapping. LA0716 was shown to accumulate methylated kaempferol and quercetin compounds instead of Me-Ms (personal communication with Feng Shi and
Chao Li). These chemical profiles suggest that LA0716 does not have functional OMTs
or myricetin substrate for the production of Me-Ms. In addition, epistatic interactions

	
  

121

between LA0716 and JP117 mutant backgrounds might have changed biochemistry in
OC F2 in unpredicted manners. Another challenge in genetic mapping is that genetic
markers are not saturated for the cultivated tomato and its wild relatives. Given the
circumstance, although still incomplete, the genomic sequence of S. lycopersicum and S.
pennellii guided to the identification of candidate genes by enabling gene homology
search and providing polymorphic sequences based on which genetic markers can be
generated.

SlMOMT1a and 1b genes were found to be closely linked within the last 1.5Mb region of
chromosome 6. Previous studies reported examples where specialized metabolite
biosynthesis genes are clustered. Clustering of genes for terpene production is a wellknown example (Falara et al. 2011). These clustered gene families are shown to merge
during gene duplication and obtain novel functions (Pichersky and Gang 2000, Sasaki et
al. 2002, Vision et al. 2000).

In addition to the biochemistry of trichome metabolites, understanding their biological
roles is also important. Although the impact of various flavonoids on insect feeding
(Onyilagha et al. 2004) and insect enzyme activities (Mitchell et al. 1993) were studied,
little is known about the role of Me-Ms in plant-herbivore insect interactions. Field and
green house experiments revealed the higher plant damage in JP117 than M82 and IL6-4
against M. sexta and E. cucumeris insects. Because plants produce and release cocktails
of a wide range of chemical compounds (Baldwin 2001), it is hard to test the role of
specific metabolites in plant defense. Two potential reasons are speculated to explain

	
  

122

variable plant defense among the genetic variants, JP117, M82 and IL6-4. First, given
that the impact of specialized metabolites on plant defense is concentration-dependent
(Mitchell et al. 1993), reduced total Me-Ms in JP117 can be thought to impair plant
resistance. Otherwise, based on the reports that chemical modification of hydroxyl groups
on flavonoids alters their chemical activities (Cao et al. 1997, Heim et al. 2002),
differential accumulation of Me-Ms can be considered to vary plant defense in JP117,
M82 and IL6-4.

Despite intensive efforts, our knowledge of trichome methylated myricetin biochemistry
is still limited. This study revealed that two or more MOMTs on the bottom of
chromosome 6 are involved in the production of trichome Me-Ms in S. lycopersicum
cultivated tomato and suggested the impact of Me-Ms on plant defense against herbivore
insects. In that little is known about trichome metabolites, especially about non-volatile
compounds, this study helps improve our understanding about trichome metabolism.

Materials and methods
Plant growth conditions
M82 cultivated tomato seeds were obtained from the Tomato Genetic Resource Center
(http://tgrc.ucdavis.edu/). Tomato EMS M2 mutant population and S. pennellii ILs were
provided from Dr. Dani Zamir (The Hebrew Univeristy of Jerusalem, Faculty of
Agriculture, Rehobot, Israel). Plants were grown for 3 weeks under the growth conditions
as described in Schilmiller et al. 2010a.

	
  

123

A single leaf dip extraction and negative mode LC-ToF-MS analysis
For chemical extraction of non-volatile trichome metabolites, a single leaf dip method
was used as described in Schilmiller et al. 2010a. Negative mode LC-ToF-MS analysis
was performed following previous studies (Schilmiller et al. 2010a). To quantify the
accumulation of Me-Ms, individual peak areas of mono-, di- and tri-Me-M were obtained
from LC-ToF-MS analysis and normalized to internal standard and leaf dry weight. Total
Me-M amount was calculated by summing the peak areas of all three Me-Ms.

Harvesting types 1 and 6 trichome extracts and positive mode LC-MS/MS analysis
A total of 50 glands from each type of trichomes were collected with micropipettes and
extracted in 50 µl of the extraction solvent used for LC-ToF-MS analysis. Types 1 and 6
trichomes were harvested from 3 week-old plants grown under the identical growth
conditions as used for LC-ToF-MS analysis. LC-MS/MS analysis and data interpretation
including metabolite annotation and quantification was performed by Chao Li as
described in (Schmidt et al. 2011).

Generation of genetic cross progeny for JP117 mutant and IL6-3/4
JP117 mutant was backcrossed to M82 wild type plant in a reciprocal manner, resulting
in M82 x JP117 and JP117 x M82 plants. F1 progeny plants from both reciprocal BC
were chemically analyzed and they revealed identical Me-M profiles. BC F1 and F2
progeny from a M82 x JP117 cross were used for the genetic studies.

	
  

124

IL6-3/4 plants were crossed to M82 in a reciprocal manner. Consistent Me-M profiles
were observed from all F1 progeny from various crosses, M82 x IL6-3, IL6-3 x M82,
M82 x IL6-4 and IL6-4 x M82. BC F1 and F2 progeny from a M82 x IL6-4 cross were
used for the genetic studies.

Genetic mapping of JP117 mutant
JP117 mutant was outcrossed to S. pennellii (LA0716) to generate mapping population,
OC F2 progeny. Publicly available markers (markers 1~4 and 13 from Figure 3.5) were
obtained from Sol Genomics Network (http://solgenomics.net/). When markers were not
available for the location of interest, CAPS markers were generated in this study (markers
5~12 from Figure 3.5). The polymorphic sequences between S. lycopersicum and S.
pennellii were identified based on genomic sequence and trichome ESTs for two species,
S. lycopersicum (http://solgenomics.net/, SL2.31 chromosome sequences) and S. pennellii
(http://mapman.mpimp-golm.mpg.de/blast/blast.html) (Schilmiller et al. 2010b). CAPS
markers

were

generated

by

CAPS

designer

program

(http://solgenomics.net/tools/caps_designer/caps_input.pl/). Primer sequences and the
types of restriction enzymes for individual markers are listed in Table 3.1.

Genotyping was performed by genomic DNA (gDNA) PCR followed by digestion with
suitable restriction enzymes. For PCR, 0.5 µg of gDNA was used with 10 µl of RedTaq
mix (Sigma-Aldrich, http://www.sigmaaldrich.com/) and 5 µl of forward and reverse
primers (2pmol/µl), respectively. The thermocycling conditions were as follows: 40 times
for 45 s at 94°C, 45 s at 55°C and 1 min at 72 °C. The PCR products were incubated with

	
  

125

restriction enzymes for 4 hrs at 37°C. These restriction enzymes were chosen to
distinguish S. lycopersicum or S. pennellii genetic background by recognizing
polymorphic sequence between two different species S. lycopersicum and S. pennelli and
cutting DNA fragment amplified from one species not from the other.

Mapping the location of MOMTs on chromosome 6
The generation of CAPS markers and genotypes was performed as described above.
Primer sequences and the types of restriction enzymes for individual markers are listed in
Table 3.2.

Search for candidate OMT genes
BLAST analysis was performed to obtain genomic DNA sequence for MOMT homologs
in S. lycopersicum (http://solgenomics.net/, SL2.31 chromosome sequences) and S.
pennellii (http://mapman.mpimp-golm.mpg.de/blast/blast.html) using shMOMT gene
sequences (Schmidt et al. 2011). The domain characteristic for plant OMT was searched
(http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml). Gene structures were predicted
by Softberry program (www.softberry.com).

Generation of phylogeny tree and validation of its reliability
Phylogeny tree was generated by Neighbor joining method (Saitou and Nei 1987) and its
reliability was tested by bootstrap test (Felsenstein 1985) with n=500. Phylogeny tree
generation and statistical test was performed by MEGA5 program (Tamura et al. 2011).

	
  

126

The field - and green house insect feeding assays
Plants were grown for 3 weeks under growth chamber conditions as discussed above.
Three week-old plants were transferred to a green house and grown for another week.
The 4 week-old plants were used for the field - and green house insect feeding assays.
Different genetic variants were placed with randomized orders.

Field experiments were performed at the Department of Plant Pathology Research Farm,
in the summer of 2010 and 2011. Plants were monitored once a week to observe
herbivore insects and plant damage. The monitoring was performed twice a week once
plant damage occurred by insects. To estimate plant damage by M. sexta, the damage
degree was assigned to the scale of 0 to 3 based on how much plant tissues were lost. For
example, when the loss was 0~10% compared to whole plant mass, value '0' was scored.
Values from 1 to 3 were assigned when the loss was 10~30, 30~60 and 60~100%
respectively. Damage scoring was performed with a blind approach (i.e. the genotype of
plants were not provided to observers when plant damage was scored). To estimate plant
damage by E. cucumeris, the number of holes was counted.

Green house test was performed to repeat M. sexta feeding assay. The eggs were obtained
from the Department of Entomology, North Carolina State Universtiy in Raleigh and
hatched at 26°C following supplier's recommendation. Two newly hatched larvae were
placed on the consistent positions of leaf tissue for individual plants. This aims to
minimize the variation of insect feeding due to different plant developmental stages.
Thirty replicate plants were tested for each genotype of JP117, M82 and IL6-4. Four

	
  

127

Table 3.2. The list of co-dominant amplified polymorphic sequence (CAPS) markers
used to map MOMTs on chromosome 6.
Primer sequences and restriction enzymes are shown for individual CAPS markers used
to map the location of MOMTs on chromosome 6. See 'Materials and Methods' for details
about genotyping procedures.

MOMT
Name

Primer
Type

Primer Sequence
(5'-3')

Restriction
Enzyme

1a

F

AGGTGGACTTGGAA
TATCATTAGCT
TGTATTTGTTCAATT
ACAACTACTTTACC
ATCA
GAACTTGACCTATT
TGAGATTATAGCA
AATATTCTAAGTAA
GGTGAGTGTGAAAT
TTC
ATGACTTCTTTTCAT
CCGTTAA
TTATTGTTAATTTTC
TAAGTTA
ATGCCAACGCAATC
TTACTCAAGG
TAAAACCAGCCTCC
AAAAAGAGTTTTTC
CAACATGTATGCTA
ATGAGCTAATAG
TTAAGGATAAACTT
CCATTAGAGACC

EcoRV

200,
500

700

ApoI

1146

366,
780

KpnI

341

117,
224

None

1300

1100

DraI

1210

177,
1033

R
1b

F
R

1c

F
R

2a

F
R

2b

F
R

	
  

128

Expected Product
Size (bp)
L
P

replicate plants were tested for control plants od-2 and CM respectively. Living M. sexta
larvae were collected after 14 days of feeding on the individual genetic variants and
insect weight was measured.

One-way ANOVA test was performed to validate the variation of insect feeding in JP117
mutant, M82 and IL6-4 by using JMP8 program (http://www.jmp.com/software/jmp8/).

Acknowledgement
I thank Chao Li, a former graduate student in Dr. Jones' group, and Dr. Jones for LCMS/MS performance and data analysis including metabolite identification and
quantification for LC-MS/MS analysis of the isolated types 1 and 6 trichome extracts.

	
  

129

References

	
  

130

References
Baldwin, I.T. (2001) An ecologically motivated analysis of plant-herbivore interactions
in native tobacco. Plant Physiol., 127, 1449-1458.
Baldwin, I.T., Halitschke, R., Paschold, A., von Dahl, C.C. and Preston, C.A. (2006)
Volatile signaling in plant-plant interactions: "talking trees" in the genomics era.
Science, 311, 812-815.
Braca, A., Bilia, A.R., Mendez, J. and Morelli, I. (2001) Myricetin glycosides from
Licania densiflora. Fitoterapia, 72, 182-185.
Buchana, B., Gruissem, W. and Jones, R. (2002) Biochemistry & Molecular Biology of
Plants: Wiley.
Buer, C.S., Imin, N. and Djordjevic, M.A. (2010) Flavonoids: new roles for old
molecules. J. Integr. Plant. Biol, 52, 98-111.
Cao, G., Sofica, E. and Priora, R.L. (1997) Antioxidant and prooxidant behavior of
flavonoids: structure-activity relationships. Free Radical Biology and Medicine,
22, 749-760.
Eshed, Y. and Zamir, D. (1994) A genomic library of Lycopersicon pennellii in
Lycopersicon esculentum - A tool for fine mapping of genes EUPHYTICA, 79,
175-179.
Eshed, Y. and Zamir, D. (1995) An introgression line population of Lycopersicon
pennellii in the cultivated tomato enables the identification and fine mapping of
yield-associated QTL. Genetics, 141, 1147-1162.
Falara, V., Akhtar, T., Nguyen, T.T., Spyropoulou, E.A., Bleeker, P.M.,
Schauvinhold, I., Matsuba, Y., Bonini, M.E., Schilmiller, A.L., Last, R.L.,
Schuurink, R.C. and Pichersky, E. (2011) The tomato (Solanum lycopersicum)
terpene synthase gene family. Plant Physiol., 157, 770-789.
Felsenstein, J. (1985) Confidence limits on phylogenies: An approach using the
bootstrap. Evolution, 39, 783-791.
Fontanarrosa, M.S., Collantes, M.B. and Bachmann, A.O. (2009) Seasonal patterns of
the insect community structure in urban rain pools of temperate Argentina. J.
Insect. Sci., 9, 10.
Gang, D.R. (2005) Evolution of flavors and scents. Annu. Rev. Plant. Biol., 56, 301-325.

	
  

131

Heim, K.E., Tagliaferro, A.R. and Dennis J Bobilyaa, A. (2002) Flavonoid
antioxidants: chemistry, metabolism and structure-activity relationships. The
Journal of Nutritional Biochemistry, 13, 572-584.
Ibrahim, R.K., Bruneau, A. and Bantignies, B. (1998) Plant O-methyltransferases:
molecular analysis, common signature and classification. Plant Molecular
Biology, 36, 1-10.
Jander, G., Norris, S.R., Rounsley, S.D., Bush, D.F., Levin, I.M. and Last, R.L.
(2002) Arabidopsis map-based cloning in the post-genome era. Plant Physiol.,
129, 440-450.
Kang, J.H., Liu, G., Shi, F., Jones, A.D., Beaudry, R.M. and Howe, G.A. (2010a) The
tomato odorless-2 mutant is defective in trichome-based production of diverse
specialized metabolites and broad-spectrum resistance to insect herbivores. Plant
Physiol., 154, 262-272.
Kang, J.H., Shi, F., Jones, A.D., Marks, M.D. and Howe, G.A. (2010b) Distortion of
trichome morphology by the hairless mutation of tomato affects leaf surface
chemistry. J. Exp. Bot., 61, 1053-1064.
Kim, D.H., Kim, B.G., Lee, Y., Ryu, J.Y., Lim, Y., Hur, H.G. and Ahn, J.H. (2005)
Regiospecific methylation of naringenin to ponciretin by soybean Omethyltransferase expressed in Escherichia coli. J. Biotechnol., 119, 155-162.
Liu, B., Wang, H., Du, Z., Li, G. and Ye, H. (2010) Metabolic engineering of
artemisinin biosynthesis in Artemisia annua L. Plant. Cell. Rep.
McDowell, E.T., Kapteyn, J., Schmidt, A., Li, C., Kang, J.H., Descour, A., Shi, F.,
Larson, M., Schilmiller, A., An, L., Jones, A.D., Pichersky, E., Soderlund,
C.A. and Gang, D.R. (2011) Comparative functional genomic analysis of
Solanum glandular trichome types. Plant Physiol., 155, 524-539.
Mitchell, M.J., Keogh, D.P., Crooks, J.R. and Smith, S.L. (1993) Effects of plant
flavonoids and other allelochemicals on insect cytochrome P-450 dependent
steroid hydroxylase activity. Insect Biochem. Mol. Biol., 23, 65-71.
Muzac, I., Wang, J., Anzellotti, D., Zhang, H. and Ibrahim, R.K. (2000) Functional
expression of an Arabidopsis cDNA clone encoding a flavonol 3'-Omethyltransferase and characterization of the gene product. Arch. Biochem.
Biophys., 375, 385-388.
Onyilagha, J.C., Lazorko, J., Gruber, M.Y., Soroka, J.J. and Erlandson, M.A.
(2004) Effect of flavonoids on feeding preference and development of the crucifer
pest Mamestra configurata Walker. J. Chem. Ecol., 30, 109-124.

	
  

132

Pearson, W.R. and Lipman, D.J. (1988) Improved tools for biological sequence
comparison. Proc. Natl. Acad. Sci. U.S.A, 85, 2444-2448.
Pichersky, E. and Gang, D.R. (2000) Genetics and biochemistry of secondary
metabolites in plants: an evolutionary perspective. Trends. Plant. Sci., 5, 439-445.
Rodriguez-Saona, C., Vorsa, N., Singh, A.P., Johnson-Cicalese, J., Szendrei, Z.,
Mescher, M.C. and Frost, C.J. (2011) Tracing the history of plant traits under
domestication in cranberries: potential consequences on anti-herbivore defences. J
Exp Bot, 62, 2633-2644.
Saitou, N. and Nei, M. (1987) The neighbor-joining method: a new method for
reconstructing phylogenetic trees. Mol. Biol. Evol., 4, 406-425.
Sasaki, T., Matsumoto, T., Yamamoto, K., Sakata, K., Baba, T., Katayose, Y., Wu,
J., Niimura, Y., Cheng, Z., Nagamura, Y., Antonio, B.A., Kanamori, H.,
Hosokawa, S., Masukawa, M., Arikawa, K., Chiden, Y., Hayashi, M.,
Okamoto, M., Ando, T., Aoki, H., Arita, K., Hamada, M., Harada, C.,
Hijishita, S., Honda, M., Ichikawa, Y., Idonuma, A., Iijima, M., Ikeda, M.,
Ikeno, M., Ito, S., Ito, T., Ito, Y., Iwabuchi, A., Kamiya, K., Karasawa, W.,
Katagiri, S., Kikuta, A., Kobayashi, N., Kono, I., Machita, K., Maehara, T.,
Mizuno, H., Mizubayashi, T., Mukai, Y., Nagasaki, H., Nakashima, M.,
Nakama, Y., Nakamichi, Y., Nakamura, M., Namiki, N., Negishi, M., Ohta,
I., Ono, N., Saji, S., Sakai, K., Shibata, M., Shimokawa, T., Shomura, A.,
Song, J., Takazaki, Y., Terasawa, K., Tsuji, K., Waki, K., Yamagata, H.,
Yamane, H., Yoshiki, S., Yoshihara, R., Yukawa, K., Zhong, H., Iwama, H.,
Endo, T., Ito, H., Hahn, J.H., Kim, H.I., Eun, M.Y., Yano, M., Jiang, J. and
Gojobori, T. (2002) The genome sequence and structure of rice chromosome 1.
Nature, 420, 312-316.
Schilmiller, A., Shi, F., Kim, J., Charbonneau, A.L., Holmes, D., Daniel Jones, A.
and Last, R.L. (2010a) Mass spectrometry screening reveals widespread
diversity in trichome specialized metabolites of tomato chromosomal substitution
lines. Plant J., 62, 391-403.
Schilmiller, A.L., Last, R.L. and Pichersky, E. (2008) Harnessing plant trichome
biochemistry for the production of useful compounds. Plant J., 54, 702-711.
Schilmiller, A.L., Miner, D.P., Larson, M., McDowell, E., Gang, D.R., Wilkerson, C.
and Last, R.L. (2010b) Studies of a biochemical factory: tomato trichome deep
expressed sequence tag sequencing and proteomics. Plant Physiol., 153, 12121223.
Schilmiller, A.L., Schauvinhold, I., Larson, M., Xu, R., Charbonneau, A.L.,
Schmidt, A., Wilkerson, C., Last, R.L. and Pichersky, E. (2009) Monoterpenes
in the glandular trichomes of tomato are synthesized from a neryl diphosphate

	
  

133

precursor rather than geranyl diphosphate. Proc. Natl. Acad. Sci. U. S. A., 106,
10865-10870.
Schmidt, A., Li, C., Shi, F., Jones, A.D. and Pichersky, E. (2011) Polymethylated
myricetin in trichomes of the wild tomato species Solanum habrochaites and
characterization of trichome-specific 3'/5'- and 7/4'-myricetin Omethyltransferases. Plant Physiol., 155, 1999-2009.
Stamp, N.E. and Yang, Y. (1996) Response of Insect Herbivores to Multiple
Allelochemicals Under Different Thermal Regimes. Ecology, 77, 1088-1102.
Stevens, J., Hart, H. and Wollenweber, E. (1995) The systematic and evolutionary
significance of exudate flavonoids in Aeonium. Phytochemistry, 39, 805-813.
Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. and Kumar, S. (2011)
MEGA5: Molecular Evolutionary Genetics Analysis using maximum likelihood,
evolutionary distance, and maximum parsimony methods. Mol. Biol. Evol.
Tian, Q., Li, D. and Patil, B.S. (2002) Identification and determination of flavonoids in
buckwheat (Fagopyrum esculentum Moench, Polygonaceae) by high-performance
liquid chromatography with electrospray ionisation mass spectrometry and
photodiode array ultraviolet detection. Phytochem Anal, 13, 251-256.
van Der Heijden, R., Jacobs, D.I., Snoeijer, W., Hallard, D. and Verpoorte, R.
(2004) The Catharanthus alkaloids: pharmacognosy and biotechnology. Curr.
Med. Chem., 11, 607-628.
Vingron, M. and Waterman, M.S. (1994) Sequence alignment and penalty choice.
Review of concepts, case studies and implications. J Mol Biol, 235, 1-12.
Vision, T.J., Brown, D.G. and Tanksley, S.D. (2000) The origins of genomic
duplications in Arabidopsis. Science, 290, 2114-2117.
Wagner, G. (1991) Secreting glandular trichomes: more than just hairs. Plant Physiol.,
96, 675-679.
Wagner, G.J., Wang, E. and Shepherd, R.W. (2004) New approaches for studying and
exploiting an old protuberance, the plant trichome. Ann. Bot., 93, 3-11.
Wang, G., Tian, L., Aziz, N., Broun, P., Dai, X., He, J., King, A., Zhao, P.X. and
Dixon, R.A. (2008) Terpene biosynthesis in glandular trichomes of hop. Plant
Physiol., 148, 1254-1266.
Weinhold, A. and Baldwin, I.T. (2011) Trichome-derived O-acyl sugars are a first meal
for caterpillars that tags them for predation. Proc. Natl. Acad. Sci. U. S. A, 108,
7855-7859.

	
  

134

Chapter 4. Mass spectrometry reveals strong diversity in glandular trichome
acylsugars of the wild tomato Solanum habrochaites

Abstract
Acylsugars are polyesters of short to medium length acyl chains on sucrose or glucose
produced in secretory glandular trichomes of many solanaceous plants, including
cultivated tomato. Despite their abundance and wide taxonomic distribution, there is
relatively little information about the diversity of these compounds and nearly nothing is
known about the genes responsible for their biosynthesis. In this study, acylsugar
diversity was assessed for 80 accessions of the wild tomato species Solanum
habrochaites from throughout the Andes Mountains. Trichome metabolites were
analyzed by liquid chromatography-time of flight mass spectrometry, revealing the
presence of at least 34 structurally diverse acylsucroses and two acylglucoses. Trichome
acylsugar profiles were compared to each other and to that of the cultivated tomato (S.
lycopersicum M82) using hierarchical clustering and principal component analyses.
These approaches revealed distinct phenotypic classes that varied based on the presence
of glucose or sucrose, the numbers and lengths of acyl chains and relative total amounts
of acylsugars. Chemically similar accessions clustered geographically. The finding of
such strong S. habrochaites acylsugar diversity suggests the existence of a relatively
large repertoire of biosynthetic enzymes in the acylsugar metabolic network.

	
  

135

Introduction
Trichomes are specialized epidermal cells that protrude from the surface of a variety of
plant tissues. They are thought to protect against environmental stresses such as
herbivory (Kang et al. 2010a, Weinhold and Baldwin 2011), loss of water through
transpiration and UV irradiation (Zhou et al. 2007). In particular, secreting glandular
trichomes (SGTs) serve as ‘chemical factories’ where specialized metabolites are
produced, stored or volatized (Schilmiller et al. 2010a, Schilmiller et al. 2008, Wagner
1991). In addition, SGTs produce and secrete proteins on the plant surface for insect
protection (Thipyapong et al. 1997, Yu et al. 1992) and pathogen defense (Shepherd et
al. 2005). SGTs also contribute to the taste and smell of plants by releasing volatile
metabolites. For example, the distinctive aromas of many Mediterranean herbs of the
Lamiaceae (mint family) derive from SGTs (Schilmiller et al. 2008) and compounds
from the glands of hops (Humulus lupulus L. in the Cannabaceae) contribute to beer
flavor and aroma (Wang et al. 2008). Furthermore, a number of SGT-borne metabolites
are commercially valuable, especially for pharmaceutical purposes. For example,
artemisinin, a widely used antimalarial, is a sesquiterpene lactone from the trichomes of
Artemisia annua L. (Liu et al. 2011). In addition to their value in foods and medicines,
trichomes provide excellent models for analyzing biosynthetic enzymes and pathways
(Bohlmann and Gershenzon 2009, Sallaud et al. 2009, Schilmiller et al. 2008, Schilmiller
et al. 2009).

Plants in the genus Solanum include important crop species such as potato, eggplant and
tomato. Previous studies reported that SGTs of cultivated tomato (Solanum lycopersicum)

	
  

136

and its wild relatives accumulate high levels of exudates containing a variety of
specialized metabolites, for example flavonoids, alkaloids and terpenoids (McDowell et
al. 2011, Schilmiller et al. 2010a, Schilmiller et al. 2008, Wagner 1991).
Cultivated tomato and its wild relatives have morphologically and chemically diverse
trichomes. For example, Luckwill (Luckwill 1943) defined seven morphologically
distinguishable types of trichomes in plants of this genus including four glandular types
(type 1, 4, 6, and 7) (Figure 4.1; for more examples of Solanum trichome images, see
Figures 3 and 1 from Kang et al. 2010a and Kang et al. 2010b, respectively). The
presence of specific types of trichomes and their densities vary across species and even
within a single plant according to tissue types and developmental stages (Werker 2000).
These morphologically distinct SGTs vary in the amounts and types of metabolites that
they produce, accumulate and/or secrete (Werker 2000). For example, S. lycopersicum
M82 leaf type 6 SGTs accumulate the sesquiterpenes β-caryophyllene and α-humulene
while the glands on the stem lack these metabolites (Schilmiller et al. 2010b). There are
also species- and accession-specific differences in SGT metabolite profiles. For instance,
methylketones accumulate in type 6 glands of a subset of S. habrochaites accessions
(Fridman et al. 2005, Yu et al. 2010). Similarly, acylglucoses are highly abundant in type
4 glands of S. pennellii LA0716 while acylsucroses predominate in S. lycopersicum and
S. habrochaites (Kennedy 2003, McDowell et al. 2011). The chemical and morphological
diversity of trichomes in different Solanum species and accessions makes the genus an
attractive target for identification of diverse trichome-borne metabolites and the major
biosynthetic pathways responsible for their synthesis operating in each trichome type.

	
  

137

(A)

(B)

Short
Long

Long
Short

(C)

(D)

Short

Short
Long
Long

(E)

(F)

Long

Short

Long
Short

Figure 4.1. Light microscope images for leaf trichomes and whole leaf from three S.
habrochaites accessions.
A-B. LA2101; C-D. LA1777; E-F. LA2975. Images were taken with three week-old
plants. ‘Long’ indicates long SGTs, either type 1 or 4 and ‘short’ short SGTs such as type
6 or 7.

	
  

138

The value of the comparative metabolomics approach in trichomes was recently
demonstrated in the studies of Solanum trichome mono- and sesquiterpene biosynthesis
(Bohlmann and Gershenzon 2009, Sallaud et al. 2009, Schilmiller et al. 2009). It was
discovered that S. lycopersicum trichomes synthesize monoterpenes from the cisprenyldiphosphate intermediate neryl diphosphate (Schilmiller et al. 2009). This is
contrary to the previous paradigm where the trans-prenyldiphosphate geranyldiphosphate
was considered the universal intermediate for monoterpene biosynthesis. An analogous
example of biosynthetic innovation was reported for SGTs of S. habrochaites LA1777
(Sallaud et al. 2009), shown to produce sesquiterpenes in the plastid using the all cisprenyldiphosphate substrate Z,Z-farnesyldiphosphate (Z,Z-FPP). This is counter to the
commonly described cytosolic sesquiterpene pathway, which uses the all trans
sesquiterpene synthase substrate E,E-farnesyldiphosphate. These observations suggest
that trichome specialized metabolism is evolutionarily plastic, perhaps due to selective
pressure from insects or other environmental stress agents.

Acylsugars are sticky exudates that are thought to physically or chemically improve plant
defense (Mirnezhad et al. 2010). Results from the literature indicate strong acylsugar
diversity in various Solanum trichomes (McDowell et al. 2011, Schilmiller et al. 2010a,
Schilmiller et al. 2010b). Acylsugars are categorized as either sucrose- or glucose esters
based on the type of sugar core (Figure 4.2) and they also have varying number of acyl
chains decorating the sugar moiety. In particular, S. pennellii accumulates enormous
amounts of acylsugars, up to 20% of leaf dry weight (Fobes et al. 1985). In addition,
previously published data showed that total acylsugars in geographically distinct S.

	
  

139

(A)

R

R

R

(B)

R

R

R
R

Figure 4.2. Structural classes of acylsugars in Solanum species.
(A) Schematic structure of an acylglucose. The structure shown depicts a glucose triester
comprised of glucose and 3 acyl chains with various numbers of carbons represented as
'R'.

	
  

140

Figure 4.2. Continued
(B) The schematic structure of an acylsucrose. The proposed structure of a sucrose
tetraester with three acyl chains on the glucose ring and one on the fructose ring. If the
sugar moiety is decorated with either three or five acyl chains, it is referred to as either a
sucrose triester or pentaester, respectively. The positions of the acyl chains are currently
unknown, with the exception of the most abundant acylsugar in cultivated tomato (M82)
that was structurally characterized by NMR (Schilmiller et al. 2010a).

	
  

141

pennellii accessions vary in quantity, the proportion of sucrose or glucose backbones and
the overall types of fatty acid esters (FAs) on the sugars (Shapiro et al. 1994). However,
this study did not identify specific acylsugar types. To explore the detailed acylsugar
chemotypes within accessions of one species, we focused on 80 accessions collected
throughout the geographical range of S. habrochaites in Peru and Ecuador (Table 4.1).

Results
Overview of the approach
Acylsugars are produced in SGTs of a variety of Solanaceous species, sometimes to such
high levels that they make the surface of the plant sticky (Fobes et al. 1985, Wagner
1991). Despite an interest in the possible antimicrobial or insecticidal roles of these
metabolites, relatively little is known about their structural diversity or biosynthesis
(Ghangas and Steffens 1993). To document acylsugar diversity within a single species,
we analyzed 80 Solanum habrochaites accessions collected from throughout the range of
the species (Table 4.1 and Sifres et al. 2011) and M82 cultivated tomato (Solanum
lycopersicum).

As diagrammed in Figure 4.3, extracts of trichome and surface metabolites were analyzed
by liquid chromatography-time of flight mass spectrometry (LC-TOF-MS) (Schilmiller et
al. 2010a). In this method the combination of LC and high mass accuracy TOF-MS
provides good metabolite selectivity in a relatively rapid assay. The leaf dip sampling
method was chosen to maximize extraction of glandular trichome metabolites, while
minimizing contamination by epidermal pavement cells (Schilmiller et al. 2010a). Mass

	
  

142

Table 4.1. The location of original collection site for S. habrochaites accessions.
The geographic information was obtained from TGRC (Tomato Genetics Resource
Center, http://tgrc.ucdavis.edu/). Accessions were listed in geographical order from north
to south. Accessions lacking latitude and longitude information in the TGRC database are
listed in the order of accession numbers. See Figure 4.6 for more details about HCA
result.

	
  

143

Table 4.1.

Site

Accession
LA1625
LA1624
LA1266
LA1223
LA0407
LA2119
LA2861
LA2128
LA2105

LA1253
LA2106
LA3863
LA2100
LA2864
LA2101
LA2860
LA2098
LA2110

	
  

Site
S of
Jipijapa
Jipijapa
Pallatanga
Alausi
Mirador
Guayaquil
Saraguro
Las
Juntas
Zumbi
Jardin
Botanico,
Loja
Pueblo
NuevoLandangue
Yambra
Sozoranga
Sozorango
Sozorango
Cariamanga
Cariamanga
Sabianga
Yangana
#2

Dept./
Prov

Country

Latitude

Longitude

Elevation

HCA

Manabi

Ecuador

-1.33

-80.58

-

C

Manabi
Chimbo
-razo
Chimbo
-razo

Ecuador

-1.33

-80.57

-

C

Ecuador

-1.98

-78.45

1,000

D

Ecuador

-2.2

-78.83

2,200

D

Guayas

Ecuador

-2.98

-79.77

40

D

Loja

Ecuador

-3.6

-79.22

2,600

D

Loja

Ecuador

-3.82

-79.27

2,090

D

Zamora
Chinchi
-pe

Ecuador

-3.89

-78.78

1,000

D

Loja

Ecuador

-4

-79.22

2,200

D

Loja

Ecuador

-4.14

-79.2

2,000

D

Loja

Ecuador

-4.22

-79.23

1,700

D

Loja

Ecuador

-4.33

-79.78

-

D

Loja

Ecuador

-4.33

-79.78

1,200

E

Loja

Ecuador

-4.33

-79.78

1,650

E

Loja

Ecuador

-4.33

-79.55

1,800

D

Loja

Ecuador

-4.33

-79.55

1,700

D

Loja

Ecuador

-4.37

-79.8

700

E

Loja

Ecuador

-4.38

-79.18

2,000

D

144	
  

Table 4.1. Continued
LA3864
LA2109
LA2650
LA2175
LA1718
LA2204
LA2196
LA2314
LA2156
LA2155
LA2812
LA0386
LA4137

LA2167

Yangana
Yangana
#1
Ayabaca
Timbaruca
Huancabamba
Balsapata
Caclic
San
Francisco
Ingenio
Montan
Maydasbamba
Lambayeque
Cajamarca
Barrio
Delta,
Cajamarca
Cementerio
Cajamarca

LA1352

Rupe

LA3854

Llaguen

LA2329

Aricapam
pa

LA1393
LA1362
LA1775
LA2574

	
  

Caraz
Chacchan
Rio
Casma
Cullaspungro

Loja

Ecuador

-4.38

-79.18

-

D

Loja

Ecuador

-4.38

-79.18

2,000

D

Piura
Cajamarca

Peru

-4.63

-79.95

800

B

Peru

-5.14

-79.01

1,150

E

Piura

Peru

-5.24

-79.45

-

D

Peru

-5.77

-77.87

-

E

Peru

-6.18

-77.91

1,600

E

Peru

-6.42

-77.87

1,650

E

Peru

-6.53

-78.79

1,700

E

Peru

-6.65

-78.54

2,600

E

Peru

-6.7

-79.92

-

E

Peru

-7.16

-78.5

2,500

B

Cajamarca

Peru

-7.17

-78.35

2,900

B

Cajamarca

Peru

-7.18

-78.52

2,300

B

Peru

-7.3

-78.81

2,100

E

Peru

-7.72

-78.73

1,800

B

Peru

-7.81

-77.7

2,450

B

Peru

-9.05

-77.81

A

Ancash

Peru

-9.31

-77.98

Ancash

Peru

-9.52

-77.88

2,280
2,1002,300
9001,800

Ancash

Peru

-9.54

-77.8

2,000

A

Amazonas
Amazonas
Amazonas
Cajamarca
Cajamarca
Lambayeque
Cajamarca

Cajamarca
Chicam
a
La
Libertad
Ancash

145	
  

A
A

Table 4.1. Continued
LA1361
LA2975
LA1777
LA2976
LA1392
LA1779
LA1978
LA3796
LA3794
LA1557
LA1764
LA0094
LA1559
LA1648
LA0361
LA1298
LA1295
LA1753
LA1560
LA2409
LA1691
LA2722
LA1681
LA1695
LA1696

	
  

Pariacoto
Coltao
Rio
Casma
Huangra
Huaraz to
Casma
Rio
Casma
Colca
Anca,
Marca
Alta
Fortaleza
Rio
Huara
W of
Canta
CantaYangas
DesvioHuamantanga
AboveYaso
Canta
Yaso
Surco
Surco
Matucana
Miraflores
Yauyos
Puente
Auca
Mushka
Cacachuhuasin,
Canete
Huanchu
y-Cacra

Ancash
Ancash

Peru
Peru

-9.55
-9.55

-77.89
-77.68

1,490
3,200

A
A

Ancash

Peru

-9.55

-77.59

3,150

A

Ancash

Peru

-9.56

-77.7

3,000

A

Ancash

Peru

-9.56

-77.7

3,120

A

Ancash

Peru

-9.56

-77.66

2,600

A

Ancash

Peru

-10.11

-77.48

2,450

B

Ancash

Peru

-10.12

-77.49

2,810

B

Ancash

Peru

-10.15

-77.36

-

B

Lima

Peru

-11.22

-76.63

-

A

Lima

Peru

-11.47

-76.62

2,400

B

Lima

Peru

-11.52

-76.68

2,100

B

Lima

Peru

-11.53

-76.69

-

B

Lima

Peru

-11.55

-76.72

1,500

B

Lima
Lima
Lima
Lima

Peru
Peru
Peru
Peru

-11.57
-11.58
-11.88
-11.88

-76.73
-76.75
-76.44
-76.44

1,600
1,600
2,100
2,000

B
B
B
B

Lima

Peru

-11.89

-76.44

-

B

Lima

Peru

-12.27

-75.83

3,000

D

Lima

Peru

-12.45

-75.93

2,074

B

Lima

Peru

-12.59

-75.95

2,500

B

Lima

Peru

-12.75

-75.85

2,450

B

Lima

Peru

-12.81

-75.78

-

B

Lima

Peru

-12.82

-75.79

2,100

B

146	
  

Table 4.1. Continued

LA1721
LA1918
LA1928
LA1927
LA1033
LA1255
LA1265
LA1347
LA1391
LA1731
LA2099
LA2104
LA2107
LA2144
LA2552
LA2869

Ticrapo
Viejo
Llauta
Ocana
Ocobamba
Hacienda Taulis
Pedistal
Rio
Chimbo
Empalme
Otusco
Bagua to
Olmos
Rio San
Juan
Sabianga
to
Sozorango
Pena
Negra
Los Lirios
Chanchan
Las
Flores
MatolaLa Toma

Huanca
velica
Ica
Ica

Peru

-13.38

-75.44

2,100

B

Peru
Peru

-14.27
-14.38

-74.92
-74.82

2,600
2,660

B
B

Peru

-14.47

-74.81

2,540

B

Peru

-

-

-

E

Ecuador

-

-

-

D

Ecuador

-

-

-

C

Peru

-

-

-

A

Peru

-

-

-

E

Peru

-

-

-

B

Loja

Ecuador

-

-

-

D

Loja

Ecuador

-

-

-

D

Loja
Chimborazo
Cajama
rca

Ecuador

-

-

-

D

Ecuador

-

-

-

D

Peru

-

-

-

B

Loja

Ecuador

-

-

-

E

Ica
Lambayeque
Loja
Chimborazo
La
Libertad
Cajamarca
Huanca
velica

	
  

	
  

147	
  

spectra were collected at multiple collision energies (multiplexed collision-induced
dissociation; Gu et al. 2010, Schilmiller et al. 2010a), with spectra stored separately for
each collision energy. The MS spectra obtained at low collision energies provide
molecular mass information, and those generated at higher energies yield successively
more fragment ions whose masses are useful for structural annotation (for an example see
Figure 4.4). Each metabolite elutes from the HPLC column at a characteristic retention
time, and the lowest collision energy yields ions of mass to charge ratio (m/z) indicative
of its molecular mass. The combination of retention time and m/z is referred to as
‘analytical signals’. The mass defect of the ions, which is the digit of the mass value that
follows the decimal place, was also used to categorize the type of compound represented
by each analytical signal. The relative mass defect, which is the defect normalized to the
measured mass, largely reflects the hydrogen content of the ionized molecule (Last et al.
2007, Stagliano et al. 2010).

Thirty-six analytical signals were characterized as associated with acylsugars from one or
more of the 80 S. habrochaites accessions analyzed, with 34 acylsucroses and two
acylglucoses (Table 4.2). Extracted ion chromatograms for formate adducts of acylsugars
-

(labeled '[M+HCOO] m/z' in Table 4.2) at the lowest collision energy were used for
quantification of the analytical signal for each acylsugar. The peak area for each extracted
ion chromatogram was integrated and normalized to an internal standard to control for
variation in injection volume. The value was further normalized to leaf dry weight to
account for differences in the size of the plant tissues extracted. These normalized peak
areas were used as a measure of acylsugar quantity (Figure 4.3).

	
  

148

Growth and harvest of
plant material

Leaf trichome metabolite extraction

LC-TOF-MS metabolite analysis

Peak area integration

Normalization to internal standard
and dry leaf weight

Quantitative analysis of acylsugar
chemistry

Figure 4.3. Acylsugar analysis workflow.
Plant extracts from the single leaf dip method were analyzed by LC-TOF-MS
(Schilmiller et al. 2010a). Peak areas for extracted ion chromatogram corresponding to 36
acylsugar analytical signals were integrated using QuanLynx software. These signals
were normalized to an internal standard and dry leaf weight. The normalized data were
used for the quantitative comparison of acylsugar profiles in 80 S. habrochaites
accessions.

	
  

149

(A)

100
XIC for all acylsugars
detected

m/z 807!

%

0

30.00

20.00

10.00

40.00

100
(B)

m/z 807

XIC for m/z 807
%

0

(C)

40.00

30.00

20.00

10.00

100

807
[M+HCOO]-

MS spectrum at low energy
condition!
%

0
100

200

300

400

500

700

800 (m/z)!

677

100
(D)

600

509

MS spectra at high energy
condition

807 [M+HCOO]-

425

%
341

593
-C5!

-C5

-C5
761 !
[M-H]-!

-C11!

0
100

200

300

400

500

Figure 4.4

	
  

150

600

700

800 (m/z)!

Figure 4.4.

An example of the use of LC-TOF with multiplexed collision-induced

dissociation to annotate acylsugar S4:26 (5,5,5,11).
Extracted ion chromatograms (XICs) and mass spectra (MS) from S. habrochaites
accession LA2869 were shown to describe how metabolite annotation was characterized
by using MS spectra from different energy conditions. This analytical signal with m/z
807 and RT 35.6 was annotated as S4:26 (5,5,5,11) based on MS spectra obtained under
fragmenting and non-fragmenting energy conditions (Table 4.2).
(A) XIC for all acylsugars detected in accession LA2869 under low collision energy
(aperture 1 voltage of 10 V) reveals a complex population of related metabolites. A peak
with nominal m/z 807 eluting at 35.6 min. is indicated.
(B) XIC for m/z 807 shows a single peak at 35.6 min elution time (aperture 1 voltage of
10 V).
(C) For the peak of m/z 807, the MS spectrum was obtained under low energy nonfragmenting conditions (aperture 1 voltage of 10 V), revealing a single major molecular
-

formate adduct ion [M+HCOO] .
(D) For the peak of m/z 807, MS spectra were obtained from conditions that promote
fragmentation (aperture 1 voltage of 55 V). The results revealed major peaks due to loss
-

of three C5 and one C11 group from m/z 761 [M-H] , consistent with the annotation of
S4:26 (5,5,5,11) for this peak. The m/z 425 peak is interpreted as loss of C11 from m/z
593 peak, S2:16 (5,11), to yield S1:5 (5).

	
  

151

Acylsugar complexity and diversity revealed by comparison of three Solanum
accessions
Examination of the acylsugar analytical signals (Table 4.2) revealed highly diverse
acylsugar profiles among S. lycopersicum M82 and the two S. habrochaites accessions
LA1777 and LA2106 as shown in Figure 4.5. Acylsugar profiles similar to published
results were found for S. lycopersicum M82 (Schilmiller et al. 2010a), with the sucrose
tetraester S4:17 (2,5,5,5) as the most abundant compound, as shown in Figure 4.5a. Note
that in this nomenclature, ‘S’ or 'G' indicates a sucrose or glucose backbone, respectively,
‘4’ indicates the total number of acyl chains, ‘17’ is the sum of the number of carbon
molecules in the acyl chains and '(2,5,5,5)' describes the length of individual acyl chains.
The next most abundant acylsugars in M82 are S4:16 (2,4,5,5), followed by S3:22
(5,5,12) and S4:24 (2,5,5,12).

S. habrochaites LA1777 trichome acylsugar profiles were more complex than S.
lycopersicum M82, with 10 major acylsucrose analytical signals detected (Figure 4.5b).
The most striking difference was the length of acylsugar chains. The major acylsugars in
M82 consist predominantly of shorter chain FAs, especially C2 and C5. In addition to
these short chain FAs, LA1777 accumulated acylsugars with FA chains whose length
ranged from C8 to C12 (Figure 4.5a,b). M82 and LA1777 even had differences in types
of short chain FA lengths: while C4 and C5 acyl chains were of similar abundance in
LA1777, M82 mostly accumulated C5. LA1777 extracts included acylsugars with acyl
chain lengths not found at all in M82; for example, S3:18 appears to have a C8 acyl chain

	
  

152

Table 4.2. The list of analytical signals characterized as acylsugars in LC-TOF-MS
analysis.
36 analytical signals corresponding to acylsugars were selected based on retention time in
LC and accurate mass obtained from low energy condition in negative ion mode MS.
Fragment ions generated from high energy condition were used to distinguish isomers
that possess the same mass-to-charge ratio (m/z), resulting in identification of 34
acylsucroses and 2 acylglucoses. The metabolite annotation includes the type of sugar
backbone, the number of acyl chains, total number of carbons in acyl chains and expected
length for each acyl chain. For example, S3:19 (4,5,10) indicates an acylsucrose with 3
acyl chains of 4,5 and 10 carbons respectively, for a total of 19 carbons. [M+HCOOH]

-

(m/z) was obtained from the non-fragmenting condition (aperture 1 voltage of 10V) and
fragment ions from fragmenting environment (aperture 1 voltage of 55V) in MS. S4:25
(2,5,6,12)* was the most likely annotation for the corresponding analytical signal. This
was also distinct from the other isomer of m/z 793 in LC retention time. However, the
annotation is of relatively low confidence due to its low abundance and the stronger
signal intensity of m/z 723, S3:21 (5,5,11) whose retention time was very close to that of
S4:25 (2,5,6,12)*.

	
  

153

Table 4.2. Continued

Metabolite
annotation
S3:13 (4,4,5)

[M+HCOO](m/z)
611

S3:14 (4,5,5)

18.1

625

S3:15 (5,5,5)

20.5

639

S4:14 (2,4,4,4)
S4:15 (2,4,4,5)

16.5
18.2

639
653

S4:16 (2,4,5,5)

20.1

667

S3:18 (5,5,8)

27.5

681

S4:17 (2,5,5,5)

22.5

681

S3:19 (5,5,9)

29.1

695

S3:19 (4,5,10)

30.0

695

S4:18 (4,4,5,5)

24.7

695

S3:20 (5,5,10)

31.4

709

S3:20 (4,5,11)

32.5

709

S4:19 (4,5,5,5)

25.5

709

S3:21 (5,6,10)

32.2

723

S3:21 (5,5,11)

33.5

723

S3:21 (4,5,12)

34.4

723

S4:20 (2,4,4,10)

28.1

723

S4:20 (5,5,5,5)

27.1

723

S3:22 (5,5,12)

34.4

737

S4:21 (2,4,5,10)

	
  

Retention
time (min)
16.5

30.6

737

154

Fragment ions (m/z)
341, 323, 143
579, 495, 425, 411, 341,
101, 87
593, 425, 407, 341, 323,
179, 101
593, 551, 87
607, 565, 101, 87
621, 579, 537, 495, 341,
323, 101, 87
635, 551, 425, 407, 341,
179, 143, 101
635, 551, 509, 425, 407,
341, 323, 101
649, 565, 509, 425, 341,
323, 157, 101
565, 495, 425, 407, 341,
171, 101
649, 579, 509, 425, 341
663, 579, 509, 425, 407,
341, 323, 171, 101
663, 579, 495, 411, 341,
323, 185, 101, 87
663, 579, 495, 425, 341
667, 593, 579, 509, 425,
407, 341, 323, 171, 115,
101
593, 509, 425, 407, 341,
323, 185, 101
723, 677, 593, 523, 411,
325,199,101, 87
677, 635, 481, 411, 393,
341, 323, 101, 87
677, 593, 509, 425, 341,
101
691, 607, 509, 425, 407,
341, 323, 199, 179, 101
691, 649, 495, 393, 341,
323, 171, 101, 87

Table 4.2. Continued

Metabolite
annotation

Retention
time (min)

[M+HCOO](m/z)

S3:23 (5,6,12)

751

S4:22 (2,4,4,12)

31.6

751

S4:22 (2,5,5,10)

32.3

751

S3:24 (6,6,12)

38.1

765

S4:23 (2,4,5,12)

33.7

765

S4:23 (2,5,5,11)

33.4

765

S4:24 (2,5,5,12)

35.5

779

S4:24 (4,5,5,10)

33.9

779

S4:25 (5,5,5,10)

30.9

793

S4:25 (2,5,6,12)*

33.6

793

S5:25 (5,5,5,5,5)

32.1

807

S4:26 (5,5,5,11)

35.6

807

S4:27 (5,5,5,12)

37.2

821

G3:15 (5,5,5)

23.5

477

G3:22 (5,5,12)

	
  

36.3

35.3

575

155

Fragment ions (m/z)
705, 621, 607, 523,
421, 341, 323, 199,
101
709, 663, 481, 393,
323, 305, 199, 87
705, 621, 579, 425,
341, 323, 171, 101
719, 537, 439, 341,
199, 115
719, 677, 495, 411,
323, 199, 101, 87
719, 677, 509, 425,
101
733, 691, 509, 425,
341, 323, 199, 101
677, 593, 509, 425,
341, 171, 101
663, 579, 425, 341,
101
663, 579, 395, 199,
115
761, 677, 593, 509,
407, 341, 323, 101
761, 677, 593, 509,
341, 185, 101
775, 691, 607, 509,
341, 323, 199, 101
375, 329, 227, 143,
101
375, 329, 227, 199,
101

439,
115,
341,
407,
323,
341,
185,
407,
407,
171,
101,
425,
425,
425,
125,
143,

B!
100!

M82 (LA3475) !
(S. lycopersicum)!

(A)!
%!

E! F!

A!
D!
C!

0!
10.00!

20.00!

30.00!

40.00!

J!
100!

K!

(B)!

LA1777 !
(S. habrochaites)!

H!
A!
%!

D!

G!

B!

L!
M!

I!
0!
10.00!

20.00!

30.00!

40.00!
S!

100!
(C)!

LA2106!
(S. habrochaites)!

Q!
O!

%!

C!
E!
P!

N!

T!

R!

0!
10.00!

20.00!

Figure 4.5

	
  

156

30.00!

40.00!

Figure 4.5. Extracted ion chromatograms of three accessions illustrating chemical
diversity in acylsugar profiles.
Extracted ion chromatograms (XICs) were obtained from leaf dip extraction of M82 (A),
LA1777 (B) and LA2106 (C) for acylsugar-associated analytical signals (Table 4.2). The
metabolite annotations for the peaks detected with relatively high abundance are listed.
A, S4:16 (2,4,5,5); B, S4:17 (2,5,5,5); C, S3:20 (5,5,10); D, S4:22 (2,5,5,10); E, S3:22
(5,5,12); F, S4:24 (2,5,5,12); G, S4:14 (2,4,4,4); H, S4:15 (2,4,4,5); I, S3:18 (5,5,8); J,
S4:20 (2,4,4,10); K, S4:21 (2,4,5,10); L, S4:22 (2,4,4,12); M, S4:23 (2,4,5,12); N, S3:14
(4,5,5); O, S3:15 (5,5,5); P, S4:19 (4,5,5,5); Q, S4:20 (5,5,5,5); R, S3:19 (5,5,9); S, S3:21
(5,5,11); T, S3:23 (5,6,12). In M82, early-eluting high intensity peaks correspond to
acylsugars with short FA chains. LA2106 has high abundance late-eluting peaks
associated with longer FA chain-containing acylsugars. LA1777 accumulated both types
of acylsugars.

	
  

157

(5,5,8), and three have C10 acyl chains: S4:20 (2,4,4,10), S4:21 (2,4,5,10) and S4:22
(2,5,5,10). LA1777 not only produced additional major acylsugars not abundant in M82,
but there were also differences in the relative ratio of acylsugars identified in both
species. For instance, both S4:16 (2,4,5,5) and S4:17 (2,5,5,5) were found in M82 and
LA1777, but their relative abundance was opposite in the two accessions. In contrast,
S3:22 (5,5,12) and S4:24 (2,5,5,12) were detected in M82, but not in LA1777. This
comparative analysis revealed that acylsugars of two accessions from different species, S.
lycopersicum M82 and S. habrochaites LA1777, vary in the length and numbers of acyl
chains and in their relative abundance.

Comparison of trichome metabolites in the two S. habrochaites accessions LA1777 and
LA2106 also revealed dramatic differences in acylsugar composition despite the fact that
they are in the same species (compare Figure 4.5b and c). In fact, no overlap in major
acylsugars was found between these two accessions. This was in large part due to the low
abundance of acylsugars containing C2-chains in LA2106. Instead, S3:21 (5,5,11), S4:20
(5,5,5,5) and S3:15 (5,5,5) were the most abundant metabolites. In addition, LA2106
contained acylsugars with C9 and C11 acyl chains, which were not detected in LA1777
or M82. Taken together, these results show major differences in trichome acylsugar
accumulation both between and within species. This observation led us to consider
acylsugar accumulation in a larger set of S. habrochaites accessions.

	
  

158

Systematic comparison of acylsugar composition in 80 S. habrochaites accessions
To explore S. habrochaites metabolite diversity more fully, the complete set of
normalized analytical signals from trichome extracts of 80 accessions was analyzed using
hierarchical clustering analysis (HCA; Figure 4.6) and principal component analysis
(PCA; Figures 4.7 and 4.8). HCA provides an efficient approach to summarize the
detailed chemical phenotypic relationships among accessions and information about the
influence of specific acylsugars on the phenotypic clustering. PCA helps make these
trends easier to visualize.

In this analysis, the fractional peak area (analogous to mol%) for the 36 individual
analytical signals was used rather than the normalized peak areas for two reasons. First,
plant total specialized metabolite accumulation is sensitive to the environment.
Furthermore, because we do not have authentic standards for the target acylsugar
analytes, it is not possible to control for differences in relative detector response to each
of the 36 metabolites or changes in MS detector performance over time. Despite the
inherent challenge to perform absolute quantification with LC-MS, the ratio of individual
analytical signals to the total detector response was consistent across identical samples
run at different times (Figure 4.9b). In addition, the fold differences in total acylsugars in
S. habrochaites accessions and M82 tomato were modest (Figure 4.10) compared to two
previously characterized nearly isogenic S. pennellii x M82 introgression lines (ILs 5-3
and 11-3) that accumulate approximately 15-fold reduced amount of total acylsugars than
that of M82 (Schilmiller et al. 2010a; see ‘Variation of total acylsugar levels in 80 S.
habrochaites accessions’ below for details). Taken together, comparative analysis with

	
  

159

(a)
Subgroup A

*
Subgroup B
Supercluster 1!

**

Subgroup C
Subgroup D

Supercluster 2!

Subgroup E

(II)

(I)

Figure 4.6

	
  

160

Figure	
  4.6.	
  Continued	
  
	
  
	
  
(b)	
  
	
  
SubAccession
group
LA1347
A
LA1361
A
LA1362
A
LA1557
A
LA1393
A
LA1779
A
LA1392
A
LA1775
A
LA1777
A
LA2574
A
LA2976
A
LA2975
A
LA0094
B
LA1681
B
LA1691
B
LA1295
B
LA1753
B
LA1695
B
LA1696
B
LA1560
B
LA2650
B
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

Accession
LA1731
LA2722
LA1559
LA1648
M82
LA0361
LA1721
LA1918
LA1928
LA1927
LA0386
LA2552
LA1978
LA3796
LA3794
LA4137
LA1298
LA1764
LA2167
LA2329
LA3854

Subgroup
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B

161	
  

Accession
LA1265
LA1625
LA1624
LA1253
LA1255
LA2110
LA2106
LA2107
LA2104
LA2861
LA3864
LA3863
LA2101
LA2105
LA2109
LA2128
LA2119
LA2144
LA1223
LA2860
LA0407

Subgroup
C
C
C
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D

Accession
LA1266
LA1718
LA2099
LA2409
LA1033
LA2098
LA1391
LA2175
LA2812
LA1352
LA2314
LA2196
LA2204
LA2155
LA2156
LA2100
LA2864
LA2869

Subgroup
D
D
D
D
E
E
E
E
E
E
E
E
E
E
E
E
E
E

	
  
	
  
Figure	
  4.6.	
  Continued	
  
	
  
(c)	
  
	
  
Acylsugar
Group
annotation
S3:13 (4,4,5)
II
G3:22 (5,5,12)
II
S3:14 (4,5,5)
II
S3:15 (5,5,5)
II
S3:22 (5,5,12)
II
S4:27 (5,5,5,12)
II
S3:20 (5,5,10)
II
S4:25
II
S3:21 (5,6,10)
II
S4:26 (5,5,5,11)
II
S4:25 (5,5,5,10)
II
S5:25
II
(5,5,5,5,5)
S3:19 (4,5,10)
II
S3:20 (4,5,11)
II
S3:24 (6,6,12)
II
S3:21 (4,5,12)
II
G3:15 (5,5,5)
II
S4:19 (4,5,5,5)
II
S4:20 (5,5,5,5)
II
S3:21 (5,5,11)
II
S3:23 (5,6,12)
II

Acylsugar
annotation
S4:14 (2,4,4,4)
S4:15 (2,4,4,5)
S4:20 (2,4,4,10)
S4:21 (2,4,5,10)
S3:18 (5,5,8)
S4:16 (2,4,5,5)
S4:22 (2,4,4,12)
S3:19 (5,5,9)
S4:17 (2,5,5,5)
S4:24 (4,5,5,10)
S4:18 (4,4,5,5)

I
I
I
I
I
I
I
I
I
I
I

S4:23 (2,4,5,12)

I

S4:24 (2,5,5,12)
S4:22 (2,5,5,10)
S4:23 (2,5,5,11)

I
I
I

	
  

	
  

Group

162	
  

Figure 4.6. Hierarchical clustering analysis (HCA) of acylsugar chemistry in Solanum
accessions.
(a) Normalized signal intensities were used to calculate the percentage composition of 36
individual acylsugars and the dataset analyzed by HCA. Ward’s minimum variance
method was used with standardization. A red square means the % composition of a
specific acylsugar in the given accession was higher than the average for entire samples,
a gray square equal level as the average, and a blue square lower than the average. The
relative intensities of color indicate the degree of difference between the average signal
intensity for entire samples and intensity in the specific accession for the acylsugar of
interest. That is, the more dramatic difference exists, the darker color is presented in the
square. HCA resulted in 2 superclusters that consist of 2 and 3 subgroups, respectively
(subgroups A and B, and subgroups C - E). LA1777 (*) and M82 (**), two relatively
well-studied accessions, were marked. See ‘Results’ for details.
(b) Accessions numbers are listed with subgroup name. Accessions from the top of (a)
are listed first followed by ones at the bottom. For example, LA1347 is the first accession
on the top and LA2869 is the last one at the bottom in (a).
(c) Acylsugar annotations are listed following the order from left to right in (a). That is,
S3:13(4,4,5) and S4:23(2,5,5,11) are the very left and right respectively in (a).

	
  

163

fractional peak area of each acylsugars is an efficient approach to understanding diverse
acylsugar profiles across accessions.

As shown in Figure 4.6, HCA revealed two major clusters of trichome acylsugar
chemistry ('superclusters' 1 and 2) that consist of 2 and 3 main subgroups, respectively
(subgroups A and B and subgroups C, D and E). Supercluster 1 includes M82 tomato and
S. habrochaites LA1777, and is characterized by accessions with tetraacylsucroses
containing C2 chains (acylsugar group I listed on the bottom of Figure 4.6; see Figure
4.5a-b for LC-TOF extracted ion chromatograms of M82 and LA1777, respectively).
PCA also revealed that this supercluster accumulates tetraacylsucroses (subgroups A and
B in Figure 4.7) with C2 FA chains in relatively high abundance (Figure 4.8). In contrast,
supercluster 2 is chemically more diverse, and a distinguishing feature is accumulation of
triacylsucroses without C2-chains in most of these lines; this is readily seen in the
extracted ion chromatogram for LA2106 (Figure 4.5c), and more broadly for accessions
in this supercluster in Figures 4.7c and 4.8c. The general trend observed progressing from
subgroup A to E is that the proportion of 4S decreases and 3S increases (Figure 4.7).

Many of the dominant differences in trichome metabolites driving the subgroupings
within each supercluster are visible in the HCA (Figure 4.6). The two subgroups within
the tetraacylsucrose-dominant supercluster 1 are largely distinguished by the presence
(subgroup A) or absence (subgroup B) of one to two C4 acyl chains in most of the
abundant molecules. These trends in acyl chain lengths are also quite evident by PCA, as

	
  

164

Composition of glucose and
sucrose esters (%) !

	
  
0%!
0

Subgroups:
A!

LA1347!
LA1361!
LA1362!
LA1557!
LA1393!
LA1779!
LA1392!
LA1775!
LA1777!
LA2574!
LA2976!
LA2975!
LA0094!
LA1681!
LA1691!
LA1295!
LA1753!
LA1695!
LA1696!
LA1560!
LA2650!
LA1731!
LA2722!
LA1559!
LA1648!
M82!
LA0361!
LA1721!
LA1918!
LA1928!
LA1927!
LA0386!
LA2552!
LA1978!
LA3796!
LA3794!
LA4137!
LA1298!
LA1764!
LA2167!
LA2329!
LA3854!
LA1265!
LA1625!
LA1624!
LA1253!
LA1255!
LA2110!
LA2106!
LA2107!
LA2104!
LA2861!
LA3864!
LA3863!
LA2101!
LA2105!
LA2109!
LA2128!
LA2119!
LA2144!
LA1223!
LA2860!
LA0407!
LA1266!
LA1718!
LA2099!
LA2409!
LA1033!
LA2098!
LA1391!
LA2175!
LA2812!
LA1352!
LA2314!
LA2196!
LA2204!
LA2155!
LA2156!
LA2100!
LA2864!
LA2869!

Principal component 2 (26.47%) !

(A)!
4

0
Subgroup B!

* Subgroup A!
X Subgroup D!

Subgroup C!

-4
-4
Subgroup E!

0
4

Principal component 1 (57.32%) !

B!
C!

165

Principal component 2 (26.47%) !

(B)!
1
5S

0
3S

4S
3G

-1
-1
0

D!

Figure 4.7. Types of glucose and sucrose esters in Solanum accessions.
1

Principal component 1 (57.32%) !

(C)!

120%!

100%!
100

80%!

60%!

40%!

5S!
4S!
3S!
3G!

20%!

E!

Figure 4.7. Continued
A-B. Principal component analysis (PCA) was performed based upon the percentage
composition of glucose triesters (3G) as well as sucrose triesters (3S), tetraesters (4S) and
pentaester (5S), and compared with the acylsugar composition of individual accessions,
shown in part C. A. Score plots; B. Loading plots; C. The average % compositions of 4
types of acylsugars are plotted with standard error in individual accessions. The
accessions were listed in the identical order as in HCA (Figure 4.6).

	
  

166

0
0%!
Subgroups: A!

	
  
LA1347!
LA1361!
LA1362!
LA1557!
LA1393!
LA1779!
LA1392!
LA1775!
LA1777!
LA2574!
LA2976!
LA2975!
LA0094!
LA1681!
LA1691!
LA1295!
LA1753!
LA1695!
LA1696!
LA1560!
LA2650!
LA1731!
LA2722!
LA1559!
LA1648!
M82!
LA0361!
LA1721!
LA1918!
LA1928!
LA1927!
LA0386!
LA2552!
LA1978!
LA3796!
LA3794!
LA4137!
LA1298!
LA1764!
LA2167!
LA2329!
LA3854!
LA1265!
LA1625!
LA1624!
LA1253!
LA1255!
LA2110!
LA2106!
LA2107!
LA2104!
LA2861!
LA3864!
LA3863!
LA2101!
LA2105!
LA2109!
LA2128!
LA2119!
LA2144!
LA1223!
LA2860!
LA0407!
LA1266!
LA1718!
LA2099!
LA2409!
LA1033!
LA2098!
LA1391!
LA2175!
LA2812!
LA1352!
LA2314!
LA2196!
LA2204!
LA2155!
LA2156!
LA2100!
LA2864!
LA2869!

Composition of !
fatty acid chains (%) !

Principal component 2 (11.48%) !

(A)!

0
Subgroup
* Subgroup A!
B!
Subgroup C!

X Subgroup D!

Subgroup E!

-50
-50
0
50

Principal component 1 (84.53%)!

B!

C!

167

Principal component 2 (11.48%)!

50
(B)!
1
C2
C5

0
C12

C4 C8
C9
C10

-1

D!

Figure 4.8. The composition of fatty acids in trichome acylsugars.
C11

C6

-1
Principal component 1 (84.53%) !

0
1

(C)!
120%!

100%!
100

80%!

60%!

40%!

20%!

C12!
C11!
C10!
C9!
C8!
C6!
C5!
C4!
C2!

E!

Figure 4.8. Continued
The percentage composition of the 9 acyl chains, C2, 4, 5, 6, 8, 9, 10, 11 and 12, was
calculated from acylsugar profiles and used for PCA. A. Score plots; B. Loading plots; C.
The averaged composition of each acyl chain in individual accessions is shown with
standard error. The accessions were listed in the identical order as in HCA (Figure 4.6).
S4:25 (2,5,6,12) was not included in this analysis due to its relatively low confident
annotation (Table 4.2), but it only altered the second place of decimal points compared to
the calculation including S4:25 (2,5,6,12) as source of corresponding FAs.

	
  

168

70!

(A)!

Peak area IS-1 gdw-1!

60!
50!
40!
30!
20!
10!
0!
1!
1a

2!
1b

3!
1c

4!
2a

2nd batch !

Fractional composition (%) !

1st batch !

(B)!

5!
2b

6!
3a

7!
3b

8!
3c

3rd batch !

1

70!

1a!

2

80!

1b!

3

60!

4

50!

5

40!

6

30!

7

20!

8

10!
0!

1!
1

1c!
2a!
2b!
3a!
3b!
3c!

2! 3! 4! 5 6! 7! 8! 9!10
5!
10!11!12!13!14!15!16!17!18!19!20!21!22!23!24!25!26!27!28!29!30!31!32!33!34!35!36!
25
30
35
15
20

Figure 4.9. Test of reproducibility in acylsugar quantity and composition in eight M82
replicate plants.

	
  

169

!

Figure 4.9. Continued
Eight M82 replicate plants were grown and chemically analyzed in 3 separate batches
with n = 3, 2, 3, respectively.
(A) Total acylsugar amounts in individual M82 plants. The amount of total acylsugars
was obtained by summing normalized peak areas for 36 acylsugars. Even within replicate
plants from identical batch showed variation in total acylsugar amounts, for example,
M82 - 3a, 3b and 3c in the third batch.
(B) Percentage composition of 36 individual acylsugars in each M82 plants. The
fractional composition of 36 acylsugars was very similar among 8 replicate plants.

	
  

170

seen in Figure 4.8. Another distinguishing feature of major acylsugars in subgroup B
accessions is the abundance of short-chain esters, typically C2, C4 and C5. In contrast, a
broader size range of esters (C2, 4, 5, 10 and 12) are found in subgroup A with abundant
S4:14 (2,4,4,4), S4:15 (2,4,4,5), S4:20 (2,4,4,10), S4:21 (2,4,5,10) and S4:22 (2,4,4,12).

Triacylsucrose-dominated accessions in supercluster 2 split into three subgroups (C to E)
based upon various acyl chain characteristics. For example, plants in the small group C
accumulated characteristic acylsugars including S3:24 (6,6,12), while the C6 FA chain
was not commonly observed in other S. habrochaites accessions. Group C accessions,
LA1285, LA1624 and LA1625, also contained G3:15 (5,5,5) in relatively high abundance
(See Figure 4.7 for the trend of G3 accumulation in PCA and its proportion in each
accession). Group D showed a shift toward high relative abundance of acylsugars with
longer acyl chains ranging from C6 to C12. Although the accessions in this group still
accumulated minor amounts of acylsugars containing C2, major acylsugars consisted of
longer FA chains without C2 chains: S3:20 (4,5,11), S3:21 (5,5,11), S3:23 (5,6,12) and
S4:26 (5,5,5,11). Group E showed some variation within the group. For example,
LA2155 and LA2156 from group E accumulated a pentaacylsucrose (Figure 4.7).

Variation of total acylsugar levels in 80 S. habrochaites accessions
The accumulation of total acylsugars was analyzed to assess variation across accessions.
S. habrochaites accessions accumulated acylsucroses in high abundance and relatively
low or undetectable amounts of acylglucoses (Figures 4.7 and 4.10). This analysis also

	
  

171

revealed the variation of total acylsugar levels across this species, with the total quantity
obtained by summing the normalized peak areas for the 36 analytical signals in Table 4.2
as described in the methods section (Figure 4.10). With the exception of LA2101, this
analysis showed a maximum eight-fold range of total acylsugars across all accessions,
ranging from half to four-fold that of M82.

LA2101 was an exceptionally low accumulator, with approximately one-tenth as much
acylsugar detected as M82. This led to the hypothesis that LA2101 had dramatically
reduced numbers of SGTs or abnormal trichome development. To test this hypothesis,
trichomes from three S. habrochaites accessions, LA2101, LA1777 and LA2975, were
analyzed by light microscopy (Figure 4.1). These three lines were chosen as accessions
that accumulate total acylsugars in relatively low, intermediate and high levels,
respectively (Figure 4.10). This analysis did not reveal abnormally low SGT density or
aberrant morphology in LA2101 compared to the intermediate and high acylsugar level
accessions. This suggests that differences in acylsugar metabolism or storage, rather than
relative numbers of SGTs, are responsible for the reduced metabolite accumulation in
LA2101.

The accumulation of total acylsugars was also assessed to test variation among accessions
in different HCA clusters (Figure 4.6). Visual inspection of the data in Figure 4.10
suggested that lines in supercluster 1 (subgroups A and B) have higher total acylsugars
than supercluster 2. In fact, this analysis revealed that the mean accumulation of total

	
  

172

Peak area IS-1 gdw-1

350!
300!
250!
200!
TAS!
TAG!

150!
100!

0!

LA1347!
LA1361!
LA1362!
LA1557!
LA1393!
LA1779!
LA1392!
LA1775!
LA1777!
LA2574!
LA2976!
LA2975!
LA0094!
LA1681!
LA1691!
LA1295!
LA1753!
LA1695!
LA1696!
LA1560!
LA2650!
LA1731!
LA2722!
LA1559!
LA1648!
M82!
LA0361!
LA1721!
LA1918!
LA1928!
LA1927!
LA0386!
LA2552!
LA1978!
LA3796!
LA3794!
LA4137!
LA1298!
LA1764!
LA2167!
LA2329!
LA3854!
LA1265!
LA1625!
LA1624!
LA1253!
LA1255!
LA2110!
LA2106!
LA2107!
LA2104!
LA2861!
LA3864!
LA3863!
LA2101!
LA2105!
LA2109!
LA2128!
LA2119!
LA2144!
LA1223!
LA2860!
LA0407!
LA1266!
LA1718!
LA2099!
LA2409!
LA1033!
LA2098!
LA1391!
LA2175!
LA2812!
LA1352!
LA2314!
LA2196!
LA2204!
LA2155!
LA2156!
LA2100!
LA2864!
LA2869!

50!

Subgroup
A!

B

C

D

E

Figure 4.10. Normalized total acylsugar levels in S. habrochaites accessions.
The amount of total acylsugars was obtained by summing the normalized peak areas for
34 acylsucroses and 2 acylglucoses. The mean level was calculated from replicates to
obtain total acylsucrose (TAS) and total acylglucose (TAG) results, with standard error
[n=2-3 for individual accessions (except n=1 for LA1253, LA1764, LA2812 and LA2976
and n=8 for M82)]. Then, the values were further normalized so that total acylsugar
amounts can be compared across accessions including those grown and analyzed on
different batches. See ‘Materials and Methods’ for the details. The plants are listed in the
same order as in Figure 4.6.

	
  

173

0!
Subgroup A!

	
  
LA1347!
LA1361!
LA1362!
LA1557!
LA1393!
LA1779!
LA1392!
LA1775!
LA1777!
LA2574!
LA2976!
LA2975!
LA0094!
LA1681!
LA1691!
LA1295!
LA1753!
LA1695!
LA1696!
LA1560!
LA2650!
LA1731!
LA2722!
LA1559!
LA1648!
M82!
LA0361!
LA1721!
LA1918!
LA1928!
LA1927!
LA0386!
LA2552!
LA1978!
LA3796!
LA3794!
LA4137!
LA1298!
LA1764!
LA2167!
LA2329!
LA3854!
LA1265!
LA1625!
LA1624!
LA1253!
LA1255!
LA2110!
LA2106!
LA2107!
LA2104!
LA2861!
LA3864!
LA3863!
LA2101!
LA2105!
LA2109!
LA2128!
LA2119!
LA2144!
LA1223!
LA2860!
LA0407!
LA1266!
LA1718!
LA2099!
LA2409!
LA1033!
LA2098!
LA1391!
LA2175!
LA2812!
LA1352!
LA2314!
LA2196!
LA2204!
LA2155!
LA2156!
LA2100!
LA2864!
LA2869!

Peak area IS-1 gdw-1
300!

250!

200!

150!

100!
TAS!
TAG!

50!

B
C

174
D

E

Figure 4.11. Total acylsugar levels in S. habrochaites accessions without additional

normalization for batch-to-batch comparison.

Total acylsugar levels were calculated using identical set of data used for Figure 4.10.

However, while the values for Figure 4.10 were further normalized for batch-to-batch

comparison, the additional normalization was not applied in this analysis. However, two

analyses, with and without additional normalization, mean results were within ±15%.

acylsugars in accessions from supercluster 1 was approximately 2-fold higher than those
of supercluster 2 (p <10-7, Student's t-Test).

The correlation between geographic distribution and acylsugar chemistry in S.
habrochaites
We investigated whether collection sites of S. habrochaites accessions are related to
acylsugar chemistry (Figure 4.12). In fact, the C2 sucrose tetraester-containing
supercluster1 plants (subgroups A and B) were collected from southern Peru, and
supercluster 2 plants (subgroups C, D and E) were distributed in northern Peru and
Ecuador.

Discussion
This study revealed that S. habrochaites acylsugars are a structurally diverse group of
glucose and sucrose-based polyesters with short- to medium-length (C2-C12) acyl chains.
These metabolites are produced in the SGTs of plants in the Solanaceae, sometimes
exuded in large enough amounts to cause the surface of the tissue to become sticky.
Understanding the biosynthesis and diversity of these molecules is of interest because
they have documented roles in plant defense (Kang et al. 2010a, Mirnezhad et al. 2010,
Weinhold and Baldwin 2011), yet little is known about their biosynthesis.

We are taking a multi-faceted approach to understand the biosynthesis and chemical
diversity

of

SGT

specialized

metabolites

in

tomato

and

wild

relatives

(http://www.trichome.msu.edu). Previous work with a limited number of accessions of

	
  

175

several Solanum species revealed differences in the sugar backbone and overall types of
acyl chains (McDowell et al. 2011, Schilmiller et al. 2010a). More recently we chose S.
habrochaites for intensive study because of the availability of dozens of accessions, with
geographic

information

about

collection

sites

for

many

of

these

lines

(http://tgrc.ucdavis.edu/).

Despite the lack of availability of authentic standards for all but one of the acylsugars
being studied, reverse-phase LC separation coupled with high mass accuracy negativeion-mode TOF-MS provides information about the sugar backbone as well as acyl chain
number and lengths for 36 SGT acylsugars detected in this study (Figure 4.2, Table 4.2,
Figure 4.4). All but two are sucrose esters; the glucose triesters, G3:15 (5,5,5) and G3:22
(5,5,12), found in a total of 5 accessions in groups C and E represented ≤ 7% of total
acylsugars in these lines (Figures 4.6 and 4.7). Total acylsugar amount was estimated to
vary by 8-fold across the 80 accessions (Figures 4.10 and 4.11) and diversity in numbers
and lengths of acyl chains was found that was beyond our expectations for accessions of
plants in the same species (Figures 4.6-8).

Acylsugar chemotypes
Based on HCA and PCA of the acylsugar profiles, the accessions define two
superclusters and at least five subgroups, and these results are helpful in revealing
patterns in the complex dataset. One general observation is that the M82 cultivated
tomato had less diverse acylsugar profiles than most of the wild species accessions (see

	
  

176

Figure 4.12. Correlation between geographic distribution and acylsugar chemistry in S.
habrochaites accessions.

	
  

177

Figure 4.12. Continued
Accessions are coded based upon subgroups indentified by HCA (Figure 4). There is
little overlap between the two superclusters: Supercluster 1 plants (subgroups A and B),
which have C2-containing sucrose tetraesters, were collected in southern Peru, and
supercluster 2 plants (subgroups C, D and E) were distributed in northern Peru and
Ecuador. Although all accessions whose collection site information was available (Table
1) were plotted, some accessions are almost superimposed with others in this map due to
their close geographic location.

	
  

178

Figures 4.5 and 4.6). This is presumably due to genetic bottlenecks introduced during
domestication and breeding for S. lycopersicum varieties with desirable agronomic
properties (Rodriguez-Saona et al. 2011). Our analysis revealed several major qualitative
differences in SGT acylsugar profiles. A main distinction is the presence or absence of
acylsugars containing C2 acyl chains, which is a key feature separating superclusters 1
and 2 (Figures 4.6-8). Within the tetraacylsucrose-dominated supercluster 1, subgroup B
accessions mainly have C2, C4 and C5 acylchains, whereas subgroup A acylsugars have
a more varied range of acylchain lengths. In addition to the general lack of C2 esters,
supercluster 2 has a diverse array of acylsugars, including pentaacylsucrose S5:25
(5,5,5,5,5), C6-containing esters and the glucose triesters G3:15 (5,5,5) and G3:22
(5,5,12) (Figures 4.6-8).

Interestingly, four of six of the S. lycopersicum M82 acylsugars detected contain C2 acyl
chains, in common with the chemotypes of nearly 30 S. habrochaites supercluster 1
accessions. This is in contrast to the supercluster 2 accessions, which either do not have
detectable C2-containing acylsugars, or only accumulate measurable amounts of one C2containing metabolite: S4:25 (2,5,6,12) based on its current annotation (Table 4.2). This
makes the acylsugar chemotype of the supercluster 1 S. habrochaites accessions more
similar to that of S. lycopersicum M82 than to other conspecific accessions.

Genetic differences in presence or absence of acylsucrose C2 chains were previously
found in a comparison between S. lycopersicum M82 and the overlapping nearly isogenic
chromosome substitution lines IL1-3 and IL1-4. The ILs, accumulated S3:15 (5,5,5) in

	
  

179

high abundance while S4:17 (2,5,5,5) was predominant in M82 parent. This result is
consistent with the hypothesis that an acetyltransferase active in S. lycopersicum is
defective or modified in the S. pennellii IL1-3 region, causing lack of enzymatic activity
or altered substrate specificity. It is interesting to speculate that the differences in
acylsucrose acetylation in S. habrochaites supercluster 1 and 2 accessions may be caused
by presence or absence of the same acetyltransferase(s). In this scenario, SGT expression
of the enzyme responsible for adding the C2 chain to sucrose esters would predate the
split leading to S. habrochaites and S. lycopersicum, and have been lost in some S.
habrochaites lines.

Furthermore, HCA results helped us speculate possible biosynthetic pathways based on
relationships of particular acylsugars in clustering (Figure 4.6). For instance, four major
acylsugars, S4:14 (2,4,4,4), S4:15 (2,4,4,5), S4:20 (2,4,4,10) and S4:21 (2,4,5,10), were
closely clustered in LA1777. Two acylsugars of high abundance, S4:22 (2,5,5,10) and
S4:23 (2,5,5,11), are another example for close chemical relationships (acylsugar group I
in Figure 4.6). Based on their close relationships in the clusters, these compounds are
thought to behave in a similar manner in LA1777. The results suggest one possibility that
these acylsugars share common substrates and/or enzymes for the production of the
compounds in LA1777 as proposed in Figure 4.13. Although characterization of enzymes
and substrate intermediates is required to test the proposed biosynthetic pathways, our
acylsugar profiling study helped us generate hypotheses for further elucidation of
acylsugar biochemistry.

	
  

180

Sucrose!

S2:6 (2,4)
C4 AT

C5 AT

S3:10 (2,4,4)!
C4 AT!

S3:12 (2,5,5)!

S3:11 (2,4,5)!

C10 AT!

C4 AT

C10 AT

C10 AT!

C11 AT

C5 AT
S4:14 !
(2,4,4,4)!

S4:15 !
(2,4,4,5)!

S4:20 !
(2,4,4,10)!

S4:15
(2,4,4,5)!

S4:21
(2,4,5,10)!

S4:22
(2,5,5,10)!

S4:23 !
(2,5,5,11)!

Figure 4.13. Hypothetical biosynthetic pathways to produce a subset of major acylsugars
in LA1777.
HCA revealed chemical relationship of acylsugars in S. habrochaites accessions (Figure
4.6). In particular, two subsets of highly abundant acylsugars in LA1777 showed close
relationships in chemical clustering, S4:14 (2,4,4,4), S4:15 (2,4,4,5), S4:20 (2,4,4,10) and
S4:21 (2,4,5,10), and S4:22 (2,5,5,10) and S4:23 (2,5,5,11), respectively (group I in
Figure 4.6). Their close relationships in the clustering led to a hypothesis that they share
common substrates and/or enzymes for the production. Hypothetical biosynthetic
pathways leading to the production of these acylsugars in LA1777 were proposed,
assuming the reactions are sequential. For chemical compounds, identical nomenclature
was used as Table 4.2. ‘AT’ indicates acyltransferase.

	
  

181

Geographical Context
Our results showing geographical differences in S. habrochaites acylsugar profiles across
Peru and Ecuador extend published studies of S. pennellii accessions (Shapiro et al.
1994). Shapiro and coworkers reported that accessions from the southern part of the
range, including the widely-studied LA0716, predominately accumulate glucose esters
with C4-C12 chains. As S. pennellii accessions were sampled going from south to north
of their range in the Andes Mountains, they found an increase in the ratio of sucrose to
glucose esters. Our results show that there is a strong prevalence of C2-containing
tetraacylsucroses from southern S. habrochaites accessions (supercluster 1), whereas
supercluster 2 accessions predominated from northern Peru and Ecuador. These findings
raise the question of whether the geographic clustering of plants with similar acylsugar
chemistry results primarily from incomplete population dispersal or is the result of
adaptation to the local abiotic or biotic environments.

Summary
LC-TOF-MS screening of S. habrochaites SGT acylsugars revealed diverse acylsugar
chemotypes. The varied lengths of acyl-chains and combinations of numbers and types of
esters in the 36 acylsugar structures suggests the interplay of a varied set of enzymes in S.
habrochaites acylsugar biosynthesis. Knowing which metabolites accumulate in specific
accessions − for instance presence or absence of C2 esters, accumulation of unusual
length acyl chains such as C6 or C8 or synthesis of penta-acylsucrose S5:25 (5,5,5,5,5) −
will inform selection of accessions to study the biosynthesis and functions of acylsugars.
Once the genes for these enzymes are known, it should be possible to infer the changes in

	
  

182

gene sequences, gene expression and enzyme activities that lead to some or all of the
metabolic diversity. Furthermore, because there is information about the collection sites
of many of these accessions, it should become possible to place the chemical diversity
into an ecological context.

Materials and methods
Plant growth, metabolite extraction and identification.
Germplasms for M82 cultivated tomato and 80 S. habrochaites accessions were obtained
from the Tomato Genetic Resource Center (http://tgrc.ucdavis.edu/). Plant seedlings were
grown and non-volatile leaf surface and trichome metabolites were extracted as
previously described in Schilmiller et al. 2010a. Two to three biological replicates
(individual plants grown in the same growth chamber) were prepared for most S.
habrochaites accessions and 8 replicates for M82 cultivated tomato were analyzed; only
one plant was screened for accessions of S. habrochaites that had poor germination and
growth (n=1 for LA1253, LA1764, LA2812 and LA2976).

Chemical analysis by LC-TOF-MS
LC-TOF-MS (Shimadzu LC-20AD pumps; Waters LCT Premier mass spectrometer) was
used for the chemical analysis of non-volatile metabolites by modification of published
methods (Gu et al. 2010, Schilmiller et al. 2010a). Ten microliters of extract was injected
into a BetaBasic C18 HPLC column (1 x 150 mm with 3.0 µm particles, Thermo
Scientific, http://www.thermoscientific.com/) maintained at 35°C. Two solvents were
used as mobile phase (solvent A – 0.15% formic acid in MilliQ water and solvent B –

	
  

183

methanol), with a 0.05ml/min flow rate. The gradient profile was as follows: 5% B for
initial step; linear gradient to 50% B at 5 min; linear gradient to 95% B at 33 min; 100%
B from 33-38 min; 5% B at 43 min. TOF-MS with multiplexed collision-induced
dissociation was performed by switching Aperture 1 voltages (10, 25, 40, 55 and 80 V)
for quasi-simultaneous production of spectra without and with fragment ions. Lowest
energy spectra were used for quantification of analytical signals, while spectra generated
at 55V were used for metabolite structural information.

Characterization of analytical signals for acylsugars and their metabolite annotation
Analytical signals obtained from negative ion mode LC-TOF-MS analysis were
investigated to identify acylsugar-associated signals by using a series of criteria: relative
mass defect, LC retention time and mass-to-charge ratio (m/z) under low ionization
conditions. Fragment ions produced from high collision energies in MS were used for
metabolite annotation.

Specifically, the analytical signals were obtained from LC-TOF-MS analysis with 43min
LC separation and m/z 50-1,500 MS detection range. The signals were collected using
MarkerLynx software (Waters, http://www.waters.com/). The relative mass defect of
each resulting signal was calculated to identify acylsugar metabolites. Mass defect was
computed with accurate mass information obtained from TOF following the equation,
[(experimental m/z – nominal value of m/z) ÷ experimental m/z] x 106 (ppm). As relative
mass defect is correlated with the number of hydrogen atoms in the compound, it is
helpful in categorizing the class of chemical compounds (Last et al. 2007, Stagliano et al.

	
  

184

2010). In this study compounds with relative mass defect between 200-700ppm, as an
inclusive cut-off, were used to exclude analytical signals that are not associated with
acylsugars. Known acylsugars from the Solanacea have relative mass defects in the range
of 300-600 ppm, whereas earlier eluting compounds with < 300ppm often correspond to
flavonoids that have lower hydrogen content. In contrast, membrane lipids
characteristically have signals with relative mass defects > 600ppm.

LC retention times were used in addition to accurate mass to define analytical signals.
Acylsugars eluted after glycosylated alkaloid tomatine from the reverse-phase analysis in
HPLC due to lower aqueous solubility of acylsugars. Furthermore, a previous study
reported a list of m/z for acylsugar-associated analytical signals detected in relatively
high abundance in M82 cultivated tomato (Schilmiller et al. 2010). The combination of
LC retention time and m/z list enabled identification of 36 analytical signals that
correspond to acylsugars in S. habrochaites accessions. From the previous study, m/z 821
was the largest value found for acylsugar-associated signals in M82 cultivated tomato.
Thus, analytical signals with successive increase of m/z 14 (corresponding to methylene
groups) were searched from m/z 821 to query whether there are any potential signals for
novel acylsugars detected in S. habrochaites accessions. Signals with incremental
decrease of m/z 14 were also searched from the minimal value reported for cultivated
tomato. However, neither approach found additional acylsugar-associated analytical
signals.

	
  

185

For the 36 analytical signals identified as acylsugars, metabolite annotation was obtained
based on co-eluting fragment ions generated at aperture 1 higher voltage conditions.
Information obtained from this analysis included the number of acyl chains and lengths
of individual acyl chains. As a result, analytical signals for 34 acylsucroses and 2
acylglucoses were identified (Table 4.2).

Quantitative data analysis and statistical tests
For quantitative analysis, peak areas were integrated for extracted ion chromatograms for
the

individual

analytical

signals

using

QuanLynx

software

(Waters,

http://www.waters.com/). All peaks that have the same m/z and retention time were
confirmed to be identical compounds by assessing fragment ion masses generated using
elevated collision energies. However, when a number of peaks with same m/z were seen
within relatively small retention time window, minor peaks whose abundance is nearly
above detection level were not chosen for peak integration. This approach resulted in 36
distinct acylsugars for this chemical analysis. The integrated peak area for each specific
signal was then normalized to internal standard (propyl-4-hydroxybenzoate, SigmaAldrich, http://www.sigmaaldrich.com/) and leaf dry weight.

An estimate of the amount of total acylsugars was obtained by summing the normalized
peak areas for 36 analytical signals (Table 4.2). As the chemical analysis for 80
accessions was performed in 3 different batches with separate plant growth, harvests and
LC-MS performances, the value was further normalized to make a batch-to-batch
comparison possible (Figure 4.10). Every batch included M82 cultivated tomato plants as

	
  

186

control (n=8 in total; 3, 2, 3 per planting and analysis, respectively), and their profiles
were qualitatively similar with minor variation of quantity (Figure 4.9). Thus, the average
of all 8 M82 plants was used to normalize values from the individual batches.

All statistical tests were performed by JMP 8.0 software (SAS Institute,
http://www.jmp.com/). For HCA, the percent composition of 36 analytical signals was
computed individually for every plant extract and the average was obtained for the
individual accessions with available replicates. The averaged value was standardized by
JMP 8.0 software before HCA so that all variables would have equal impact in the
clustering process. Ward’s minimum variance method was used for HCA (Milligan
1980).

For the analysis of sucrose- and glucose ester composition by PCA, the same data set was
used as for HCA. The normalized average peak areas for 36 analytical signals were
categorized to acylsucroses or acylglucoses. Acylsucroses were further grouped to 3S, 4S
or 5S categories according to the number of acyl chains. 3G was only type of acylglucose
identified in this study. The peak areas for each category were summed, their percent
composition was calculated and PCA performed (Figure 4.7).

The normalized average peak areas were also used to calculate the composition of FA
chains. The level of FAs in acylsugars was calculated based on metabolite annotation for
each analytical signal. For instance, if the normalized peak area of S4:20 (2,4,4,10) is 1,
the level of C2, 4 and 10 were weighted 1, 2 and 1 fold, respectively. In this way, the

	
  

187

level of FAs was estimated from the individual analytical signals. The resulting values
were summed for individual FAs, ranging from C2 to 12, and then the percent of each
FAs was calculated. Then, the composition was analyzed by PCA (Figure 4.8).

The geographic distribution of S. habrochaites accessions
For the original collection site of S. habrochaites accessions, latitude and longitude
information was obtained from the Tomato Genetic Resource Center, except for the
accessions lacking the information in the database (Table 4.1). Locations were mapped
using DIVA-GIS software (http://www.diva-gis.org/).

	
  

188

References

	
  

189

References
Bohlmann, J. and Gershenzon, J. (2009) Old substrates for new enzymes of terpenoid
biosynthesis. Proc. Natl. Acad. Sci. U S A, 106, 10402-10403.
Fobes, J.F., Mudd, J.B. and Marsden, M.P.F. (1985) Epicuticular lipid accumulation
on the leaves of Lycopersicon pennellii (Corr.) D'Arcy and Lycopersicon
esculentum Mill. Plant Physiol., 77, 567-570.
Fridman, E., Wang, J., Iijima, Y., Froehlich, J.E., Gang, D.R., Ohlrogge, J. and
Pichersky, E. (2005) Metabolic, genomic, and biochemical analyses of glandular
trichomes from the wild tomato species Lycopersicon hirsutum identify a key
enzyme in the biosynthesis of methylketones. Plant Cell, 17, 1252-1267.
Ghangas, G.S. and Steffens, J.C. (1993) UDPglucose: fatty acid transglucosylation and
transacylation in triacylglucose biosynthesis. Proc. Natl. Acad. Sci. U S A, 90,
9911-9915.
Gu, L., Jones, A.D. and Last, R.L. (2010) Broad connections in the Arabidopsis seed
metabolic network revealed by metabolite profiling of an amino acid catabolism
mutant. Plant J., 61, 579-590.
Kandra, G., Severson, R. and Wagner, G.J. (1990) Modified branched-chain amino
acid pathways give rise to acyl acids of sucrose esters exuded from tobacco leaf
trichomes. Eur. J. Biochem., 188, 385-391.
Kang, J.H., Liu, G., Shi, F., Jones, A.D., Beaudry, R.M. and Howe, G.A. (2010a) The
tomato odorless-2 mutant is defective in trichome-based production of diverse
specialized metabolites and broad-spectrum resistance to insect herbivores. Plant
Physiol., 154, 262-272.
Kang, J.H., Shi, F., Jones, A.D., Marks, M.D. and Howe, G.A. (2010b) Distortion of
trichome morphology by the hairless mutation of tomato affects leaf surface
chemistry. J. Exp. Bot., 61, 1053-1064.
Kennedy, G.G. (2003) Tomato, pests, parasitoids, and predators: tritrophic interactions
involving the genus Lycopersicon. Annu. Rev. Entomol., 48, 51-72.
Last, R.L., Jones, A.D. and Shachar-Hill, Y. (2007) Towards the plant metabolome and
beyond. Nat. Rev. Mol. Cell. Biol., 8, 167-174.
Liu, B., Wang, H., Du, Z., Li, G. and Ye, H. (2011) Metabolic engineering of
artemisinin biosynthesis in Artemisia annua L. . Plant. Cell. Rep., 30, 689-694.

	
  

190

Luckwill, L. (1943) The genus Lycopersicon: A historical, biological, and taxonomic
survey of the wild and cultivated tomatoes Aberdeen, Scotland: Aberdeen
University Press.
McDowell, E.T., Kapteyn, J., Schmidt, A., Li, C., Kang, J.H., Descour, A., Shi, F.,
Larson, M., Schilmiller, A., An, L., Jones, A.D., Pichersky, E., Soderlund,
C.A. and Gang, D.R. (2011) Comparative functional genomic analysis of
Solanum glandular trichome types. Plant Physiol., 155, 524-539.
Milligan, G.W. (1980) An examination of the effect of 6 types of error perturbation on
15 clustering algorithms. Psychometrika, 45.
Mirnezhad, M., Romero-Gonzalez, R.R., Leiss, K.A., Choi, Y.H., Verpoorte, R. and
Klinkhamer, P.G. (2010) Metabolomic analysis of host plant resistance to thrips
in wild and cultivated tomatoes. Phytochem. Anal., 21, 110-117.
Rodriguez-Saona, C., Vorsa, N., Singh, A.P., Johnson-Cicalese, J., Szendrei, Z.,
Mescher, M.C. and Frost, C.J. (2011) Tracing the history of plant traits under
domestication in cranberries: potential consequences on anti-herbivore defences.
J. Exp. Bot., 62, 2633-2644.
Sallaud, C., Rontein, D., Onillon, S., Jabes, F., Duffe, P., Giacalone, C., Thoraval, S.,
Escoffier, C., Herbette, G., Leonhardt, N., Causse, M. and Tissier, A. (2009)
A novel pathway for sesquiterpene biosynthesis from Z,Z-farnesyl pyrophosphate
in the wild tomato Solanum habrochaites. Plant Cell, 21, 301-317.
Schilmiller, A., Shi, F., Kim, J., Charbonneau, A.L., Holmes, D., Daniel Jones, A.
and Last, R.L. (2010a) Mass spectrometry screening reveals widespread
diversity in trichome specialized metabolites of tomato chromosomal substitution
lines. Plant J., 62, 391-403.
Schilmiller, A.L., Last, R.L. and Pichersky, E. (2008) Harnessing plant trichome
biochemistry for the production of useful compounds. Plant J., 54, 702-711.
Schilmiller, A.L., Miner, D.P., Larson, M., McDowell, E., Gang, D.R., Wilkerson, C.
and Last, R.L. (2010b) Studies of a biochemical factory: tomato trichome deep
expressed sequence tag sequencing and proteomics. Plant Physiol., 153, 12121223.
Schilmiller, A.L., Schauvinhold, I., Larson, M., Xu, R., Charbonneau, A.L.,
Schmidt, A., Wilkerson, C., Last, R.L. and Pichersky, E. (2009) Monoterpenes
in the glandular trichomes of tomato are synthesized from a neryl diphosphate
precursor rather than geranyl diphosphate. Proc. Natl. Acad. Sci. U S A, 106,
10865-10870.

	
  

191

Shapiro, J.A., Steffens, J.C. and Mutschler, M.A. (1994) Acylsugars of the wild
tomato Lycopersicon pennellii in relation to geographic distribution of the
species. Biochem. Syst. Ecol., 22, 545-561.
Shepherd, R.W., Bass, W.T., Houtz, R.L. and Wagner, G.J. (2005) Phylloplanins of
tobacco are defensive proteins deployed on aerial surfaces by short glandular
trichomes. Plant Cell, 17, 1851-1861.
Sifres, A., Blanca, J. and Nuez, F. (2011) Pattern of genetic variablity of Solanum
habrochaites in its natural area of distribution. Genet. Resour. Crop. Ev., 58, 347360.
Slocombe, S.P., Schauvinhold, I., McQuinn, R.P., Besser, K., Welsby, N.A., Harper,
A., Aziz, N., Li, Y., Larson, T.R., Giovannoni, J., Dixon, R.A. and Broun, P.
(2008) Transcriptomic and reverse genetic analyses of branched-chain fatty acid
and acyl sugar production in Solanum pennellii and Nicotiana benthamiana. Plant
Physiol., 148, 1830-1846.
Stagliano, M.C., DeKeyser, J. G., Omiecinski, C. J., and Jones, A. D. (2010) 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, 3578-3584.
Thipyapong, P., Joel, D.M. and Steffens, J.C. (1997) Differential expression and
turnover of the tomato polyphenol oxidase gene family during vegetative and
reproductive development. Plant Physiol., 113, 707-718.
Wagner, G.J. (1991) Secreting glandular trichomes: more than just hairs. Plant Physiol.,
96, 675-679.
Walters, D.S. and Steffens, J.C. (1990) Branched chain amino acid metabolism in the
biosynthesis of Lycopersicon pennellii glucose esters. Plant Physiol., 93, 15441551.
Wang, G., Tian, L., Aziz, N., Broun, P., Dai, X., He, J., King, A., Zhao, P.X. and
Dixon, R.A. (2008) Terpene biosynthesis in glandular trichomes of hop. Plant
Physiol., 148, 1254-1266.
Weinhold, A. and Baldwin, I.T. (2011) From the cover: trichome-derived O-acyl sugars
are a first meal for caterpillars that tags them for predation. Proc. Natl. Acad. Sci.
U S A, 108, 7855-7859.
Werker, E. (2000) Trichome diversity and development. In Advances in Botanical
Research Incorporating Advances in Plant Pathology, pp. 1-35.

	
  

192

Yu, G., Nguyen, T.T., Guo, Y., Schauvinhold, I., Auldridge, M.E., Bhuiyan, N., BenIsrael, I., Iijima, Y., Fridman, E., Noel, J.P. and Pichersky, E. (2010)
Enzymatic functions of wild tomato methylketone synthases 1 and 2. Plant
Physiol., 154, 67-77.
Yu, H., Kowalski, S.P. and Steffens, J.C. (1992) Comparison of polyphenol oxidase
expression in glandular trichomes of Solanum and Lycopersicon species. Plant
Physiol., 100, 1885-1890.
Zhou, B., Li, Y., Xu, Z., Yan, H., Homma, S. and Kawabata, S. (2007) Ultraviolet Aspecific induction of anthocyanin biosynthesis in the swollen hypocotyls of turnip
(Brassica rapa). J. Exp. Bot., 58, 1771-1781.

	
  

193