T’ *‘r
l (”[851

_ ‘ LlBRARY
s" ’3?th MiChigi-i‘ :State
UniverSItY

 

 

 

 

This is to certify that the

dissertation entitled

CLONING AND CHARACTERIZATION
OF CAROTENOID HYDROXYLASES
IN ARABIDOPSIS THALIANA

presented by

LI TIAN

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in _ELam;_B;Langy

M64;

Major professor

Date fi/ll/OB

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

PLACE IN RErURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

JUN 0 5 2008

 

 

J IN 0 A gnns

 

' '3‘ 2 3V8

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/Dateouopes-p. 15

CLONING AND CHARACTERIZATION OF CAROTENOID HYDROXYLASES IN
ARABIDOPSIS THALIANA

By

Li Tian

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

2003

ABSTRACT
CLONING AND CHARACTERIZATION OF CAROTENOID HYDROXYLASES IN
ARABIDOPSIS THA LIA NA
By

Li Tian

Carotenoids are isoprenoid-derived pigments that are synthesized by bacteria,
fungi, and plants. Oxygenated carotenoids, xanthophylls, are essential components of the
light harvesting complexes and play critical roles in photoprotection. Lutein and
zeaxanthin are dihydroxy xanthophylls produced from their corresponding carotene
precursors by the action of [3- and s-ring carotenoid hydroxylases. B-hydroxylases add
hydroxyl groups to the B-rings of carotenes and have been cloned from several bacteria
and plants. s-hydroxylase hydroxylates on e-rings and was genetically defined by the lutI
mutation in Arabidopsis, but not yet cloned. The main goal of this dissertation was to
clone and characterize the Arabidopsis e-hydroxylase gene by molecular, genomic, and
genetic approaches. The in viva functions of B- and e-hydroxylases as well as
xanthophylls were studied in parallel.

We proposed three hypotheses for the identity of the Arabidopsis s-
hydroxylase/L UT 1 locus: 1) it is a hydroxylase evolved from and related to the [3-
hydroxylase. 2) it is an independently evolved hydroxylase and divergent from the B-
hydroxylase. 3) it is an ancillary protein that modifies the existing B-hydroxylase to
function towards e-rings. In an attempt to test the first hypothesis, 1 screened cDNA and

genomic libraries from Arabidopsis using the Arabidopsis B-hydroxylase cDNA as a

probe. As a result of the screening, I cloned and characterized a second B-hydroxylase in
Arabidopsis (Chapter 2). The encoded protein shares high identity to the previously
reported Arabidopsis B-hydroxylase 1 and functions efficiently on B-rings (but not 8-
rings). Neither B-hydroxylase cDNA maps to the LUTI locus, and the genomic region
encompassing the LUTI locus does not contain a third related hydroxylase. In order to
test the second and third hypotheses, I identified the LUTI gene by map-based cloning
and found that it encodes a cytochrome P450 type monooxygenase (Chapter 3).

In order to address the individual and overlapping functions of the three
Arabidopsis carotenoid hydroxylase activities in vivo, T-DNA knockout mutants
corresponding to B-hydroxylase l and 2 (b1 and b2, respectively) were isolated and all
possible hydroxylase mutant combinations were generated together with lut] (Chapter 4).
Carotenoid composition in leaf and seed, NPQ induction and amplitude were analyzed in
these mutants. Overall, the hydroxylase mutants provide insight into the unexpected
overlapping activity of carotenoid hydroxylases in vivo.

I have also used mutations in Arabidopsis thaliana that selectively eliminate (and
substitute) specific xanthophylls in order to study their function(s) in vivo (Chapter 5).
These include two lutein-deficient mutants, lutl and Iut2, the epoxy xanthophyll-deficient
abal mutant, and the luIZabaI double mutant. Photosystem stoichiometry, antenna sizes,
and xanthophyll cycle activity have been related to alterations in non-photochemical
quenching of chlorophyll fluorescence (N PQ). Altered NPQ in xanthophyll biosynthetic
mutants is explained by disturbed macro-organization of LHC II and reduced PS II-

antenna size in the absence of the optimal, wild type xanthophyll composition.

This dissertation is dedicated to my parents, J inxiang Tian and Baoqin Chang.

ACKNOWLEDGMENTS

First and foremost, I would like to thank my mentor, Dr. Dean DellaPenna, for his
guidance, encouragement, and patience during my graduate study. Dean has helped me to
become a better scientist and a mature person. It has always been a privilege to work in
his lab. I would also like to thank my committee members, Drs. Katherine Osteryoung,
Christoph Benning, and John Ohlrogge for their advises, suggestions, and stimulating
discussions.

Barry, Heiko, Dave, Max, and Steve, former postdocs in the lab, have guided me
through the early years of my graduate school and taught me useful molecular biology
and photobiology techniques. I would like to thank Dr. Tom Newman for instructions on
Real Time PCR and Dr. Jan Zeevaart for helping me with ABA analysis. Maria and
Valeria have provided excellent technical support for my research.

I would also like to thank Joe, Craig, and the rest of the DellaPenna lab members
for teaching me interesting American slang. I will always remember the sunny summer
days when I went rock-climbing with Craig, Dave, Max, and Steve. I also enjoyed
collecting fossils and rocks with Heiko. I missed the fun times and good food that I had
with Molly, Haiming, and Xiaoxia in Reno. Woonbong and Xiaohua have been my
inspiration for playing tennis. Dean showed me the way to the putting green. Uwe gave
me my very first sailing lesson. I want to thank Elena for inviting me to her wonderful

parties, Laura for sharing books with me, and Eva for always being a friend.

There has been emotional ups and downs over the past few years, but I am so
grateful that I have my loving parents and sister (and my very cute nephew Tongtong)
who have always truly believed in me and offered me their love and support.

I am pleased to finally see a closure for my graduate study and am excited to

welcome a new beginning.

vi

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................... x
LIST OF FIGURES ............................................................................................................ xi
CHAPTER 1
INTRODUCTION ............................................................................................................... l
1.1 Carotenoid biosynthesis ................................................................................................. 2
1.1.1. Terpenoid biosynthesis .................................................................................. 2
1.1.2. The non-mevalonate pathway for carotenoid biosynthesis in plastids .......... 2
1.1.3. Carotenoid biosynthesis in higher plants ....................................................... 7
1.1.4. Cytochrome P450 monooxygenase .............................................................. 26
1.1.4.1. Cytochrome P450 monooxygenase and carotenoid biosynthesis.26
1.1.4.2. Localization of the plant cytochrome P4505 ................................. 28
1.1.4.3. Cytochrome P450 functions in plants and monooxygenation
reaction mechanism ....................................................................... 29
1.1.4.4. Arabidopsis NADPH-Cytochrome P450 reductases (CPR) ......... 31
1.1.4.5. Current genomics approach for cytochrome P450 characterization
........................................................................................................ 34
1.1.5. Model for possible carotenoid biosynthetic enzyme complexes ................. 35
1.2. Regulation and metabolic engineering of the carotenoid pathway ............................. 39
1.2.1. Regulation of carotenoid biosynthesis ......................................................... 39
1.2.2. Metabolic engineering ................................................................................. 41
1.3. Carotenoid functions in higher plants ......................................................................... 45
1.3.1. Zeaxanthin as a membrane stabilizer ........................................................... 45
1.3.2. Carotenoid functions in light harvesting and photoprotection ..................... 46
1.3.2.1. Light harvesting ............................................................................ 47
1.3.2.2. The role of carotenoids in photoprotection ................................... 48
1.3.2.3. The role of carotenoids in NPQ .................................................... 48
1.3.2.4. NPQ is measured by way of chlorophyll fluorescence ................. 50
1.3.2.5. The mechanism of qE ................................................................... 51
1.4. Approaches taken in this dissertation for characterization of carotenoid hydroxylases
............................................................................................................................................ 55
1.4.1. Map-based cloning of the lat] mutation ...................................................... 56
1.4.2. T-DNA insertional mutants as a tool for in vivo function study .................. 58
References .......................................................................................................................... 61
CHAPTER 2
CHARACTERIZATION OF A SECOND CAROTENOID B-HYDROXYLASE GENE
FROM ARABIDOPSIS AND ITS RELATIONSHIP TO THE LUTI LOCUS ............... 73
Abstract .............................................................................................................................. 74
Introduction ........................................................................................................................ 75
Materials and methods ....................................................................................................... 77
Results ................................................................................................................................ 84

vii

Discussion .......................................................................................................................... 94

Acknowledgements ............................................................................................................ 99
References ........................................................................................................................ 100
CHAPTER 3

THE ARABIDOPSIS LUTI LOCUS ENCODES A CAROTENOID e-HYDROXYLASE
THAT IS A MEMBER OF THE CYTOCHROME P450 FAMILY ............................... 102
Abstract ............................................................................................................................ 103
Introduction ...................................................................................................................... 104
Results .............................................................................................................................. 107
Discussion ........................................................................................................................ 1 19
Methods ............................................................................................................................ 129
Acknowledgements .......................................................................................................... 1 34
References ........................................................................................................................ 1 3 5
CHAPTER 4

FUNCTION ANALYSIS OF B- AND 8- RING CAROTENOID HYDROXYLASES IN
ARABIDOPSIS T HALIANA ............................................................................................. 138
Abstract ............................................................................................................................ 139
Introduction ...................................................................................................................... 141
Results .............................................................................................................................. 145
Discussion ........................................................................................................................ 1 63
Methods ............................................................................................................................ 176
Acknowledgements .......................................................................................................... 1 82
References ........................................................................................................................ 1 83
CHAPTER 5

XANTHOPHYLL BIOSYNTHETIC MUTANTS OF ARABIDOPSIS T HALIANA.
ALTERED NONPHOTOCHEMICAL QUENCHIN G OF CHLOROPHYLL
FLUORESCENCE IS DUE TO CHANGES IN PHOTOSYSTEM II ANTENNA SIZE

AND STABILITY ........................................................................................................... 186
Abstract ............................................................................................................................ 187
Introduction ...................................................................................................................... 1 88
Materials and Methods ..................................................................................................... 191
Results .............................................................................................................................. 193
Discussion ........................................................................................................................ 208
Acknowledgements .......................................................................................................... 213
References ........................................................................................................................ 214
CHAPTER 6
FUTURE DIRECTIONS ................................................................................................. 217
6.1. Developmental responses to carotenoid deficiency in Arabidopsis hydroxylase
mutants ............................................................................................................................. 218
6.1.1 ABA biosynthesis in the hydroxylase deficient mutants ............................ 218
6.1.2. Chloroplast ultrastructures in the hydroxylase mutants ............................. 220
6.1.3. Lipid composition in the hydroxylase mutants .......................................... 222

viii

6.2. Mutants screening ..................................................................................................... 222

6.3. Structural and mechanistic basis of the B- and e— hydroxylases ............................... 224
6.4. Hydroxylase mutants response to environmental stresses ........................................ 225
References ........................................................................................................................ 226
APPENDIX A

ANALYSIS OF CAROTENOID BIOSYNTHETIC GENE EXPRESSION DURING
MARIGOLD PETAL DEVELOPMENT ........................................................................ 228
Abstract ............................................................................................................................ 229
Introduction ...................................................................................................................... 230
Materials and methods ..................................................................................................... 233
Results .............................................................................................................................. 237
Discussion ........................................................................................................................ 257
Acknowledgements .......................................................................................................... 262
References ........................................................................................................................ 263

LIST OF TABLES

Table 1.1. Classification of terpenoids in higher plants ..................................................... 3
Table 1.2. Arabidopsis cytochrome P4505 with predicted chloroplast localization ......... 12
Table 1.3. Arabidopsis cytochrome P4503 with predicted chloroplast localization ......... 30
Table 2.1. TaqMan probes and primers used in this study ................................................ 80
Table 3.1. Wild type and mutant carotenoid composition in leaf tissue quantified by
HPLC ............................................................................................................................... 116
Table 4.1. Wild type and mutant carotenoid composition in leaf tissue quantified by
HPLC ............................................................................................................................... 153
Table 4.2. Wild type and mutant carotenoid composition in seed quantified by

HPLC ............................................................................................................................... 157
Table 4.3. Chl a/b ratios and maximum PSII efficiencies (F v/F m) in wild type and
mutant leaves ................................................................................................................... 160
Table 4.4. Sequences of TaqMan probes and primers .................................................... 180

Table 5.1. Chl a/b-ratios, photosystem stoichiometries, antenna sizes and maximum PS 11
efficiencies (Fv/Fm) in A. thaliana WT and xanthophyll mutants .................................. 194

Table 5.2. Carotenoid composition of A. thaliana WT and xanthophyll mutant leaves.l97

Table 5.3. Pigment composition of A. thaliana WT and xanthophyll mutant light-

harvesting complexes as resolved by nondenaturing electrophoresis as displayed in
Figure 5.4 ........................................................................................................................ 202

Table A.1. Information about the genes discussed in this report including the number of
nucleotides of each clone, the number of amino acids encoded by each cDNA, and the
accession number is presented ......................................................................................... 240

LIST OF FIGURES

Figure 1.1. Isoprenoid biosynthesis in higher plants .......................................................... 5
Figure 1.2. Carotenoid biosynthesis in higher plants ........................................................ 1 1
Figure 1.3. Map locations of carotenoid biosynthetic genes in Arabidopsis .................... 14
Figure 1.4. Carotenoid desaturation reactions in plants and bacteria ............................... 18
Figure 1.5. Lycopene cyclization mechanism ................................................................... 21
Figure 1.6. Cytochrome P450 reaction mechanism .......................................................... 33
Figure 1.7. Proposed model for carotenogenic complexes in chloroplasts of plants ........ 38

Figure 2.1. Alignment of proteins encoded by the Arabidopsis B-hydroxylase 1 and 2
cDNAs ................................................................................................................................ 86

Figure 2.2. HPLC elution profiles of B-hydroxylase 1 and 2 cDNAs expressed in B or 8-
ring producing E. coli ........................................................................................................ 89

Figure 2.3. A. Schematic diagrams of B-hydroxylase 1 and 2 genomic clones. B. Map
positions of B-hydroxylase 1, B-hydroxylase 2 and LUTI locus ....................................... 91

Figure 2.4. B-hydroxylase 1 and 2 mRNA levels in Arabidopsis leaf, flower, stem, root
and silique tissues .............................................................................................................. 93

Figure 2.5. Neighbor-joining tree for deduced amino acid sequences of B-hydroxylase

genes from plants and bacteria ........................................................................................... 97
Figure 3.1. Carotenoid hydroxylation reactions .............................................................. 105
Figure 3.2. Positional cloning of the LUTI gene ............................................................ 109
Figure 3.3. A. Identification of the [w] mutation. B. Deduced amino acid sequence of

LUTI . C. Predicted membrane topology of LUTl .......................................................... 1 13
Figure 3.4. HPLC elution profiles of wild type, lutI-Z, and lut1-3 plants ...................... 115
Figure 3.5. RT-PCR reactions using LUTl specific primers .......................................... 117

Figure 3.6. A. The hydroxylation reactions of B- and e- ring.3-D structures of a- and B-
carotene substrates for hydroxylation reactions ............................................................... 123

xi

Figure 3.7. Oxidizing intermediates for cytochrome P450 and non-heme di-iron
monooxygenases .............................................................................................................. 125

Figure 4.1. Xanthophyll biosynthesis in Arabidopsis ..................................................... 142

Figure 4.2. A. [Li-hydroxylase 1 and B-hydroxylase 2 insertional mutations (b1 and b2,
respectively). B. RT-PCR reactions with RNA extracted from wild type (WS) and

homozygous b1 and b2 knockout mutants ....................................................................... 147
Figure 4.3. Six—week old wild type and mutant Arabidopsis plants ............................... 149
Figure 4.4. TaqMan Real Time PCR analysis of B-hydroxylase 1 and 2, lycopene B- and

e-cyclase mRNA levels in wild type and mutant genotypes ............................................ 152
Figure 4.5. NPQ analysis of wild type and mutant genotypes ........................................ 162
Figure 4.6. Hydroxylation levels of B- and 8- rings in all genotypes ............................. 170

Figure 4.7. Accumulation of B,B-carotene and [Le-carotene branch carotenoids in leaves
of wild type and selected hydroxylase mutants .............................................................. 174

Figure 5.1. A. Nonphotochemical fluorescence quenching (NPQ) in leaves of A. thaliana
WT and xanthophyll biosynthetic mutants, as a function of time of exposure to 500 umol

photons m'zs". B. NPQ as a function of actinic light intensity ........................................ 195

Figure 5.2. Kinetics of V de-epoxidation via A to Z in leaves of A. thaliana WT (A) and
the lut] (B) and lut2 (C) mutants exposed to a PFD of 1000 umol photons m'zs'l ........ 199

Figure 5.3. Light activation of NPQ in leaves of A. thaliana WT, the [at] and lut2
mutants ............................................................................................................................. 203

Figure 5.4. Nondenaturing ("green") gel electrophoretic separation of pigment-protein
complexes from A. thaliana thylakoid membranes of the indicated genotypes .............. 205

Figure A.1. Non-mevalonate isoprenoid and carotenoid biosynthetic pathway in plants
.......................................................................................................................................... 239

Figure A.2 A. Stages of marigold petal development are depicted. B. Fully expanded
flowers from the four marigold cultivars discussed in this report are shown .................. 246

Figure A.3. Carotenoid levels in leaves and during various stages of petal development
in the different marigold cultivars .................................................................................... 248

Figure A.4. HPLC analysis of pigments from marigold leaves and petals of the cultivar
Dark Orange Lady ............................................................................................................ 249

xii

Figure A.5. Northern blot of RNA from stage 3 and 4 of petals of each of the indicated
cultivars and from leaves of the Dark Orange Lady cultivar ........................................... 253

Figure A.6. Northern blots of RNA from stage 3 and 4 petals of each of the indicated
cultivars and from leaves of the Dark Orange Lady cultivar ........................................... 255

Figure A.7. Northern blots of RNA from stage 3 and 4 petals of each of the indicated
cultivars and from leaves of the Dark Orange Lady cultivar ........................................... 258

xiii

CHAPTER 1

Introduction

1.1 Carotenoid biosynthesis

1.1.1 Terpenoid biosynthesis

Terpenoids are one of the major groups of secondary metabolites that are
produced by plants. All terpenoids, including those derived from both primary and
secondary metabolism, are constructed from the five-carbon isoprene unit, isopentenyl
pyrophosphate (IPP) (Table 1.1). Terpenoids are also known as isoprenoids because of
these isoprene building blocks. The common precursors for biosynthesis of all terpenoids
are IPP and dimethylallyl pyrophosphate (DMAPP). Two major biosynthetic pathways
that operate in different subcellular compartments lead to IPP production for different
groups of terpenoids in plants (Figure 1.1). Cytoplasmic IPP is derived from the
mevalonate (MVA) pathway and leads to the synthesis of sesquiterpene (C15) and
triterpene (C30). The plastidic IPP biosynthetic pathway begins with condensation of
glyceraldehyde-3-phosphate (G3 P) and pyruvate, and gives rise to monoterpene (Clo),
diterpene (C20), and tetraterpene (C40) classes of terpenoids, the latter of which includes

tocopherols, phylloquinones, and carotenoids.

1.1.2 The non-mevalonate pathway for carotenoid biosynthesis in plastids

The classic mevalonate pathway was discovered in the 1950's and has been

studied in detail in fungi, animals, and plants (Goodwin, 1971; Schroepfer, 1981). The

mevalonate pathway was considered to be the sole source of IPP for plant isoprenoid

 

Table 1.1. Classification of terpenoids in higher plants.

 

lsoprene
1

(30)th

Carbon

5
1 0
1 5
20
30
40

>40

Class name
hemiterpene
monoterpene
sesquiterpene
diterpene
triterpene
tetraterpene
polyterpene

Examples
isoprene
Iimonene, menthol
phytoalexin, abscisic acid
taxol, casbene, phytol, gibberellin
brassinosteroid
carotenoid, tocopherol, phylloquinone

rubber, plastoquinone, ubiquinone

 

 

Figure 1.1. Isoprenoid biosynthesis in higher plants. Biosynthetic enzymes of both
mevalonate and DXP pathways and their locations in the cell are indicated. HMG-CoA,
hydroxymethylglutaryl CoA; MVA, mevalonate; MVP, 5-phosphomevalonate; MVPP, 5-
diphosphomevalonate; IPP, isopentenyl pyrophosphate; DMAPP, dimethylallyl
pyrophosphate; F PP, farnesyl diphosphate; GGPP, geranylgeranyl pyrophosphate; DXP,
l-deoxyxylulose 5-phosphate; MEP, 2-C-methyl-D-erythritol 4-phosphate; CDP-ME, 4-
diphosphocytidyl-Z-C-methyl-D-erythritol; CDP-MEP, 4-diphosphocytidyl-2-C-methyl-
D-erythritol 2-phosphate; ME-cPP, 2-C-methyl-D-erythritol 2,4-cyclodiphosphate ;
HMBPP, hydroxymethyl-butenyl 4-diphosphate.

AACT, acetoacetyl CoA thiolase (EC 2.3.1.9); HMGS, HMG-CoA synthase (EC
4.1.3.5); HMGR, HMG-CoA reductase (EC 1.1.1.88); MVK, MVA kinase (EC 2.7.1.36);
PMK, MVP kinase (EC 2.7.4.2); PMD, MVPP decarboxylase (EC 4.1.1.33); IDI, IPP
isomerase (EC 5.3.3.2); GPS, GPP synthase (EC 2.5.1.1); FPS, F PP synthase (EC
2.5.1.10); GGPS, GGPP synthase (EC 2.5.1.29); DXS, DXP synthase (EC 4.1.3.37);
DXR, DXP reductoisomerase (EC 1.1.1.267); CMS, CDP-ME synthase (EC 2.7.7.60);
CMK, CDP-ME kinase (EC 2.7.1.148); MCS, ME-2,4cPP synthase (EC 4.6.1.12); HDS,
HMBPP synthase; IDS, IPP/DMAPP synthase.

This figure is adapted from Figure 1 in Rodriguez-Concepcion and Boronat (2002).

Plant Physiol. 130, 1079-1089.

 

 

 

mgobfiogmmEm Alim§£m2>cm

 

Recs—coina—

 

0558.

 

m:_=0..on£0 ‘/

maOG

/ moccasgammi
m 8:295
macaw" u 0

226380... b\

          

Baa...

£05520."—

£328.25
nae

mx ma

 

 

k

/<

  

 

 

 

 

   

E33030

3.599533

622.00

an“. i mEouoa ho co=m§c9a

\

mssxoso

an: Allin—<20

«x
man.

an“.

F

8:05:55:

     

£2225:

 

 

Figure 1.1

biosynthesis for almost four decades until the discovery of non-mevalonate pathway in
bacteria and plants in the late 80's and early 90's (F lesch and Rohmer, 1988; Rohmer et
a1., 1993; Schwender et a1., 1996). Incorporation of l3C-labelled glucose, acetate,
pyruvate, or erythrose into the hopanoids and/or ubiquinones in several bacteria
(Zymomonas mobilis, Methylobacteriumfujisawaense, Escherichia coli, and
Alicyclobacill acidoterrestris) yielded a labeling pattern not compatible with the MVA
pathway and suggested a novel route for the early steps of isoprenoid biosynthesis. It was
suggested that the C5 isoprene unit was produced from the condensation of a C2 unit
derived from pyruvate decarboxylation and a triose phosphate (Rohmer et a1., 1993;
Horbach et a1., 1993). Subsequently, similar labeling studies were undertaken in plants
(Schwender et a1., 1996) and indicated the presence of a pathway other than the MVA
pathway for [PP biosynthesis in higher plants. This new IPP biosynthetic pathway has
been called the "Rohmer pathway" after the scientist who first discovered the pathway, or
the "Alternative pathway" to indicate that it is an alternative to the classic mevalonate
pathway. Following identification of the key metabolites in the pathway, it is also called
the "Pyruvate/glyceraldehyde 3-phosphate (G3P) pathway" (the substrate), the
"Deoxyxylulose 5-phosphate (DXP) pathway" (the first pathway intermediate), or the "2-
C-methyl-D-erythritol 4-phosphate (MEP) pathway" (the second pathway intermediate).
Combined molecular biology, genetics, biochemistry, organic chemistry, and
bioinformatics approaches have been applied to first define the DXP pathway in plants
and then identify the genes of the pathway in bacteria and plants. The first step in the
DXP pathway is the head to head condensation of G3P and pyruvate to form l-deoxy-D-

xylulose-S-phosphate (DXP) by DXP synthase (DXPS). DXP is then reduced by DXP

reductoisomerase (DXPR) to form MEP. DXPS and DXPR have been cloned from a
number of plants and bacteria (Mandel, 1996; Kuzuyama et a1., 1998; Lois et a1., 1998;
Takahashi et a1., 1998). MEP is then converted to 2-C-methyl-D-erythritol 2,4-
cyclodiphosphate (ME-2,4cPP) in three enzymic steps catalyzed by CDP-ME synthase,
(CMS), CDP-ME kinase (CMK), and ME-2,4cPP synthase (MCS), encoded by yng,
ychB, and ygbB, respectively, in E. coli (Figure 1.1.; Rohdich et a1., 1999; Herz et a1.,
2000; Ltittgen et a1., 2000). CMS condenses MEP and CTP to form 4-diphosphocytidyl-
2-C-methyl-D-erythritol (CDP-ME). CDP-ME is phosphorylated by CMK to form 4-
diphosphocytidyl-Z-C-methyl-D-erythritol 2-phosphate (CDP-ME2P). The CMP
molecule is then eliminated from CDP-ME2P to produce ME-2,4cPP by MCS.

As a biochemically and genetically well studied organism, E. coli has been a target
for DXP pathway research and all the pathway genes have been initially cloned from
E. coli (reviewed in Rohmer, 1999; Lichtenthaler, 2000). Comparative genomics has
facilitated the isolation of DXP pathway genes in plants. All plant DXP pathway enzymes
have putative signal peptides for chloroplastic localization (reviewed in Lichtenthaler,
1999; Rodriguez-Concepcién and Boronat, 2002). The DXP pathway is the source of IPP
for the biosynthesis of carotenoids and other chloroplastic metabolites, such as
tocopherols and gibberellins. Therefore, the regulation of the DXP pathway may be an
important control point in quantitative production of carotenoids and other plastidic.

isoprenoids.

1.1.3 Carotenoid biosynthesis in higher plants

 

Carotenoids are best known as colorants that are widespread in plants and
microorganisms. Carotenoids are also present in animals where they are not synthesized
but rather obtained from food. More than 700 different carotenoids are now known
(Britton, 1998). This part of the introduction will focus on the reactions and enzymes that
belong to the main carotenoid biosynthetic pathway common to all higher plants, namely
leading from phytoene to neoxanthin synthesis. Additional reactions and enzymes are
involved in the production of a large variety of carotenoids that accumulate in
microorganisms and will not be reviewed in this chapter.

The light absorbing properties of carotenoids derive from their conjugated double
bond system in which n-electrons are delocalized over the entire polyene chain. This
conjugated double bond system is also called a chromophore due to its firnction in light
absorption. Carotenoids with 3-5 conjugated double bonds absorb light in the UV region
and are colorless. Carotenoids with seven or more conjugated double bonds absorb light
in the visible region, which give such carotenoids their yellow, orange, or red colors.

In plants, carotenoids are essential components of the photosystem reaction
centers and the light harvesting complexes. B-carotene is mainly associated with the
reaction centers while lutein, violaxanthin, and neoxanthin are associated with the light
harvesting complexes. Carotenoid biosynthesis is localized in the chloroplast and the
pathway enzymes are present at very low abundance. Most carotenoid enzymes are
associated with the thylakoid membrane and are sensitive to detergent treatment, which
makes purification of carotenoid biosynthetic enzymes difficult and in only a few cases
has been successful (Camara and Moneger, 1981; Beyer et a1., 1985). Using a molecular

biology approach and elegant E. coli based screening and complementation systems

 

(Cunningham et a1., 1994; Lotan and Hirschberg, 1995), genes for carotenoid
biosynthesis have been isolated from various sources in recent years, including E. coli,
cyanobacteria, green algae, and higher plants (Figure 1.2; Table 1.2).

The first committed step in carotenoid biosynthesis is the formation of phytoene
from two geranylgeranyl pyrophosphate (GGPP) molecules by phytoene synthase (PSY)
(Figure 1.3). The condensation of two GGPP molecules requires the abstraction of one
hydrogen from the C(l) position of each GGPP and results in a lS-cis double bond in
phytoene in higher plants. The PSY gene has been cloned and found to be structurally and
functionally conserved in plants and bacteria (Sandmann, 1994; Schledz et a1., 1996). The
enzymes from both organisms contain conserved prenyl transferase domains that
resemble squalene synthase, an enzyme that catalyzes the condensation of two farnesyl
pyrophosphate (FPP) molecules to form squalene (Sandmann, 1994; Schledz et a1.,
1996). In vitro analysis with daffodil PSY showed that the enzyme requires galactolipids
for catalytic activity. In addition, both soluble and membrane-bound forms of PSY are
detected in the daffodil chromoplast stroma (Schledz et a1., 1996). The soluble form of
PSY is inactive and becomes activated upon membrane attachment. The active
membrane-bound form of PSY is consistent with the presence of lipophilic phytoene and
the existence of subsequent carotenoid biosynthetic steps in the thylakoid membrane.
Although the plastid HSP70 was highly induced during flower development and
previously found to be associated with phytoene desaturase, its association with PSY was
not observed in daffodil chromoplasts (Al-Babili et a1., 1996; Schledz et a1., 1996).

Phytoene is colorless because of its short (triene) chromophore. The formation of

colored carotenoids requires the extension of the conjugated double bond system by a

 

 

Figure 1.2. Map locations of carotenoid biosynthetic genes in Arabidopsis. PS Y,
phytoene synthase; PDS, phytoene desaturase; ZDS, Q-carotene desaturase; C C R 1,
carotenoid isomerase 1; CCRZ, carotenoid isomerase 2; LYCfl, lycopene B-cyclase; LYCg,
lycopene e-cyclase; ,BCHI, B-carotene hydroxylase 1; ,BCHZ, B-carotene hydroxylase 2;

LUTI , s-hydroxylase; ZEP, zeaxanthin epoxidase; VDE, violaxanthin deepoxidase.

Figure 1.2

VDE

CCR2

CCR1

N

ZDS
LYCfl
PSY
PDS
flCH 1
LUT1
flCH2
LYC a
ZEP

 

 

Table 1.2. Arabidopsis carotenoid biosynthetic genes.

 

Gene

PSY
PDS
ZDS
CCR1
CCR2
LYCB
LYCe
BCH1
BCH2
eCH
ZEP
VDE

Enzyme

phytoene synthase
phytoene desaturase
q-carotene desaturase
carotenoid isomerase 1
carotenoid isomerase 2
lycopene B-cyclase
lycopene e-cyclase
carotenoid B-hydroxylase 1
carotenoid B-hydroxylase 2
carotenoid a-hydroxylase
zeaxanthin epoxidase

violaxanthin deepoxidase

Identifier

At5917230
At4gl4210
At3904870
n . a.
At1906820
At3g10230
At5957030
At4g257oo
At5952570
At3953130
At5967030
At1908550

Location
(Chromosome)

5

4

3

Mutant

n.a.
n.a.
n.a.
ccr1
ccr2
n.a.
Iut2
b1
b2
Iut1
npq2 (abet)

npq1

 

n.a., not available.

 

 

Figure 1.3. Carotenoid biosynthesis in higher plants. The Arabidopsis mutations that
block specific steps of the pathway are indicated. lutI, lut2, abaI (npq2), npq1, ccrI and
cch are mutations in a-hydroxylase, lycopene e-cyclase, zeaxanthin epoxidase,
violaxanthin deepoxidase, carotenoid isomerase 1, and carotenoid isomerase 2,
respectively. pets] and pdsZ disrupt phytoene desaturation reactions but are not the

mutations in the phytoene desaturase gene.

GGPP X 2

W

pds1. Pd82 T Phytoene

 

Phytofluene
ccrt,ccr2 *

i C-carotene

l Neurosporene
Lycopene

lut2
l y-ca rotene
8-carotene

M ii 't' .1..." :3
l ...........
MM ”fit ' '

B-cry ptoxanthln

Iutt + Zeinoxanthin ~
”OOH
abet Zeaxanthin
(um?) +4— npq1

Lutein

abet Antheraxanthln
(npq2) npqt

m
0

:0 ‘3

l Violaxanthin

Neoxanthln

OH

Figure 1.3

 

series of desaturation reactions. The phytoene desaturation reactions proceed with
elimination of two hydrogen atoms and formation of a double bond. Four successive
desaturations take place alternatively on each half of phytoene to yield phytofluene, @-
carotene, neurosporene, and lycopene, respectively. The desaturation reactions are
catalyzed by a single enzyme phytoene desaturase (CRT I) in bacteria and fungi. In
higher plants, these four double bonds are introduced by the combined action of two
enzymes, phytoene desaturase (PDS) and C-carotene desaturase (ZDS) (Figure 1.4). PDS
requires flavin adenine dinucleotide (FAD) as a cofactor and plastoquinones are
intermediate electron acceptors in the reaction in higher plants (Norris et a1., 1995;
Schledz et a1., 1996). Bleaching herbicides like norflurazon, can inhibit PDS directly and
therefore prevent the synthesis of colored carotenoid and cause accumulation of colorless
phytoene (Britton, 1998). Another class of herbicides, triketones, also causes
accumulation of phytoene and bleaching of tissue by inhibiting the biosynthesis of the
electron transporter, plastoquinone (Britton, 1998) (Figure 1.4). It was proposed that
bacteria share a common desaturation mechanism with higher plants, although bacteria
use ubiquinones rather than plastoquinones as electron acceptors. This hypothesis was
supported by the experimental evidence that the plant PDS gene can be expressed and is
active in E. coli; bacterial CRT I can be transformed into and still be active in plants
(Bartley et a1., 1999; Ye et a1., 2000).

The Arabidopsis immutans (im) mutant disrupts a plastidic terminal oxidase that
is homologous to mitochondrial alternative oxidase and shows a variegated phenotype
(Carol et a1., 1999). The albino sectors of the im mutants contain decreased levels of

carotenoids and increased levels of phytoene suggesting a role for IM protein in the

 

phytoene desaturation reactions. The alternative oxidase in mitochondria oxidizes the
ubiquinone pool using molecular oxygen as a terminal acceptor. Because PDS requires
plastoquinone for activity, it was proposed that IM protein acts as a terminal oxidase and
is associated with the transport of electrons from PDS (Carol et a1., 1999) (Figure 1.4).

In nature, lycopene and many other colored carotenoids exist in the all-trans
configurations. However, the biosynthetic precursor phytoene is synthesized in thel 5-cis
configuration. Therefore, isomerization of the C (15,15') double bond must occur at a
certain stage of carotenoid biosynthesis. The existence of a potential carotenoid
isomerase enzyme was suggested from the phenotype of recessive mutations in tomato
(Tomes et a1., 1953) and Scenedesmus (Ernst and Sandmann, 1988), which accumulate
prolycopene (7Z,9Z,7'Z,9'Z tetra-cis-lycopene), poly-cis-isomers of phytofluene, @-
carotene, and neurosporene. This hypothesis was confirmed by cloning of isomerase
genes from both tomato and Arabidopsis (Isaacson et a1., 2002; Park et a1., 2002). The
tangerine mutant of tomato (Lycopersicon esculentum) produces fruits that are orange
instead of red and accumulate prolycopene (72, 92, 7'2, 9'Ztetra—cis-1ycopene) instead of
all-trans lycopene. Using map-based cloning, the gene encoding tangerine, CRTISO, was
isolated. It encodes a redox-type enzyme that is structurally related to the bacterial PDS.
E. coli cells transformed with the Erwinia GGPP synthase (GGPS) and PSY genes and
Synechococcus PCC7942 PDS gene accumulated lS-cis-C-carotene. When the tomato
ZDS gene was introduced into this E. coli background, prolycopene (72, 92, 7'2, 9'2
tetra-cis-lycopene) was produced. Expressing both of the tomato ZDS and CR TISO genes
in the 15-cis-Q-carotene producing E. coli strain led to significant accumulation of all-

trans-lycopene. This result fiirther confirmed the isomerization activity of CRTISO.

 

Figure 1.4. Carotenoid desaturation reactions in plants and bacteria. Proposed

integration of desaturation reactions with electron transport carriers in plants is indicated.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ecoaou>_

III}.

_ 2.203232.

 

 

 

 

 

 

 

.. ..r
m P
o I can—2:502
5. 05:59.3...» cogs—to:
No a: 1/ I—l
3053688.."— AI ram—«u Al moN
tcnncah £05020 \
ozofigeoaonm
_w.
of ma.
CONE—.532 .
.2. 23:23.5 canes—to: w
No S / I—I I—I .m
1 {Ma r. W
tone—.2. c9820 3
0305:3205.
i 25:

30: :2620 3323.

$2

2.89.3.0

Cit/C(rjé

”cwzccgcn

{$3.31

2.08%...

TIE/3243a

”3'"

tfifis
int."
7

“."V "CN'V’ IVIa". -‘ - . _ ‘8:\»

m?
‘.'.

:10

“I b 1 --"‘t‘1'.\." . _‘
311-". N-Xi 33.1 ,4

v as
‘ ‘.-‘ 5

3"»! \ K'-.-._. ‘.
' -'.7'.’A‘3§:-‘\'~\\‘,~; -,

32%???

l'0"|“""~.- \‘J N ,5
F'R-wzesassur

2.20am

Figure 1.4.

18

 

Figure 1.4. Carotenoid desaturation reactions in plants and bacteria. Proposed

integration of desaturation reactions with electron transport carriers in plants is indicated.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

econ—00>.

\(Irikéé

_ 0:22.352.

 

 

 

 

 

 

 

 

a ...t....,.
O I canoes—be: .
E. o:o...~ao_:m nouns—to:
No a: / Lo<z
3053623... Al 9:. AI moN
ton—men: 9.2820 \
ozefigmouozn
.p
a .
O I 35550: are
E. ecBhaoEm ecu—23:32
No a: v/ 1.32 I...“
oeosauoummi Al 9E Al won.
toamcmb c2820 \
03055320....
A 355

30:. £95020 0>3N~3l

\r\/\r\/\r\’f\’\J\’\J}/\J\

ocououm9U

g

econ—$235.

u

r (12..

- s9
. . .

“at“mw‘w

. n -., -. .
. "832‘.” 2“,: .

_hmo

:5?i'§¢?-EE-§L¥~.ifii

a ‘ ""'~V\-"~~ 4“,. a V,‘ '(‘u' '\"o"‘. V'Q'(‘- -. ‘y‘.
1‘. _,. v . . . g4 . _. .
'-‘~\. . sister} 1.1923? rt‘, ,‘,".<'°':‘:.3 .«t at

.23“
.. :b.

,. ..«-
43.5"

t‘it.

.W

.3.

2.30.3

Figure 1.4.

18

 

Orthologs of CRTISO were also identified in Arabidopsis and cyanobacteria (Masamoto
et a1., 2001; Park et a1., 2002). Characterization of the carotenoid isomerase in tomato and
Arabidopsis also raised the possibility of photoisomerization of prolycopene to form all-
trans-lycopene (Isaacson et a1., 2002; Park et a1., 2002). This hypothesis was supported
by the fact that carotenoid isomerase deficiency in Synechocystis mutant was
compensated by illumination (Masamoto et a1., 2001). Evolutionarily, the occurrence of
carotenoid isomerase has been correlated with the presence of the poly-cis-desaturation
pathway that utilizes PDS and ZDS enzymes in higher plants and cyanobacteria.

The subsequent cyclization of lycopene is an important branching point for
carotenoid biosynthesis. It is thought that the straight chain end group of lycopene is first
folded to the desired shape upon binding of the substrate by the cyclase enzyme, followed
by electrophilic attack at the C(2) position and formation of the C(1)-C(6) bond to close
the ring (Figure 1.5). Loss of a proton at the C(6) position leads to the formation of the 0-
ring; alternatively, loss of a proton at C(4) forms the c-ring (Figure 1.5). Carotenoids with
two B-end groups (6. g. B-carotene and zeaxanthin) and one B- and one 8- end group (e. g.
lutein) are most abundant in plants and perform critical functional roles. The formation of
two symmetrical B-end groups is catalyzed by lycopene B-cyclases. Lycopene e-cyclases
are related enzymes that only catalyze the formation of one e-ring and in combination
with lycopene B-cyclase, a-carotene (B,e-carotene) is produced in higher plants
(Cunningham et a1., 1996). Carotenoids with two e-rings are uncommon in plants and
only accumulate to substantial levels in lettuce.

Lycopene B- and e- cyclases have been cloned from various plants; both enzymes

show high similarity at the protein sequence level and likely evolved from the same

19

Figure 1.5. Lycopene cyclization mechanism. The flexible acyclic end group of the
substrate is first folded into the required shape and then a concerted electrophilic attack
occurs at the C2 position, which closes C1-C6 bond to form the ring. A loss of HA results

in the formation of the B-ring and an alternative loss of HB leads to the formation of the 8-

ring.

20

“as“
H +

 

 

[3 ring a ring

Figure 1.5

21

 

ancestor (Cunningham et a1., 1994; Hugueney et a1., 1995; Cunningham et a1., 1996).
The e-cyclase that catalyzes the formation of two e-rings has been cloned from Romaine
lettuce (Krubasik and Sandmann, 2000). By using a domain swapping approach, a six-
amino acid domain was identified for determining whether one or two e-rings are
produced (Cunningham and Gantt, 2001 ). Further mutagenesis studies showed that a
single histidine residue in the lettuce e-cyclase is the determining factor for the formation
of two e-rings (Cunningham and Gantt, 2001).

In addition to cyclization of end groups, another key control point for carotenoid
biosynthesis is the introduction of various oxygen functional groups into carotenes to
produce the oxygenated derivatives of carotenes, xanthophylls. Oxygenation reactions are
catalyzed by a large and diverse group of enzymes, such as hydroxylases, ketolases, and
epoxidases. A wide range of xanthophylls is formed from these reactions, including the
most abundant carotenoids in photosynthetic plant tissues: lutein, neoxanthin, and
violaxanthin. The most common oxygenation reactions are the hydroxylation of B-
carotene (B,B-carotene) and or-carotene (B,e-carotene) to form zeaxanthin (B,B-carotene-
3,3'-diol) and lutein (B,e-carotene-3,3’-diol), respectively. Early isotope labeling studies
with 180 showed that the hydroxyl groups in these reactions originated from molecular
oxygen rather than from water (Yamamoto et a1., 1962). Both 0- and e-ring
hydroxylations take place by direct replacement of the hydrogen atom at the C(3)
position of the ring structure by a hydroxyl group. Hydroxylations on the C(3) carbon of
the ionone ring is stereospecific in all cases. All B-ring hydroxyl groups (e.g. B-
cryptoxanthin, zeaxanthin, and the B-ring of lutein) have the same chirality and are

designated as 3R. The chirality of C(3) on the a-ring is opposite to B-ring, but is also

22

 

designated as 3R according to the sequence rule (Britton, 1998).

B-hydroxylase enzymes have been cloned from a range of bacteria, fungi, and
higher plants (Misawa et a1., 1990; Hundle et a1., 1993; Sun et a1., 1996; Bouvier et a1.,
1998; Masamoto et a1., 1998; Tian and DellaPenna, 2001). Although all B-hydroxylases
carry out the same reaction, they can be clearly categorized into three groups based on
their primary structures. Eukaryotic B-hydroxylases are highly conserved and share less
than 30% protein identity with B-hydroxylases from non-photosynthetic bacteria. The B-
hydroxylase of Synechocystis PCC6803 (a cyanobacteria) lacks similarity to plant and
non-photosynthetic bacterial B-hydroxylases with the exception of conserved histidine
residues (Masamoto et a1., 1998). Synechocystis B-hydroxylase is more closely related to
bacterial B-carotene ketolases (B-ketolase), suggesting that it may have evolved
independently of the plant and bacterial B-hydroxylases.

As a group, the three classes of B-hydroxylases have overall low protein identity
but contain conserved histidine regions that have been previously identified in the
membrane fatty acid desaturases (Shanklin et a1., 1994). Like the fatty acid desaturases,
the B-hydroxylases are non-heme di-iron proteins that require iron, ferredoxin, and
ferredoxin oxido-reductase for activity. They are able to directly oxidize the carbon
atoms on the ionone ring structures. In contrast to the well-studied B-hydroxylases, the
enzyme involved in e-ring hydroxylation has only been genetically identified by the [w]
mutation in Arabidopsis (Pogson et a1., 1996), and at the time of this writing is the only
remaining gene in the main carotenoid biosynthetic pathway in plants to be cloned
(Figure 1.3). The publication of the Arabidopsis genome in 2000 allowed thorough

database searches with known B-hydroxylase sequences from bacteria, fungi, and higher

23

plants, but did not identify a possible LUT1 ortholog. Furthermore, IutI did not map to
either of the two cloned B-hydroxylases in Arabidopsis (Tian and DellaPenna, 2001). The
isolation and characterization of the LUT1 gene is a major focus of this dissertation and
will be discussed in chapter 3.

In photosynthetic and non-photosynthetic bacteria, green algae, fungi, and higher
plants, there are also carotenoids with keto functional groups, the ketocarotenoids, such
as echininone (B,B-carotene-4-one), canthaxanthin (B,B-carotene—4,4'-dione), and
astaxanthin (3,3'-dihydroxy-B,0-carotene—4,4'-dione). As with B-hydroxylases, B-C-4-
oxygenases that catalyze the formation of the keto-groups have been cloned from
different phyla (Misawa et a1., 1994; Lotan and Hirschberg, 1995; Femandez—Gonzalez et
a1., 1997). Green algal and bacterial B-C-4-oxygenases are similar in sequence while B-C-
4-oxygenases from plants have high similarity to plant and bacterial B-hydroxylases. B-C-
4-oxygenases from cyanobacteria define a third distinct group of ketolase enzymes that
are similar to bacterial PDS (Fernandez-Gonzalez et a1., 1997). Despite this wide
sequence diversity, the B-C-4-oxygenases from bacteria, green algae, and higher plants
are non-heme di-iron oxygenases, analogous to the B-hydroxylases in higher plants.

The xanthophyll cycle consists of a cyclic sequence of interconversions between
zeaxanthin, antheraxanthin (5,6-epoxy-5,6-dihydro-B,B-carotene-3,3'-diol), and
violaxanthin (5,6,5',6'-diepoxy-5,6,5',6'-tetrahydro-B,B-carotene-3,3'-diol). These
reactions are catalyzed by two enzymes, zeaxanthin epoxidase (ZEP) and violaxanthin
deepoxidase (VDE), which localize to different chloroplast subcellular compartments.
ZEP localizes to the stromal side of the thylakoid membrane and is constitutively active.

Genes encoding ZEP have been cloned from various plants (Bouvier et a1, 1996; Marin et

24

 

 

 

 

a1., 1996). ZEP is a monooxygenase that requires 02 and is active at approximately pH
7.5. The ZEP gene from Capsicum annuum was expressed in E. coli and shown to require
NADPH, ferredoxin, and ferredoxin: NADP oxido-reductase for activity (Bouvier et a1.,
1996), similar to the B-hydroxylases. ZEP specifically utilized zeaxanthin and
antheraxanthin as substrates, while the B-ring of lutein was not a substrate in vitro
(Bouvier et a1., 1996). This result is consistent with the fact that lutein epoxide does not
normally accumulate in viva. VDE catalyzes the reverse reactions of ZEP. Deepoxidation
of violaxanthin to form zeaxanthin and antheraxanthin is induced by excessive light and
is catalyzed by VDE (reviewed in Niyogi, 1999).

When absorption of light exceeds the ability of plants to fix C02, a higher than
normal pH gradient builds up across the thylakoid membrane due to photosynthetic
electron transport. VDE is located within the thylakoid membrane and is most active at a
pH around 5.2, which occurs with the acidification of the thylakoid lumen due to the
buildup of large H+ gradient under excessive light. VDE was purified from lettuce, and
ascorbate and monogalactosyldiacylglyceride (MGDG) are required for VDE activity in
vitro (Rockholm and Yamamoto, 1996). ZEP and VDE were the first plant enzymes
identified that belong to the lipocalin family, a group of proteins that bind small
lipophilic molecules and have conserved B-barrel tertiary structures (Bugos et a1., 1998).

Further modifications of antheraxanthin and violaxanthin in pepper lead to
rearrangement of the epoxy end groups to give rise to the allenic end group of neoxanthin
and 6-OXO-K end group of capsanthin and capsorubin. The neoxanthin synthase gene has
been cloned from both tomato and potato and shows high sequence identity to lycopene

cyclases (Al-Babili et a1., 2000; Bouvier et a1., 2000). The formation of capsanthin and

25

capsorubin is catalyzed by capsanthin capsorubin synthase (CCS). CCS from Capsicum
annuum also belongs to the lycopene cyclase family. This enzyme converts
antheraxanthin or violaxanthin to capsanthin or capsorubin, respectively, by a mechanism
similar to lycopene cyclases (Bouvier et a1., 1994). Other modifications of carotenoids
include the degradation of C40 carotenoid and generation of apocarotenoids (degradation
from one end), e. g. abscisic acid, and diapocarotenoids (degradation from both ends), e. g.

bixin and crocetin.

1.1.4 Cytochrome P450 monooxygenase

1.1.4.1 Cytochrome P450 monooxygenase and carotenoid biosynthesis

Cytochrome P450 is a superfamily of proteins that metabolizes a large variety of
physiological important compounds, e. g. steroids and fatty acids (reviewed in Ortiz de
Montellano, 1995). In plants, cytochrome P4503 play important roles in secondary
metabolism, such as in the synthesis of phenylpropanoids. Cytochrome P4503 also
catalyze the degradation of exogenous compounds (reviewed in Chapple, 1998). About
300 cytochrome P450 enzymes are identified in the Arabidopsis thaliana genome, which
constitute one of the largest families of enzymes in this model organism.

There is some circumstantial evidence that cytochrome P450 enzymes may be
involved in carotenoid biosynthesis in photosynthetic organisms. In the green algae
Haematococcus pluvialis, biosynthesis of astaxanthin is carried out by two different

pathways: the addition of two keto-groups followed by two hydroxyl-groups or two

26

hydroxylation steps followed by two ketolation steps. Astaxanthin biosynthesis in
Haematococcus requires oxygen, NADPH, and Fez: common cofactors for cytochrome
P450 type oxygenases, suggesting that the biosynthesis of astaxanthin might involve a
cytochrome P450 enzyme. Treatment of cells with the cytochrome P450 specific inhibitor
ellipticine inhibited the carotenoid hydroxylase activity and only the diketo product
canthaxanthin accumulated (Schoefs et a1. 2001). Although these data strongly suggested
that carotenoid synthesis in Haematococcus involves a cytochrome P450 type enzyme,
direct demonstration of a cytochrome P450 carotenoid biosynthetic activity in the
organism has not been shown.

Cytochrome P450 enzymes are structurally a diverse group but contain some
conserved motifs involved in catalysis. The most highly conserved regions in cytochrome
P4503 are the PERF motif, the K-helix, and the heme-binding domain (Poulos, 1995).
The consensus sequence that binds the proximal face of the heme, F xxGxxxCxG,
contains a conserved cysteine that is the fifth ligand of the heme iron. The non-conserved
region of the cytochrome P450 proteins usually binds the substrate and one of the several
known redox partners. By definition, cytochrome P4503 are grouped into a farnily if their
amino acid sequences are >40% identical and are further grouped into subfamilies if the
identity is >55%. In the standard cytochrome P450 nomenclature, cytochrome P450
enzymes start with CYP followed by a number designating the family, then a capital
letter designating subfamily (e. g. CYP86B). Within subfamilies, allelic variants that are
less than 3% different are designated with an additional number after the subfamily

designation, e.g. CYP86Bl.

27

1.1.4.2 Localization of the plant cytochrome P4503

Cytochrome P450 monooxgenases have molecular mass ranging from 45 to 62
kDa with an average of 55 kDa. The majority of cytochrome P4503 reside in or are
associated with the endoplasmic reticulum. Although a few mitochondrial cytochrome
P4503 have been found in the mammalian adrenal cortex, no plant mitochondrial-
localized cytochrome P4503 have been discovered thus far. Recently, several lines of
evidence have indicated that plant cytochrome P4503 also localize to the chloroplast.

The first direct evidence for a chloroplast-localized cytochrome P450 was the
identification of CYP86B1. CYP86Bl has an unknown function but the protein sequence
shows a putative chloroplast transit peptide. Using import assays with isolated pea
chloroplast, CYP86B1 has been localized to the outer envelope of chloroplast (Watson et
a1., 2001). Two cytochrome P450 enzymes involved in oxylipin (oxygenated fatty acids)
metabolism are allene oxide synthase (AOS, CYP74A) and fatty acid hydroperoxide
lyase (HPL, CYP74B). AOS catalyzes the synthesis of 12,13-epoxyoctadecatrienoic acid
(EOT), an intermediate in jasmonic acid biosynthesis, from l3-hydroperoxide linoleinic
acid (l3-HPOT). HPL alternatively converts 13-HPOT to 12-oxo-cis-9-dodecenoic acid
(ODA), which leads to the production of C6 aldehydes. Although both A08 and HPL
belong to the CYP74 family, they are targeted to different membranes of the chloroplast
envelope by different pathways (Froehlich et al. 2001). Tomato (Lycopersicon
esculentum) AOS contains a putative chloroplast targeting peptide and inserts into the
inner envelope of chloroplast. Its targeting process requires both ATP and protease-

sensitive components on the outer face of the chloroplast. HPL, on the other hand, does

28

not have an N-terminal chloroplast targeting peptide and its insertion to the outer
envelope does not require ATP and is not protease sensitive (Froehlich et a1. 2001).
CYP86B1 and tomato A08 and HPL are the only known activities of chloroplast
cytochrome P4503 in higher plants. Out of the 272 known Arabidopsis cytochrome
P4503, only nine are predicted to be chloroplast targeted, most of these have unknown

functions (Table 1.3).

1.1.4.3 Cytochrome P450 functions in plants and monooxygenation reaction

mechanism

Based on their functions, cytochrome P4503 in plants can be generally divided
into two classes: those involved in biosynthesis and those involved in detoxification and
degradation. Cytochrome P4503 play important roles in the biosynthesis of a bewildering
array of secondary metabolites, such as lignins, terpenoids, flavonoids, sterols,
isoflavonoids, and furanocoumarins. Cytochrome P4503 are also involved in the synthesis
of several plant hormones, such as jasmonic acid (JA), gibberellic acid (GA),
brassinosteroid, and auxin (reviewed in Schuler, 1996). The first characterized plant
cytochrome P450 was cinnamate-4-hydroxylase, which converts trans-cinnamic acid to
p-coumaric acid (Russell and Conn, 1967). Besides the indispensable roles in
biosynthesis, cytochrome P4503 are also essential in detoxifying herbicides (Schuler,
1996)

Cytochrome P450 enzyme catalyzes monooxygenation in a multistep mechanism.

The substrate first binds to the active site of the cytochrome P450 enzyme close to the

29

 

Table 1.3. Arabidopsis cytochrome P4503 with predicted
chloroplast localization.

 

Gene Subfamily Function
At1g31800 97A3 unknown
At1g74550 98A9 unknown
At39531 30 9701 unknown
At4g13310 71A20 unknown
At4g36220 84A1 ferulate-5-hydroxylase
At4g37340 81 D3 unknown
At5gO5260 79A2 unknown
At5925120 71 B11 unknown
At5942650 74A allene oxide synthase

 

30

 

iron center but not directly ligated to the iron. Molecular oxygen then binds to the
reduced heme iron, which resulted from electron transfer. A hydrogen atom from the
substrate is subsequently abstracted by an oxoiron species (F ew=O), and finally, a rapid
collapse of the resulting carbon radical-hydroxyferryl complex to give the hydroxylated
product and the resting-state enzyme (Figure 1.6). NADPH-cytochrome P450 oxido-
reductase (CPR), a membrane-bound flavoprotein, is required and essential for the

transfer of electrons during the cytochrome P450 catalytic cycle.

1.1.4.4 Arabidopsis NADPH-Cytochrome P450 reductases (CPR)

Cytochrome P4503 require CPR for function. There is only one CPR in animals
(Ortiz de Montellano, 1995), while in plants there are multiple forms of CPRs present.
This is maybe because of the requirement of specific cytochrome P4503 in plants or
because cytochrome P4503 and hence CPRs localize to different subcellular
compartments in plant cells. It is also possible that plant CPRs are developmental stage or
tissue specific. Arabidopsis CPRs were cloned using a functional approach by
complementing yeast deficient of CPR with Arabidopsis cDNAs (Urban et a1. 1997).
Two Arabidopsis CPRs were identified from screening and designated as ATRl
(At4g24520) and ATR2 (At4g30210), respectively. Both reductases contain the signature
FMN-, F AD-, and NADPH- binding domains for cytochrome P450 and support the
activity of cinnamate 4-hydroxylase (CYP73A5) in vitro (Urban et a1. 1997). ATRl and
ATR2 are 63% identical at the amino acid level and are conserved in their C-terminal

sequences but not their N-termini. ATR2 contains a stretch of amino acids that are

31

Figure 1.6. Cytochrome P450 reaction mechanism. A. Generalized scheme for
Cytochrome P450 enzyme reaction. B. The proposed hydrogen abstraction-oxygen

rebound mechanism for cytochrome P450 hydroxylation.

952.2 counabenu

 

 

 

 

   

 

 

 

     

 

«>0 go go
m\ m\ m\
2 liq Z Ill—IN?- Z IQ:
\ _;/\+ \\V._/\A \\/z.._/\
z z z Wz zi|_ _VI 2
\ :o 0
ii cox: FD
/ / /
m
:9. .2325 83.81 to<z
of
$8262
flame“. oEoEuosoH
No
:m 388m 2.52.6 Ea<z m
< W.
F

33

 

serine/threonine rich and predicted to be chloroplast-targeted (Urban et a1., 1997; Hull
and Celenza, 1999). A third putative CPR was identified from database searching and
designated as ATR3 (At3 g02280). However, its expression and activity have not yet been

tested.

1.1.4.5 Current genomics approach for cytochrome P450 characterization

The sequencing of the Arabidopsis genome has identified 272 cytochrome P45 03
(http://www.biobase.dk/P450/p450.shtml), most with unknown functions. Genome
sequencing and high-throughput functional genomics are tools that have been developed
for Arabidopsis and allow scientists to specifically mutagenize genes of interest and study
their functions in viva. Using a reverse genetics approach, T-DNA insertional lines for
several cytochrome P450 genes have been isolated by several groups, but only one
mutant showed a visible phenotype that affected normal growth. The others still need to
be analyzed more closely for possible biochemical alterations or changes in primary
and/or secondary metabolism.

Another approach for cytochrome P450 characterization has been to correlate the
expression of specific cytochrome P4503 with physiological and developmental states.
The impact of developmental stage on cytochrome P450 expression was investigated
using microarray technology (Xu et a1., 2001). Many cytochrome P4503 on the
microarray showed tissue specific expression patterns, which may facilitate selective
study of the role of certain cytochrome P4503 in tissue-specific metabolic pathways.

However, most of the highly expressed cytochrome P4503 still have unknown fimctions.

34

 

A larger scale microarray experiment with different conditions and clustering analysis
(e. g. self-organizing map) will be more informative because cytochrome P450 with
known fiinction in a cluster may suggest a potential function of the other unknown
cytochrome P4503. Chapter 3 of this dissertation described the isolation and functional
characterization of CYP97C 1 , a cytochrome P450 enzyme involved in carotenoid

biosynthesis.

1.1.5 Model for possible carotenoid biosynthetic enzyme complexes

Carotenoids are hydrophobic compounds, therefore the carotenoid biosynthetic
enzymes are predicted to be membrane-bound or membrane-associated. A "phytoene
synthase" complex, which contains IPP isomerase, GGPS, and PSY, was isolated from
chromoplasts of higher plants (e. g. daffodil flowers and bell peppers) (Camara and
Moneger, 1981; Kleinig and Beyer, 1985). The complex was readily solubilized
suggesting that it was only loosely (peripherally) attached to the thylakoid membrane.
Biochemical analysis also showed that imported PSY is present in both soluble and
membrane-bound forms in daffodil chromoplasts (Schledz et a1., 1996). The soluble
protein is inactive while the membrane-bound PSY is active. The presence of active PSY
in the thylakoid membrane is consistent with the lipophilic character of the end product,
phytoene.

Based on the above experimental evidence, the physical properties of carotenoids,
and predicted localization and hydrophobicity of carotenoid biosynthetic enzymes, an

overall model for carotenogenic complexes was proposed by Cunningham and Gantt

35

 

(1998) (Figure 1.7). In this hypothetical model, a single IPP isomerase, two GGPS, and
one PSY form a peripheral complex that associates with the thylakoid membrane. IPP
substrates are supplied to and phytoene are produced from this peripheral "phytoene
synthase complex". The phytoene produced is then transferred to PDS, which is a part of
a proposed integral membrane complex including PDS, ZDS, and lycopene cyclases.
Because only [LB-carotene and B,e-carotene are normally formed in most of the plants, it
was proposed that there are two types of complexes in the thylakoid membrane, one
containing two lycopene B-cyclases to produce B,B-carotene (the [LB-complex) and one
containing one B- and one s- cyclase to form B,e-carotene (the [Le-complex). A complex
with two e-cyclases is not hypothesized due to the lack of 8,8-carotene in most of the
plants.

This model suggested a fundamental mode for carotenogenic enzyme interactions
in product generation. However, cis-isomers of phytoene need to be converted to all-trans
lycopene before they could be converted to the cyclized products. A more refined model
(Cunningham, 2002) added the isomerase enzyme to both 0,0- and [La-complexes
(Figure 1.7). However, there is no direct evidence for the existence of membrane
carotenoid complexes hence the hypothesis has not been proven. A key factor lacking in
the model is the accommodating ZEP, VDE, and the hydroxylases. How are these
enzymes in the later part of the pathway organized? Are they involved in the same
complexes as the cyclases? The lack of accumulation of reaction intermediates (such as
phytoene, Q-carotene, and lycopene) has been considered as a consequence of the
substrate channeling in the hypothetical enzyme complexes. However, it is also likely

that the carotenogenic enzymes have low Kms and high affinity for various substrates,

36

 

 

Figure 1.7. Proposed model for carotenogenic complexes in chloroplasts of plants.
IPI, isopentenyl pyrophosphate isomerase; GGPS, geranylgeranyl pyrophosphate
synthase; PSY, phytoene synthase; PDS, phytoene desaturase; ZDS, Q-carotene
desaturase; LHCB, lycopene B-cyclase; LHCE, lycopene e-cyclase; CRTISO, carotenoid
isomerase. This figure is adapted after Cunningham and Gantt. (1998) Ann. Rev. Plant

Physiol. Plant Mol. Biol. 49, 557-583.

37

onEou ad

m>U._ mDN

  

02.5.0

 

man. man.

mnN m>o._

23323

/&. a

.8383 ad

OMEN—o

 

m>o.._ mDN

 

r

won.

14‘

 

mon—

L

k

 

 

 

mn—N m>0.._

«Eobm

Figure 1.7.

38

and this could also explain why no accumulation of intermediates are observed. A more
refined model for carotenogenic complex formation is expected with a better
understanding of carotenoid enzyme properties. Chapter 4 of this dissertation will address

this question in reference to hydroxylases from a molecular genetic perspective.

1.2 Regulation and metabolic engineering of the carotenoid pathway

1.2.1 Regulation of carotenoid biosynthesis

Carotenoids share the common biosynthetic precursor, IPP, with the synthesis of
many other biologically important terpenoids. Therefore, the regulation of carotenoid
formation must also involve the branching or channeling of isoprene tuiits into the
carotenoid pathway. The non-mevalonate pathway provides the isoprene building blocks
in plastids and may regulate the production of carotenoids by apportioning carbon
through carotenoid biosynthesis. Mutation in the CLAI gene from Arabidopsis, a gene
that encodes the transketolase DXPS, impeded the development of chloroplasts, caused
an albino phenotype, and led to a severe reduction in carotenoid and chlorophyll levels
(Mandel et a1., 1996). The CapTKT 2 gene cloned from pepper fruit encodes a plastid
transketolase. It catalyzes the formation of DXP from pyruvate and G3P in vitro, and
therefore proved to be a DXPS (Bouvier et al., 1998a). CapTKTZ expression was up-
regulated during the chloroplast to chromoplast transition, which correlated with the
massive accumulation of carotenoids in ripening bell peppers (Bouvier et al., 1998a). In

addition, it has been observed in tomato fruits that the expression of the DXPS gene

39

:orrelates with the ripening-associated massive accumulation of carotenoids (Lois et a1.,
2000). Expression analysis in tomato also showed that the DXPS gene has a similar
:xpression pattern as PS Y1 , a key regulatory gene in the carotenoid biosynthetic pathway.
vloreover, feeding of DXP induced the expression of both DXPS and PSY] in tomato
’ruit (Lois et a1., 2000). The combined results from Arabidopsis, pepper, and tomato
:uggest that the non-mevalonate pathway may be involved in the regulation of flux into
he carotenoid pathway and therefore should be taken into consideration in studying of
:arotenoid biosynthesis. It has also been recognized that PSY is a major quantitative
:ontrol point for carotenoid biosynthesis, which will be discussed in detail in the
bllowing sections on metabolic engineering.

Carotenoids are essential components of the photosystem reaction centers and the
i ght harvesting complexes in chloroplasts. The regulation of carotenoid biosynthesis in
green tissue is therefore likely related to and integrated with chloroplast development,
:hlorophyll formation, the level of carotenoid binding proteins, and thylakoid membrane
ipid composition. The mechanism for this complicated regulation is still not well
inderstood. However, light is known to be a key factor in the regulation of carotenoid
>iosynthesis in chloroplasts. Transferring Arabidopsis or tomato plants from low light to
iigh light induced a shifi in the ratio between lycopene B- and s— cyclase mRNA levels
Hirschberg, 2001). This result not only suggests that xanthophyll composition is
'egulated by light but also indicates a modulation of carotenoid biosynthesis at the branch
)f a- and B- carotene in the carotenoid pathway (Hirschberg, 2001). A key regulatory
itep in xanthophyll composition at the [3,0- and [3,8- branches (at the level of the e-

'yclase enzyme) has previously been suggested by the Arabidopsis lut2 mutant. Iut2 is a

40

 

disruption in e-ring cyclization in which lutein biosynthesis is abolished and B-carotene
derived xanthophylls accumulated to a much greater amount than wild type (Pogson et
a1., 1996). Finally, transgenic Arabidopsis overexpressing the lycopene e-cyclase gene
produced almost twice as much lutein as wild type providing direct evidence for the key
regulatory role of the cyclization enzymes and the e-cyclase in partitioning carbon
between a- and [3- branches (Pogson and Rissler, 2000).

The regulation of carotenoid biosynthesis in chromoplasts has been more
extensively studied than chloroplasts (Brarnley, 2002). Tomato and pepper fruits and
marigold flowers have been used as model systems for studying chromoplast carotenoid
biosynthesis. Experimental data from all three systems have shown that several of the
pathway genes are transcriptionally up—regulated during the large increase in carotenoid
biosynthesis in each system. For example, the mRNA levels for PSY and the desaturases
were induced 10— to 20- fold during tomato ripening (Giuliano et a1., 1993); carotenoid
accumulation in cultivars of marigold that differ in their carotenoid levels by 100 fold is
correlated with the mRNA levels of the pathway genes (Moehs et a1., 2001, Appendix A).
Post-transcriptional regulation of carotenoid genes was reported in daffodil chromoplasts
where PSY was inactive in its soluble form but became active when it was bound to the
membrane (Schledz et a1., 1996). In addition, carotenoid-associated and binding proteins
play important roles in the sequestration of carotenoids and affect the accumulation of

carotenoids within chromoplasts post-transcriptionally (V ishnevetsky et a1., 1999).

1.2.2 Metabolic engineering

41

Carotenoids have diverse functions related to human health; they are required for
normal vision and are associated with a reduction in the occurrence and severity of
several degenerative diseases including some cancers. Certain carotenoids are limiting in
some diets and associated with increased risk of age-related macular degeneration. While
carotenoid supplements are available, the efficiency of absorption of supplements varies.
Carotenoids in food are facilitated in their absorption during digestion where dietary
lipids enhance the solubility of carotenoids. Engineering of the carotenoid pathway in
plants has been proposed as a method to elevate specific carotenoids in the food supplies I
or to provide plant-based production of specific carotenoids for industrial or agricultural
purposes.

One strategy for metabolic engineering of carotenoids is to increase the total
amount of existing carotenoids in plants or a specific plant tissue. PSY catalyzes the first
committed step for carotenoid biosynthesis and may direct carbon flow from the "bulk"
IPP pool into the carotenoid pathway. Therefore, it has been considered as a regulatory
point for the pathway and used as a target for metabolic engineering by several groups.
The differing outcomes of apparently similar experimental approaches highlight the large
gaps in our understanding of carotenoid synthesis in plants. Overexpressing PSY gene in
tomatoes had little impact on fruit carotenoid levels and resulted in dwarfism (Fray et a1.,
1995). This is because GGPP is also the precursor for gibberellin biosynthesis (Figure
1.1). Overexpressing PSY led to a decrease in gibberellin synthesis in green tissue, and
hence dwarfism (Fray et a1., 1995).

One of the most successful examples of carotenoid engineering to date is using a

bacterial PSY gene (Erwinia uredavara; crtB) fused to a seed specific promoter and

42

expressed in the canola (Brassica napus) seeds. This led to a more than 50-fold increase
in total seed carotenoids (Shewmaker et a1., 1999). This increased level of carotenoids is
comparable or even higher than certain high carotenoid producing flowers or fruits.
However, the increase in the carotenoid pool impacted other compounds that also use IPP
and GGPP as precursors, such as tocopherols and chlorophylls in the developing seeds.
Sterols are synthesized through the cytosolic mevalonate pathway and their levels were
not changed in the transgenic seeds. The fatty acid composition in transgenic canola
seeds was also slightly changed with relative more oleic acid (18:1) than linoleic acid
(18:2) and linolenic acid (18:3).

Metabolic engineering of the carotenoid pathway was also accomplished by
shifting to another carotenoid gene product in tomato. A bacterial (Erwinia uredavara)
phytoene desaturase gene was constitutively expressed in tomato (Riimer et a1., 2000). B-
carotene content was increased about three-fold at the expense of lycopene but total
carotenoid levels were reduced, suggesting a feedback inhibition of the pathway. Levels
of other GGPP-derived isoprenoids did not significantly change in this experiment.

Another general metabolic engineering strategy is to transfer the entire carotenoid
biosynthetic pathway into a tissue that is devoid of carotenoids. Mammals do not
synthesize carotenoids and depend solely on dietary intake for carotenoids. B-carotene
(provitamin A) is the precursor of retinal, a critical visual pigment for humans (Simpson
and Chichester, 1981). There is no B-carotene present in the endosperm of rice (Oryza
sativa), therefore provitarnin A deficiency is a severe problem in the countries where rice
is the major staple food (Ye et a1., 2000). Immature rice endosperm does produce GGPP,

the precursor for phytoene biosynthesis. Therefore, the entire B—carotene biosynthetic

43

pathway was engineered into rice endosperm by transforming vectors containing daffodil
(Narcissus pseudanarcissus) PSY and lycopene B-cyclase open reading frames, and a
bacterial phytoene desaturase (Erwinia uredavara; crtl) gene (Ye et a1., 2000). As a
result, B-carotene was produced to a substantial amount in the transgenic lines. In
addition, lutein and zeaxanthin were also unexpectedly accumulated in the transformed
endosperm indicating lycopene s-cyclase and the hydroxylases may have been
constitutively expressed or induced in the transgenic rice lines. B-carotene accumulating
rice endosperm sets a good example of engineering crop plants for improved nutritional
value.

Besides providing plant food with increased carotenoid content for humans,
engineering of the carotenoid biosynthetic pathway can also enhance stress tolerance of
plants. The Arabidopsis B-hydroxylase 1 gene was overexpressed under the constitutive
358 promoter in wild type Arabidopsis (Davison et al., 2002). This led to accumulation
of xanthophyll cycle carotenoids at twice the levels in wild type plants and an increase in
the total carotenoids without affecting the lutein levels. Sucrose gradient fractionation of
pigment protein complexes indicated that the extra xanthophylls produced reside mainly
in the LHC II complex. The transgenic plants were tested for stress tolerance by
switching growth condition to high light (1000 umol photon m'zs'l) and high temperature
(40°C) for two weeks. Two to four times more zeaxanthin were accumulated in the
transgenic plants when compared to wild type controls. As a consequence, the transgenic
plants showed less negative effects from stress than wild type as indicated by decreased
anthocyanin accumulation and a 30% decrease in lipid peroxidation, as measured by

malondialdehyde (MDA) production. Although zeaxanthin is involved in

44

nonphotochemical quenching (N PQ), the constitutive accumulation of zeaxanthin did not
lead to an increase in the NPQ capacity in the engineered plants. Therefore, increased
NPQ is not likely the cause for increased stress tolerance. Earlier findings showed that
zeaxanthin plays a role in preventing lipid peroxidation (Havaux and Niyogi, 1999).
Therefore, the transgenic plants most probably tolerate stress by preventing lipid
peroxidation through the action of zeaxanthin. Genetic manipulation of a single enzyme
(B-hydroxylase 1) in the carotenoid biosynthetic pathway leads to stress tolerance and

suggests it may be a route for producing more stress tolerant crops.

1.3 Carotenoid functions in higher plants

1.3.1 Zeaxanthin as a membrane stabilizer

Thylakoid membranes are bilayers of arnphipathic di-acyl galactolipids in which
proteins are imbedded. Thylakoid membranes contain high amounts of polyunsaturated
galactolipids and lack cholesterol or other sterols. Therefore, thylakoid membranes are
relatively fluid and thermolabile system, however, they are also sensitive targets of
singlet oxygen and reactive oxygen species during high light stress. Terpenoid molecules
in the thylakoid membrane include tocopherols, chlorophylls, carotenoids, and
plastoquinones. It has been suggested that at-tocopherol plays a role in membrane
stabilization that is analogous to cholesterol (Fukuzawa et a1., 1977), although no direct
evidence for this hypothesis has been shown.

In vitro studies have shown that the orientation of zeaxanthin in lipid bilayers is

45

with the long axis of zeaxanthin perpendicular to the thylakoid membrane surface and the
polar 1.3-end groups anchored at both sides of the membrane (Gruszecki and Sielewiesiuk,
1991). It is suggested that zeaxanthin acts as a "rivet" spanning the entire thylakoid
membrane bilayer and linking the two halves of the bilayer. Zeaxanthin accumulates to
high levels in excess light conditions through the deepoxidation of the violaxanthin.
Therefore, it is possible that the operation of xanthophyll cycle correlates the light stress
and regulation of thylakoid membrane fluidity and thermostability by zeaxanthin.

Barley leaves were exposed to a photon flux density (PFD) of 1,000 umol m'zs'l
for several minutes, which leads to a conversion of violaxanthin to zeaxanthin without the
induction of lipid peroxidation (Tardy and Havaux, 1997). The light treatment led to a
decrease in thylakoid membrane fluidity as measured by electron spin resonance (ESR)
spectroscopy. The rigidification of the thylakoid membrane correlated with an increase in
zeaxanthin and was blocked by dithiothreitol (an inhibitor of violaxanthin deepoxidase).
Membrane fluidity recovered in the dark when zeaxanthin was re-epoxidized to
violaxanthin. Such a phenomenon was not observed in the abaI mutant of Arabidopsis,
which lacks the xanthophyll cycle and constitutively accumulates zeaxanthin (Tardy and
Havaux, 1997). There is also an increased thermostability in the light treated barley
leaves, possibly because the stabilization effect by zeaxanthin counteracts the effect of
heat induced fluidity increase of membrane lipids. Plants may have developed this
mechanism to quickly protect from high light induced membrane destruction by

producing membrane stabilizing zeaxanthin molecules through the xanthophyll cycle.

1 .3.2 Carotenoid functions in light harvesting and photoprotection

46

1.3.2.1 Light harvesting

The earth's biosphere derives its energy from sunlight through photosynthesis.
Plants, algae, and photosynthetic bacteria have evolved efficient systems to harvest the
light from sun and use this energy to drive metabolic reactions required for life. The main
photosynthetic pigments in higher plants are chlorophylls (Chls) and carotenoids. In the
chloroplasts of plants, Chls and carotenoids are embedded in the photosystems (PS) that
are localized in the thylakoid membranes. Two different photosystems (PS I and PS 11)
are linked together by an electron transport chain. PS 11 is a large membrane complex
containing pigments for light harvesting and pigment binding proteins. It oxidizes water
to produce oxygen and transfers electrons by photochemical reactions. PS 1 functions as a
light-dependent plastocyanin-ferredoxin oxido-reductase that accepts the electrons-from
the electron transport chain and reduces NADP+ to generate NADPH. Each PS consists of
two closely linked components: a reaction center and a light harvesting complex (LHC).
LHCs consist of antenna pigments (Chls and carotenoids). Upon absorption of light,
antenna pigments are converted into the excited singlet state. Energy from excited singlet
antenna pigments can be transferred to neighboring pigment molecules via resonance
transfer and eventually delivered to the reaction center Chl a. The absorbed light energy
is then converted into chemical energy and provides organic chemicals for plants
(photochemistry, qP). In addition to photochemistry, the excited light energy absorbed by
antenna pigments can be emitted as chlorophyll fluorescence or dissipated as heat (non-

radiative decay).

47

1.3.2.2 The role of carotenoids in photoprotection

In addition to being used in photosynthesis or decaying to the ground state, triplet
chlorophyll (3Chl) can be formed from singlet chlorophyll (lChl) through intersystem
crossing. 3Chl has a longer lifetime than lChl (ms time scale vs. ns time scale) therefore is
able to react with 02 to produce singlet Oz ('02), posing a high potential for
photooxidative damage to the PSs. Excessive light energy can prolong the half-life of
lChl and consequently increase the probability of producing more 3Chl and '02 in the
LHCs. Generation of '02 in the LHCs is harmful because the thylakoid membrane (where
LHCs are located ) is enriched in polyunsaturated fatty acid (PUFA), which are highly
susceptible to 102 attack and can initiate peroxyl radical chain reactions. The damage
caused by 102 and its reactive products can decrease the efficiency of the PS and lead to
photoinhibition.

Carotenoids in the LHCs have excited singlet and triplet energy states that are
lower than Chls and are able to accept excitation energy directly from 3Chl to prevent the
formation of 102 (Niyogi, 1999). The excited triplet carotenoid has very low energy
therefore is unable to generate other reactive species and instead dissipates energy as heat
to convert to the ground state harmlessly. In addition to quenching '02, carotenoids are
also involved in other mechanisms that protect the photosystem from photodamage, the

most notable being nonphotochemical quenching (N PQ).

1.3.2.3 The role of carotenoids in NPQ

48

Thermal dissipation of excessive light energy is measured as nonphotochemical
quenching of chlorophyll fluorescence (NPQ or qN). NPQ is a major mechanism
protecting plants from photooxidative damage. Both qP and NPQ can minimize
chlorophyll fluorescence, decrease chlorophyll fluorescence life time and excitation
energy decay by the 3Chl pathway. When reactive molecules are inevitably generated,
despite the operation of NPQ and qP, carotenoids, plastid-localized tocopherols,
ascorbate, glutathione, and other antioxidative enzymes can protect the photosynthetic
apparatus from the various reactive oxygen species. Other protective mechanisms that
prevent plants from photodamage include: alteration in the leaf surface to increase light
reflection, decreases in transmittance of light to the chlorophyll containing cells, and leaf
and chloroplast movements to decrease the light absorption. However, zeaxanthin-
associated NPQ is both more ubiquitous and rapid compared to these mechanisms.

NPQ consists of three major components, qE, qT, and q1. qE is the principal and
most rapidly inducible and reversible component of NPQ. qE is energy dependent,
activated by lowered thylakoid lumen pH, and can rapidly relax within seconds to
minutes. qT relies on state transitions of LHC 11 between PSII and PSI and is a slower
component than qE. It relaxes in tens of minutes. A phosphokinase activated under high
light phosphorylates the LHC lIb polypeptides, which uncouples LHC II from PS 11 and
causes association of LHC II with PS 1. Experimental evidence has suggested that qT is
not essential for photoprotection in higher plants but is required for NPQ in algae
(N iyogi, 1999). The slowest component of NPQ is qI, which is also called

photoinhibition. qI is involved in longer term adaptation to light stress and relaxes within

49

a range of hours. qI could be a combination of photodamage and photoprotection and is
the least characterized NPQ component.

It has been firmly established from combined molecular, genetic, and biochemical
studies that qE requires a lower pH in the thylakoid lumen, the xanthophyll cycle
carotenoid (zeaxanthin), and the PsbS protein (Miiller et a1., 2001). The mechanism of qE
will be described in the following sections. Lutein has also been proposed to be involved
in NPQ based on studies with the Arabidopsis lut2 mutation, which affects the lycopene
e-cyclase gene and is unable to synthesize lutein (Pogson et a1., 1996; Pogson et a1.,
1998). Detailed characterization of LHC size and stability established that lutein
deficiency in the lut2 mutant also affects the assembly and stability of LHC II. The
conclusion from such studies is that although lutein is necessary for efficient NPQ, it is

not directly involved in the mechanism of NPQ (Lokstein et a1., 2002).

1.3.2.4 NPQ is measured by way of chlorophyll fluorescence

Excitation energy is consumed by three processes: photochemistry, chlorophyll
fluorescence, and thermal dissipation. Photochemistry includes charge separation at the
reaction center and electron transport via a series of carriers. Chlorophyll fluorescence is
the emission of photons by radiative de-excitation of the excited chlorophyll molecules,
while non-radiative thermal dissipation of excitation energy occurs via heat loss (NPQ).
Either increased consumption of excited energy by the photochemical pathway or
increased dissipation as heat leads to increased quenching of chlorophyll fluorescence.

NPQ can therefore be determined by measuring chlorophyll fluorescence during a brief

50

pulse of light that saturates photochemistry. Chlorophyll fluorescence is measured using a
pulse amplified modulation (PAM) fluorometer. After irradiation, the chlorophyll
fluorescence changes over time, which is called Kautsky effect (Kautsky and Hirsch,
1934). The PAM-fluorescence technique allows one to obtain information on chlorophyll
fluorescence induction kinetics in viva.

Chlorophyll fluorescence within a sample is excited by strong light pulses from a
blue light emitting detector (L.E.D.) with a short wavelength red emission peak of 650
nm. The longer wavelength components of the LED. light are eliminated through a short
pass filter (A<670 nm). Electronic circuits pulse the LED. illumination and
synchronously amplify the resulting pulses and detected by a photo diode detector that

has been protected by a long pass filter (D700 nm). The filtered signal is then recorded.

1.3.2.5 The mechanism of qE

qE is associated with changes in chlorophyll fluorescence lifetime distributions
and a three-state model was proposed for PS 11 (Gilmore et a1., 1995; Gilmore et a1.,
1998). When the thylakoid lumen is not acidified and PS 11 is not active, the lifetime of
chlorophyll fluorescence is broad and designated as the W state. Upon illumination, a pH
gradient forms across the thylakoid lumen and results in the conversion of the W state to
a much shorter and narrower X state. Binding of xanthophyll (zeaxanthin or
antheraxanthin) further converts PS 11 from the X state to an active quenching Y state

(Gilmore et a1., 1995; Gilmore et a1., 1998).

51

The xanthophyll cycle

The so-called xanthophyll cycle involves reversible epoxidation/deepoxidation of
zeaxanthin to antheraxanthin and violaxanthin under normal light conditions and
deepoxidation of violaxanthin to antheraxanthin and zeaxanthin under high light stress. In
order to address the involvement of xanthophyll cycle carotenoids in NPQ, mutants that
affect xanthophyll metabolism and/or NPQ capacity have been isolated from Arabidopsis
and characterized in detail. The Arabidopsis npq1 mutation has a disrupted violaxanthin
deepoxidase and when exposed to strong light, is unable to convert violaxanthin to
antheraxanthin and zeaxanthin, even after prolonged high light treatment (N iyogi et a1.
1998). NPQ capacity is also reduced in the npq1 mutant relative to wild type plants after
high light treatment. Characterization of the npq1 mutant provided evidence that
zeaxanthin is required for most of the qE component of NPQ (Niyogi et a1., 1998). As
mentioned previously, lutein is indirectly involved in NPQ and the introduction of the
lut2 mutation into the npq1 background completely abolished qE (N iyogi et a1., 2001).
These studies further confirm that both zeaxanthin and lutein are necessary for maximal
NPQ.

In several algal groups, such as the diatoms, the dinophytes, and the haptophytes,
the xanthophyll cycle is replaced by conversion of diadinoxanthin and diatoxanthin, the
diadinoxanthin cycle (Lohr and Wilhelm, 1999). Because only one of the ionone ring of
diadinoxanthin carries an epoxy group, this cycle only contains one
deepoxidation/epoxidation step. The coexistence of both the xanthophyll cycle and

diadinoxanthin cycle was observed under high light stress conditions in the diatom

52

Phaeadactylum tricarnutum (Lohr and Wilhelm, 1999). Recently, it has been shown that
in the parasitic plant C uscuta reflexa, neoxanthin is replaced by lutein-5,6-epoxide
(Bungard et a1., 1999). Under high light, lutein-5,6-epoxide is deepoxidized to lutein.
This reaction is probably catalyzed by violaxanthin deepoxidase, which has been shown
to use lutein-5,6-epoxide as a substrate. This interesting discovery also supports a role for

lutein in photoprotection.

The role of the PsbS protein in NPQ

The npq mutants in Arabidopsis have been invaluable tools for studying NPQ. As
described earlier, npq1 mutation blocks violaxanthin to zeaxanthin conversion under high
light and defines a key role of zeaxanthin in NPQ. Similarly, another npq mutant, npq4,
has defined a key role for the PsbS protein. PsbS is a pigment binding protein in
photosystem II, which when mutated in the npq4 mutant resulted in a semidominant NPQ
phenotype (Li et a1., 2000). The npq4 mutant showed compromised NPQ activity in
response to excessive light, similar to the npq1 mutant. However, npq4 has a zeaxanthin
level similar to wild type. The npq4 mutant lacks the ApH and zeaxanthin dependent
conformational change in the thylakoid membrane as measured by a light-induced
spectral change at 535nm. Although NPQ was decreased in the npq4 mutant, light
harvesting and photosynthesis were not affected (Li et a1., 2000). Global chlorophyll
fluorescence lifetime distribution analysis on wild type and npq4-1 plant indicated that
the pH-dependent formation of the X state is not present in the npq4-1 mutant where

PsbS protein is absent (Li et al., 2002b). Furthermore, overexpression of the PsbS gene

53

led to a two-fold increase in qE indicating that PsbS is a limiting factor in qE capacity (Li
et al., 2002c). The PsbS overexpressing lines showed a higher resistance to
photoinhibition compared to wild type although the xanthophyll pool size was similar in
npq4 and wild type.

While the role of PsbS protein in qE is indispensable, it is still not clear how PsbS
is involved in NPQ. It has been proposed that the PsbS protein binds H+ and/or de-
epoxidized xanthophylls and switches the LHC II to a high quenching state (Li et a1.,
2000). One effort to identify the amino acid binding sites for PsbS function is site-
directed mutagenesis (Li et al., 2002a). Predicted PsbS protein sequences from nine
different species of land plants were aligned. Eight putative H+-binding amino acids are
highly conserved in the plants and predicted to face the thylakoid lumen where PF
accumulates upon high light stress. Each individual amino acid was mutagenized and the
modified PsbS gene was transformed into a null npq4 mutant. Mutations that affect E122
and E226 had the most dramatic phenotype and all qE was abolished in the E122E226
double mutant despite the presence of wild type levels of mutant PsbS proteins. These
results indicate that the symmetrical pair E122 and E266 is essential for PsbS function
and further support that H+ binding to PsbS is critical for qE (Li et al., 2002a).

The possible involvement of other proteins in qE is still unclear. The PS 11 minor
antenna proteins CP29 and CP26 have been suggested to be involved in qE due to their
location between the inner antenna and the major LHC II (Bassi et a1., 1997). The role of
CP29 and CP26 in energy dissipation was investigated using antisense plants that
depressed the expressions of each protein independently (Andersson et a1., 2001). Both

antisense lines were capable of highly efficient photosynthesis and were still able to

54

perform qE to a significant level, although there were changes in the rate of NPQ
induction and amplitude. Thus, CP29 and CP26 are unlikely to be the sites for NPQ
(Andersson et a1., 2002).

Careful characterization of the chloroplast ycf9 gene in both Chlamydomonas and
tobacco indicated that it encodes a PS 11 core subunit, Pst, and it is involved in
maintaining the stability of the PS II-LHC II supercomplex (Swiatek et a1., 2001). Pst
deficient tobacco plants exhibited greatly reduced capacity for NPQ under excessive light
and/or decreased temperature. PsbS protein accumulation is unaltered in the Pst mutant.
The total carotenoid pool is larger in the Pst mutant relative to wild type and showed a
dramatically altered xanthophyll cycle that has a stronger increase in the deepoxidized
state and is recalcitrant to re-epoxidation dtuing dark relaxation (Swiatek et a1., 2001). In
addition, Pst is highly conserved through evolution; it is present even in organisms that
lack a xanthophyll cycle and do not accumulate zeaxanthin, PsbS protein, or CP29
protein, but are still capable of NPQ for photoprotection (Campbell et a1., 1998). This
result suggests that Pst might be the key player for NPQ in phycobillisome containing

organisms.

1.4 Approaches taken in this dissertation for characterization of carotenoid

hydroxylases

As described in section 1.1.3, the e-ring hydroxylase has been genetically defined
by the lat] mutation in Arabidopsis but has not been cloned (Pogson et a1., 1996).

Attempts to isolate a gene that is related to B-hydroxylases but functions with e-rings by

55

screening of Arabidopsis cDNA and genomic libraries and database searching using [3-
hydroxylase as a query were not successful (Tian and DellaPenna, 2001). Though a
second B-hydroxylase gene was identified and characterized in these studies, its activity
was similar to B-hydroxylase 1 in that it effectively hydroxylated B-rings but was poorly
active toward s-ring. Furthermore, neither of the B-hydroxylase genes mapped to the [w]
interval. Map-based cloning was therefore adopted to isolate the LUT1 gene during the
course of this dissertation research (Chapter 3). The in viva fimctions of the [3- and e-
hydroxylases and possible interactions and overlapping functions were studied by
isolating T-DNA knockout mutants of the two Arabidopsis B-hydroxylase genes (b1 and
b2), generating and characterizing various mutant combination lines defective in b1, b2,
and lutI . The effects of the resulting carotenoid composition modifications on NPQ
induction kinetics were measured by chlorophyll fluorescence in the different carotenoid

hydroxylase mutants (Chapter 4).

1.4.1 Map-based cloning of the Iutl mutation

When no additional information is available, the gene disrupted by the mutation
can be cloned by physically mapping the mutant locus to a specific chromosomal location
and using functional complementation to demonstrate the gene identified encodes the
targeted protein. In order to map a novel mutation to a well-defined chromosomal region,
linkage analysis is carried out between the mutation and the previously mapped markers.
The markers being used include traditional morphological markers that give a visible

phenotype and molecular markers, such as Cleaved Amplified Polymorphic Sequences

56

(CAPS) and Simple Sequence Length Polymorphisms (SSLP) markers. It has been
estimated that C01 and Ler differ by 0.5% to 1% at the DNA level (Hauser et al. 1998),
due to point mutations, insertions, or deletions that randomly occurred in one ecotype but
not in the other. Arabidopsis contains tandem repeats of one-, two-, or three- nucleotide
motifs. These repeat sequences are usually polymorphic in different ecotypes because of
variations in the number of repeats introduced by DNA polymerase. All these natural
differences between ecotypes can be conveniently utilized to generate molecular markers.
CAPS is a method using PCR to detect the restriction site polymorphorism between two
ecotypes. The PCR amplified fragments are digested with specific restriction enzymes
and the polymorphism is then revealed by agarose gel electrophoresis. SSLP method
takes advantage of the variation in the numbers of repeat units in different ecotypes and
the polymorphic fragments are directly detected by PCR followed by agarose gel
electrophoresis.

To map a mutation that was generated in one ecotype, the mutant is crossed to a
plant of another ecotype that is wild type for that locus and the F1 progeny is allowed to
self-pollinate. The segregating F2 generation is then used to analyze for linkage between
the mutation and the molecular markers for the two different ecotypes. The
recombination frequency between a mutation and a molecular marker is the number of
the chromosomes scored as mutant ecotype divided by the total number of chromosomes
analyzed. However when converting the recombination frequency to map distance, two
factors need to be considered for calculation. One is that when two independent
recombination events occurred between two markers, no recombination event is actually

scored. The other one is that a recombination event can influence the probability of a

57

second recombination event occurring in its vicinity, which is called interference. In
Arabidopsis, a map distance is estimated by the formula (Kosambi function)
D=25xln[(100+2r)/(100-2r)], where r is the recombination rate expressed in its
percentage form and D is the map distance in centiMorgans (cM) (Koomneef and Stam,
1992).

The lutl mutation that genetically defines the e-hydroxylase in Arabidopsis is an
EMS mutation in the Columbia ecotype background (Pogson et a1., 1996). A
homozygous lut] mutant was crossed with a wild type (L UT 1 ) Lansberg erecta plant. F1
progeny were self-pollinated and the homozygous lutl plants in the F2 population were
used to analyze for linkage between [at] and various molecular markers (CAPS and
SSLP). The LUT1 locus was initially mapped at 67 cM :t 3 cM on chromosome 3 (Tian
and DellaPenna, 2001). Fine mapping of the LUT1 locus is subsequently carried out to

clone the LUT1 gene and is discussed in detail in Chapter 3.

1.4.2 T-DNA insertional mutants as a tool for in viva function study

During the course of this dissertation, the genome of Arabidopsis was fully
sequenced (The Arabidopsis Genome Initiative, 2000) and led to the development of
many important tools to study gene function. Once the genome sequence is known, large
scale T-DNA insertional mutagenesis becomes a convenient tool for "reverse genetics".
T-DNA insertion often causes a null mutation and offers an efficient way to study the
function of gene products in viva. T-DNA insertion allows targeting of individual genes

in a gene family and the stable maintenance of the mutant genotype and phenotype

58

because the mutant can be kept in the heterologous state if the homozygous plant is
lethal. With the availability of large population of T-DNA insertional mutants and the
effectiveness of the PCR-based knockout screening approaches and DNA sequencing, T-
DNA insertional mutagenesis has been a preferred approach for gene silencing over
antisense.

Briefly, the screening process for T-DNA insertional mutants starts with vacuum
infiltration of the wild type Arabidopsis plants with Agrobacterium harboring a binary T-
DNA vector. Transformed plants are selected by antibiotic resistance. Transformed T.
plants are grouped and T2 seeds are collected from grouped T1 plants. The samples are
further consolidated to even smaller number of pools, which is more convenient for
management and screening. Genomic DNA is extracted from pooled plants. DNA is
further pooled together to form a limited number of superpools that are convenient for
PCR reactions. DNA from superpools are used as templates and screened with one T-
DNA border primer and one gene specific primer. The T-DNA knockout mutants are
verified with gene specific primers. Once a positive hit is identified, it can be sequenced
with the T-DNA border primer to identify the insertion site. The same pair of primers is
used to screen the DNA sub-pools that constitute the superpool until an individual mutant
is identified. The homozygous mutant is isolated and the genotype and phenotype
confirmed with various analytical methods. Several T-DNA populations have been made
available to public through the Arabidopsis Biological Resource Center
Grttp://www.arabidopsis.org/abrc/tdna_lines.html). With the advance of high-throughput
sequencing, different groups have also sequenced large numbers of T-DNA insertions

from the available T-DNA knockout mutant populations. Researchers can simply search

59

 

the T-DNA mutant population with a DNA sequence of interest to identify potential T-
DNA mutant. One such site used in this research is the SIGAL website at the Salk
Institute (http://signal.salk.edu/cgi-bin/tdnaexpress).

Two B-hydroxylase genes have been cloned from Arabidopsis and represent the
only identifiable B-hydroxylases in the genome (Sun et a1., 1996; Tian and DellaPenna,
2001). During the course of this research, a single antisense B-hydroxylase transgene was
employed to attempt to study the function of both B-hydroxylase 1 and 2 in viva (Rissler
and Pogson, 2001). The two B—hydroxylases are 79% identical at the nucleic acid level
and a concern with this approach is that a single antisense gene might not be able to
effectively eliminate expression of both genes. As discussed in Chapter 4, this appears to
be the case because the T-DNA knockout mutants differ significantly in phenotype from
the B-hydroxylase 1 antisense plants. T-DNA knockout mutants that correspond to
disruptions of the respective B-hydroxylases were isolated during this dissertation study
and crossed to each other and to the lat] mutant. The effects of multiple hydroxylase

deficiency on carotenoid composition and possible interactions and overlapping functions

of the hydroxylases will be discussed in Chapter 4.

 

 

References

Al-Babili, S., v. Lintig, J., Haubruck, H., and Beyer, P. (1996). A novel, soluble form
of phytoene desaturase from Narcissus pseudanarcissus chromoplasts is Hsp70-
complexed and competent for flavinylation, membrane association and enzymatic
activation. Plant J. 9, 601-612.

Al-Babili, S., Hugueney, P., Schledz, M., Welsch, R., Frohnmeyer, H., Laule, 0., and
Beyer, P. (2000). Identification of a novel gene coding for neoxanthin synthase from
Salanum tuberasum. FEBS Lett. 485,168-172.

Andersson, J., Walters, R.G., Horton, P., and Jansson, S. (2001). Antisense inhibition
of the photosynthetic antenna proteins CP29 and CP26: implications for the mechanisms
of protective energy dissipation. Plant Cell 13, 1193-1204.

Bak, S., and Feyereisen, R. (2001). The involvement of two P450 enzymes, CYP83B1
and CYP83A1, in auxin homeostasis and glucosinolate biosynthesis. Plant Physiol. 127,
108-118.

Bak, 8., Tax, F.E., Feldmann, K.A., Galbraith, D.W., and Feyereisen, R. (2001).
CYP83Bl, a cytochrome P450 at the metabolic branch point in auxin and indole
glucosinolate biosynthesis in Arabidopsis. Plant Cell 13, 101-111.

Barnes, H. J. (1996). Maximizing expression of eukaryotic cytochrome P4503 in
Escherichia coli. Methods Enzymol. 272, 3-14.

Barnes, H.J., Arlotto, M.P., and Waterman, M.R. (1991). Expression and enzymatic
activity of recombinant cytochrome P450 17 alpha-hydroxylase in Escherichia coli. Proc.
Natl. Acad. Sci. USA 88, 5597-5601.

Bartley, G.E., Scolnik, P.A., and Beyer, P. (1999). Two Arabidopsis thaliana carotene
desaturases, phytoene desaturase and zeta-carotene desaturase, expressed in Escherichia
coli, catalyze a poly-cis pathway to yield pro-lycopene. Eur. J. Biochem. 259, 396-403.

Bassi, R., Sandona, D., and Croce, R. (1997). Novel aspects of chlorophyll a/b-binding
proteins. Physiol. Plant 100, 769-779.

Bertea, C.M., Schalk, M., Karp, F., Maffei, M., and Croteau, R. (2001).
Demonstration that menthofuran synthase of mint (Mentha) is a cytochrome P450
monooxygenase: cloning, functional expression, and characterization of the responsible
gene. Arch. Biochem. Biophys. 390, 279-286.

Beyer, P., Weiss, G., and Kleinig, H. (1985). Solubilization and reconstitution of the
membrane-bound carotenogenic enzymes from daffodil chromoplasts. Eur. J. Biochem.

61

153, 341-346.

Bouvier, F., Hugueney, P., D'Harlingue, A., Kuntz, M., and Camara, B. (1994).
Xanthophyll biosynthesis in chromoplasts: isolation and molecular cloning of an enzyme
catalyzing the conversion of 5,6-epoxicarotenoids into ketocarotenoid. Plant J. 6, 45-54.

Bouvier, F., D'Harlingue, A., Hugueney, P., Marin, E., Marion-Poll, A., and
Camara, B. (1996). Xanthophyll biosynthesis: cloning, expression, functional
reconstitution, and regulation of beta-cyclohexenyl carotenoid epoxidase from pepper
(Capsicum annuum). J. Biol. Chem. 271, 28861-28867.

Bouvier, F., D'Harlingue, A., Suire, C., Backhaus, R.A., and Camara, B. (1998a).
Dedicated roles of plastid transketolases during the early onset of isoprenoid biosynthesis
in pepper fi'uits. Plant J. 117, 1423-1431.

Bouvier, F., Keller, Y., D'Harlingue, A., and Camara, B. (1998b). Xanthophyll
biosynthesis: molecular and functional characterization of carotenoid hydroxylases from
pepper fruits (Capsicum annuum L.). Biochirn. Biophys. Acta. 1391, 320-328.

Bouvier, F., D’Harlingue, A., Backhaus, R.A., Kumagai, M.H., Camara, B. (2000).
Identification of neoxanthin synthase as a carotenoid cyclase paralog. Eur. J. Biochem.
267, 6346-6352.

Bramley, P.M. (2002). Regulation of carotenoid formation during tomato fruit ripening
and development. J. Exp. Bot. 53, 2107-2113.

Britton, G. (1998). Overview of carotenoid biosynthesis. In Carotenoids: Biosynthesis
and Metabolism, Vol. 3, G. Britton, S. Liaaen-Jensen, and H. Pfander, eds (Basel,
Switzerland: Birkhauser), pp. 13-147.

Bugos, R.C., Hieber, A.D., and Yamamoto, H.Y. (1997). Xanthophyll cycle enzymes
are members of the lipocalin family, the first identified from plants. J. Biol. Chem. 273,
15321-15324.

Bungard, R.A., Ruban, A.V., Hibberd, J.M., Press, M.C., Horton, P., and Scholes,
JD. (1999). Unusual carotenoid composition and a new type of xanthophyll cycle in
plants. Proc. Natl. Acad. Sci. USA 96, 1135-1139.

Burkhardt, P.K., Beyer, P., Wunn, J., Kliiti, A., Armstrong, G.A., Schledz, M., van
Lintig, J., and Potrykus, I. (1997). Transgenic rice (Oryza sativa) endosperm
expressing daffodil (Narcissus pseudanarcissus) phytoene synthase accumulates
phytoene, a key intermediate of provitamin A biosynthesis. Plant J. 11, 1071-1078.

Camara, B., and Moneger, R. (1981). Carotenoid biosynthesis. In vitro conversion of

antheraxanthin to capsanthin by a chromoplast enriched fraction of capsicum fruits.
Biochem. Biophys. Res. Commun. 99, 1 117-1122.

62

Campbell, D., Hurry, V., Clarke, A.K., Gustafsson, P., and Oquist, G. (1998).
Chlorophyll fluorescence analysis of cyanobacterial photosynthesis and acclimation.
Microbiol. Mol. Biol. Rev. 62, 667—683.

Carol, P., Stevenson, D., Bisanz, C., Breitenbach, J., Sandmann, G., Mache, R.,
Coupland, G., and Kuntz, M. (1999). Mutations in the Arabidopsis gene IMMUTANS
cause a variegated phenotype by inactivating a chloroplast terminal oxidase associated
with phytoene desaturation. Plant Cell 11, 57-68.

Chapple, C. (1998). Molecular-genetic analysis of plant cytochrome P450-dependent
monooxygenase. Ann. Rev. Plant Physiol. Plant Mol. Biol. 49, 311-343.

Cunningham, F.X. Jr., Sun, Z.R., Chamovitz, D., Hirschberg, J., and Gantt, E.
(1994). Molecular structure and enzymatic function of lycopene cyclase from the
cyanobacterium Synechococcus sp. strain PCC7942. Plant Cell 6, 1107-1121.

Cunningham, F.X. J r., Pogson, 8., Sun, Z.R., McDonald, K.A., DellaPenna, D., and
Gantt, E. (1996). Functional analysis of the B and e lycopene cyclase enzymes of
Arabidopsis reveals a mechanism for control of cyclic carotenoid formation. Plant Cell 8,
1613-1626.

Cunningham, F.X. Jr., and Gantt, E. (1998). Genes and enzymes of carotenoid
biosynthesis in plants. Ann. Rev. Plant Physiol. Plant Mol. Biol. 49, 557-583.

Cunningham, F.X. Jr., and Gantt, E. (2001). One ring or two? Determination of ring

number in carotenoids by lycopene s-cyclases. Proc. Natl. Acad. Sci. USA 98, 2905-
2910.

Cunningham, F.X. Jr. (2002). Regulation of carotenoid synthesis and accumulation in
plants. Pure Appl. Chem. 74, 1409-1417.

Davison, P.A., Hunter, C.N., and Horton, P. (2002). Overexpression of B-carotene
hydroxylase enhances stress tolerance in Arabidopsis. Nature 418, 203-206.

Elrad, D., Niyogi, K.K., and Grossman, AR. (2002). A major light-harvesting
polypeptide of photosystem 11 functions in thermal dissipation. Plant Cell 14, 1801-1816.

Ernst, S., and Sandmann, G. (1988). Poly-cis carotene pathway in the Scenedesmus
mutant C-6D. Arch. Microbiol. 150, 590—594.

Estabrook, R. W. and Werringloer, J. (1978). The measurement of difference spectra:
application to the cytochromes of microsomes. Methods Enzymol. 52, 212-220.

Feldmann, K. A. (2001). Cytochrome P4503 as genes for crop improvement. Curr. Opin.
Plant Biol. 4, 162-167.

63

Fernandez-Gonzalez, B., Sandmann, G., and Vioque, A. (1997). A new type of
asymmetrically acting beta-carotene ketolase is required for the synthesis of echinenone
in the cyanobacterium Synechocystis sp. PCC 6803. J. Biol. Chem 272, 9728—9733.

Flesch, G., and Rohmer, M. (1988). Prokaryotic hopanoids: the biosynthesis of the
bacteriohopan skeleton. Eur. J. Biochem. 175, 405-411.

Fray, R.G., Wallace, A., Fraser, P.D., Valera, D., Hedden, P., Bramley, P.M., and
Grierson,D. (1995). Constitutive expression of a fruit phytoene desaturase gene in
transgenic tomatoes causes dwarfism by redirecting metabolites from the gibberelin
pathway. Plant J. 8, 693-701.

Froehlich, J.E., Itoh, A., and Howe, G.A. (2001). Tomato allene oxide synthase and
fatty acid hydroperoxide lyase, two cytochrome P4503 involved in oxylipin metabolism,
are targeted to different membranes of chloroplast envelope. Plant Physiol. 125, 306-317.

Fujita, V.S., Thiele, D.J., and Coon, M.J. (1990). Expression of alcohol-inducible
rabbit liver cytochrome P-450 3a (P-45011E1) in Saccharamyces cerevisiae with the
copper-inducible CUP] promoter. DNA Cell Biol. 9, 111-118.

Fukuzawa, K., Hayashi, K., and Suzuki, A. (1977). Effects of alpha-tocopherol analogs
on lysosome membranes and fatty acid monolayers. Chem. Phys. Lipids 18, 39-48.

Gilmore, A.M., Hazlett, T.L., and Govindjee, G. (1995). Xanthophyll cycle-dependent
quenching of photosystem II chlorophyll a fluorescence: formation of a quenching
complex with a short fluorescence lifetime. Proc. Natl. Acad. Sci. USA 92, 2273-2277.

Gilmore, A.M. (1997). Mechanistic aspects of xanthophyll cycle-dependent
photoprotection in higher plant chloroplasts and leaves. Physiol. Plant 99, 197-209.

Gilmore, A.M., Shinkarev, V.P., Hazlett, T.L., and Govindjee, G. (1998).
Quantitative analysis of the effects of intrathylakoid pH and xanthophyll cycle pigments
on chlorophyll a fluorescence lifetime distributions and intensity in thylakoids.
Biochemistry 37, 13582-13593.

Giuliano, G., Bartley, G.E., and Scolnik, P. (1993). Regulation of carotenoid
biosynthesis during tomato development. Plant Cell 5, 379-3 87.

Goodwin, T.W. (1971). Biosynthesis of carotenoids and plant triterpenes. Biochem. J.
123, 293-329.

Graham, S. E., and Peterson, J. A. (1999). How similar are P4503 and what can their
differences teach us? Arch. Biochem. Biophys. 369, 24-29.

Gruszecki, W.I., and Sielewiesiuk, J. (1991). Galactolipid multibilayers modified with
xanthophylls: orientational and diffractometric studies.Biochim. Biophys. Acta. 1069, 21-

26.

Guengerich, F. P. (1991). Reactions and significance of cytochrome P-450 enzymes. J.
Biol. Chem. 266, 10019-10022.

Hansen, C.H., Du, L., Naur, P., Olsen, C.E., Axelsen, K.B., Hick, A.J., Pickett, J.A.,
and Halkier, B.A. (2001). CYP83b1 is the oxime-metabolizing enzyme in the
glucosinolate pathway in Arabidopsis. J. Biol. Chem. 276, 24790-24796.

Hauser, M.T., Adhami, F., Dorner, M., Fuchs, E., and Glossl, J. (1998). Generation of
co-dominant PCR-based markers by duplex analysis on high resolution gels. Plant J. 16,
117-125.

Havaux, M., and Niyogi, KK. (1999). The violaxanthin cycle protects plants from
photooxidative damage by more than one mechanism. Proc. Natl. Acad. Sci. USA 96,
8762-8767.

Helliwell, C.A., Poole, A., Peacock, W.J., and Dennis, E.S. (1999). Arabidopsis ent-
kaurene oxidase catalyzes three steps of gibberellin biosynthesis. Plant Physiol. 119, 507-
5 10.

Helliwell, C.A., Chandler, P.M., Poole, A., Dennis, E.S., and Peacock, W.J. (2001).
The CYP88A cytochrome P450, ent-kaurenoic acid oxidase, catalyzes three steps of the
gibberellin biosynthesis pathway. Proc. Natl. Acad. Sci. USA 98, 2065-2070.

Helliwell, C.A., Sullivan, J.A., Mould, R.M., Gray, J.C., Peacock, W.J., and Dennis,
E.S. (2001). A plastid envelope location of Arabidopsis ent-kaurene oxidase links the
plastid and endoplasmic reticulum steps of the gibberellin biosynthesis pathway. Plant J.
28, 201-208.

Herz, S., Wungsintaweekul, J., Schuhr, C.A., Hecht, S., Liittgen, H., Sagner, S.,
Fellermeier, M., Eisenreich, W., Zenk, M.H., Bacher, A., and Rohdich, F. (2000).
Biosynthesis of terpenoids: ngB protein converts 4-diphosphocytidyl-ZC-methyl-D-
erythritol 2-phosphate to 2C-methyl-D-erythritol 2,4-cyclodiphosphate. Proc. Natl. Acad.
Sci. USA 97, 2486-2490.

Hirschberg, J. (2001). Carotenoid biosynthesis in flowering plants. Curr. Opin. Plant
Biol. 4, 210-218.

Horbach, S., Sahm, H., and Welle, R. (1993). Isoprenoid biosynthesis in bacteria: two
different pathways? FEMS Microbiol. Lett. 111, 135-140.

Hugueney, P., Badillo, A., Chen, H.C., Klein, A., and Hirschberg, J. (1995).

Metabolism of cyclic carotenoids: a model for the alteration of this biosynthetic pathway
in Capsicum annuum chromoplasts. Plant J. 8, 417-424.

65

Hull, A.K., and Celenza, J.L. (1999). Bacterial expression and purification of the
Arabidopsis NADPH-Cytochrome P450 reductase ATR2. Protein Expr. Purif. 18, 310-
315.

Hull, A. K., Vij, R., and Celenza, J.L. (2000). Arabidopsis cytochrome P4503 that
catalyze the first step of tryptophan-dependent indole-3-acetic acid biosynthesis. Proc.
Natl. Acad. Sci. USA 97, 2379-2384.

Humphreys, J.M., Hemm, M.R., and Chapple, C. (1999). New routes for lignin
biosynthesis defined by biochemical characterization of recombinant ferulate 5-
hydroxylase, a multifunctional cytochrome P450-dependent monooxygenase. Proc. Natl.
Acad. Sci. USA 96, 10045-10050.

Isaacson, T., Ronen, G., Zamir, D., and Hirschberg, J. (2002). Cloning of tangerine

from tomato reveals a carotenoid isomerase essential for the production of B-carotene and
xanthophylls in plants. Plant Cell 14, 333-342.

Itoh, A., and Howe, G.A. (2001). Molecular cloning of a divinyl ether synthase.
Identification as a CYP74 cytochrome P-450. J. Biol. Chem. 276, 3620-3627.

Kleinig, H., and Beyer, P. (1985). Carotene synthesis in spinach (Spinacia aleracea)
chloroplasts and daffodil (Narcissus pseudanarcissus) chromoplasts. Methods Enzymol.
110, 267-273.

Koornneef, M., and Stam, P. (1992). Genetic analysis. In Methods in Arabidopsis
research. C. Koncz, N.H., Chua, and J. Schell, eds (Singapore: World Scientific
Publishing Co), pp. 83-99.

Krubasik, P., and Sandmann, G. (2000). Molecular evolution of lycopene cyclases
involved in the formation of carotenoids with ionone end groups. Biochem. Soc. Trans.
28, 806-810.

Krysan, P.J., Young, J.C., Sussman, M.R. (1999). T-DNA as an insertional mutagen in
Arabidopsis. Plant Cell 11, 2283-2290.

Kuzuyama, T., Takahashi, S., Watanabe, H., and Seto, H. (1998). Direct formation of
2-C-methyl-D-erythritol 4-phosphate from l-deoxy-D-xylulose 5-phosphate by l-deoxy-

D-xylulose 5-phosphate reductoisomerase, a new enzyme in the non-mevalonate pathway
to isopentenyl diphosphate. Tetrahedron Lett. 39, 4509-4512.

Li, X.P., Bjr’irkman, 0., Shih, C., Grossman, A.R., Rosenquist, M., Jansson, S., and
Niyogi, KK. (2000). A pigrnent-binding protein essential for regulation of
photosynthetic light harvesting. Nature 403, 391-395.

Li, X.P., Phippard, A., Pasari, J., and Niyogi, K.K. (2002a). Structure-function
analysis of photosystem II subunit (PsbS) in viva. Func. Plant Biol. 29, 1131-1139.

66

Li, X.P., Gilmore, A.M., and Niyogi, K.K. (2002b). Molecular and global time-resolved

analysis of a psbS gene dosage effect on pH- and xanthophyll cycle-dependent
nonphotochemical quenching in photosystem II. J. Biol. Chem. 277, 33590-33597.

Li, X.P., Miiller-Moulé, P, Gilmore, A.M., and Niyogi, K.K. (2002c). PsbS-dependent
enhancement of feedback de-excitation protects photosystem II from photoinhibition.
Proc. Natl. Acad. Sci. USA 99, 15222-15227.

Lichtenthaler, H.K. (1999). The l-deoxy-D-xylulose-5-phosphate pathway for
isoprenoid biosynthesis in plants. Ann Rev Plant Physiol Plant Mol Biol 50, 47-65.

Lichtenthaler, H.K. (2000). Non-mevalonate isoprenoid biosynthesis: enzymes,
genes and inhibitors. Biochem. Soc. Trans. 28, 785-789.

Lohr, M., and Wilhelm, C. (1999). Algae displaying the diadinoxanthin cycle also
possess the violaxanthin cylce. Proc. Natl. Acad. Sci. USA 96, 8784-8789.

Lois, L.M., Campos, N., Rosa Putra, S., Danielsen, K., Rohmer, M., and Boronat, A.
(1998). Cloning and characterization of a gene from Escherichia coli encoding a
transketolase-like enzyme that catalyzes the synthesis of D-l-deoxyxylulose 5-phosphate,
a common precursor for isoprenoid, thiamin, and pyridoxol biosynthesis. Proc. Natl.
Acad. Sci. USA 95, 2105-2110.

Lois, L.M., Rodriguez-Concepcion, M., Gallego, F., Campos, N., and Boronat, A.
(2000). Carotenoid biosynthesis during tomato fruit development: regulatory role of 1-
deoxy-D-xylulose 5-phosphate synthase. Plant J. 22, 503-513.

Lokstein, H., Tian, L., Polle, J.E., and DellaPenna, D. (2002). Xanthophyll
biosynthetic mutants of Arabidopsis thaliana: altered nonphotochemical quenching of
chlorophyll fluorescence is due to changes in Photosystem II antenna size and stability.
Biochim. Biophys. Acta. 1553, 309-319.

Lotan, T., and Hirschberg, J. (1995). Cloning and expression in Escherichia coli of the
gene encoding B-C-4-oxygenase, that converts B-carotene to the ketocarotenoid
canthaxanthin in Haematococcus pluvialis. FEBS Lett 364, 125-128.

Liittgen, H., Rohdich, F., Herz, S., Wungsintaweekul, J., Hecht, S., Schuhr, C.A.,
Fellermeier, M., Sagner, S., Zenk, M.H., Bacher, A., and Eisenreich, W. (2000).
Biosynthesis of terpenoids: YchB protein of Escherichia coli phosphorylates the 2-
hydroxy group of 4-diphosphocytidyl-2-C-methyl-D-erythritol. Proc. Natl. Acad. Sci.
USA 97, 1062-1067.

Mandel, M.A., Feldmann, K.A., Herrera-Estrella, L., Rocha-Sosa, M., and Leon, P.

(1996). CLA I , a novel gene required for chloroplast development, is highly conserved in
evolution. Plant J. 9, 649-658.

67

Mann, V., Harker, M., Pecker, I., and Hirschberg, J. (2000). Metabolic engineering of
astaxanthin production in tobacco flowers. Nat. Biotechnol. 18, 888-892.

Marin, E., Nussaume, L., Quesada, A., Gonneau, M., Sotta, B., Hugueney, P., Frey,
A., and Marion-Poll, A. (1996). Molecular identification of zeaxanthin epoxidase of
Nicatiana plumbaginifalia, a gene involved in abscisic acid biosynthesis and
corresponding to the ABA locus of Arabidopsis thaliana. EMBO J. 15, 2331-2342.

Masamoto, K., Wada, H., Kaneko, T., and Takaichi, S. (2001). Identification of a
gene required for cis-to-trans carotene isomerization in carotenogenesis of the
cyanobacterium Synechocystis 3p. PCC 6803. Plant Cell Physiol. 42, 1398-1402.

Misawa, N., Masamoto, K., Hori, T., Ohtani, T., Biiger, P., and Sandmann, G.
(1994). Expression of an Erwinia phytoene desaturase gene not only confers multiple
resistance to herbicides interfering with carotenoid biosynthesis but also alters
xanthophyll metabolism in transgenic plants. Plant J. 8, 481—489.

Moehs, C.P., Tian, L., Osteryoung, K.W., and DellaPenna, D. (2001). Analysis of
carotenoid biosynthetic gene expression during marigold petal development. Plant Mol.
Biol. 45, 281-293.

Miiller, P., Li, X.P., and Niyogi, K.K. (2001). Non-photochemical quenching. A
response to excess light energy. Plant Physiol. 125, 1558-1566.

Nelson, D. R. (1999). Cytochrome P450 and the individuality of species. Arch. Biochem.
Biophys. 369, 1-10.

Niyogi, K.K., Grossman, A.R., and Bjfirkman, O. (1998). Arabidopsis mutants define
a central role for the xanthophyll cycle in the regulation of photosynthetic energy
conversion. Plant Cell 10, 1 121-1 134.

Niyogi, K.K. (1999). Photoprotection revisited: genetic and molecular approaches. Annu.
Rev. Plant Physiol. Plant Mol. Biol. 50, 333-359.

Niyogi, K.K., Shih, K., Chow, W.S., Pogson, B.J., DellaPenna, D., and Bjorkman, O.
(2001). Photoprotection in a zeaxanthin- and lutein-deficient double mutant of
Arabidopsis. Photosynth. Res. 67, 139-145.

Norris, S.R., Barrette, T.R., and DellaPenna, D. (1995). Genetic dissection of
carotenoid synthesis in Arabidopsis defines plastoquinone as an essential component of

phytoene desaturation. Plant Cell 7, 2139-2149.

Ortiz de Montellano, P.R. (1995). Structure, mechanism, and inhibition of cytochrome
P450. Drug Metab. Dispos. 23, 1181-1187.

68

Paquette, S.M., Bak, S., and Feyereisen, R. (2000). Intron-exon organization and
phylogeny in a large superfamily, the paralogous cytochrome P450 genes of Arabidopsis
thaliana. DNA Cell Biol. 19, 307-317.

Park, H., Kreunen, S.S., Cuttriss, A.J., DellaPenna, D., and Pogson, B.J. (2002).
Identification of the carotenoid isomerase provides insight into carotenoid biosynthesis,
prolarnellar body formation, and photomorphogenesis. Plant Cell 14, 321-332.

Pogson, B., McDonald, K.A., Truong, M., Britton, G., and DellaPenna, D. (1996).
Arabidopsis carotenoid mutants demonstrate that lutein is not essential for photosynthesis
in higher plants. Plant Cell 8, 1627-1639.

Pogson, B.J., Niyogi, K.K., Bjfirkman, 0., and DellaPenna, D. (1998). Altered
xanthophyll compositions adversely affect chlorophyll accumulation and
nonphotochemical quenching in Arabidopsis mutants. Proc. Natl. Acad. Sci. USA 95,
13324-13329.

Pogson, B.J., and Rissler, H.M. (2000). Genetic manipulation of carotenoid
biosynthesis and photoprotection. Phil. Trans. Roy. Soc. Lond. B 355, 1395-1403.

Pompon, D., Louerat, B., Bronine, A., and Urban, P. (1996). Yeast expression of
animal and plant P4503 in optimized redox environments. Methods Enzymol. 272, 51-64.

Pompon, D., and Nicolas, A. (1989). Protein engineering by cDNA recombination in
yeasts: shuffling of mammalian cytochrome P-450 functions. Gene 83, 15-24.

Poulos, T. L. (1995). Cytochrome P450. Curr. Opin. Struct. Biol. 5, 767-774.

Ralston, L., Kwon, S.T., Schoenbeck, M., Ralston, J., Schenk, D.J., Coates, R.M.,
and Chappell, J. (2001). Cloning, heterologous expression, and functional
characterization of 5-epi-aristolochene-l,3-dihydroxylase from tobacco (Nicatiana
tabacum). Arch. Biochem. Biophys. 393, 222-235.

Rissler, H.M., and Pogson, B.J. (2001). Antisense inhibition of the beta-carotene
hydroxylase enzyme in Arabidopsis and the implications for carotenoid accumulation,
photoprotection and antenna assembly. Photosynth. Res. 67, 127-137.

Ro, D.K., Mah, N., Ellis, BE, and Douglas, C.J. (2001). Functional characterization
and subcellular localization of poplar (Populus trichocarpa x Populus deltoides)
cinnamate 4-hydroxylase. Plant Physiol. 126, 317-329.

Rockholm, D.C., and Yamamoto, H.Y. (1996). Violaxanthin de-epoxidase. Plant
Physiol. 110, 697-703.

Rohdich, F., Wungsintaweekul, J., Fellermeier, M., Sagner, S., Herz, S., Kis, K.,
Eisenreich, W., Bacher, A., and Zenk, M.H. (1999). Cytidine 5'-triphosphate-

69

dependent biosynthesis of isoprenoids: ngP protein of Escherichia coli catalyzes the
formation of 4-diphosphocytidyl-2-C-methylerythritol. Proc. Natl. Acad. Sci. USA 96,
11758-11763.

R6mer, 8., Fraser, P.D., Kiano, J.W., Shipton, C.A., Misawa, N., Schuch, W., and
Bramley, P.M. (2000). Elevation of the provitamin-A content of transgenic tomato
plants. Nat. Biotechnol. 18, 666-669.

Rohmer, M., Knani, M., Simon, P., Sutter, B., and Sahm, H. (1993). Isoprenoid
biosynthesis in bacteria: a novel pathway for the early steps leading to isopentenyl
diphosphate. Biochem. J. 295, 517-524.

Rohmer, M. (1999). The discovery of a mevalonate-independent pathway for isoprenoid
biosynthesis in bacteria, algae and higher plants. Nat. Prod. Rep. 16, 565-574.

Russell, D.W., and Conn, E.E. (1967). The cinnamic acid 4-hydroxylase of pea
seedlings. Arch. Biochem. Biophys. 122, 256-258.

Sandmann, G. (1994). Carotenoid biosynthesis in microorganisms and plants. Eur. J.
Biochem. 223, 7-24.

Sarry, J.E., Montillet, J.C., Sauvaire, Y., and Havaux, M. (1994). The protection
fimction of the xanthophyll cycle in photosynthesis. FEBS Lett. 353, 147-150.

Schledz, M., Al-Babili, S., Lintig, J., Haubruck, H., Rabbani, S., Kleinig, H., and
Beyer, P. (1996). Phytoene synthase from Narcissus pseudanarcissus: functional
expression, galactolipid requirement, topological distribution in chromoplasts and
induction during flowering. Plant J. 10, 781-792.

Schoefs, B., Rmiki, N., Rachadi, J., and Lemoine, Y. (2001). Astaxanthin
accumulation in Haematococcus requires a cytochrome P450 hydroxylase and an active
synthesis of fatty acids. FEBS Lett. 500, 125-128.

Schroepfer, G.J. Jr. (1981). Sterol biosynthesis. Annu. Rev. Biochem. 50, 585-621.

Schuler, M. A. (1996). Plant Cytochrome P450 Monooxygenase. Crit. Rev. Plant Sci.
15, 235-284.

Schwender, J., Seemann, M., Lichtenthaler, H.K., and Rohmer, M. (1996).
Biosynthesis of isoprenoids (carotenoids, sterols, prenyl side chain of chlorophylls and
plastoquinone) via a novel pyruvate/glyceraldehyde 3-phosphate non-mevalonate
pathway in the green Scenedesmus abliquus. Biochem. J. 316, 73-80.

Shanklin, J., Whittle, E., and Fox, B.G., (1994). Eight histidine residues are
catalytically essential in a membrane-associated iron enzyme, stearoyl-CoA desaturase,
and are conserved in alkane hydroxylase and xylene monooxygenase. Biochemistry 33,

70

12787-12794.

Shewmaker, C.K., Sheehy, J.A., Daley, M., Colburn, S., and Ke, D.Y. (1999). Seed-
specific overexpression of phytoene synthase: increase in carotenoids and other metabolic
effects. Plant J. 20, 401 -412.

Siminszky, B., Corbin, F.T., Ward, E.R., Fleischmann, T.J., and Dewey, RE. (1999).
Expression of a soybean cytochrome P450 monooxygenase cDNA in yeast and tobacco
enhances the metabolism of phenylurea herbicides. Proc. Natl. Acad. Sci. USA 96, 1750-
1755.

Simpson, K.L., and Chichester, CO. (1981). Metabolism and nutritional significance
of carotenoids. Ann. Rev. Nutr. 1, 351-374.

Sun, 2., Gantt, E., and Cunningham, F.X. (1996). Cloning and functional analysis of
the B-carotene hydroxylase of Arabidopsis thaliana. J. Biol. Chem. 271, 24349-24352.

Swiatek, M., Kuras, R., Sokolenko, A., Higgs, D., Olive, J., Cinque, G., Muller, B.,
Eichacker, L.A., Stern, D.B., Bassi, R., Herrmann, R.G., and Wollman, EA. (2001).
The chloroplast gene ycf9 encodes a photosystem II (PSII) core subunit, Pst, that
participates in PSII supramolecular architecture. Plant Cell 13, 1347-1367.

Takahashi, S., Kuzuyama, T., Watanabe, H., and Seto, H. (1998). A l-deoxy-D-
xylulose S-phosphate reductoisomerase catalyzing the formation of 2-C-methyl-D-
erythritol 4-phosphate in an alternative nonmevalonate pathway for terpenoid
biosynthesis. Proc. Natl. Acad. Sci. USA 95, 9879-9884.

Tardy, F., and Havaux, M. (1997). Thylakoid membrane fluidity and thermostability
during the operation of the xanthophyll cycle in higher-plant chloroplasts. Biochim.
Biophys. Acta. 1330, 179-193.

The Arabidopsis Genome Initiative (2000). Analysis of the genome sequence of the
flowering plant Arabidopsis thaliana. Nature 408, 796-815.

Tian, L., and DellaPenna, D. (2001). Characterization of a second carotenoid [3-
hydroxylase gene from Arabidopsis and its relationship to the LUT1 locus. Plant Mol.
Biol. 47, 379-388.

Tomes, M.L., Quackenbush, F.W., Nelson, O.E., and North, B. (1953). The
inheritance of carotenoid pigment system in the tomato. Genetics 38, 117—127.

Tracewell, C.A., Vrettos, J.S., Bautista, J.A., Frank, H.A., and Brudvig, G.W.
(2001). Carotenoid photooxidation in photosystem II. Arch. Biochem. Biophys. 385, 61-
69.

Urban, P., Mignotte, C., Kazmaier, M., Delorme, F., and Pompon, D. (1997).

71

Cloning, yeast expression, and characterization of the coupling of two distantly related
Arabidopsis thaliana NADPH-cytochrome P450 reductases with P450 CYP73A5. J. Biol.
Chem. 272, 19176-19186.

Vishnevetsky, M., Ovadis, M., Zuker, A., and Vainstein, A. (1999). Molecular
mechanisms underlying carotenogenesis in the chromoplast: multilevel regulation of
carotenoid-associated genes. Plant J. 20, 423-431.

Watson, C.J., Froehlich, J.E., Josefsson, C.A., Chapple, C., Durst, F., Benveniste, I.,
and Coolbaugh, RC. (2001). Localization of CYP86B1 in the outer envelope of
chloroplasts. Plant Cell Physiol. 42, 873-878.

Winkler, R.G., Frank, M.R., Galbraith, D.W., Feyereisen, R., and Feldmann, KA.
(1998). Systematic reverse genetics of transfer-DNA-tagged lines of Arabidopsis.
Isolation of mutations in the cytochrome P450 gene superfamily. Plant Physiol. 118, 743-
750.

Wang, L. L. (1998). Cytochrome P450 monooxygenases. Curr. Opin. Chem. Biol. 2,
263-268.

Xu, W., Bak, S., Decker, A., Paquette, S.M., Feyereisen, R., and Galbraith, D.W.
(2001). Microarray-based analysis of gene expression in very large gene families: the
cytochrome P450 gene superfamily of Arabidopsis thaliana. Gene 272, 61-74.

Yamamoto, H.Y., Chichester, C.O., and Nakayama, T.O.M. (1962). Arch. Biochem.
Biophys. 96, 645.

Ye, X., Al-Babili, S., Kliiti, A., Zhang, J., Lucca, P., Beyer, P., and Potrykus, I.

(2000). Engineering the provitarnin A (Ii-carotene) biosynthetic pathway into
(carotenoid-free) rice endosperm. Science 287, 303-305.

72

CHAPTER 2

Characterization of a second carotenoid B-hydroxylase gene from Arabidopsis and

its relationship to the LU T 1 locus

The work presented in this chapter has been published:

Tian, L. and DellaPenna, D. (2001) Plant Mol. Biol. 47, 379-388.

73

Abstract

Xanthophylls are oxygenated carotenoids that perform critical roles in plants. B-
carotene hydroxylases (Ii-hydroxylases) add hydroxyl groups to the B-rings of carotenes
and have been cloned from several bacteria and plants, including Arabidopsis. The IutI
mutation of Arabidopsis disrupts s-ring hydroxylation and has been suggested to identify
a related carotene hydroxylase that functions specifically on e-ring structures. We have
used library screening and genomics-based approaches to isolate a second B-hydroxylase
genomic clone and its corresponding cDNA from Arabidopsis. The encoded protein is
70% identical to the previously reported Arabidopsis B-hydroxylase 1. Phylogenetic
analysis indicates a common origin for the two proteins, however, their different
chromosomal locations, intron positions and intron sizes suggest their duplication is not
recent. Although both hydroxylases are expressed in all Arabidopsis tissues analyzed, 0-
hydroxylase 1 mRNA is always present at higher levels. Both cDNAs encode proteins
that efficiently hydroxylate the C-3 position of B-ring containing carotenes and are only
weakly active towards e-ring containing carotenes. Neither B-hydroxylase cDNA maps to
the LUT1 locus, and the genomic region encompassing the LUT1 locus does not contain a
third related hydroxylase. These data indicate that the LUT1 locus encodes a protein
necessary for s-ring hydroxylation but unrelated to B-hydroxylases at the level of amino

acid sequence.

74

Introduction

Carotenoids are isoprenoid-derived pigments that are synthesized by bacteria,
fungi and plants. In plants, carotenoids are synthesized and localized in plastids. In
chloroplasts, carotenoids are essential components of the photosynthetic antenna and
reaction center complexes and serve as precursors for abscisic acid biosynthesis (Rock
and Zeevaart, 1991). In chromoplasts of nonphotosynthetic tissues, carotenoids can
accumulate to high levels where they provide color to flowers and fruits and act as
attractants for pollinators (Goodwin, 1986). In addition to these roles in plants, the
carotenoid pathway is a target for herbicides (Bramley, 1994) and certain carotenoids
play a role in human nutrition as provitamin A (Solomons and Bulux, 1993).

Carotenoids can be divided into two major groups: carotenes and xanthophylls.
Carotenes are linear or cyclized hydrocarbons such as lycopene, or-carotene and [3-
carotene. Xanthophylls are oxygenated derivatives of carotenes and can contain epoxy,
keto or hydroxy groups. Lutein (B,a-carotene-3,3'-diol), a dihydroxy at-carotene
derivative, is the most abundant carotenoid in photosynthetic plant tissue while the
analogous dihydroxy B-carotene derivative, zeaxanthin, is an important component of the
xanthophyll cycle (Demmig-Adams et al. , 1996).

Carotenoid hydroxylases catalyze the formation of lutein and zeaxanthin from or-
carotene and B-carotene, respectively. B-hydroxylases add hydroxyl groups to the number
3 carbon of both B-rings and have been cloned from several photosynthetic eukaryotes,
including Arabidopsis, tomato and algae (reviewed in Cunningham and Gantt, 1998).
Biochemical studies of pepper fruit B-hydroxylase show that the enzyme is a nonheme

diiron protein containing conserved histidine motifs (Bouvier et al. , 1998) also found in

75

the membrane fatty acid desaturase family (Shanklin et al. , 1994). In vitro reconstitution
experiments indicate that pepper B-hydroxylase requires iron, ferredoxin and ferredoxin
oxido-reductase for activity and uses activated oxygen to break the C-H bond and
introduce a hydroxyl group (Bouvier et al. , 1998).

B-hydroxylases have also been cloned from various photosynthetic and non-
photosynthetic bacteria (Masamoto et al., 1998; Misawa et al., 1990; Hundle et al.,
1993). Although all B-hydroxylases carry out the same reaction, they can be clearly
categorized into three groups based on their primary structure. Eukaryotic B-hydroxylases
are highly conserved and share less than 30% protein identity with B-hydroxylases from
non-photosynthetic bacteria. The B-hydroxylase of Synechocystis PCC6803 (a
cyanobacteria) shares little similarity to plant and non-photosynthetic bacterial B-
hydroxylases with the exception of conserved histidine residues (Masamoto et al., 1998).
Synechocystis B-hydroxylase is more closely related to bacterial B-carotene ketolases (B-
ketolase), suggesting that it may have evolved independently of the plant and bacterial

hydroxylases.

One feature distinguishing plants and green algae from other carotenoid
containing organisms is the synthesis of e-ring containing carotenoids, for example, a-
carotene ([3,3-carotene). a-Rings differ from B-rings in the position of the double bond on
the ring structure (Figure 2.2). a-Carotene is converted to lutein by hydroxylation at
carbon 3 on both the B and 8 rings. Cloned Arabidopsis B-hydroxylase efficiently
hydroxylates carbon 3 of B-rings but showed almost no activity towards carbon 3 of 8-
rings (Sun et al. , 1996). In addition, the chirality of 8 ring hydroxylation is opposite to

that of B-ring hydroxylation (Britton, G., 1990), suggesting that a distinct activity is

76

involved. Genetic data further support the existence of a second hydroxylase as the 1m]
mutation of Arabidopsis disrupts e-ring hydroxylation without affecting B-hydroxylation
(Pogson et al. , 1996), suggesting that the LUT1 locus encodes a protein necessary for s-
hydroxylase activity.

We propose three hypotheses for the identity of the Arabidopsis 8-
hydroxylase/L UT 1 locus: 1) it is a hydroxylase evolved from and related to the [3-
hydroxylase, 2) it is a hydroxylase evolved independently and therefore divergent from
the B-hydroxylase, or 3) it is an ancillary protein that modifies the existing B-hydroxylase
to function towards e-rings. To test these hypotheses, we attempted to clone the
Arabidopsis e-hydroxylase by molecular, genomic and genetic approaches. In this paper,
we report the isolation and characterization of a second, related B-hydroxylase from

Arabidopsis and its relationship to the LUT1 locus.

Materials and methods

Screening of Arabidopsis thaliana genomic and cDNA libraries

An Arabidopsis CD4-8 Landsberg genomic library and CD4-13, CD4-l4 size
selected cDNA libraries were obtained from the Arabidopsis Biological Resource Center
(Columbus, Ohio). Approximately 200,000 plaques from the genomic library were
screened using the Arabidopsis B-hydroxylase 1 cDNA as a probe (Sun et al., 1996).
Prehybridization was performed in 6X SSPE, 0.5% nonfat dry milk, 0.5% SDS at 60°C

for four hr and hybridization with a-32P labeled B-hydroxylase 1 insert (500,000 cpm/ml

77

hybridization solution) at the same temperature overnight. Filters were washed twice at
55°C in 1X SSC, 0.1% SDS (low stringency wash) for 30 min and exposed to X-ray film
for 16 hr. Following film development, the membranes were washed twice in 0.1X SSC,
0.1% SDS at 65°C (high stringency wash) for 30 min and again exposed to X-ray film for
16 hr. Plaques showing strong signals at low stringency and partial loss of signal at high
stringency were picked and purified. Plaques showing equal signals at both low and high
stringency were used as positive controls. Phage DNA was purified according to standard
methods (Sambrook et al., 1988). The cDNA library was screened in a similar manner
using the B-hydroxylase 1 insert as a probe and putative clones were isolated and

sequenced.

DNA sequencing and analysis

DNA sequencing was performed using Prism Ready Reaction DyeDeoxy
Terminator Cycle Sequencing Kit and an ABI 310 automated DNA sequencer (Perkin
Elmer Applied Biosystems, Foster City, CA). cDNA and genomic clones were fully
sequenced by primer walking. DNA sequence analyses were carried out with

MacVectorTM 7.0 software (Eastman Kodak Company, Rochester, NY).

RNA isolation

One-month-old Arabidopsis leaf, stem, root, flower and silique tissues were
harvested, frozen in liquid nitrogen and ground to a fine powder. Three ml of

homogenization buffer/gm tissue (0.2 M sodium borate, 30 mM EGTA, 1% SDS, 1%

78

deoxycholate, 2% polyvinylpyrrolidone 40,000, 10 mM DTT) was preheated to 65°C and
one volume 1:1 (vzv) phenol:chloroform was added. The rrrixture was vortexed for
several min and the phases were separated by centrifuging at 10,000g for 20 min at room
temperature. The aqueous phase was extracted twice with one volume of chloroform,
adjusted to 2 M LiCl and incubated at 4°C overnight. The resulting precipitate was
collected by centrifuging at 10,000g for 20 min at room temperature. The pellet was
washed once with 2 M LiCl and resuspended in lml TE, 0.5%SDS and reprecipitated by
adding 0.1 volume 2 M potassitun acetate (pH 5.5) and two volumes of ethanol at room
temperature for 10 min. The precipitate was pelleted at 10,000g for 20 min at 4°C then
resuspended in 1 ml DEPC treated water. RNA was quantified spectrophotometrically
and stored at —80°C. Polysaccharides from total RNA preparation were removed as

described (Ausubel et al., 1995).

cDNA synthesis and TaqMan PCR assay

Reverse transcription reactions were conducted in triplicate using the TaqMan
Gold RT-PCR kit (Perkin Elmer Applied Biosystems, Foster City, CA) according to the
manufacture's protocol. Ten ug of total RNA was used as template in each reverse
transcription reaction. cDNA products from triplicate reactions were pooled together for
TaqMan analysis.

TaqMan probes and primers (shown in Table 2.1) were designed from the cDNA

sequences of B-hydroxylase 1, B-hydroxylase 2 and a housekeeping gene, Elongation

Factor la (EF-la), using Primer Express software (Perkin Elmer Applied Biosystems,

79

 

Table 2.1. TaqMan probes and primers used in this study.

 

B-hydroxylase 1 Forward
Reverse
Probe

B-hydroxylase 2 Forward
Reverse
Probe

EF-1 (it Forward
Reverse

Probe

CTCGTGCACAAGCGTTTCC

GGCGACCTTI'CGGAGGTAA

TGTAGGTCCCATCGCCGACGTCC

TI'CTCCGCAAACCACCCTATA

ACGTCGGAAGCCGTTGAAT

CCACCGCAGI I | ICCCTCCATCTCT

CGAACTTCCATAGAGCAATATCGA

GCATGGGTGTI'GGACAAACTT

ACCACGGTCACGCTCGGCCT

 

80

Foster City, CA). Reporter dye and quencher dye for the TaqMan probes were F AM (6-
carboxyfluorescein) and TAMRA (6-carboxytetramethylrhodamine), respectively.
TaqMan probe and primer concentrations were optimized by determining the minimum
probe and primer concentrations that gave the highest amplification for each probe target.
The optimal TaqMan probe, forward primer and reverse primer concentrations were: EF-
lor, 175 nM, 300 nM, 300 nM; B-hydroxylase 1, 225 nM, 50 nM, 50 nM; B-hydroxylase
2, 225 nM, 300 nM, 300 nM. TaqMan PCR assays for each gene target were performed
in triplicate on cDNA samples from reverse transcription reactions or plasmids standards
containing genes of interest using an ABI Prism 7700 Sequence Detection system (Perkin
Elmer Applied Biosystems, Foster City, CA). For each 30 ul TaqMan reaction, 2 pl
cDNA template (out of a 100 pl reverse-transcription reaction containing 10 ug total
RNA) was mixed with 15 ul TaqMan Universal PCR Master Mix (2X), 3 ul forward
primer, 3 ul reverse primer, 3 ul TaqMan probe and 4 ul H2O. PCR parameters were

50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s and 60°C for l min.

Data analysis of TaqMan PCR assay

TaqMan PCR results were captured and analyzed using Sequence Detector
Software (SDS version 1.7; Perkin Elmer Applied Biosystems, Foster City, CA). After
analyzing the raw data, an amplification plot for each sample was generated, showing the
increase in the reporter dye fluorescence (ARn) with each cycle. In the initial cycles of
PCR, there is little change in fluorescence signal, which defines the baseline above which

a fixed fluorescence threshold is set. The threshold cycle (CT), representing the cycle

81

number at which the fluorescence passed the fixed threshold, was then calculated for each
amplification plot. CT values of all the samples were exported into Microsoft Excel
worksheets for further analysis.

Standard curves were generated using serial dilutions of known molar quantities
of B-hydroxylase 1 or B-hydroxylase 2 cDNA. Quantification of B-hydroxylase 1 and B-
hydroxylase 2 transcripts in each tissue were accomplished by comparing the C1— values
to their standard curves, respectively. The molar amount of both transcripts (per 0.2 pg
total RNA) from the same tissue could then be directly compared (Figure 2.4). The
house-keeping gene EF-lor has its lowest expression (largest CT value) in floral tissue.
Relative EF -10t message in different tissues was calculated by arbitrarily setting floral

EF-l or equal to 1.

Computer analyses of B-hydroxylase sequences

Deduced amino acid sequences of B-hydroxylases from plants and bacteria were
aligned using the ClustalW (version 1.4) algorithm (Thompson et al., 1994). This
alignment was used to construct a neighbor-joining tree with the Synechocystis PCC6803
B-hydroxylase as the outgroup (MacVectorTM software, version 7.0; East Kodak
Company, Rochester, NY). Residues prior to the start codon of Agrobacterium
aurantiacum were not used for calculation of distance, i.e. chloroplast targeting
sequences of plant B-hydroxylases were not included for distance calculation. Positions
with gaps were also excluded from the calculation. Distances were corrected for multiple

substitutions (Kimura, 1980). GenBank accession numbers of the B-hydroxylases are

82

given in the legend to Figure 2.5. Bootstrap analysis was performed to test the reliability

of the branches (10,000 replicates).

HPLC analysis of enzymatic activities in carotenoid-producing Ecoli

[El-Carotene, the-carotene and B-carotene producing E. coli strains were generously
provided by Dr. Cunningham (University of Maryland, College Park, Maryland;
Cunningham et al., 1994; Cunningham et al., 1996). Plasmids expressing the open
reading frames of the B-hydroxylase 1 and 2 cDNAs were transformed into a B-carotene
producing E. coli strain (Sun et a1. 1996). Fifty ml of E. coli culture was grown in LB
media at 37°C to an OD. 550 of 0.8, at which time the expression of fusion protein was
induced with 0.4 mM IPTG at 28°C for an additional twelve hr. The cells were collected
by centrifiigation and resuspended in 400 pl H2O to which 750 pl 2:1 methanol:
chloroform (vzv) was added. After centrifugation, the chloroform phase containing
carotenoid pigments was recovered, transferred to a fresh eppendorf tube and dried under
vacuum. The carotenoid containing residue was resuspended in ethyl acetate, clarified by
centrifugation and resolved by chromatography on a C13 reverse phase HPLC column
(Waters, Milford, MA). Mobile phases were A: acetonitrile: H2O: triethylamine
(900:9921, v:v:v) and B: ethyl acetate. The column was equilibrated at 1 ml/min flow rate
with buffer A and the following gradient initiated after sample injection: 0 to 31.0 min:
0% B to 66.6% B; 31.0 to 31.2 min: 66.6% B to 100% B. 100% B buffer was maintained
for an additional two min at which time the column was reequilibrated with 100% buffer

A. Fifty ul samples were injected and data was collected at 440 nm, 477 nm and 296 nm.

83

Mapping of the lat] mutation

Homozygous lut1 mutants (Columbia) were crossed with wild type Landsberg
erecta and F2 progeny homozygous for the IutI mutation were selected by HPLC analysis
of leaf carotenoids as described (Pogson et a1., 1996). Genomic DNA was isolated from
individual F2 plants (Dellaporta, 1983) and further purified by phenolzchloroform
extraction and precipitation prior to PCR amplifications. The lutl mutation was mapped
relative to CAPS (Cleaved Amplified Polymorphic Sequences) markers (Konieczny and
Ausubel, 1993). CAPS PCR reactions were performed in 50 ul volumes containing PCR
buffer, 3 mM MgCl2, 0.125 mM dNTP mix, 2.5 U of Taq DNA polymerase, 30 pmol of
each primer and 50-100 ng genomic DNA template. Following preincubation at 94°C for
3 min, a PCR program (annealing at 55°C for 15 3, extension at 72°C for 30 s,

denaturation at 94°C for 15 s) was performed for 40 cycles.

Results

Isolation of cDNAs encoding a novel Arabidopsis B-carotene hydroxylase

In plants, B-hydroxylase adds hydroxyl groups to the 3 and 3' ring carbons of B-
carotene to form zeaxanthin. A cDNA encoding this enzyme has been previously cloned
from Arabidopsis by color complementation of a B-carotene containing E. coli strain (Sun
et al. 1996). This cDNA, B-hydroxylase 1, was used as a probe to screen an Arabidopsis

cDNA library for related clones. Of the 48 clones identified, most were truncated

84

versions of the B-hydroxylase 1 cDNA except for two, which appeared to encode a
second related hydroxylase in Arabidopsis. The newly identified clone was designated [3-
hydroxylase 2.

Figure 2.1 shows a comparison of the deduced protein sequences of B-
hydroxylase l and 2. The proteins share 70% identity at the primary sequence level. The
chloroplast transit peptide prediction software ChloroP v1.1 (Center for Biological
Sequence Analysis, The Technical University of Denmark, Denmark) predicts a
chloroplast transit peptide cleavage site between Val-52 and Glu-53 in both [3-
hydroxylase sequences, consistent with the presumed chloroplastic location of carotenoid
biosynthetic enzymes (Cunningham and Gantt, 1998). The putative transit peptides show
the highest divergence and their removal increases the primary sequence identity of the

two hydroxylases to 81%.

Activity and substrate specificity of the B-hydroxylase 2 protein

The B-hydroxylase 1 cDNA was previously shown to be highly active toward [3
rings and weakly active toward 8 rings in E. coli strains engineered to produce carotenoids
containing these ring structures (Sun et al. , 1996). The B-hydroxylase 2 cDNA was
similarly introduced into E. coli strains engineered to accumulate [3 or a ring containing
carotenes and the activity towards each ring structure was determined by HPLC.

When expressed in a B-carotene accumulating background, the B-hydroxylase 2
cDNA produced 90% zeaxanthin (B,B-carotene-3,3’-diol), and 6% of the

monohydroxylated intermediate, B-cryptoxanthin (B,B-carotene-3-ol) (Figure 2.2A). The

85

 

  

      
  
 

Milne 1 1 writers? “5117??! Facet: rm - m s s s r or R L R Lit! met: 3 ear 3 P
p-Ol-lase 2 1 IIAAG 1.8.1 I All? thK PL: N23 Sis-.155 A N H Pl 8 T A V Fifi Fifi-Li R ' if: N G
B—OHase 1 41 3 LR Fir-t RF saw: when Ram IE was mszr NA 1 DAE
[30113392 41 F RRR;x IL rvc P.MV EE RK.QS§ P MD onN KP E.S‘RT‘, s - - - s E r
B-Ol-lase1 :1 Y L A LiR‘L'f A E311 Lg REWKS'TE‘R smrrqu me‘rm vm
13011886 2 " LMT SLRL: L KK A-fi K;.KKS.-E..RFez‘-I. YALA AVM :55 F .51 1.5.-MAI NAVN
B-OHase 1 121 meatm seem 33 ML MWA

B-Ol-lase 2 131?:YRF swan KGGE v av L EMF or F AL 3 veAAv GME FWARw
fi-OHase 1 m fitmxisum; , a PM van-norm out
B—OHasez 161 Ah. us LINN-IE-ES PREeAsFEL NDVFA t- r NAVPAI eL-t.
B-OHase 1 201 Siren FNKGL V’PGL cr= emote Mr Vii-=31 armr— . ”in GI.‘ m
3.011;,” 201 YYGF LNKGL VEGL cP oAGLer IMF eMAvMPvfll -‘_cL KRR
ri-onase 1 241 trim-era ovmrmmz LE on MY L
B—OHaseZ 241 new A‘NvPvLRKvAA oL TDKFKGVPYGLF Ler V
B-OHase 1 231 AEEV‘G N

augment NSXK “68353-85

B-OHGSOZ 211.;EEVG KtEE-l‘.‘ t-x-EISRR-A. Isl-AI"

Figure 2.1. Alignment of proteins encoded by the Arabidopsis B-hydroxylase 1 and 2
cDNAs. Residues identical between B-hydroxylase l and 2 (B-OHase 1 and B—OHase 2,
respectively) are shaded. Conserved histidine residues are boxed. The black arrow

indicates the predicted cleave site of the transit peptide in both sequences.

86

B-hydroxylase 1 cDNA gave a similar conversion in this background (data not shown).
When the B-hydroxylase 2 cDNA was expressed in a 8-carotene accumulating E. coli
background (ES-carotene, emu-carotene, which has a single s-ring structure), only a small
amount (<6%) was converted to the hydroxylated product (8,w-carotene-3-ol) (Figure
2.23), again similar to what is observed with the B-hydroxylasel cDNA (data not
shown). Another e-ring containing carotenoid (gs-carotene which has a-rings at both
ends of the polyene chain) was also tested and similarly low levels of e-ring

hydroxylation were observed with both B-hydroxylase cDNAs (data not shown).

Structural comparisons of B-hydroxylase l and B-hydroxylase 2 genes

The B-hydroxylase 1 cDNA was also used to screen an Arabidopsis genomic
library and clones encoding both the B-hydroxylase 1 and B-hydroxylase 2 genes were
identified and fully sequenced. Their structures are compared in Figure 2.3A. Both
genomic clones contain six introns of differing size and sequence that are conserved in
position in the two genes. Subsequent to sequencing by our laboratory, BAC clones
containing both B-hydroxylase genes were sequenced by the Arabidopsis Genome
Initiative Project, hence their chromosomal locations are known (Figure 2.3B). The B-
hydroxylase 1 gene is located on chromosome 4 at 75 cM (GenBank accession number
AF 125577), while the B-hydroxylase 2 gene is located on chromosome 5 at 105 cM

(GenBank accession number ABOZS606).

Expression of B-hydroxylase 1 and 2 genes in different Arabidopsis tissues

87

Figure 2.2. HPLC elution profiles of B-hydroxylase l and 2 cDNAs expressed in [3
or a ring-producing E. coli. A. HPLC elution profiles of carotenoids in the B-
carotene-accumulating E. coli strain (gray trace) and after transformation with the [3-
hydroxylase 2 expression plasmid (black trace). B. HPLC elution profiles of
carotenoids in the B-carotene-accmnulating E. coli strain (gray trace) and after
transformation with the B-hydroxylase 2 expression plasmid (black trace). The

chemical structures of the compounds of each peak are shown below their names.

\\\\\\\\\
R:

C. The chemical structures of B and a-rings.

88

 

 

 

 

 

A
zeaxanthin
1mm won
E B-carotono
3 ii
1
~
a: 1;
g 11’
11
3'3 11
e
3 B-cryptounthln
.n
<
l I l

 

 

 

l
1 o 1 5 20 25 30
minutes

 

fi-urotom

 

Absorption at 440 nm

 

 

 

 

 

 

minutes
1
2 8
3 5
4
efing

 

 

 

 

Figure 2.2

89

Figure 2.3. A. Schematic diagrams of B-hydroxylase 1 and 2 genomic clones. Exons and
introns are drawn to scale with their positions and sizes indicated by straight lines and
triangles extracted from the exon sequences, respectively. B. Map positions of B-
hydroxylase 1, B-hydroxylase 2 and LUT1 locus. B-hydroxylase l and 2 map positions
were obtained from the Arabidopsis thaliana Database. The map position of lutl was

determined by linkage to CAPS markers. The diagram is not to scale.

90

 

 

B-OHase 1 I '
ATG TGA
B—OHase 2 - '
ATG TM
100 bp
B
III IV V
- —LUT 1
72.8 - —F3Ha 75 — — B_0Hase 1
78.4 - - 98300

105--B-OHase 2
115--LFY3

 

 

 

Figure 2.3

91

Initial attempts to determine B-hydroxylases transcripts level in various
Arabidopsis tissues by Northern blotting were not successful due to the fact that both
genes are expressed at extremely low levels. We therefore adopted the more sensitive
TaqMan RT-PCR method for expression analyses (Heid et a1 ., 1996). Quantification was
performed using a standard curve of each target sequence.

Both hydroxylase mRNAs are most abundant in Arabidopsis leaf tissue with B-
hydroxylase 1 mRNA approximately twenty fold higher than B-hydroxylase 2 mRNA
(Figure 2.4). In the other tissues examined, B-hydroxylase 1 mRNA was always more
abundant, ranging from seventeen to seventy fold higher than B-hydroxylase 2 mRN A.
EF -1a transcript levels varied less than 50% between all tissues with the exception in

roots, where EF-la mRNA levels were 3.4 fold higher than in flowers.

Mapping of the LU T I locus relative to the B-hydroxylase l and 2 genes

The lut1 mutation defines a locus whose disruption reduces a ring hydroxylation
more than 90% relative to wild type (Pogson et al., 1996). To determine if either B-
hydroxylase gene showed linkage to the LUT1 locus, linkage analyses were performed
using CAPS markers (Konieczny and Ausubel, 1993). The location of both [3-
hydroxylase genes is known from the Arabidopsis genome sequencing program and the
CAPS markers nearest to each (g8300 and LF Y3 for B-hydroxylase l and 2, respectively)
were used to test for linkage in a F2 population segregating for the lutl mutation.
Recombination rates between the LUT1 locus and both B-hydroxylase linked CAPS

markers exceeded 50% indicating that neither B-hydroxylase is linked to the [ml

92

 

  
    

 
 

 
    

 

 

 

 

< E 'B-OHase1
z : DB-OHaseZ
5 h 871

‘3 100° 407

"" : 259
O)

1 F 69

N _

O- 100

2 E

z _

m 6 7
‘5 10 5 ’ '
g E 3 4 I
6 ‘ I I I

E 1

" leaf flower stem root silique
EF-1

'evef‘l 1.2 1.0 1.5 3.4 1.2 |

 

Figure 2.4. B-hydroxylase 1 and 2 mRNA levels in Arabidopsis leaf, flower, stem,
root and silique tissues. The amol amount of target RNAs per 0.2 ug total RNA for each
gene in different tissues are shown on top of the bar graphs. Relative EF-la expression

levels are indicated below the corresponding tissues.

93

mutation. Mapping of the [w] mutation was carried out with the same F2 population
using twelve pairs of CAPS markers distributed over the five Arabidopsis chromosomes.

The LUT1 locus maps at 67 cM +/- 3 cM on chromosome 3.

Discussion

In plants, carotene hydroxylases are key enzymes responsible for the formation of
xanthophyll cycle carotenoids and lutein, the most abundant carotenoid in higher plant
photosystems. Carotene hydroxylases add a hydroxyl group to the 3 position of [3- and 8-
rings. B-ring hydroxylases have been cloned and analyzed from several organisms and
shown to be highly active against B-rings and only weakly active against s-rings. All
carotene hydroxylases are nonheme diiron proteins, contain the conserved histidine
residues indicated in Figure 2.1 and, at least for plants, have been shown to require
ferredoxin oxidoreductase for their activity (Bourier et al. 1998).

Despite the fact that hydroxylated a-ring carotenoids (e. g. lutein) are the most
abundant xanthophylls in plants, a hydroxylase enzyme specific for s-rings has not yet
been cloned. It has been proposed from considerations of reaction stereochemistry that
there exists a hydroxylase enzyme that acts specifically on s-rings (Britton, 1990). In
Arabidopsis, this activity has been genetically identified by the [w] mutation, which
disrupts s-ring hydroxylation and causes accumulation of the monohydroxy pathway
intermediate, zeinoxanthin (lie-carotene, with only the B-ring hydroxylated), at the

expense of the normal dihydroxy product, lutein (Pogson et al. , 1996).

94

We propose three hypotheses for the nature of the e-ring hydroxylase activity in
plants. Given the similarity of the B and 8 ring substrates and hydroxylation enzyme
reaction mechanisms, one possibility would be that the Arabidopsis e-hydroxylase is an
enzyme related to and evolved from the B-hydroxylase. By analogy, Arabidopsis B and e
cyclases both function on lycopene, form similar ring structures (B and 8 rings) and show
58% similarity at the amino acid level. An alternative hypothesis is that though the s-
hydroxylase has a function similar to the B-hydroxylase, it evolved independently and
therefore would show little homology with the B-hydroxylase. This has precedent in that
B-hydroxylases from bacteria and plants have similar activities but little sequence
homology. A final possibility is that the “e-ring hydroxylase” defined by the lutl
mutation does not have hydroxylation activity per se but is an ancillary protein that
modifies the endogenous B-hydroxylase enzyme activity to efficiently function on 8-
rings.

In an attempt to test the first hypothesis and isolate an s-hydroxylase cDNA
related to the B-hydroxylase enzyme, we screened cDNA and genomic libraries from
Arabidopsis using the Arabidopsis B-hydroxylase cDNA as probe. A second, novel
hydroxylase cDNA class with 70% protein identity to the Arabidopsis B-hydroxylase 1
cDNA was isolated. Although phylogenetic analysis (Figure 2.5) and protein sequence
alignment (Figure 2.1) indicate that the two hydroxylases were derived from a common
ancestor, comparison of their gene structures and sequences indicate they have been
diverging for some time. Like all known carotene-hydroxylases, the new hydroxylase has

ten conserved histidine residues localized in the hydrophilic domains of the protein,

95

Figure 2.5. Neighbor-joining tree for deduced amino acid sequences of B-
hydroxylase genes from plants and bacteria. B-hydroxylase sequence from
Synechocystis PCC6803 was used as the outgroup. Branch lengths are drawn to scale.
Bootstrap values are indicated (10,000 replicates). GenBank accession numbers of the
sequences are as follows.

A.a.: Agrobacterium aurantiacum (D58420) A.l.: Alcaligenes sp. (D58422)
A.t.: Arabidopsis thaliana (B-OHase l, U58919; B-OHase 2, ABO25606)

C .a.: Capsicum annuum (B-OHase l, Y09225; B-OHase 2, Y09722)

C .u.: Citrus unshiu (AF296158) E.h.: Erwinia herbicola (852982)

E.u.: Erwinia uredovora (D90087) F.a.: Flavobacterium ATCC 21588 (U62808)
H.p.: Haematococcus pluvialis (AF 082326)

L.e: Lycopersicon esculentum (B-OHase 1, Y14809; B-OHase 2, Y14810)

N.p.: Narcissus Pseudonarcissus (A1278882) P.m.: Paracoccus marcusii (Y15112)

SS: Synechocystis sp. PCC6803 (D90906, 5111468) T.e.: Tagetes erecta (AF251018)

96

 

 

8.8. B-OHase

F.a. fl-OHase

100 M. B-OHase

P.m. B-OHase

 

 

0.2

Figure 2.5

 

92

 

 

 

10° A.a. B-OHase

, ill—4 Jiu B-Ol-lase
.h. B-Ol-lase

C.u. B-OHase

Le. B-OHase 2

C.a. B-Ol-lase1

At. B-OHase 2

M. B-OI-lase1

T.e. B-Ol-lase

Le. B-OHase1

98 (2.3. B-OI-lase 2

L—N.p. B-OHase

H.p. B-OHase

 

 

 
 
 
  
 
 
  

76

 

 

97

which are thought to coordinate the reactive iron atoms (Bouvier et al. , 1998). Although
there is an overall high identity between the new hydroxylase and the Arabidopsis [3-
hydroxylase 1 protein, significant amino acid changes do exist between the proteins
(Figure 2.1), which could be sufficient to alter substrate specificity and enzyme activity.

To determine whether the new hydroxylase was an e- or B-ring hydroxylase, its
activity, expression and chromosomal location relative to the lat] mutation were
examined. When expressed in E. coli strains producing various cyclic carotenoids, the
new hydroxylase showed high activity against B-ring substrates but very low activity
against e-ring substrates, similar to the B-hydroxylase 1 cDNA. Both hydroxylases are
expressed at low levels in root, leaf, flower, silique and stem tissues and could only be
detected by PCR based methods. B-hydroxylase 1 transcripts were more abundant than B-
hydroxylase 2 transcripts in all tissues analyzed. Finally, neither of the B-hydroxylase
genes in Arabidopsis maps to the LUT1 locus indicating neither encodes the activity
defined by the LUT1 locus.

The presence of two B-hydroxylases in a single plant system has been reported in
pepper, where both enzymes were shown to be active towards B-rings (Bouvier et al.,
1998). Pepper B-hydroxylase 1 expression was induced during the chloroplast to
chromoplast transition in fruit, however, tissue-specific expression patterns of the two
hydroxylases were not examined. Our data show that both Arabidopsis hydroxylases
were expressed in all the tissues, and B-hydroxylase 1 has a steady-state message level
ten to fifty fold higher than B-hydroxylase 2 (Figure 2.4). We are currently investigating

the expression patterns of the two hydroxylases under stress conditions.

98

We have cloned and characterized a second member of the B-hydroxylase gene
family in Arabidopsis. Several lines of evidence allow us to conclude this second gene
encodes an active B-hydroxylase enzyme and not the e-hydroxylase. Most notably, the
encoded protein is highly active against B-rings, but not e-rings, and does not map to the
LUT1 (e-hydroxylase) locus. Our data are consistent with the LUT1 locus encoding either
a third, divergent hydroxylase in Arabidopsis or an ancillary protein that modifies
existing B-hydroxylase(s) to be active toward e-rings. Our exhaustive library screening
and database analyses suggest that if LUT1 does encode an enzyme with e-hydroxylase
activity, it likely evolved independently of the B-hydroxylases as no additional B-
hydroxylase paralogs are present in the recently completed Arabidopsis genome (The
Arabidopsis Genome Initiative, 2000). Indeed, it is most likely that LUT1 encodes an
ancillary protein that modifies existing B-hydroxylase activities, as the genome sequence
encompassing the LUT1 locus does not contain proteins with the conserved histidine
motifs characteristic of all known carotene-hydroxylases. We are currently fine mapping

the [w] mutation in preparation for map-based cloning of this locus.

Acknowledgements

We would like to thank Dr. Thomas Newman (DNA sequencing facility,
Michigan State University) for assistance with TaqMan procedures. We would also like
to thank Dr. Alan Prather (Department of Botany and Plant Pathology, Michigan State

University) for suggestions on phylogenetic analysis.

99

References

Arabidopsis Genome Initiative (2000). Analysis of the genome sequence of the
flowering plant Arabidopsis thaliana. Nature 408, 796-815.

Ausubel, F.M., Brent, R., Kingston, RE, Moore, D.B., Seidman, J.G., Smith, J.A.,
and Struhl, K. (1995). Current Protocols in Molecular Biology. John Wiley, New York,
pp. 2.4.1-2.4.2.

Bouvier, F., Keller, Y., d’Harlingue, A., and Camara, B. (1998). Xanthophyll
biosynthesis: molecular and functional characterization of carotenoid hydroxylases from
pepper fruits (Capsicum annuum L.). Biochim. Biophys. Acta. 1391, 320-328.

Bramley, P.M. (1994). Carotenoid biosynthesis: a target site for bleaching herbicides.
Biochem. Soc. Trans. 22, 625-629.

Britton, G. (1990). Carotenoid biosynthesis-an overview. In: N.I., Krinsky, M.,
Mathews-Roth, and RR, Taylor (Eds) Carotenoids: Chemistry and Biology, Plenum,
New York, pp. 167-184.

Cunningham, F .X. Jr., Sun Z., Chamovitz, D., Hirschberg, J., and Gantt E. (1994).
Molecular structure and enzymatic function of lycopene cyclase from the cyanobacterium
Synechococcus sp strain PCC7942. Plant Cell 6, 1107-1121.

Cunningham, F.X.Jr., Pogson, B., Sun, 2., McDonald, K.A., DellaPenna, D., and
Gantt, E. (1996). Functional analysis of the B and e lycopene cyclase enzymes of
Arabidopsis reveals a mechanism for control of cyclic carotenoid formation. Plant Cell 8,
1613-1626.

Cunningham, F.X. J r., and Gantt E. (1998). Genes and enzymes of carotenoid
biosynthesis in plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49, 557-583.

Dellaporta, S.L., Wood, J., and Hicks, J.B.l983. A plant DNA minipreparation:
version 11. Plant Molecular Biology Reporter 1(4): 19-21.

Demmig-Adams, B., Gilmore, A.M. and Adams 111, W.W. 1996. In vivo functions of
carotenoids in higher plants. The FASEB Journal 10: 403-412.

Goodwin, T.W. 1986. Metabolism, nutrition, and function of carotenoids. Annu Rev
Nutr 6:273-297.

Heid, C.A., Stevens, J., Livak, K.J., and Williams, P.M. 1996. Real time quantitative
PCR. Genome Methods 6:986-994.

100

Hundle, B.S., O'Brien, D.A., Beyer, P., Kleinig, H. and Hearst J.E. 1993. In vitro
expression and activity of lycopene cyclase and beta-carotene hydroxylase from Erwinia
herbicola. F EBS Lett 315(3):329-34.

Kimura, M. (1980). A simple method for estimating evolutionary rates of base
substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16, 111-
120.

Konieczny, A., and Ausubel, F. M. (1993). A procedure for mapping Arabidopsis
mutations using co-dominant ecotype-specific PCR-based markers. Plant J. 4, 403-410.

Masamoto, K., Misawa, N., Kaneko, T., Kikuno, R., and Tab, H. (1998). B-carotene
hydroxylase gene from the Cyanobacterium Synechocystis sp. PCC6803. Plant Cell
Physiol. 39, 560-564.

Misawa, N., Nakagawa, M., Kobayashi, K., Yamano, S., Izawa, Y., Nakamura, K.,
and Harshima, K. (1990). Elucidation of the Erwinia uredovora carotenoid biosynthetic
pathway by functional analysis of gene products expressed in Escherichia coli. J.
Bacteriol. 172, 6704-6712.

Pogson, B., McDonald, K.A., Truong, M., Britton, G., and DellaPenna, D. (1996).
Arabidopsis carotenoid mutants demonstrate that lutein is not essential for photosynthesis
in higher plants. Plant Cell 8, 1627-1639.

Rock, C.D., and Zeevaart, J.A.D. (1991 ). The aba mutant of Arabidopsis thaliana is
impaired in epoxy-carotenoid biosynthesis. Proc. Natl. Acad. Sci. USA 88, 7496-7499.

Sambrook, J., Fritsch, E.F., and Maniatis, T. (1988). Molecular Cloning: a Laboratory
Manual. 2 nd ed. Cold Spring Harbor Laboratory Press, Plainview, New York.

Shanklin, J., Whittle, E., and Fox, B.G. (1994). Eight histidine residues are
catalytically essential in a membrane-associated iron enzyme, stearoyl-CoA desaturase,
and are conserved in alkane hydroxylase and xylene monooxygenase. Biochemistry 33,
12787-12794.

Solomons, N.W., and Bulux, J. (1993). Plant sources of provitamin A and human
nutrition. Nutr. Rev. 51, 199-204.

Sun, Z., Gantt, E., and Cunningham, F.X. (1996). Cloning and functional analysis of
the B-carotene hydroxylase of Arabidopsis thaliana. J. Biol. Chem. 271, 24349-24352.

Thompson, J.D., Higgins, D.G., and Gibson, T.J. (1994). Improving the sensitivity of

progressive multiple sequence alignment through sequence weighting, positions-specific
gap penalties and weight matrix choice. Nucl. Acids Res. 22, 4673-4680.

101

CHAPTER 3

The Arabidopsis LUT1 locus encodes a carotenoid e-hydroxylase that is a member

of the cytochrome P450 family

102

Abstract

Lutein is the most abundant carotenoid in photosynthetic tissues and its
occurrence is highly conserved in higher plants. Lutein acts as a structural component of
the light harvesting complexes and plays an essential role in photoprotection. B- and 8-
hydroxylases catalyze the formation of lutein from a-carotene (Bx-carotene). In the
EMS-derived IutI mutants, lutein accumulation is decreased ~80% and the immediate
biosynthetic precursor zeinoxanthin (B,s-carotene, with only the B-ring hydroxylated)
accumulates. These biochemical characteristics suggest that [w] is a mutation affecting s-
hydroxylase. We cloned the LUT1 gene by map-based cloning and found that it encodes a
cytochrome P450 type monooxygenase. An additional null lutI T-DNA knockout allele
(IutI-3) was identified. The phenotype of Iut1-3 confirmed that the EMS-derived 1w] -1
and [ml -2 alleles are leaky. The deduced LUT1 protein sequence has a putative
chloroplast transit peptide and a transmembrane domain, consistent with its predicted
thylakoid membrane localization. Putative orthologs of LUT1 are present in monocots,
dicots, and diatoms. LUT1 represents a novel class of carotenoid hydroxylase that has
evolved independently from and uses a different mechanism than the non-heme di-iron B-
hydroxylases. We suggest that the evolution of these different hydroxylation mechanisms
may be due to differing redox chemistry and/or substrate specificity. Characterization of
LUT1 activity will provide additional information on carotenoid hydroxylation reaction
mechanisms and possible carotenogenic complex formation on the thylakoid membrane.
Isolation of LUT1 will also facilitate metabolic engineering for lutein production in plants

and other organisms.

103

.4-

Introduction

Plant carotenoids are C40 compounds derived from eight isoprene units.
Carotenoids are essential structural and functional components of the photosynthetic
antenna and reaction center complexes and specific carotenoids serve as precursors for
the synthesis of abscisic acid. In addition to their roles in plants, some carotenoids play
an essential role in human nutrition as provitamin A (reviewed in Cunningham and Gantt,
1998). Based on their structures, carotenoids can be grouped into two classes: the
carotenes, which are linear, or cyclized hydrocarbons and the xanthophylls, which are
various oxygenated derivatives of carotenes. One xanthophyll, lutein (3R, 3'R-B,s-
carotene-3,3’-diol), is the most abundant carotenoid in all plant photosynthetic tissues,
where it plays an important role in light harvesting complex II (LHC II) assembly and
functions in photoprotection. Zeaxanthin (3 R, 3'R-B,B-carotene-3,3'-diol) is a structural
isomer of lutein and is a critical component in the mechanism of non-photochemical
quenching (reviewed in Niyogi, 1999; Hirschberg, 2001).

The committed step in lutein and zeaxanthin biosynthesis is the cyclization from
lycopene to form 01- and B- carotene, respectively (Figure 3.1). a-carotene has one s-ring
and one B-ring, while B-carotene has two B—rings. The e- and [5- rings differ from each
other by the position of double bond on each ring structure (Britton, 1998). This small
difference has profound consequences for the structural and photochemical properties of
01- and B- carotene and their xanthophyll derivatives (N iyogi, 1999). Lutein and
zeaxanthin are produced from 01- and B- carotene, respectively, by a class of enzymes

known as carotenoid hydroxylases. B-hydroxylases add hydroxyl groups to carbon 3 of

104

WWW
lycopene A/\

MW e.

 

a-carotene B-carotene
zeinoxanthin B-cryptoxanthin
lutein zeaxanthin

Figure 3.1. Carotenoid hydroxylation reactions.

105

the B-rings of B-carotene to form zeaxanthin, while hydroxylations of carbon 3 on the B-

and e-ring of a-carotene catalyzed by B- and s- hydroxylase, respectively, yield lutein

(Figure 3.1).

Based on the configuration of carbon 3 (C-3) and the requirement for molecular
oxygen, carotenoid hydroxylation reactions were predicted to be catalyzed by mixed
function oxygenase, such as cytochrome P4505 (Walton et al., 1969; Britton, 1998). B-
hydroxylases have been cloned from a variety of photosynthetic and non-photosynthetic
bacteria, fungi, and higher plants (reviewed in Cunningham and Gantt, 1998) through
functional complementation and heterologous hybridization. In all three phyla, the
deduced B-hydroxylase amino acid sequences do not encode cytochrome P450
monooxygenases but rather are non-heme di-iron proteins and have a different
hydroxylation mechanism than heme-binding cytochrome P450 enzymes (Shanklin et al.,
1994). Biochemical and mutagenesis studies with pepper (Capsicum annum) B-
hydroxylases have confirmed that the enzymes require iron, ferredoxin, and ferredoxin
oxido-reductase for activity and all of the ten conserved iron-coordinating histidines are
required for activity (Bouvier et al., 1998). Although all B-hydroxylase enzymes
efficiently hydroxylate B-rings, they work poorly on e-rings in vitro (Sun et al., 1996).

Early isotope labeling studies have shown that hydroxylation reactions are
stereospecific (Walton et al., 1969; Milborrow et a1., 1982). The chirality of C-3 carbon
in the s-ring of lutein is opposite to that in the B-ring. This difference in chirality may
explain why B-hydroxylase functions poorly on a-ring and was an initial suggestion that

two different hydroxylases are needed for the [3- and 8- ring hydroxylations. Genetic

106

studies in Arabidopsis have demonstrated the presence of an e-ring specific hydroxylase
based on mutant analysis and biosynthetic precursor accumulation (Pogson et al., 1996).
Mutation of the LUT1 locus decreases the production of lutein by 80% and results in
accumulation of the monohydroxyl immediate precursor zeinoxanthin, a classic
phenotype for a mutation affecting a biosynthetic enzyme. e-ring hydroxylation was
specifically blocked in IutI because production of B-carotene derived xanthophylls was
unaffected. It was therefore proposed that [at] is a mutation specific for s-hydroxylation
(Pogson et al., 1996).

Attempts to clone an e-ring specific hydroxylase by sequence-based homology to
B-hydroxylase in Arabidopsis were not successful, though a paralog for the B-
hydroxylase 1 gene (IS-hydroxylase 2) was identified and characterized (Tian and
DellaPenna, 2001). Like cloned B-hydroxylases from all three phyla, B-hydroxylase 2 is
highly active toward the B-ring but only weakly active toward the e-ring. In addition, a
thorough search of the fully sequenced Arabidopsis genome did not identify an additional
gene that bears significant homology to the two cloned B-hydroxylases, suggesting that
the a-hydroxylase defines a structurally distinct carotenoid hydroxylase family.
Furthermore, neither of the two Arabidopsis B-hydroxylases mapped to the LUT1 locus in
Arabidopsis (Tian and DellaPenna, 2001). We report here identification of the LUT1
locus by positional cloning and show that LUT1 defines a new class of carotenoid

hydroxylases in nature.

Results

107

Fine mapping of the LUT1 locus

We previously mapped LUT1 to the bottom arm of chromosome 3 at 67 i3 cM
(Tian and DellaPenna, 2001). For fine mapping of the LUT1 locus, 530 plants
homozygous for the Iutl mutation were identified from approximately 2,000 plants in a
segregating F2 mapping population. Using both CAPS and SSLP markers, LUT1 was
initially localized to an interval spanning two BAC clones (F8J2 and T4D2) and was
further delineated to an interval of 100 kb containing 30 predicted proteins (Figure 3.2).
As with all other carotenoid biosynthetic enzymes, the LUT1 gene product is predicted to
be targeted and localized to the chloroplast. Within the 100 kb interval containing LUT1,
six proteins were predicted as being chloroplast-targeted by the TargetP prediction
software (http://www.cbs.dtu.dk/services/TargetP). One of these chloroplast-targeted
proteins, At3g53130, was a member of the cytochrome P450 monooxygenase family
(CYP97C1). Cytochrome P4503 are heme-binding proteins that insert a single oxygen
atom into substrates, e. g. hydroxylation reactions. Therefore we considered At3g53130 to

be a strong candidate for LUT1.

Identification of L U T I

The identity of At3g53130 as LUT1 was demonstrated by complementation

analysis. Homozygous lutI mutants were transformed with a 4116 bp genomic fragment

from wild type Columbia (the background of lutl) containing the At3g53130 coding

region, 959 bp upstream of the start codon, and 690 bp downstream of the stop codon.

108

At3g§3120
At3953130

8 S 8 8
Chloroplast a § § §
Targeted ‘3 3 ‘3 «3
< < < <
BACs
Recombinants 31530 01530 01530 11530
19, 00 kb 19,8 0 kb
Chr 3 t (l

 

Figure 3.2. Positional cloning of the LUT1 gene. 530 homozygous IutI F2 plants were
used as mapping population. [at] was delineated to a region on two BAC clones F8J2 and
T4D2 within 100 kb on chromosome 3. The number of recombinants between [at] and
molecular markers are shown below the BAC clones. Predicted chloroplast targeted

proteins within this interval are shown on top of the BAC clones.

[09

Eight independent transformants were selected by Basta resistance and all showed a wild
type lutein level when analyzed by HPLC (data not shown). These data indicate that
At3g53130 genomic DNA can complement the lutl mutation.

To determine the nature of the lat] mutations, we sequenced both original EMS-
derived alleles of [at] (Pogson et al., 1996). The lut1-2 allele contains a G to A mutation
at the highly conserved exon/intron splice junction (5' AG/GT, frequency 67%,
67%/100%, 100%) that would cause an error in RNA splicing and lead to production of a
non-functional protein (Figure 3.3A). The [at] -1 allele was sequenced but no mutations
were identified in the coding region. However, a rearrangement was identified by
southern analysis in the upstream region of the lutI-l allele but was not characterized
further (data not shown). A third [at] allele, lut1-3, was identified by screening of a T-
DNA knockout population using At3g53130-specific primers. A homozygous knockout
mutant was obtained that had a T-DNA insertion in the sixth intron of the LUT1 gene
(Figure 3.3A). HPLC analysis of leaf carotenoids fiom Iut1-3 showed a complete lack of
lutein and accumulation of the monohydroxyl intermediate zeinoxanthin (Figure 3.4).
This result indicates that the previously EMS-derived lutI-I and IutI-2 alleles are indeed
leaky for s-hydroxylase function. Overall, the combined complementation of the lat]
mutation, splice site mutation of the lutI-Z allele, and phenotype of an At3g53130 T-

DNA knockout mutant confirm At3g53130 is the LUT1 locus.

LU T I encodes a chloroplast-targeted cytochrome P450 with a single transmembrane

domain

llO

Figure 3.3B shows the deduced amino acid sequence of LUT1. The chloroplast
transit peptide prediction software ChloroP v 1.1
(http://www.cbs.dtu.dk/services/ChloroP/) predicts a chloroplast transit peptide that is
cleaved between Arg-36 and Ser-37 (Figure 3.3 B). The prediction of a chloroplast
localization for LUT1 is consistent with the subcellular localization of carotenoid
biosynthesis in higher plants (Cunningham and Gantt, 1998).

The deduced amino acid sequence of LUT1 also contains several features
characteristic of cytochrome P450 enzymes. Cytochrome P450 monooxygenases have a
consensus sequence of (A/G)GX(D/E)T(T/S) that forms a binding pocket for molecular
oxygen with the invariant Thr residue playing a critical role in oxygen binding in both
prokaryotic and eukaryotic cytochrome P4505 (Chapple, 1998). In the deduced LUT1
protein sequence, this oxygen-binding pocket is highly conserved (single underline in
Figure 3.3 B). The conserved sequence around the heme-binding cysteine residue for
cytochrome P450 type enzymes is FXXGXXXCXG, which is also observed in LUT1
(double underline in Figure 3.38).

B-hydroxylases have four hydrophobic regions that are predicted to be
transmembrane helices. A model by Cunningham and Gantt (1998) suggested that the
four helices cross the thylakoid membrane and the di-iron binding sites facing the stroma.
The topology of LUT1 was predicted from its deduced amino acid sequence and a
hydropathy plot (Figure 3.3C). Unlike the B-hydroxylases, LUT1 only has a single
predicted transmembrane domain (25 aa) located in the C-terminus of the protein (Figure

3.3C).

lll

Figure 3.3. A. Identification of the [at] mutations. Overview of the intron-exon
organization of LUT1 and the mutations in Iut1-2 and 1ut1-3. B. Deduced amino acid
sequence of LUT1. The putative cleavage site for chloroplast targeting sequence is
indicated by arrow. The molecular oxygen binding pocket and the cysteine motif are
single and double underlined, respectively. C. Predicted membrane topology of LUT1.
The hydrophilicity plot is drawn with Kyte/Doolite method (window size = 19 aa). The

predicted transmembrane domain is shaded in gray.

112

AG

0

Hydropholicity

Figure 3.3

51
101
151
201
251
301
351
401
451
501

T-DNA

Iut1-3
M 1-2
G->A mutation
at the exonnntron spice site 100 bP

a:
msuspssssrssanmrannspxmrrs mssuxpxmms
smomspnnrnrarnssemzserprmmowannnoemnp
Lumen?unmepmvrvsnnmavnmrmrncnvmsn
ruescrnmrmnmmsumnsvzvnnvrcmamvnm.
gnu-eamxmrsgunovzcmmernsmmswznmzu.
mamamnmrmrmcxrwngvmnvrumv-meur
mamamourwnmrsrLarnnsannvssvonmnmsmvicfl
gesvnrmannsmssmmzmavnmmannInflux-rec
Imsmnawvunmgwn Imerm-roqomrsvmrnsesz
EMBD‘LPZRI'D Inonpunnmrn'rprsmpnxcvungrmuva
mvrmnmvnvmgu era-rear rarmcumsm

 

100 200 300 400
amino acid

113

Characterization of [at] mutant alleles

Carotenoids were extracted from four-week old wild type, lut1-2, and Iut1-3
plants, and separated on HPLC (Figure 3.4). The T-DNA derived lutI-3 mutant shows a
complete absence of lutein, consistent with a null mutation in the e-hydroxylase gene.
The EMS-derived IutI-2 allele contains 5% of wild type levels suggesting the mutation is
leaky (Table 3.1).

The leakiness of lull-2 allele was further characterized by RT-PCR reactions
using primers spanning the mutation site (Figure 3.5). lut1-2 accumulates the same size
amplification product as the wild type, suggesting that a certain amount of message is
spliced correctly and results in functional protein. Additional amplification products are
not present in the lut1-2 mutant. The leakiness of the [at] -1 allele can be explained by a
rearrangement upstream from the start codon (not shown), which possibly affects the

promoter sequence and transcriptional efficiency.

Enzymatic activities of LUT1

Two approaches have been adopted to express the LUT1 gene and analyze its
activities towards different substrates in vitro, although neither has been successful to
date. One approach is to express the LUT1 gene in a yeast expression system that
contains the Arabidopsis cytochrome P450 reductase (ATR2) (Pompon et al., 1996). An
alternative approach being performed is to express both LUT1 and Arabidopsis

cytochrome P450 reductase (Hull and Celenza, 1999) in carotenoid producing E. coli

114

 

 

 

b
V'
Zei
Zea
N u .A La)
lut1-2
b
V H B
Zei
N A A SI Zea n
a
b

 

Absorption at 440 nm

11

—
r 7 T Y r 1 Y

10 15 20 25 30 35 40 45

 

 

 

 

 

 

minutes

Figure 3.4. HPLC elution profiles of wild type, lull-2, and lut1-3 plants. N,

neoxanthin; V, violaxanthin; A, antheraxanthin; L, lutein; Zea, zeaxanthin; b, chlorophyll

b; a, chlorophyll a; Zei, zeinoxanthin; B, B-carotene.

115

 

0.06031
.0: :0: .600 An. .00... 8:00.50. 5:200 0 55.3 .050 :000 ES. 20:05.0 5:85:90 .0: 0:0 2080.05 0600 05 5.5, 09:05 00:.0> 000055.00 :. c020 20:20:00
@0030 00050000 .30.: 02.0.0. 05 FEB ..o.m« 3:05.096 .00 .0 :006 05 m. 0:.0> 50m 0+0 20 6655690 .05.: m0 00000.90 0. 30:06:00 .0 .::0E0 0..»

 

 

. . . . . . . . . . . . as :00 . . -
0 :3 an :0. m 90 m: .0. v 3. m a: v «A mm .00. o as 00 a. 0 3 on 5.00.50 08.0.00 0 c m 2:.
0,000.05 .03 0.0 $2.: .0. 0000.0 .0. 3030 .00. 0.00.8. .3. 0.03.3 as 0.008 . .0. 0.00.8 .8 o. .03 «is.
. . . . . . . . . . . . :0. . . an.
N 20 in .00. 0 ~00 E. u c u c .0: 0 30 B .0: m 0.0 «0 3.00.00 0 : «0.00.02 .00
20:8... .5. 330.00 .0... .0: 8: 0.03.00 .0. c.0030 2%.; .0... .00. 0.03.8. 0.5
36:06:00 .08... 0:385:01. :.5:0x00N 55:90:05.2 55:90.05 :.5:0xooz 0:20.00... :_5:0x0:_0N £05..

 

.040... .3 005E000 0:00.. .00. :. 85.000500 20:06:00 505:: 0:0 00.: 25> fin 0.00..

 

116

 

Figure 3.5. RT-PCR reactions using LUT1 specific primers. Total RNA templates
used in lane 1,2 are wild type, lut1-2, respectively. Lane 3 is a blank control for the PCR

reactions.

117

strains. Modifications are currently being made in order to obtain e-hydroxylase activities
in vitro.

In order to test the activity of LUT1 (e-ring hydroxylase) towards B- and 8- rings,
the full length LUT1 cDNA was subcloned into the yeast pYeDP60 vector (Pompon et al.
1996) and expressed in the yeast WATll strain that has been engineered to express an
Arabidopsis cytochrome P450 reductase (Urban et al. 1997). Empty pYeDP6O vector
transformed WATll cells were used as negative controls throughout the expression
studies. The microsomal fraction was prepared from WATll cells expressing either the
pYeDP60-L UT 1 construct or the empty pYeDP6O vector. Carbon monoxide (CO)
difference spectrum was measured for both microsomal preparations. We were unable to
obtain CO spectra for LUT1. In retrospect, this is likely due to the absence of an ER
targeting sequence on LUT1. The mature protein is not targeted to the membranes
therefore it is not present in the microsomal fraction prepared. Currently, experiments are
carried out to engineer an ER targeting sequence onto the LUT1 coding sequence.

A second approach being taken is to engineer both full length and truncated LUT1
(without the chloroplast targeting sequence) into an Arabidopsis cytochrome P450
reductase containing plasmid for expression in E. coli (Hull and Celenza, 1999). These
constructs are being transformed into E. coli strains engineered to accumulate carotenoids
with different ring structures (Cunningham et al., 1994). Hydroxylation of carotene
substrates has not yet been observed in these E. coli strains. It is possible that LUT1 also
requires a specific lipid (e. g. MGDG) for activity analogous to that reported for the
violaxanthin deepoxidase enzyme (Rockholm and Yamamoto, 1996). We are currently

incorporating various lipids into the reaction mixtures.

118

Putative LU T I orthologs

LUT1 was designated as CYP97C] according to the cytochrome P450
nomenclature (http://www.biobase.dk/P450). By definition, cytochrome P4505 are
grouped into a family if their amino acid sequences are >40% identical and are filrther
grouped into subfamilies if the identity is >55%. There are two additional Arabidopsis
proteins that belong to the CYP97 family: CYP97A3 (At1g31800) and CYP97B3
(At4g15110) with 49% and 42% identity to LUT1, respectively. CYP97A3 is predicted to
be chloroplast targeted while CYP97B3 is apparently cytosolic.

Various EST and genomic databases were searched using the deduced amino acid
sequence of LUT1 as a query. There are two other dicot proteins that belong to different
CYP97 subfamiles: CYP97Bl from pea (Pisum Sativum) and CYP97B2 from soybean
(Glycine max) are 44% and 42% identical to At3g53130, respectively, suggesting that
they are possible LUTlorthologs. We were also able to identify an ortholog for LUT1
(AAK20054) in the monocot plant Oryza sativa. This rice protein is 70% identical to
LUT1. In addition, there is also a CYP97 family member (3 7% identical to LUT1;
AAL73435) in diatom (Skeletonema costatum, CYP97E1), the function of which is

unknown.

Discussion

Lutein is the most abundant carotenoid in photosynthetic tissues of plants and

119

plays essential roles in light harvesting and photoprotection (N iyogi, 1999). B- and e-
hydroxylases add hydroxyl groups to the B- and a— ring of a-carotene to form lutein. B-
hydroxylases have been extensively studied and characterized in the past several years. In
contrast, the e-hydroxylase is only genetically defined by the IutI mutation in
Arabidopsis (Cunningham and Gantt, 1998). Several lines of evidence have confirmed
that At3g53130 is the LUT1 gene and encodes the carotenoid e-ring hydroxylase that is a
member of the cytochrome P450 family. Cytochrome P450 heme-binding
monooxygenases carry out a mechanistically distinct type of hydroxylation reaction from
non-heme di-iron B-hydroxylases, which suggests that LUT1 is a novel hydroxylase that

has evolved independently from B-hydroxylases.

EMS-derived lull mutants are leaky

lutI-I and lut-2 were previously isolated from a screen of EMS-generated mutants
(Pogson et al., 1996). Both alleles still accumulated lutein, suggesting that these EMS-
derived mutants might be leaky. Cloning of the LUT1 gene enabled the isolation of a
knockout mutant (1w! -3) using a reverse genetics approach. The IutI-3 mutant
completely lacks lutein (Figure 3.4), confirming that the EMS-derived 1w] -1 and lut1-2
alleles are leaky and that LUT1 is the only activity that carries out e-ring hydroxylation in
viva. Sequencing analysis has shown that IutI-2 is a mutation in an intron-exon splicing
site (Figure 3.3A). This intron is 98 bp long and would shift the opening reading frame
and cause a premature stop codon if not correctly spliced. RT-PCR reactions using

primers spanning the lutl-Z mutation site showed that LUT1 transcripts still accumulate

120

in the lut1-2 mutant (Figure 3.5). This result indicates that some of the LUT1 mRNA is
correctly spliced and functional protein is synthesized, which explains the leakiness of
LUT1 function in the lutI-2 mutant. An "unspliced mRNA" (with the fourth intron) was
not observed in the lut1-2 amplification products. A possible splicing site is identified in
the fourth intron, which may cause an alternative splicing and result in a product with
similar size to LUT1. Southern analysis has shown that there is a rearrangement upstream
of the LUT1 gene in the [at] -1 mutant (data not shown), which may affect the efficiency

of the promoter and result in reduced LUT1 mRNA and activity in IutI-I .

Different hydroxylation mechanisms for B- and e-ring hydroxylase

Carotenoid B- and e-hydroxylase add hydroxyl groups to the B- and 8- rings,
respectively. B- and e-ring are very similar and differ from each other only in the
placement of a double bond on the ring structure (Figure 3.6A). While both enzymes use
nearly identical substrates, why are two different types of monooxygenases required for
the hydroxylation reactions?

Both B- and 3- ring hydroxylations require the abstraction of a hydrogen atom
from the C-3 position of the ring structure (Figure 3.6A). However, it is relatively easier
to withdraw a hydrogen atom from the C-3 of an e-ring than from a B-ring (dissociation
energy 86 kcal/mol vs. 100 kcal/mol; Berkowitz et al., 1994), because the C-3 on e-ring
is an allylic carbon while the C-3 carbon on B-ring is not (Figure 3.6A). Therefore,
cleavage of C-H bond at the C-3 position of an e-ring requires less energy due to the

resonance stabilization of the allylic radical product. The B-ring requires a stronger

121

Figure 3.6. A. The hydroxylation reactions of B- and 8- ring. R, polyene chain. B. 3-D

structures of a- and B- carotene substrates for hydroxylation reactions.

122

 

 

 

 

 

 

 

B-ca rote ne

 

cat-carotene

Figure 3.6

123

oxidant for hydrogen abstraction.

This raises the possibility that non-heme di-iron B-hydroxylases are stronger
oxidants than the cytochrome P450 enzymes. The currently accepted mechanism for
hydrocarbon hydroxylation involves two reactions: first, abstraction of a hydrogen atom
from the substrate by an oxoiron species (F e'V=O), and secondly, a rapid collapse of the
resulting carbon radical-hydroxyferryl complex to give the hydroxylated product and the
resting-state enzyme (Groves et al., 1978). Both cytochrome P450 and di-iron
monooxygenases hydroxylate substrates by a hydrogen atom abstraction/oxygen rebound
mechanism with a short-lived iron-oxo intermediate. The key cytochrome P450
intermediate is an Few=O porphyrin n-radical cation (Ortiz de Montellano, 1995), while
the key di-iron monooxygenase intermediate is a di-Fe'v unit with bridging oxos
(Shanklin and Cahoon, 1998) (Figure 3.7). Experimental evidence have shown that both
enzymes are able to add oxygen to inactivated C-H bonds (Sono et al., 1996; Yoshizawa,
2000) and it is not clear which enzyme produces a stronger oxidant in the "activated"
intermediate. Therefore, the requirement of two fimdamentally different types of
hydroxylases for the two rings can not be simply explained by their oxidation capacities.

Another key factor that determines enzyme catalysis is access to substrates.
Neither the B- nor the e- hydroxylase has been crystallized, therefore the precise
conformations of their substrate binding sites are not known. However, when comparing
the carotene substrate structures, the B-ring still has a double bond conjugated to the
hydrocarbon backbone, thus, the B-ring is restrained to the same plane as the hydrocarbon
backbone. On the other hand, the double bond in the a-ring is not conjugated to the

backbone and has relatively free rotation along the C6'-C7' carbon (Figure 3.63). One

124

 

 

 

Home binding cytoch romo P450
monooxygenase

Fe'( F e”
\

Non-homo (ii-iron
monooxygenase

Figure 3.7. Oxidizing intermediates for cytochrome P450 and non-heme di-iron

monooxygenases.

125

explanation for the hydroxylase substrate specificity is that the substrate binding pocket
of B-hydroxylase can only accommodate a straight chain hydrocarbon, such as the [3-ring
structure; an e-ring that is tilted from the hydrocarbon backbone can not fit in the B-
hydroxylase binding site. It is vise versa for the e-hydroxylase. Another explanation for
the substrate specificity is that both hydroxylases can bind to the B- and a— ring substrates
and abstract a hydrogen atom from C-3 carbon, but the hydroxylation reaction is
stereospecific, i.e., the ring substrate has to be at the "correct" conformation for the
oxidation reaction to occur. In addition, [3- and a-hydroxylases may differ in their
membrane topologies (Cunningham and Gantt, 1998; Figure 3.3C), which may contribute

to their substrate specificity.

Evolution of B- and s-hydroxylases

Over 700 carotenoids have been identified in nature and constitute a structurally
diverse group of compounds. The polyene chain in carotenoids has been conserved
through evolution while the end groups have not and show a wide variety of
modifications. Specific carotenoid end groups are necessary for their recognition and/or
binding to proteins and orientation in the thylakoid membrane, which further determine
the biological diversities of carotenoid functions in different organisms.

Carotenoids with [3,13- and [3,a-end groups are prevalent in plants and perform
critical roles in light harvesting and photoprotection. One of the most common
modifications on [3,B- and [Le-end groups are hydroxylation reactions to form various

xanthophylls. Carotenoid B-hydroxylases have been cloned from various photosynthetic

126

and non-photosynthetic bacteria, green algae, and higher plants. They have invariably
been shown to be di-iron enzymes. From phylogenetic studies based on the deduced
amino acid sequences, B-hydroxylases can be clearly grouped into three clads: plant and
nonphotosynthetic bacterial B-hydroxylases are related but share less than 30% protein
identity, while the cyanobacterial B-hydroxylases are more closely related to the bacterial
B-ketolases (Tian and DellaPenna, 2001).

The present data have shown that plant e-hydroxylases evolved independently of
B-hydroxylases. Putative e-hydroxylase orthologs have been identified from both
monocot (rice) and dicots (soybean and pea) plants via database search. There is also a
cytochrome P450 protein in diatom (Skeletonema costatum, CYP97E1) that belongs to
the same family as LUT1 and is possibly related to s-hydroxylase. The question remains:
what is the evolutionary origin of the e-hydroxylase in plants?

The substrate for the e-hydroxylase is a-carotene (ifs-carotene) and e-
hydroxylase may have evolved in parallel to a-carotene synthesis. a—carotene is unique to
land plants and green algae that gave rise to land plants. In prokaryotes, a-carotene has
only been found in Prochlorococcus and Acaryochloris (Hess et al., 2001; Miyashita et
al., 1997). The development of photosystem II (PSII) marked an important evolution of
carotenoid biosynthesis. Chlorophyll b (Chl b) is mainly associated with PSII.
Prochlorococcus is a Chl b-containing cyanobacteria that lives in the ocean. Two
Prochlorococcus strains, MED4 (low Chl b/a ratio) and MIT9313 (high Chl b/a ratio) are
representatives of hi gh- and low- light adapted ecotypes. This genus is distinct from other
cyanobacteria by the presence of high concentration of a-carotene. Both MED4 and

MIT9313 genomes contain two genes encoding for putative lycopene B- and e- cyclases.

127

B-hydroxylase orthologs have also been identified from these two prochlorococcus.
However, no lutein was detected in the pigment profile and no potential e-hydroxylase

identified in database searches.

Future perspectives

Current models (Cunningham and Gantt, 1998) suggests that two types of
desaturase/cyclase complexes exist in the thylakoid membrane, one with two B-cyclases
and the other with one B- one e-cyclase. Following this model, carotenoid hydroxylases
in higher plants may also form complexes. B-hydroxylase 1 and/or 2 may form one
complex and work specifically on [LB-rings, while e-hydroxylase may form a complex
with either B-hydroxylase 1 or 2 and preferably utilizes [Le-ring carotenoid as substrate.
Cloning of the e-hydroxylase now allows for the possibility of generating an e-
hydroxylase antibody and elucidation of possible hydroxylase complexes.

In higher plants, studies using the lutein deficient Arabidopsis [w] and lut2
mutants showed that in addition to light induced trans-thylakoid pH change and
zeaxanthin accumulation, lutein is also required for efficient nonphotochemical
quenching (Lokstein et al., 2002). Lutein is essential for optimizing photosynthetic
antenna structure, stability, therefore efficiently using harvested light under normal and
high light growth conditions, respectively. Cloning of the e-hydroxylase gene will allow
one to manipulate lutein production in vivo. Lutein is also known for protecting against
damage from reactive oxygen species and minimizing the occurrence and/or severity of

certain chronic diseases in human. Elevated lutein production will increase the nutritional

128

values of plant food and promote human health.

Methods

Positional cloning of LU T 1

lut1-2 (ecotype Columbia) was crossed to Lansberg erecta. The resulting F1
plants were self-pollinated and homozygous lutI mutants were identified in a segregating
F2 mapping population by a Thin Layer Chromatography (TLC) screening method.
Carotenoid samples were extracted as described (Tian and DellaPenna, 2001),
resuspended in ethyl acetate, spotted on a C13 TLC plate (J .T. Baker, Phillipsburg, New
Jersey), and developed in 90:10 (vzv) hexane: isopropanol. Homozygous IutI F2 plants
were identified by the characteristic appearance of an extra yellow band due to their
accumulation of zeinoxanthin that is absent in wild type plants.

[ml was initially mapped to 67 i 3 cM on chromosome 3 (Tian and DellaPenna,
2001). The lut] interval on chromosome 3 was searched for CAPS and SSLP markers
using the TAIR search engine
http://www.arabidopsis.org/servlets/Search?action=new_search&type=marker
In order to achieve additional molecular markers for fine-mapping, SSLP markers were
designed based on the INDEL information obtained from the Cereon website:
http://godot.ncgr.org/Cereon/. Linkage analysis was performed between [w] and the

molecular markers to delineate 1m] to a narrow region on chromosome 3.

129

DNA extraction and PCR reactions

A diameter of 0.5 mm leaf area of three weeks old Arabidopsis plant was excised
and pulverized in a 1.5 ml microcentrifuge tube with blue pestle (N alge Nunc
International). 150 pl DNAzol reagent (Invitrogen, Carlsbad, CA) was added and more
grinding was carried out until no green leaf tissue was visible in the solution. The tube is
incubated at room temperature with shaking for 5 min. 150 pl chloroform was added to
the tube and mixed well. The mixture was further incubated at room temperature for
another 5 min. The extracts were then centrifuged at 13,000g for 10 min. The upper phase
was transferred to a flesh microcentrifuge tube and the DNA was precipitated with 75 pl
100% ethanol at room temperature for 5 min. The DNA was centrifuged at 10,000g for 5
min. The supernatant was discarded and the pellet was washed with 300 pl 75% ethanol.
Following ethanol wash, the ethanol-DNA mix was centrifuged at 10,000g for 5 min. The
supernatant was discarded and the pellet was dried briefly under vacuum. Finally, the
DNA was resuspended in 100 pl 8 mM NaOH. The pH of the DNA solution was adjusted
to 7.5 by adding 15.9 pl of 0.1 M HEPES.

PCR reactions were performed with 1 pl of genomic DNA in a 20 pl reaction
mixture. The PCR program is 94°C for 3 min, 60 cycles of 94°C for 15 s, 50°C-60°C (the
annealing temperature is optimized for each specific pairs of primers) for 30 s, 72°C for
30 s, and finally 72°C for 10 min. A portion of the PCR product (6 pl) was then separated

on 3% agarose gel.

Complementation of [at] with wild type gene

130

An Arabidopsis cosmid library (Meyer et al., 1996) was screened and several
cosmids that span both T4D2 and F812 BAC clones were identified. At3g53130 genomic
DNA was obtained by digesting the cosmid DNA with . The resulting fragment was first
subcloned into Hind III and BamH I digested pART7 vector. The pART7-At3g53130
construct was digested with NotI and subcloned into pMLBART vector that has Basta as
a selection marker (Gleave, 1992). The identity of the construct was confirmed by both
digestion patterns and amplification with At3g53 130 specific primers. Agrobacterium
strain GV3101 was transformed with the construct using freeze-thaw method (Chen et al.,
1994). Agrobacterium DNA was extracted and its identity was confirmed by restriction
enzyme digestion followed by gel—electrophoresis. Homozygous IutI plants were

transformed with agrobacterium by Floral Dip method (Clough and Bent, 1998).

Knockout mutant isolation

At3g53130 specific primers (forward, 5'-
CTTCCTCTTCTTACTCTTCTCTCTTCACT—B'; reverse, 5'-
AAGAACGATGGATGTTATAGACTGAAATC-3') were sent to the University of
Wisconsin Arabidopsis T-DNA knockout facility for screening knockout mutants of the
LUT1 gene. A single knockout line was identified and isolated as described
(http://www.biotech.wisc.edu/Arabidopsis/). A homozygous mutant line was identified
by PCR and designated lut1-3 (Figure 3.3). A second knockout mutant (Salk_042859)

was identified through the Salk T-DNA express website (http://signal.salk.edu) and

131

designated as lut1-4. lut1-4 has a T-DNA insert 117 bp upstream of the LUT1 start codon.

HPLC and RT-PCR analyses of the lat] mutants

Wild type Columbia and WS plants as well as lut1-2 (Columbia background) and
lutI-3 (WS background) mutants were grown under 120 pmol photons m'zs'l for four
weeks. The same leaf area from each plant was excised by a cork borer. Total carotenoids
were extracted and quantified as previously described (Tian and DellaPenna, 2001). Total
RNA was extracted from Columbia and lut1-2 plants and RT-PCR reactions were
performed (Tian and DellaPenna, 2001). Primers for the PCR reactions are forward, 5'-
TTGTGTAAGATAGTCCCGAGACAGG-3' and reverse, 5'-

GGATGAGGATAGAGACGCATTGAC-3'.

Construct for yeast expression

The cDNA of At3g53130 was reverse transcribed from wild type Columbia total
RNA and gene specific primers used for PCR reactions are 5'-
AAGGATCCATGGAGTCTTCACTC'ITITCTC-3' (start codon in bold, BamHI site is
underlined) and 5'-AAGAATTCTTACCTTTGGCTCACCTTCATATAC-B' (stop codon
in bold, EcoRI site is underlined). The PCR reaction was performed using Pfu
polymerase and the PCR program was 94°C for 3 min, then 35 cycles of 94°C 30 s, 60°C
30 s, and 72°C 2 min, the PCR reaction was continued with 72°C 10 min and holded at

4°C. Yeast expression vector pYeDP60 was kindly provided by Drs. Pompon and Urban

132

(Pompon et al. 1996). Full length At3g53 130 cDNA was digested with both BamHI and

EcoRI and ligated into BamHI and EcoRI digested pYeDP6O vector.

Yeast transformation and microsomal preparation

pYeDP60-At3g53130 plasmid was transformed into yeast WATll strain as
previously described (Schiestl and Gietz, 1989). For expression studies, one colony was
picked from the pYer60-At3g53130 plate and added to 50 ml SGI media and grown at
30°C for about 24 hrs with shaking. Microsomal fraction was subsequently prepared

following procedures in Ralston et al. (2001).

Construct for E.coli expression

An N-terminal modified form (without chloroplast targeting sequence) of
Arabidopsis cytochrome P450 reductase (ATR2) open reading frame was subcloned into
the pCWori+ vector (Hull and Celenza, 1999). This plasmid was expressed in E. coli and
resulted in high yield of a functional reductase, which supports CYP79B2 activity in vitro
(Hull and Celenza, 1999). At3g53130 full length cDNA was amplified from Columbia
total RNA using Pfu polymerase with forward (5'-
AAGGATCCATGGAGTCTTCACTCTT'ITCTC-3'; BamH I site underlined and start
codon in bold) and reverse (5'-AAGAATTCTTACCT’ITGGCTCACCTTCATA-3';
EcoR I site underlined and stop codon in bold) primers. At3g53130 without chloroplast

targeting sequence was amplified with forward (5'-

133

AAGGATCCATGTCCTCCATTGAGAAACCCA-3'; BamH I site underlined and start
codon in bold) and reverse (5'-AAGAATTCTTACCTTTGGCTCACCTTCATA-3';
EcoR I site underlined and stop codon in bold) primers. The amplification products were
digested with BamH I and EcoR I and ligated into pre-digested (BamH I and EcoR I)
pBluescript SK+ vector. The resulting plasmids were digested with Pvu II and the DNA
inserts were ligated into blunt-ended pCWori+-ATR2 plasmids. Plasmids harboring both
At3 g53130 and ATR2 were transformed into carotenoid accumulating E. coli strains,
pAC-DELTA (producing carotene with a single e-ring), pAC-BETAO4 (producing [3,8-
carotene), and pAC-EPS04 Qaroducing ate-carotene) (Cunningham et al., 1994). Growth
of E. coli and carotenoid extractions were performed as described (Cunningham et al.,

1994).

Acknowledgements

We thank Drs. Dennis Pompon and Philippe Urban for providing the yeast strain
WATll and pYeDP6O vector. We also thank Dr. Clint Chapple for helpful discussion on
cytochrome P450 assays. The University of Wisconsin Arabidopsis T-DNA knockout
facility provided mutant screening service and the Salk Institute Genomic Analysis

Laboratory provided sequence-indexed Arabidopsis T-DNA insertion mutants.

134

References

Berkowitz, J., Ellison, G.B., and Gutman, D. (1994). Three methods to measure RH
bond energies. J. Phys. Chem. 98, 2744-2765.

Bouvier, F., Keller, Y., d’Harlingue, A., and Camara, B. (1998). Xanthophyll
biosynthesis: molecular and functional characterization of carotenoid hydroxylases from
pepper fruits (Capsicum annuum L.). Biochim. Biophys. Acta. 1391, 320-328.

Britton, G. (1998). Overview of carotenoid biosynthesis. In Carotenoids: Biosynthesis
and Metabolism, Vol. 3, G. Britton, S. Liaaen-Jensen, and H. Pfander, eds (Basel,
Switzerlalnd: Birkhauser), pp. 1 3- 147.

Chapple, C. (1998). Molecular-genetic analysis of plant cytochrome P450-dependent
monooxygenase. Ann. Rev. Plant Physiol. Plant Mol. Biol. 49, 311-343.

Chen, H., Nelson, R.S., and Sherwood, J.L. (1994). Enhanced recovery of
transformants of A grobacterium tumefaciens after freeze-thaw transformation and drug
selection. Biotechniques 16, 664-668, 670.

Clough, S.J., and Bent, A.F. (1998). Floral Dip: a simplified method for Agrobacterium-
mediated transformation of Arabidopsis thaliana. Plant J. 16, 735-743.

Cunningham, F.X. Jr., Sun, Z, Chamovitz, D., Hirschberg, J., and Gantt E. (1994).
Molecular structure and enzymatic function of lycopene cyclase from the
Cyanobacterium Synechococcus sp strain PCC7942. Plant Cell 6, 1107-1121.

Cunningham, F.X. Jr., and Gantt E. (1998). Genes and enzymes of carotenoid
biosynthesis in plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49, 557-583.

Cleave, A.P. (1992). A versatile binary vector system with a T-DNA organisational
structure conductive to efficient integration of cloned DNA into the plant genome. Plant
Mol. Biol. 20, 1203-1207.

Groves, J.T., McClusky, G.A., White, R.E., Coon, M. (1978). Aliphatic hydroxylation
by highly purified liver microsomal cytochrome P450 - evidence for a carbon radical
intermediate. Biochem. Biophys. Res. Commun. 81, 154-160.

Hess, W.R., Rocap, G., Ting, C.S., Larimer, F., Stilwagen, S., Lamerdin, J., and
Chisholm, S.W. (2001). The photosynthetic apparatus of Prochlorococcus: Insights
through comparative genomics. Photosynth. Res. 70, 53-71.

Hirschberg, J. (2001). Carotenoid biosynthesis in flowering plants. Curr. Opin. Plant
Biol. 4, 210-218.

135

Hull, A.K., and Celenza, J.L. (1999). Bacterial expression and purification of the
Arabidopsis NADPH-Cytochrome P450 reductase ATR2. Protein Expr. Purif. 18, 310-
315.

Lokstein, H., Tian, L., Polle, J.E.W., DellaPenna, D. (2002). Xanthophyll biosynthetic
mutants of A rabidopsis thaliana: altered nonphotochemical quenching of chlorophyll
fluorescence is due to changes in photosystem II antenna size and stability. Biochim.
Biophys. Acta. 1553, 309-319.

Meyer, K., Benning, G., and Grill, E. (1996). Cloning of plant genes based on genetic
map location. In: Genome mapping in plants. Peterson A.G. eds. (Academic Press, New
York; Landes Bioscience Publishers, Austin, TX). pp137-154.

Milborrow, B.V., Swift, LE, and Netting, A.G. (1982). The stereochemistry of
hydroxylation of the carotenoid lutein in calendula-officinalis. Phytochem. 21, 2853-
2857.

Miyashita, H., Adachi, K., Kurano, N., Ikemoto, H., Chihara, M., and Miyachi, S.
(1997). Pigment composition of a novel oxygenic photosynthetic prokaryote containing
chlorophyll d as the major chlorophyll. Plant Cell Physiol. 38, 274-281.

Niyogi, K.K. (1999). Photoprotection revisited: genetic and molecular approaches. Annu.
Rev. Plant Physiol. Plant Mol. Biol. 50, 333-359.

Ortiz de Montellano, P.R. (1995). Structure, mechanism, and inhibition of cytochrome
P450. Drug Metab. Dispos. 23, 1181-1187.

Pogson, B., McDonald, K.A., Truong, M., Britton, G., and DellaPenna, D. (1996).
Arabidopsis carotenoid mutants demonstrate that lutein is not essential for photosynthesis
in higher plants. Plant Cell 8, 1627-1639.

Pompon, D., Louerat, B., Bronine, A., and Urban, P. (1996). Yeast expression of
animal and plant P4503 in optimized redox environments. Methods Enzymol. 272, 51-64.

Ralston, L., Kwon, S.T., Schoenbeck, M., Ralston, J., Schenk, D.J. , Coates, R.M.,
and Chappell, J. (2001). Cloning, heterologous expression, and functional
characterization of 5-epi-aristolochene-l,3-dihydroxylase from tobacco (Nicotiana
tabacum). Arch. Biochem. Biophys. 393, 222-235.

Rockholm, D.C., and Yamamoto, H.Y. (1996). Violaxanthin de-epoxidase. Plant
Physiol. 110, 697-703.

Schiestl, R.H., and Gietz, RD. (1989). High efficiency transformation of intact yeast
cells using single stranded nucleic acids as a carrier. Curr. Genet. 16, 339-346.

136

Shanklin, J., Whittle, E., and Fox, B.G., (1994). Eight histidine residues are
catalytically essential in a membrane-associated iron enzyme, stearoyl-CoA desaturase,
and are conserved in alkane hydroxylase and xylene monooxygenase. Biochemistry 33,
12787-12794.

Shanklin, J., and Cahoon, E.E. (1998). Desaturation and related modifications of fatty
acids. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49, 611-641.

Sono, M., Roach, M.P., Coulter, E.D., and Dawson, J.H. (1996). Heme-containing
oxygenases. Chem. Rev. 96, 2841-2887.

Sun, Z., Gantt, E., and Cunningham, F.X. (1996). Cloning and functional analysis of
the B-carotene hydroxylase of Arabidopsis thaliana. J. Biol. Chem. 271, 24349-24352.

Tian, L., and DellaPenna, D. (2001). Cloning and characterization of a second
carotenoid B-hydroxylase gene from Arabidopsis. Plant Mol. Biol. 47, 379-388.

Urban, P., Mignotte, C., Kaznauer, M., Delmorme, F., and Pompon, D. (1997).
Cloning, yeast expression, and characterization of the coupling of two distantly related
Arabidopsis thaliana NADPH-cytochrome P450 reductases with P450 CYP73A5. J. Biol.
Chem. 272, 19176-19186.

Walton, T.J., Britton, G., and Goodwin, T.W. (1969). Biosynthesis of xanthophylls in
higher plants: stereochemistry of hydroxylation at C-3. Biochem. J. 112, 383-385.

Yoshizawa, K. (2000). Two-step concerted mechanism for methane hydroxylation on the
diiron active site of soluble methane monooxygenase. J. Inorg. Chem. 78, 23-34.

137

CHAPTER 4

Functional analysis of B- and 8- ring carotenoid hydroxylases in Arabidopsis thaliana

The work presented in this chapter has been published:
Tian, L., Magallanes-Lundback, M., Musetti, V., and DellaPenna, D. (2003). Plant

Cell, in press.

138

Abstract

Lutein and zeaxanthin are dihydroxy xanthophylls that perform critical roles in
photosystem structure and function and are produced from their corresponding carotene
precursors by the action of B- and e-ring carotenoid hydroxylases. Two genes encoding
B-ring hydroxylases (Ii-hydroxylase l and 2) have been identified in the Arabidopsis
genome and are highly active towards B-rings but only weakly active toward e-rings. A
third distinct activity required for s-ring hydroxylation has been defined by mutation of
the LU T1 locus, but not yet cloned. In order to address the individual and overlapping
functions of the three Arabidopsis carotenoid hydroxylase activities in vivo, T-DNA
knockout mutants corresponding to B-hydroxylase l and 2 (b1 and b2, respectively) were
isolated and all possible hydroxylase mutant combinations were generated. B-hydroxylase
single mutants do not exhibit obvious growth defects and have limited impact on
carotenoid compositions relative to wild type, suggesting the encoded proteins have a
significant degree of functional redundancy in vivo. Surprisingly, the b1 b2 double
mutant, which lacks both known B-hydroxylase enzymes, still contains significant levels
of B-carotene-derived xanthophylls suggesting additional B-ring hydroxylation activity
exists in vivo. The phenotype of double and triple hydroxylase mutants indicates at least a
portion of this activity resides in the LU T1 gene product. The single, double, and triple
hydroxylase mutants do not alter the total amount of carotenoids accumulated in leaf
tissue but cause progressively more severe alterations in carotenoid composition, NPQ
induction, and NPQ amplitude. Despite the severe reduction of B-carotene-derived

xanthophylls (up to 90% in the [w] b1b2 triple mutant), the double and triple hydroxylase

139

mutants still contain at least 50% of the wild type amount of hydroxylated B-rings. This
suggests it is the presence of minimal amounts of hydroxylated B-rings, rather than
minimal amounts of specific [i-carotene-derived xanthophylls, that are essential for LHC
II assembly and function in vivo. The carotenoid profiles in wild type seeds, and the
effect of single and multiple hydroxylase mutations are distinct from that in
photosynthetic tissues indicating the activities of each gene product differs in the two
tissues. Overall, the hydroxylase mutants provide insight into the unexpected overlapping

activity of carotenoid hydroxylases in vivo.

140

Introduction

Xanthophylls are oxygenated carotenoids that perform a variety of critical roles in
photosystem structure and assembly, light harvesting, and photoprotection. The
xanthophyll content of photosynthetic plant tissues is highly conserved through evolution
with lutein (L) being the most abundant followed by B-carotene, neoxanthin (N), and
violaxanthin (V). Zeaxanthin (Z) and antheraxanthin (A) accumulate to high levels in
response to high light stress. L is a critical structural component of the Light Harvesting
Complex II (LHC II) trimers, while Z, a structural isomer of L, is best known for its role
in non-photochemical quenching (N PQ). NPQ is a measurement of the dissipation of
excess light energy absorbed by the photosystems and is one of the major mechanisms
that protect plants from photooxidative damage. Energy dependent NPQ requires a pH
change in the thylakoid lumen, the PsbS protein, and accumulation of Z (and A) that
results from deepoxidation of V (Li et al., 2000; Muller et al., 2001). L has also been
found to be necessary for efficient NPQ, although its role is thought to be more indirect
than Z and A (Pogson et al., 1998; Lokstein et al., 2002).

L and Z are dihydroxy xanthophylls that are derived from (Jr-carotene ([3,8-

carotene) and B-carotene (INS-carotene), respectively, by the addition of hydroxyl groups
to the 3, 3' position of both rings (Figure 4.1). L and Z have identical chemical formulas
and similar structures but small differences in their ring structures, conjugated double
bond systems, and hydroxylation stereochemistry allow for distinct and specialized roles
in photosystem structure, light harvesting, and photoprotection (Demmig-Adams et al.,

1996; Horton et al., 1996; Niyogi et al., 1998). The enzymes that mediate carotenoid ring

141

WWW

lycopene

3“ 9
(
b1,b2 + cat-carotene B-carotone + b1,b2
D ’
H b1 2
Iut1 T zeinoxanthin B-cryptoxanthin ,b OH
E.» | | :Q’” W
¢
H
OH

zeaxanthin

lutein
:o
antheraxanthin l T
OH
'0

W

violaxanthin l
.65: “
OH

neoxanthin

OH

Figure 4.1. Xanthophyll biosynthesis in Arabidopsis. The positions of the 1m] , b1

(CrtR-bl), and b2 (CrtR-b2) mutations in the pathway are indicated.

142

hydroxylation reactions are key for the biosynthesis of these two functionally important
xanthophylls (Figure 4.1). Z is widespread in bacteria, fungi, and plants and B-
hydroxylases involved in Z synthesis have been cloned and characterized from all three
phyla. All are non-heme di-iron oxidases that contain conserved histidine motifs required
for activity (Bouvier et al., 1998). However, enzymes from the three phyla have
otherwise low protein identity, are thought to have evolved independently, but efficiently
catalyze hydroxylation of both B-rings of B-carotene to form Z (Misawa et al., 1990;
Hundle et al., 1993; Bouvier et al., 1998; Masamoto et al., 1998). The formation of L
from a-carotene requires the action of a second hydroxylase, the e-hydroxylase, in
addition to a B-hydroxylase (Britton, 1998). The e-hydroxylase has not yet been cloned
from any organism, but has been genetically identified in Arabidopsis thaliana (Pogson
et al., 1996).

With regard to the genetics and molecular biology of carotenoid hydroxylation,
Arabidopsis thaliana is the best-characterized plant system. Two genes encoding B-
hydroxylases (Ii-hydroxylase 1 and 2) are present in the Arabidopsis genome (Sun et al.,
1996; Tian and DellaPenna, 2001). The predicted mature proteins share 81% protein
identity and are coordinately expressed although B-hydroxylase 1 mRNA levels are
always much higher than B-hydroxylase 2 mRN A. When expressed in vitro, both B-
hydroxylases are highly active towards B-rings and function poorly with e-ring
containing substrates (Sun et a1., 1996; Tian and DellaPenna, 2001). Mutational studies
have also identified the LUT1 locus as being essential for hydroxylation of e-rings in
Arabidopsis. Plants homozygous for the [w] mutation show an 80% reduction in L levels

and accumulate the immediate monohydroxy precursor zeinoxanthin. The Iut] mutation

143

does not affect B-ring hydroxylation (Pogson et al., 1996), and [w] does not map to either
B-hydroxylase 1 or 2 locus. A thorough analysis of the Arabidopsis genome also failed to
identify additional paralogs similar to B-hydroxylases from bacteria, cyanobacteria, or
plants (Tian and DellaPenna, 2001). These combined data suggest that LUT1 defines a
novel class of carotenoid hydroxylase enzymes and that B-hydroxylase l, B-hydroxylase
2, and LUT1 may represent the fill] complement of carotenoid hydroxylases in
Arabidopsis.

The IutI mutation was the first demonstration that by manipulating carotenoid
hydroxylase activities in viva, one can modify both the types and amounts of
xanthophylls that accumulate in plants. More recently, the B-hydroxylase 1 gene was
constitutively overexpressed and antisensed in Arabidopsis leaf tissue. Overexpression
resulted in a two-fold increase in xanthophyll cycle carotenoids (V +A+Z) without
affecting L levels (Davison et al., 2002). Expression of a B-hydroxylase 1 antisense
construct reduced the level of V and N in leaf tissue without affecting L, however, it was
unclear which of the closely related B-hydroxylase genes were affected and to what levels
(Rissler and Pogson, 2001).

In order to elucidate the in viva functions of carotenoid hydroxylases, T-DNA
knockout mutations in the B-hydroxylase 1 and 2 genes were isolated. Homozygous B-
hydroxylase mutants were studied singly, in combination, and introduced into the
previously isolated IutI mutant background in order to determine any specific or
overlapping function(s) of the B- and e- hydroxylases in viva. The effects of various
mutant genotypes on carotenoid biosynthesis, xanthophyll composition, and NPQ

capacity are presented.

144

Results

Isolation of Arabidopsis T-DNA insertion mutants for B-hydroxylases and

generation of double and triple mutant combinations

In order to study the in viva functions of B-hydroxylase 1 and 2, plants containing
T-DNA insertion alleles of each gene were isolated. Segregation of the kanamycin
resistance marker indicated both mutant lines were segregating for a single Mendelian
locus. Lines homozygous for each insertion were identified and designated b1 and b2, for
B-hydroxylase 1 and 2, respectively (CrtR-bI and CrtR-b2 according to standard
nomenclature; Hirschberg, 2001). The genome insertion sites were determined by
sequencing products amplified from each mutant. The B-hydroxylase 1 gene (At4g25700)
contains an insertion in the first intron 375 bp downstream from the start codon, while the
B-hydroxylase 2 gene (At5g52570) contains an insertion in the fourth exon 527 bp
downstream from the start codon (Figure 4.2A).

B-hydroxylases are non-heme di-iron proteins that contain ten conserved
histidines required for iron binding and activity. The mutation of any one of the ten
histidines resulted in a complete loss of activity (Bouvier et al., 1998). In the b1 mutant,
all ten histidines are downstream of the T-DNA insertion site. Five of the ten conserved
histidines are downstream of the T-DNA insertion site in the b2 mutant. Both mutants are
therefore expected to result in a complete loss of functional protein. RT-PCR reactions

using primer pairs designed to span the T-DNA insertion site of each mutant locus were

145

Figure 4.2. A. B-hydroxylase 1 and B-hydroxylase 2 insertional mutations (b1 and b2,
respectively). Exons are represented as filled boxes and introns as lines connecting the
boxes. The sizes of the boxes and lines are drawn to scale. Locations of forward and
reverse primers for B-hydroxylase l and 2 genes as well as the T-DNA insertion site are
indicated. B. RT-PCR reactions with RNA extracted from wild type (WS) and
homozygous b1 and b2 knockout mutants. Lane 1, 4 and 7, RT-PCR products from WS.
Lanes 2, 5 and 8, RT-PCR products from b]. Lanes 3, 6 and 9, RT-PCR products from
b2. Arabidopsis s-cyclase specific primer pairs were used in lanes 1-3. B-hydroxylase 1
specific primer pairs were used in lanes 4-6. B-hydroxylase 2 specific primer pairs were

used in lanes 7-9.

146

B-hydroxylase 1

STOP

  

[S-hydroxylase 2

JL202
—>
F w STOP
‘3

100 bp

 

147

used to determine whether full-length transcripts were produced in each genotype. The
Arabidopsis lycopene e-cyclase gene (At5 g5 7030) was used as a control to demonstrate
the integrity of all RNA preparations (Figure 4.2B). Both B-hydroxylase mRN As were
detected in WS and B-hydroxylase 1 and 2 mRNA was detected in b2 and b1 mutants,
respectively. However, no amplification products were observed for B-hydroxylase 1 and
2 in the b1 and b2 mutants, respectively (Figure 4.2B). These results confirm that
transcripts spanning the T-DNA insertion sites are not present in the respective knockout
mutants.

Homozygous b1 and b2 mutants grown under moderate light conditions did not
exhibit a whole plant phenotype differing from wild type (Figure 4.3). This result
suggests significant functional redundancy exists between the two Arabidopsis 13-
hydroxylases. To further address this question, we attempted to generate a b1 b2 double
mutant. Because no additional B-hydroxylase paralogs are present in the Arabidopsis
genome, if B-hydroxylase 1 and 2 were the only gene products with B-hydroxylation
activity, one would expect a complete absence of B-carotene derived xanthophylls and
likely lethality in a b1 b2 double mutant. b1 and b2 mutants were crossed and viable,
homozygous b1 b2 double mutants were identified by PCR at the expected fi'equency in
F2 progeny. This unexpected result suggests that other gene products must also have [3-
hydroxylation activity in Arabidopsis.

Although B-hydroxylase 1 and 2 are most active towards B-rings, each can also
add a hydroxyl group to e-rings in vitro, though at much lower efficiency (Sun et al.,
1996; Tian and DellaPenna, 2001). Therefore, it is conceivable that the reciprocal could

be true for the presumed s-hydroxylase encoded by LUT1; that it can also function

148

Col WS b1 b2 Iut1

 

b1b2 Iut1 b1 Iut1b2 lut1b1b2

Figure 4.3. Six-week old wild type and mutant Arabidopsis plants. The plants shown
were grown under normal light conditions (120-150 pmol photons in2 5'1) with a 12 hr
photoperiod. Top row (from left to right): Col, WS, b1, b2, lutl. Bottom row (from left to
right): b1b2, luthI, Iut1b2, lu11b1b2.

Images in this dissertation are presented in color.

149

towards B-rings in viva. In order to determine whether and to what extent LUT1 is
functionally redundant with B-hydroxylase 1 and 2, the b1 and b2 mutations were
introduced into the 1m] mutant background, individually and in combination. lut] is in
the Columbia (Col) background while the b1 and b2 mutations are in WS. Therefore,
both C01 and WS were used as controls in all experiments containing combinations of
[at] with b] and b2. In order for a lut1b1, Iuth2, or [w] b1 b2 mutant phenotype to be
considered significantly different from wild type, it must therefore be significantly
different from both parental wild type phenotypes.

Homozygous lutI b1 and [at] b2 double mutants were selected from the
corresponding F2 populations. Neither double mutant showed a visible deleterious whole
plant phenotype compared to wild type (Figure 4.3). In order to generate a line deficient
in all three hydroxylase activities, [at] b] and lut1b2 plants were crossed and viable
homozygous lut1b1b2 triple mutants were identified in the F2 progeny. Unlike all other
mutant combinations, the lut1b1b2 triple mutant exhibited an obvious whole plant

phenotype and was paler and smaller than either wild type parental ecotype (Figure 4.3).

Expression of carotenoid hydroxylases and lycopene cyclases genes in wild type and

various mutant genotypes

To determine the effect of various mutant combinations on expression of the
remaining unmutated hydroxylases and the two lycopene cyclase enzymes, mRNA levels
for B-hydroxylase l, B-hydroxylase 2, lycopene B-cyclase, and lycopene e-cyclase were

determined in wild type and different mutant genotypes. Because expression of

150

carotenoid biosynthetic genes is quite low in photosynthetic tissues, we developed
TaqMan real time PCR assays for these studies. B-hydroxylase 1 and 2 mRNAs were not
detected in genotypes that contain disruptions in these genes (Figure 4.4A and B),
consistent with the RT—PCR analysis in Figure 4.2B. The steady-state B-hydroxylase 1
mRN A level was similar to wild type for all mutant genotypes that contain a functional
B-hydroxylase 1 gene. B-hydroxylase 2 mRNA was also similar to wild type in the lut]
mutant, but was increased approximately 50% and 30% relative to wild type in the b1 and
lut] b1 , respectively. The expression of lycopene B- and e-cyclase genes was also
analyzed to determine whether specific hydroxylase mutant combinations affected
cyclase gene expression. Both 13- and e-cyclase mRNA levels were nearly identical to
wild type in all mutant genotypes except in the Iut1b1b2 triple mutant, where the B-

cyclase mRNA level was 20% lower than wild type (Figure 4.4C and D).

Leaf pigment compositions in wild type and various mutant genotypes

Wild type and mutant genotypes were grown under 120-150 pmol photons m'2 s'1
for six weeks and their carotenoid compositions were quantified by HPLC. The level of
total carotenoids in the mutant genotypes was not significantly different from their
respective wild type controls when expressed per mole chlorophyll or fresh weight,
except that the lut1b1b2 triple mutant had 20% less total carotenoids on a fresh weight
basis (data not shown). However, though carotenoid levels were unchanged, carotenoid
compositions were significantly altered in most mutant lines (Table 4.1). The b2 mutant

was the least affected of all genotypes and did not differ significantly from WS, with the

151

 

A B-hydroxylase 1

2000 ‘
1000 ‘

B B-hydroxylaso 2

2000‘
1000‘I I I I I
01

5000 . B-cyclase

3000 y
2000 «

amol target RNA I pg total RNA
0

1000'4

5000 , D e-cyclase

 

 

«Li.

ws
Col
b1
b2
b1b2
Iut1

lut1b1

Iut1b2
lut1b1b2

Figure 4.4. TaqMan Real Time PCR analysis of B—hydroxylase 1, B—hydroxylase 2,
lycopene B—cyclase, and lycopene s-cyclase mRN A levels in wild type and mutant
genotypes. A. B-hydroxylase l. B. B-hydroxylase 2. C. lycopene B-cyclase. D. lycopene

e-cyclase. Data are shown as means + SD. (n=4).

152

 

.00... An. 0102.. 0.50020. :::28 0 0.5.3 :050 :000 80:. .:0:0...0 >..:00...:.mW.0: 0:0 0:200:05 2:00 05 5.3 02:05 000_0> 0000:2200 :. 5%
06:20:00 :000 .0 32:00:00 :0.0:: 050.0: 05 5.3 ..o.m « 2:06.090 .00 .0 :00:: 05 0. 00_0> 50w 0+0 .20 6505890 .05.: 00 0000290 0. 06:20:00 .0 .5900 0:...

 

 

00000.02 0.0.52.0. - - 0.0.0200 5.000.. 0.00.00.00.00 0.8.0.0200 0.0.0.2.: 00.020.
0000.02 0.00.0.00000 0.0.0.300 0.2.0.0000 0.0.0.300 0.0....0..«0.00 0.00. 0.02.3 0.2.0.0000 0.0.0.00... 002:.
00000.0: 0.00.0000... 0.0. 2000 0.2.0.0000 0.00.00.00.00 0.0.3.0202 0.00.0000? 0.00.0000: 0.0.0.02 .020.
00.0.0002 0.0030200 0.0. 0.5.0.0 0.3.0.0002 0.00. 200.00 0.0.2.0000 0.2.0.0000 0.00. 0000.00 0.2.0.0300 :0.
0300.0: 0.0.0.0.: - - 0.0.53.0 0.0.2000 0.2.0.0000 - 0.2.0.8002 00.0
00000.0: 0.00. 0000.3. - - 0.2.30.0 0.2.0.0300 0.2.0.0000 - 0.00. 0000.00. 00
0:202 0.2.0.0000 - - 0.0.0.020. 0.0.0.0000. 0.2.0.0200 - 0.0.0.0002. .0
00.3002 0.00. 0.3.0.0.. - - 0.0....0.0«0.00 0.0.2.0.0«:.00 0.2.0.0000 - 0.00. 0000.00. _00
00.0.0002 0.8.0.0000 - - 0.2.0.0200 0.05.0000 0.2.0.0030 - 0.8.0.0000. 0.5
0995.900 3.0.? m=>£Q05cmx¢ 55:0ng c.5cmxw:05c< 55:98—05 55:00.82 9.0.0.00... EECNXOEQN £05..—

 

.0._nI .3 0000:000 0:00.. .00. :_ :0...000::00 06:20:00 .:0.::: 0:0 00... 25> .Fé 0.00b

 

153

exception of a small decrease in N. The [)1 mutant has a much more severe phenotype
with 30% and 45% decreases in V and N, respectively, and an 18% increase in L relative
to WS. B-carotene was unchanged (Table 4.1) and the monohydroxy xanthophyll B-
cryptoxanthin (B,B-carotene-3—ol) was not detected in b1 or b2 (data not shown). With the
exception of increased L levels, the leaf carotenoid phenotype of the b1 mutant is similar
to that reported for a constitutive Arabidopsis B-hydroxylase 1 antisense transgene
(Rissler and Pogson, 2001). The [at] mutant profile is consistent with previously
published data (Pogson et al., 1996). lutl has an 80% decrease in L, accumulates the
monohydroxy xanthophyll zeinoxanthin (B,e-carotene-3-ol), and has a two-fold larger
xanthophyll cycle carotenoid pool (V +A+Z) compared to Col. B-carotene and N levels
were not significantly different in Iut] compared to C01.

The homozygous b1 b2 double mutant was quite informative with regard to the
role of various carotenoid hydroxylases in Arabidopsis. The most surprising result is that,
though both B-hydroxylase activities are eliminated in the b1 b2 mutant, B-ring
hydroxylated xanthophylls (L, V, N) are still synthesized, albeit at lower levels than WS
or either single mutant. N and V are reduced 90% and 65%, respectively, and there is a
30% increase in L relative to WS. As with the single B-hydroxylase mutants, B-carotene
is unchanged relative to wild type and B-cryptoxanthin does not accumulate. That b1 b2
has a more severe carotenoid phenotype than either single B-hydroxylase mutant
indicates that the two mutations are additive and there is significant functional
redundancy between B-hydroxylase 1 and 2. B-hydroxylase 1 and 2 are clearly the major
B-ring hydroxylation activities as their combined absence decreases the levels of [3-

carotene derived xanthophylls nearly 80%. However, B-carotene derived xanthophylls are

154

still produced at 20% of wild type levels indicating there is an additional and previously
undescribed activity in Arabidopsis that is also capable of producing B-carotene derived
xanthophylls.

The carotenoid compositions in the [w] b] and lut1b2 double mutants are quite
different from their corresponding single mutant parents. This result indicates there is
functional compensation or redundancy between LUT1 and the two B-hydroxylases. As
with the single B-hydroxylase mutants, b1 has a much greater effect on carotenoid
composition than b2 when each is introduced into the [21!] background. The level of
zeinoxanthin is increased 25%, V and L levels are reduced by 25% and 35%,
respectively, and [El-carotene is increased by 35% in [at] b1 relative to [w] . The only
carotenoids significantly changed in lut1b2 are B-carotene and L.

An even more dramatic alteration in carotenoid composition was observed in the
lut1b1b2 triple mutant. Iut1b1b2 has the highest level of B-carotene in any genotype and
more than twice the level of either wild type parent. There are severe reductions in N
(95%) and V (60%) relative to either wild type and unlike Iut], [at] b] , or Iut1b2, Z and A
are absent in the triple mutant. [l-carotene derived xanthophylls (V +Z+A+N) account for
only 5% of the total carotenoid pool in the triple mutant versus 25-30% in wild type. L is
unchanged in the triple mutant relative to the luthI and lut1b2 double mutants, but

zeinoxanthin is more than doubled relative to any other lutl containing genotype.

Seed pigment compositions of wild type and mutants

155

In order to determine the consequence of the various hydroxylase mutant
genotypes on seed carotenoid composition, carotenoids from dry seeds of wild type and
the various mutant genotypes were extracted and quantified by HPLC. In general,
Arabidopsis seeds contain five to ten-fold less total carotenoids than leaf tissue on a dry
weight basis. The major carotenoid in wild type seed is L. A and Z are present at high
levels in seeds compared to leaf tissue but unlike leaf tissue, there is little B-carotene
accumulation in seed. WS seeds have twice the level of total carotenoids as Col due to
significant increases in all carotenoids except A and Z (Table 4.2). These data indicate
significant quantitative genetic variations for seed carotenoid levels exist between
Arabidopsis ecotypes.

The level of total carotenoids in b1, b2, and IutI mutant seeds does not differ
significantly from their corresponding wild type controls. As with leaf tissue, the b1
mutation has a more significant impact on seed carotenoid composition than b2. L is
unchanged while Z, V, and N are significantly decreased and A is increased in b1 seed
relative to wild type. The carotenoid composition of b2 seed is quite different from both
b1 and wild type with a 10% decrease in L, a nearly two-fold increase in A, and 30%
increase in Z. Unlike b], V and N in b2 seed do not differ significantly from wild type.
The differential effect of the two single mutations in seeds is most obvious in the total [3-
carotene derived xanthophylls accumulated: the b1 mutation decreases while the b2
mutation increases the level of total B-carotene derived xanthophylls in seed.

The phenotype of the lut] mutation in seeds differs dramatically from that in
leaves. The most obvious difference is that L levels are only reduced 30% in seeds versus

80% in leaves and there is no zeinoxanthin accumulation in [w] seed. These surprising

156

 

.00... Am .60.-. 0.60030. 5:060 0 5.0.3 :050 :000 60:. 620.20 3:005:90 6: 0:0 0:060:05 2:00 05 5.3 0060.: 00:.0> 0000:6200 :. 0020 06:06:90
:000 .0 00060800 66.: 02.0.0: 0:. 5.3 ..o.w 0 0606.698 00:... .0 :02: 05 0. 029 :00m. .0000 960606 6:6 00 0000298 0. 06:06:00 .0 60060 0: .0

 

 

00.00000 ..0.. 0. .000. 0.0. 0000.0 0.0. 0000.0 0.0. 0.00.0 0.0. 0.000.. 0.0.0. 0.0000 0.2.. 0.00.00 00.0.3.
00. .0000 0.00. .. .0... .0 0.0. 0000.0 0.0. 0.0000 0.0.0.. 0.00.0. 0.0. 000.0 0.0.0. 00000.0 :00. 0.00000 00...:
£000.00 0.2.0.0000. 0.0. .000. 0.0. 00000 .E 00000 0.0. 0000.0 0.0.0. 000000 0.00. 0000... .020.
0.00.00 0.00. 0. .0000 0.0. 0000.0 0.0.0. 0.0000 0.0.00. 0. .000. 0.0. 0000.0 0.0.0. 00.0000 0.00. 0.00.00 .3.
00.00000 0.0.. 0.0000 0.0. 0.000 0.0. 0.0000 0.0. 0.000.. 0.0. .00.... 0.0.0. 000000 0.00. 000...... 00.0
0.000.00. 0.00. 0. .0000 0.0. 0.0000 0.0. 0.0000 0.0.. 0. .000. 0.0. 0.0000 0.0.0. 00000.0 0.00. 0,000.00 00
00.00.00. 0.0.. 0.0000 0.0. .0000 0.0.0. 0.00.0 0.0. 0.000... 0.0. 0000.0 0.00. 00.0000 0.00. .000. . .. .0
00.00.00 0.00. 0.0000. 0.0... 0000.0 0.0. 0.0000 0.0. 0000.0 0.0. 0.0000 0.0.0. 0000.0 0.00. 0.00000 .00
00.00000. 0.00. 0.00.00 0.0. .0000 0.0. 0.000... 0.... 0.0000. 0.0. 0.000... 0.0.0. 000000 0.0... 0.00000. 02.
36:06:00 .06 :. 0:285:00... :.:60x00N :_:60x0:0:6< 5:69.065 5:60.802 0:06:00... 506..

 

.000: .0 0006000 0000 0: 80:00:58 0.20.98 .0055 0:0 00>. 0:... 0.0 000.

 

157

results can be understood by considering the total amount of L produced in leaf and seed
as a percentage of dry weight. In Col leaves, the molar amount of L synthesized is ten
times that in seeds (data not shown). Thus, all things being equal, the residual or
secondary e-hydroxylase activity that allows production of 20% wild type L levels in lutI
leaf tissue is apparently sufficient to fully convert the much lower level of zeinoxanthin
in seeds to L. N and V are increased in lutI seeds while A and Z are almost unchanged.
This contrasts with the large accumulation of A and Z in lut] leaf tissue relative to wild
type (Table 4.1).

Unlike the single mutant parent lines, the carotenoid levels of b1 b2 seed are
reduced 40% relative to wild type (WS). However, as in leaf tissue, B-carotene derived
xanthophylls are still present at 10% of wild type levels despite the fact that both [3-
hydroxylase activities are eliminated in the b1 b2 double mutant. There is also a dramatic
shift in carotenoid composition in the b1b2 mutant with Z, A, V, and N reduced 40%,
50%, 90%, 80%, respectively, relative to wild type.

Interpretation of seed carotenoid data from [w] b1 , luthZ, and Iut1b1b2 is more
difficult due to the large quantitative differences in total and individual carotenoid levels
between the two parental ecotypes (C01 and WS). However, assuming the loci
responsible for these quantitative differences are segregating independently of the mutant
hydroxylase loci, F 2 progeny should segregate randomly for quantitative differences in
seed carotenoids. To minimize the impact of any quantitative loci, twenty F2 individuals
homozygous for IuthI , lut1b2, and lut1b1 b2 were selected and their seeds pooled for
analyses, which should average out quantitative differences in the progeny while

maintaining selection for the respective mutant loci. Consistent with this hypothesis,

158

lut1b1 and 1m] b2 double mutants have similar amounts of total carotenoids relative to
each other and are intermediate relative to their C01 and WS parents. [at] b2 has higher
levels of individual and total B-carotene derived xanthophylls compared to lutI b1 , while
the luthI mutant has a higher level of L than lut1b2. The total carotenoid level of the
lut1b1 b2 triple mutant seed is 35% lower than either [at] b1 or IuthZ, but L still accounts
for 80% of total carotenoids. Most B-carotene derived xanthophylls are significantly
reduced in the triple mutant relative to [w] b] or lut1b2 and there is a four to twenty-fold

increase in B-carotene relative to other genotypes (Table 4.2).

Photosynthetic efficiency and NPQ capacities in leaves of wild type and mutant

genotypes under normal light conditions

Leaf carotenoid compositional analysis showed that production of a- and [3-
carotene derived xanthophylls could be severely impacted by disrupting one or more
hydroxylase activity (Table 4.1). To determine the effect of modified xanthophyll profiles
on photosystem efficiency and NPQ, chlorophyll fluorescence was measured in leaves of
each mutant genotype grown under moderate light conditions (120-150 pmol photons rn'2
s"). The Fv/F m ratio (a measurement of maximum photosystem II photochemical
efficiency) of wild type (C01 and WS), lut], b1, and b2 mutants did not differ
significantly, but b1b2, lutl b1, lut1b2, and lut1b1b2 had lower Fv/Fm ratios compared to
wild type (Table 4.3). These results suggest that the altered carotenoid composition

resulting from elimination of more than one hydroxylase activity can affect the efficient

transfer of absorbed light energy to the photosystem II reaction center. Chl a/b ratios in

159

 

Table 4.3. Chl a/b ratios and maximum PS II efficiencies
(Fv/Fm) in wild type and mutant leaves.

 

 

Chl alb Fv/Fm
WS 2.4710.07a 0.846i0.017a
Col 2.5210.06a 0841100068
b1 2541:0088 0837100133
b2 2.50:0.053 0.847:l:0.0063
b1b2 2.63:0.05b 0.82410.011b
Iut1 2.61:0.06a 0.843i0.010a
Iut1b1 2.6410.02b 0.827i0.010b
Iut1 b2 2.67:0.03b 0.82610.015b
lut1b1 b2 2.74:0.020 0.82110.007b

 

Values marked with the same characters are not
significantly different from each other within a column
(student's t-test, P> 0.05).

I60

b1 and b2 mutants do not significantly differ from wild type but are significantly
increased in the lut] mutant as previously reported (Lokstein et al., 2002) and also in all
the double mutants. The lut1b1b2 triple mutant has the most drastically affected Chl a/b
ratio, and is significantly higher than all the other genotypes (Table 4.3).

As with wild type seed carotenoid composition, significant genetic variation was
observed for NPQ between WS and Col. WS has both a slower induction rate and lower
NPQ amplitude compared to C01. The b2 mutant has NPQ induction and amplitude
similar to WS while the b1 mutant, where the relatively highly expressed B-hydroxylase 1
gene is disrupted, has a slower induction and ~10 % reduction in NPQ amplitude. This
impact on NPQ is consistent with the greater impact of the b1 mutation on xanthophyll
cycle carotenoids (Table 4.1). Light activation of NPQ is also delayed in [at] and NPQ
amplitude is reduced, consistent with previous reports (Pogson et al., 1998; Lokstein et
al,2002)

In b1b2, there is a synergistic effect of the two B-hydroxylase mutations on both
carotenoid composition and NPQ. NPQ induction and amplitude are reduced to a much
greater extent in b1 b2 than in either single mutant parent (Figure 4.5A). In lutI b1 and
Iuth2 double mutants, the NPQ phenotypes of the single mutants are additive and likely
result from the combined deficiency of both L and xanthophyll cycle carotenoids in these
double mutants. That the [MM] phenotype is more severe than the [w] b2 phenotype can
be predicated from the respective single mutants. The lutI b1 b2 triple mutant has a very
rapid initial induction of NPQ, however the NPQ level obtained is greatly reduced

compared to wild type and all other mutant genotypes (Figure 4.5B).

16]

 

 

 

  
  
   
  

 

 

1600
B
as-
—o-—COL
’-°- —o—WS
+Iut1
015' +Iut1b1
a. . —o—Iut1b2
z 1 o- '
' . —o—Iut1b1b2
0.5
0.0 . . . . . .
o 200 400 600 son 1000

Time [s]

Figure 4.5. NPQ analysis of wild type and mutant genotypes. A. Wild type (WS and
C01) and b1 , b2 and b1 b2 mutants. B. Wild type (WS and C01) and IutI, lutI b1 , Iutlb2,
lut1b1b2 mutants. Chlorophyll fluorescence was measured during 6 min of illumination
with a PFD of 500 pmol photons m‘2 s'l followed by 9 min of dark relaxation. Each data

point is the mean of six independent experiments and all S.D. are within the symbol sizes.

162

Discussion

In plants, carotenoid hydroxylases catalyze the formation of a- and B-carotene
derived xanthophylls, which perform a variety of critical roles in the photosynthetic
apparatus. Three genes that participate in carotenoid hydroxylation reactions have been
identified in the Arabidopsis genome through combined molecular and genetic analyses.
Among these, two B-hydroxylases have been cloned and shown to be highly active
against B-rings and weakly active against e-rings (Sun et al., 1996; Tian and DellaPenna,
2001). We have isolated T-DNA knockout mutants in the B-hydroxylase 1 and 2 genes
and shown that expression of mRNA for each gene is fully eliminated. A third
Arabidopsis gene product required for e-ring hydroxylation has been genetically defined
by the lat] mutation but has not yet been cloned (Pogson et al., 1996). The effects of
these three carotenoid hydroxylase mutations individually and in various combinations
have been used to assess the roles, interactions, or functional redundancies of the

hydroxylase activities in vivo.

LUT l is the major s-ring hydroxylating activity in vivo

The effect of mutating the LU T 1 locus, the presumed e-hydroxylase, on leaf
carotenoid composition has been described previously (Pogson et al., 1996; Lokstein et
al., 2002) and has the most dramatic effects of any single hydroxylase mutation. The fact

that e-ring hydroxylation is reduced by 80% in lutI, and zeinoxanthin accumulates

indicates the LUT1 gene product is the primary e-ring hydroxylation activity in

163

photosynthetic tissue. The absence of [3,8-carotene-3’-ol (or-carotene with a single
hydroxyl group on the e-ring) in lutl or any other mutant line studied suggests that B-ring
hydroxylation precedes e-ring hydroxylation during L synthesis. Both lutI alleles Um] -I
and lutI-2) are EMS-derived and still contain 15-20% of wild type L levels. This
suggests that either both alleles are leaky and the L made is from residual LUT1 activity
or, if the alleles are null, that there is a second e-ring hydroxylation activity in
Arabidopsis. The cloning of LUT1 and identification of a definitive null mutation in the

locus is required to distinguish between these possibilities.

B-hydroxylase 1 and 2 are functionally redundant in vivo

Neither b1 nor b2 exhibits a whole plant phenotype suggesting that significant
functional redundancy exists between the two B-hydroxylases. This observation is
consistent with our inability to identify B-hydroxylase mutants in various EMS mutant
screens (data not shown). B-hydroxylase 1 and 2 are closely related at the level of protein
sequence, have similar activities in vitro and are coordinately expressed in all wild type
tissues analyzed, although B-hydroxylase 1 mRN A is always present at much higher
levels than B-hydroxylase 2 mRNA (Tian and DellaPenna, 2001). The high relative
expression level of B-hydroxylase 1 suggests it might be responsible for the majority of
B-ring hydroxylation activity in leaf tissue. The alterations to carotenoid composition in
the B-hydroxylase single mutants support this hypothesis: B-carotene-derived

xanthophylls are more severely affected in b1 than in b2 (Table 4.1; Figure 4.5).

I64

The fimction of B-hydroxylase 2 is difficult to discern by single mutant analysis,
because it is not significantly different from wild type in carotenoid composition and
NPQ function. However, B-hydroxylase 2 is clearly both functionally additive and
complementary to B-hydroxylase 1 in viva. The b1b2 double mutant has a more severe
effect on B-ring hydroxylations than the b1 mutation alone (Table 4.1) indicating that B-
hydroxylase 2 can compensate to a significant degree for the absence of B-hydroxylase 1
activity in the b1 mutant. Moreover, B-hydroxylase 2 is also able to compensate for the
loss of both LUTland B-hydroxylase 1 (in the [w] b1 double mutant) to almost the same
extent as B-hydroxylase l in the lut1b2 background (Table 4.1). These data indicate that
although B-hydroxylase 2 is expressed at a much lower level than B-hydroxylase l in
wild type, it plays an important role in functional redundancy in the carotenoid pathway.
The fimctional redundancy of B-hydroxylase 2 in the b1 and lutl b1 mutants may be
enhanced in part by a small increase in B-hydroxylase 2 expression (Figure 4.4B).

The fact that different b1 and b2 containing carotenoid mutant genotypes are
preferentially affected in the accumulation of specific carotenoids (Table 4.1), indicates
that while the two B-hydroxylases can partially compensate for each other, they are not
entirely equivalent and likely have specialized functions. This is most apparent when a B-
hydroxylase l or 2 deficiency is present in wild type or the Iut] background (Table 4.1).
While the biochemical basis for two closely related but functionally distinct [3-
hydroxylase enzymes in Arabidopsis is unknown, tomato and pepper have also
maintained two homologous B-hydroxylase genes, suggesting this may be a common

theme in carotenoid synthesis in plants.

I65

B-carotene-derived xanthophylls are still produced in the b1b2 double mutant

indicating additional B-ring hydroxylation activity exists in viva

A major and unexpected outcome of carotenoid hydroxylase double mutant
analyses is that when both known Arabidopsis B-hydroxylase activities are eliminated in
the b1b2 double mutant, B-carotene-derived xanthophylls are still produced at significant
levels. The 80% reduction in B-carotene-derived xanthophylls in b1 b2 confirms that B-
hydroxylase 1 and 2 are the predominant B-carotene hydroxylation activities in viva.
However, B-carotene-derived xanthophylls still account for 7% of the total carotenoids in
the b1 b2 mutant versus 29% in wild type. In addition, B-ring hydroxylation is also
necessary for a-carotene to be converted to L, and L is produced at even higher levels in
the b1b2 double mutant than wild type. The synthesis of B-carotene derived xanthophylls
and the accumulation of L in b1 b2 indicates an additional, previously unknown, B-ring
hydroxylation activity exists in viva. These data raise the question: what is the nature of
this additional B-ring hydroxylation activity?

Three hypotheses can be proposed for the additional B-ring hydroxylation activity
present in the b1 b2 background. (1) It is a third, novel B-hydroxylase enzyme unrelated at
the protein sequence level to known B-hydroxylases. (2) It is a secondary, intrinsic
activity of the LUT1 gene product (rs-hydroxylase) capable of B-ring hydroxylation in
viva. (3) It is a second LUT1-like enzyme that is active toward both types of rings.
Searches of the Arabidopsis genome using representative B-hydroxylase protein

sequences from all three phyla as queries failed to identify additional homologs beyond

I66

B-hydroxylase 1 and 2 (data not shown). Therefore, if the additional activity resides in a
third B-hydroxylase enzyme, it would represent a new structural class of B-hydroxylases
in nature. The possibility that LUT1 might hydroxylate B-rings in addition to e-rings in
viva has precedent as both B-hydroxylase I and 2 can also hydroxylate e-rings in vitro,
although quite poorly relative to B-ring substrates (Sun et al., 1996; Tian and DellaPenna,
2001). A reciprocal activity could also be true for LUT1. Finally, if LUT1 does not have
a secondary B-hydroxylase activity, it is possible that a LUT1 paralog exists that is active
towards both 8- and 8- rings. This would be analogous to the two highly conserved
Arabidopsis B-hydroxylases, however this LUT1 paralog would likely be a minor
component of total hydroxylase activity because it cannot fully compensate for the e-ring

hydroxylation deficiency in [w] . Clearly, isolation of LUT1 and biochemical

characterization of its product will serve to delineate these various hypotheses.

Double and triple mutants indicate overlapping functions and interactions between

the B- and e- hydroxylases

In order to investigate any overlapping functions or interactions between the B-
and e- hydroxylases of Arabidopsis, the b1 and b2 mutations were introduced into the

lull background to generate the corresponding double mutants. As with the single b1 and

b2 mutants, b1 has a more pronounced effect in the lat] background than b2. [w] b] has
decreased V and increased zeinoxanthin relative to Iut] . B-hydroxylase 1 and 2 also
appear to have a low level of s-ring hydroxylation activity in viva as elimination of either

activity in the [m] background further decreases the level of L (Table 4.1).

I67

The failure to eliminate L in luthI and lut1b2 may be due to a low level of s-ring
hydroxylation activity in viva from the single functional B-hydroxylase still present in
each double mutant. To further delineate these possibilities, we attempted to create a
lut1b1b2 triple mutant. Assuming no other carotenoid hydroxylases are present in
Arabidopsis and the lut1-2 allele is null, one would expect the triple mutant to lack all
xanthophylls and likely be lethal as a result. Surprisingly, homozygous lut1b1b2 mutants
were isolated at the expected ratio from crosses and though paler green than wild type,
they were viable and set seed. Because the L level in lut1b1b2 (where both B-hydroxylase
1 and 2 activities are eliminated) is not further reduced relative to [w] b] or lutlb2, we
conclude that B-hydroxylase l and 2 do not contribute significantly to e-ring
hydroxylation in viva. The L synthesized in Iut1b1b2 must therefore be due to either
residual a-ring hydroxylation activity from the mutant LUT1 enzyme or another a-
hydroxylase. In either case, LUT1 or a second e-hydroxylase must also be active toward

B-rings as B-ring hydroxylation still occurs at significant levels in [w] b1 b2 (Figure 4.6).

Total hydroxylation activity and the synthesis of specific xanthophylls are

differentially affected in the hydroxylase mutants

The impact of hydroxylase deficiencies has been discussed thus far based on the
relative changes to specific carotenoid (e. g. violaxanthin) or groups of a- and [3- carotene
derived xanthophylls in each genotype (Table 4.1). In the case of double and triple
mutants, the effects on individual xanthophylls can be quite extreme. However, focusing

only on the species or groups of xanthophylls affected leads to underestimation of total B-

I68

ring hydroxylation activity because hydroxylation of the B-rings of zeinoxanthin and L
are overlooked. A more accurate accounting of total hydroxylation activity is to consider
the total hydroxylation of each ring type (total moles of hydroxyl groups introduced onto
[3- and e-rings, respectively, see Figure 4.6). From this perspective, though the bl mutant
significantly decreases N and V (Table 4.1), the level of hydroxylated B-rings produced
by the mutant is hardly affected and total ring hydroxylation ([3- plus e-rings) is
unchanged relative to wild type (Figure 4.6). This is even more apparent in b1 b2 where
B-carotene derived xanthophylls are reduced 80% but B-ring hydroxylations are only
reduced 25% and total hydroxylation is not significantly changed relative to wild type.
These data suggests an even higher degree of functional redundancy among the
hydroxylases with regard to their ability to perform [3- and/or e-ring hydroxylations in the
absence of another. The ability of each hydroxylase to functionally compensate for
another in the production of specific xanthophylls is clearly much more limited (Table
4.1).

The differential effect of hydroxylase deficiencies on total hydroxylation levels
and the production of specific xanthophylls suggests a level of biochemical regulation in
the production of different xanthophylls that has not previously been considered. Whether
this regulation stems from innate differences in the substrate preference and turnover
rates of each hydroxylase or is due to the differential ability of hydroxylases to participate
in biosynthetic complexes for the production of certain xanthophylls is unknown. It is
also possible that specific molecular or biochemical regulatory mechanisms could be
activated in the mutants to preferentially provide for a minimum level of B-ring

hydroxylation.

I69

 

 

 

 

 

El Nonhydroxylated rings
.2 Hydroxylated e-rings
to I Hydroxylated B-rings
— 400
.r:
U
3 .
g 300
8
_E 200 <
h
3
E 100*
E
o _ ,
u) 3 h N N h h N N
a o e e g :5 g g g
E E F
H
E
B-hydroxylase 1 + + - + - + - + -

B-hydroxylasez + + + - - + + - .

LUT1 + + + + + - . - -

Figure 4.6. Hydroxylation levels of B- and a-rings in all genotypes. The molar amount
of hydroxylated B- and s-rings and non-hydroxylated rings were calculated from the leaf
tissue carotenoid composition for each genotype. The presence or absence of each of the

three known carotenoid hydroxylases in each genotype is indicated by a plus or minus

sign, respectively.

I70

The surprisingly high level of total B-ring hydroxylation retained in all mutant
genotypes suggests that hydroxylated B-rings, rather than specific xanthophylls, might be
the structural and/or functional minimal requirements for photosystem assembly in viva.
This is consistent with a recent report that 3—hydroxy-B-end groups, rather than B-
carotene derived xanthophylls per se, are the minimal requirements for the binding of
xanthophylls to LHC II in vitro (Phillip et al., 2002). Hydroxylated B-rings (mono- or di-
hydroxy xanthophylls) were able to promote the in vitro assembly of LHC II by
facilitating the correct folding of LHC 11 proteins. In light of this critical role for 3-
hydroxy-B-end groups in LHC II assembly, it is easier to understand why the levels of B-
ring hydroxylations, rather than specific xanthophylls, are maintained in viva in the
carotenoid hydroxylase mutants. An extreme example of this principle is [at] b1 b2, which
has the lowest level of B-carotene derived xanthophylls of any genotype (Table 4.1). In
IuthIb2, zeinoxanthin (which contains a single 3-hydroxy-[3-end group) is increased for
more than two-fold relative to lutI and accounts for 75% of the total 3-hydroxy-B-end

groups.

Modified xanthophyll composition in the hydroxylase mutants correlates with

alterations in NPQ induction kinetics

In addition to their essential roles in LHC II assembly and photosystem structure,
xanthophylls also have important fimctions in photoprotection. The severity of
xanthophyll compositional changes in the different hydroxylase mutant genotypes

coincides with their degree of compromised NPQ induction kinetics (Table 4.1; Figure

171

4.5). Other carotenoid biosynthetic mutants previously used to study xanthophyll function
in Arabidopsis in viva, lutl , lut2, npq1, aba] (npq2) and their double mutants, block the
production of one or more xanthophylls (Niyogi et al., 1998; Pogson et a1, 1998; Lokstein
et al., 2002). The carotenoid hydroxylase mutants are unique in that the synthesis of
multiple xanthophylls are attenuated to different degrees, rather than eliminated. This
allows one to modify the types and levels of specific xanthophylls involved in light
harvesting and photoprotection under normal growth conditions.

Because Chl b and N are primarily associated with LHC II (Ruban et al., 1999),
the increased Chl a/b ratio and decreased N levels in the double and triple hydroxylase
mutants (Tables 4.1 and 4.3) are consistent with a reduction in LHC II levels. These data
are also consistent with the assembly or stability of the peripheral LHC II being affected,
and are correlated with impaired NPQ induction and amplitude in the double and triple
hydroxylase mutants (Figure 4.5). In addition, the F v/F m ratios in all double and triple
hydroxylase mutants are significantly decreased, indicating the quantum efficiency of
photosystem II is also affected due to their altered carotenoid compositions. Similar
observations have been reported for the lutI, lut2, aba], and lutZabaI mutants (Tardy and
Havaux, 1996; Pogson et a1., 1998; Lokstein et al., 2002). The most severe impact on
NPQ and quantum efficiency occurs in the lut1b1b2 mutant where L and xanthophyll
cycle carotenoids are reduced over 80% relative to wild type. Given the continuum of
xanthophyll changes in the full suite of hydroxylase mutants, it will be quite informative
for our understanding of xanthophyll functions to assess the responses of the hydroxylase

mutants to high light stress.

Insights into carotenoid pathway regulation in photosynthetic tissues

172

In addition to differentiating the overlapping functions of carotenoid hydroxylases
in viva, the phenotypes of hydroxylase mutants provide further insight into levels of
regulation that exist in the carotenoid pathway. The extreme changes in xanthophyll
composition are not the result of a corresponding alteration in total hydroxylation activity
(Figure 4.6) but rather a large shift in the accumulation of carotene cyclization products
in each mutant line (Figure 4.7). The increased L level in b1 b2 relative to wild type
indicates synthesis of [Le-carotene branch carotenoids is enhanced when B-ring
hydroxylation is blocked. Likewise, syntheses of [LB-carotene branch carotenoids are
enhanced 60% relative to wild type when e-ring hydroxylation is blocked in [w] . Most
significantly, in the lat] b1 b2 triple mutant, where all three hydroxylases are absent, the
accumulation of [3,8- and [3,[3- carotenoids returns to a wild type ratio (Figure 4.7). These
results indicate that one response of the carotenoid pathway to the absence of one or more
hydroxylases is to shift synthesis between the 8,8- and [3,8- carotene branches of the
pathway; a regulation that can only take place at the level of carotenoid cyclases (Figure
4.1). This regulation appears to be translational or post-translational as mRNA levels for
both the [3- and e-lycopene cyclase are essentially unchanged in all hydroxylase mutant
genotypes (Figure 4.4C and D). A model put forth by Cunningham and Gantt (1998)
proposed that [3,8- or [343- carotene production is determined by the formation/stability of
different carotenoid desaturase/cyclase complexes in the thylakoid membrane. Our data
are consistent with the carotenoid hydroxylases participating in these proposed
complexes to determine the relative levels of B,e- or 8,8- carotene produced in leaf tissue.

An alternative explanation is that the hydroxylase deficiencies differentially affect the

173

1 1 1 1

lycopene lycopene lycopene lycopene

1111 I111

B,a-car [113-car B,s-car B,B-car B,s-car B,B-car B,e-car B,B-car

   

 

WI' b1b2 Iut1 lut1b1b2

Figure 4.7. Accumulation of 8,8-carotene and 8,8-carotene branch carotenoids in
leaves of wild type and selected hydroxylase mutants. The total relative amount of
carotenoids accumulated through lycopene and the 8,8- and [Le-cyclization branches of
the pathway are indicated by arrows. The width of an arrow is proportional to the
percentage of each carotenoid accumulated. The arrows leading to lycopene are identical
in width because the total amount of carotenoids produced in each line do not differ

significantly (refer to Table 4.1).

174

substrate specificity/affinity of the cyclases by feedback regulation due to the altered

xanthophylls accumulated in each mutant line.

The impact of hydroxylase deficiencies on seeds carotenoid compositions

The seed carotenoid content of mutants was also analyzed to determine the
penetration of each mutant phenotype in a non-photosynthetic tissue. Lutein is the main
carotenoid in both wild type leaves and seeds (50% and 80%, respectively). B-carotene
accounts for about 15% of the total carotenoids in wild type leaf tissue (Table 4.1) but is
present only in trace amounts in seeds (Table 4.2). This difference may be due to more
efficient hydroxylation of B-carotene in seeds in the absence of the carotenoid binding
proteins present in photosynthetic tissue or to enhanced turnover of B-carotene during
seed development relative to leaf tissue. In general, the effect of B-hydroxylase mutations
on seed carotenoid composition is similar to that in leaves. The b1 mutation has a more
significant impact than b2 and the b1 b2 double mutant is more severe than either single
mutant but still synthesizes B-carotene derived xanthophylls. The most surprising result
from studies of mutant seed is that unlike lut] leaves, IutI seeds do not accumulate
zeinoxanthin, which is a signature phenotype for the lat] mutation in leaves. It is possible
that unlike leaf tissue, zeinoxanthin in lutl seed is degraded or unbound zeinoxanthin is
more readily accessible to the residual s-ring hydroxylation activity of LUT1 or a second
s-ring hydroxylation enzyme.

In contrast to leaf tissue, where individual and total carotenoids do not statistically

differ between ecotypes, WS seed accumulates more than twice the carotenoid level of

I75

Col (Table 4.2). Interpreting the impact of combined lutI, b1, and b2 mutations on seed
carotenoid profiles is confounded by the segregation of other quantitative loci in the Col
(lutI) and WS (b1 and b2) parental ecotypes.The only genotype combinations that can be
directly compared are lut] b1 and lut1b2, which were taken from seeds pooled from
twenty independent homozygous F2 plants. As observed in leaf tissue, the b1 mutation
has a more significant impact in the [w] background than b2. Like leaves, B-carotene
derived xanthophylls are still synthesized at reduced levels in b1 b2 and lut1b1b2 seeds,

despite the absence of both known B-hydroxylases. These results indicate that, an

additional B-ring hydroxylation activity is also present in non-photosynthetic tissue.

Conclusions

Our data reveal the multifunctional nature of carotenoid hydroxylases, provide
insights into the overlapping functions and interactions of the three known hydroxylases
in Arabidopsis and the regulation of carotenoid ring hydroxylations in viva. The
hydroxylase mutants also present additional examples of the remarkable flexibility of
plant photosystems with regard to carotenoid composition and will be important tools for
furthering our understanding of xanthophyll functions in plants, especially in response to
high light stress. Future studies will also focus on cloning of the LUT1 gene to determine

the biochemical nature of this novel e-hydroxylase and assess its role in the synthesis of

a- and B- carotene derived xanthophylls.

Methods

176

Identification of T-DNA mutant lines and construction of double and triple mutant

combinations

The b1 and b2 plants were identified from the T-DNA insertion mutant population
available through the University of Wisconsin Biotechnology Center (Madison,
Wisconsin; http://www.biotech.wisc.edu/Arabidopsisl) using the screening strategy
previously described (Krysan et al., 1999). The T-DNA border primer was used for
sequencing amplification products to determine the site of insertion into each gene
(Figure 4.2A).

Homozygous b1 and b2 plants were crossed and individual Fl seeds were grown
and self-fertilized to obtain the F2 generation. The b1 and b2 genes are located on
chromosome 4 and 5, respectively and therefore segregate independently in an F2
population. The genotype of the F2 individuals was determined by PCR using [3-
hydroxylase 1 and 2 specific primers as well as T—DNA border primer and homozygous
b1 and b2 plants selected. The primer sequences for screening and identifying
homozygous B-hydroxylase l and 2 T-DNA knockout mutants are: B-hydroxylase 1
(forward, 5’-TTAAACGCTTTTCTGTCTGTTACGTCGTC-3'; reverse, 5'-
TTGTGATTGTAGGTCACTCCCGATCATAG-3') and B-hydroxylase 2 (forward, 5'-
AATAGAAGTGGAGTGATTCGCTGTCGATG-3'; reverse, 5'-
AAGGACACATCGGTTCCAGAAGAAATAAG-3'), respectively.

Homozygous b1 and b2 single mutants were crossed with each other and to the

EMS mutagenized lutI-2 mutant. Homozygous b1b2, lutI b1, and lut1b2 were selected

177

from the segregating F2 population. Crosses were then performed between homozygous
[w] b] and lut1b2 plants and lutlb1b2 triple mutants were isolated from the F2

population.

RT-PCR and TaqMan real time PCR assay

Total RNA was extracted from wild type (WS), b1, and b2, respectively, as
described (Tian and DellaPenna, 2001). Ten pg of total RNA was used as template and
first-strand cDNA synthesis was conducted using the SuperScript Preamplification
System (Invitrogen, Carlsbad, CA) according to manufacture's protocol. Subsequent
amplification of the first-strand cDNA was performed using Taq DNA polymerase
(Promega, Madison, WI) and a MJ Research DNA EngineTM thermal cycler (Waltharn,
MA). Primers specific for lycopene e-cyclase (forward, 5'-
CGAACAAAAGAATCTCGCC-3'; reverse, 5'-AAACCCTI"GCCACATCCT-3') were
used as a control to test the integrity of all plant RNA samples. Primer pairs spanning T-
DNA insertion sites in B-hydroxylase l and 2 open reading frames, B-hydroxylase 1
(forward, 5'—AACCGCCGTTACATTCAAACC-3'; reverse, 5'-
ACCTACAGGGAAACGCTTG-3') and B-hydroxylase 2 (forward, 5'-
TTCTCCGCAAACCACCCTATA-3'; reverse, 5'-
CCAACTCTTCTTTTCCTCCCACTTC-B') were used to determine whether full-length
transcripts were synthesized in the corresponding mutants. The PCR program for

amplification was: 94°C for 3 min, 35 cycles of 94°C for 30 sec, 55°C for 30 sec, and

178

72°C for 2 min, and a final extension at 72°C for 10 min. The PCR products were
analyzed on a 1% agarose gel.

TaqMan real time PCR assay was adopted for quantification of steady-state
mRNA levels in the wild type and hydroxylase mutant plants. First-strand cDNA
products from three independent reverse transcription reactions were pooled and used for
TaqMan analysis. TaqMan probes and primers (Table 4.4) were designed from cDNA
sequences of Arabidopsis B-hydroxylase l (AY113923), B-hydroxylase 2 (AY117225),
lycopene B-cyclase (U50739), and lycopene s-cyclase (U5073 8) using Primer Express
sofiware (Perkin Elmer Applied Biosystems, Foster City, CA). Reporter (5' end) and
quencher (3' end) dyes for the TaqMan probes are FAM and TAMRA, respectively.
TaqMan real time PCR assay conditions and analysis were performed as previously

described (Tian and DellaPenna, 2001).

Plant material and HPLC analysis of pigment content

Two wild type Arabidopsis thaliana ecotypes (C01 and WS) and homozygous b1,
b2, lutI, b1b2, [w] b] , Iutlb2, lutI b1 b2 plants were grown in soil at 120-150 pmol
photons m"2 5" under a 12hr day: 12hr night (21°C: 18°C) cycle in growth chambers.
Carotenoids from six-week old leaf tissue were extracted in a 96-well format. A well-
exposed leaf from each plant was selected and same-size leaf area was excised using a
cork borer. A leaf disk was allocated in every tube in a 96-well rack (Dot Scientific Inc.,
Burton, MI), and 300 pl of 60:40 (v:v) Acetone : Ethyl Acetate followed by 200 pl

distilled water was added. Three 4 mm glass beads (Fisher Scientific, Pittsburgh, PA)

179

 

Table 4.4. Sequences of TaqMan probes and primers.

 

 

Gene Primer/Probe 5'-3' sequence
B-hydroxylase 1 forward CTCGTGCACAAGCGTITCC
reverse GGCGACCTITCGGAGGTAA
probe TGTAGGTCCCATCGCCGACGTCC
B-hydroxylase 2 forward CCCATTGCCAACGTI'CCTI'A
reverse TGTCTGTGTGGTGTAGCTGGTG
probe C‘ITCGAAAGGTCGCCGCCGC
B-cyclase forward TCATTACTGGCACGGAWCTTG
reverse GAGCGACAACCCGAAGACC
probe TCCAGGCTG‘I'I'I’CTCCCGGAACTG
e-cyclase forward CC'ITTGGTGCTGCCGC
reverse CTTCAGACAAAGATCTCACAACTGAAT
probe AGCATGGTACATCCCGCAACAGGC

 

I80

were added to each tube, plate was sealed, fastened onto a commercial paint shaker
(H.E.R.O. Industries, Burnaby, BC, Canada) and shaken for 5 min at the maximum
speed. The 96 well plate was then centrifuged in a Sorvall RT centrifuge (Kendro
Laboratory Products, Newtown, CT) at 3,750g for 5 min to obtain phase separation. The
resulting upper Ethyl Acetate phase was transferred into an Eppendorf tube and dried
under vacuum. Once samples were dried, they were resuspended in 100 pl
Acetonitrile:H2O:Triethylamine (900:9921, v:v:v) for HPLC analysis. Seed carotenoid
extractions were performed following similar procedures except that
Methanol:Chloroform (1 :1, vzv) was used as extraction solvent and the 500 ng of Tocol
(2-methyl-2-(4,8,l2-trimethyltridecyl)chroman-6-ol) standard was added to the extraction
buffer and used to calculate recovery. HPLC separation of the carotenoids and
quantitative analysis of carotenoid compositions were performed as described (Tian and
DellaPenna, 2001).

Chlorophyll content (Chl a, Chl b, and total) of 11 cm diameter leaf discs were
quantified spectrophotometrically as described (Porra et al., 1989). Briefly, 1 ml of
dimethylformamide was added to each leaf disc and extraction was carried out at room
temperature for 2 hr in darkness. The extraction mixture was centrifuged, the supernatant
was transferred to a 1 ml cuvette, and absorption at 647 nm and 664 nm was measured
and Chl a and Chl b content were calculated as Chl a = 12 A664-3.11A647 and Chl b =

20.78A647-4.88A664, respectively.

Chlorophyll fluorescence measurements

l8l

Six-week old wild type and mutant Arabidopsis plants were dark-adapted
overnight prior to chlorophyll fluorescence measurements. In viva chlorophyll
fluorescence was measured using a pulse amplitude modulation fluorometer (F M82, PP
Systems, Haverhill, MA, USA.) using attached leaves that had not been shaded during
growth. Fv/Fm is the maximum photochemical efficiency of PSII in the dark-adapted
state. NPQ was determined as Fm-Fm'/Fm' (Lokstein et al., 2002).

Upon request, all novel materials described in this publication will be made

available in a timely manner for non-commercial research purposes.

Acknowledgements

We would like to thank the Wisconsin Knockout facility for screening the

knockout mutants and Arabidopsis Stock Center for providing us with the mutant seed.

This work was supported by the National Science Foundation (Grant IBN-0131253).

182

References

Bouvier, F., Keller, Y., D'Harlingue, A., and Camara, B. (1998). Xanthophyll
biosynthesis: molecular and functional characterization of carotenoid hydroxylases from
pepper fruits (Capsicum annuum L.). Biochim. Biophys. Acta. 1391, 320-328.

Britton, G. (1998). Overview of carotenoid biosynthesis. In Carotenoids: Biosynthesis
and Metabolism, Vol. 3, G. Britton, S. Liaaen-Jensen, and H. Pfander, eds (Basel,
Switzerland: Birkhauser), pp. 13-147.

Cunningham, F.X. Jr., and Gantt, E. (1998). Genes and enzymes of carotenoid
biosynthesis in plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49, 557-583.

Davison, P.A., Hunter, C.N., and Horton, P. (2002). Overexpression of B-carotene
hydroxylase enhances stress tolerance in Arabidopsis. Nature 418, 203-206.

Demmig-Adams, B., Gilmore, A.M., and Adams, W.W. 3rd. (1996). Carotenoids 3: in
viva function of carotenoids in higher plants. FASEB J. 10, 403-412.

Hirschberg, J. (2001). Carotenoid biosynthesis in flowering plants. Curr. Opin. Plant
Biol. 4, 210-218.

Horton, P., Ruban, A.V., and Walters, RC. (1996). Regulation of light harvesting in
green plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 47, 655-684.

Hundle, E.S., O'Brien, D.A., Beyer, P., Kleinig, H., and Hearst J.E. (1993). In vitro
expression and activity of lycopene cyclase and beta-carotene hydroxylase from Erwinia
herbicala. FEBS Lett. 315, 329-334.

Krysan, P.J., Young, J.C., and Sussman, M.R. (1999). T-DNA as an insertional
mutagen in Arabidopsis. Plant Cell 11, 2283-2290.

Li, X.P., Bjdrkman, 0., Shih, C., Grossman, A.R., Rosenquist, M., Jansson, S., and
Niyogi, K.K. (2000). A pigment-binding protein essential for regulation of
photosynthetic light harvesting. Nature 403, 391-395.

Lokstein, H., Tian, L., Polle, J.E., and DellaPenna, D. (2002). Xanthophyll
biosynthetic mutants of Arabidopsis thaliana: altered nonphotochemical quenching of
chlorophyll fluorescence is due to changes in Photosystem II antenna size and stability.
Biochim. Biophys. Acta. 1553, 309-319.

Masamoto, K., Misawa, N., Kaneko, T., Kikuno, R., and Toh, H. (1998). B-carotene

hydroxylase gene from the Cyanobacterium Synechocystis sp. PCC6803. Plant Cell
Physiol. 39, 560-564.

183

Misawa, N., Nakagawa, M., Kobayashi, K., Yamano, S., Izawa, Y., Nakamura, K.,
and Harshima, K. (1990). Elucidation of the Erwinia uredovora carotenoid biosynthetic

pathway by functional analysis of gene products expressed in Escherichia coli. J.
Bacteriol. 172, 6704-6712.

Miiller, P., Li, X.P., and N iyogi, K.K. (2001). Non-photochemical quenching. A
response to excess light energy. Plant Physiol. 125, 1558-1566.

Niyogi, K.K., Grossman, A.R., and Bjtirkman, O. (1998). Arabidopsis mutants define
a central role for the xanthophyll cycle in the regulation of photosynthetic energy
conversion. Plant Cell 10, 1121-1134.

Phillip, D., Hobe, S., Paulsen, H., Molnar, P., Hashimoto, H., and Young, A.J.
(2002). The binding of xanthophylls to the bulk light-harvesting complex of photosystem

II of higher plants: a specific requirement for carotenoids with a 3-hydroxy-B-end group.
J. Biol. Chem. 277, 25160-25169.

Pogson, B., McDonald, K.A., Truong, M., Britton, G., and DellaPenna, D. (1996).
Arabidopsis carotenoid mutants demonstrate that lutein is not essential for photosynthesis
in higher plants. Plant Cell 8, 1627-1639.

Pogson, B.J., Niyogi, K.K., Bjiirkman, 0., and DellaPenna, D. (1998). Altered
xanthophyll compositions adversely affect chlorophyll accumulation and
nonphotochemical quenching in Arabidopsis mutants. Proc. Natl. Acad. Sci. USA 95,
13324-13329.

Porra, R.J., Thompson, W.A., and Kriedemann, P.E. (1989). Determination of
accurate coefficients and simultaneous equations for assaying chlorophylls a and b
extracted with four different solvents: verification of the concentration of chlorophyll
standards by atomic absorption spectroscopy. Biochim. Biophys. Acta 975, 384-394.

Rissler, H.M., and Pogson, B.J. (2001). Antisense inhibition of the beta-carotene
hydroxylase enzyme in Arabidopsis and the implications for carotenoid accumulation,
photoprotection and antenna assembly. Photosynth. Res. 67, 127-137.

Ruban, A.V., Lee, P.J., Wentworth, M., Young, A.J., and Horton, P. (1999).
Determination of the stoichiometry and strength of binding of xanthophylls to the
photosystem 11 light harvesting complexes. J. Biol. Chem. 274, 10458-10465.

Sun, Z., Gantt, E., and Cunningham, F.X. Jr. (1996). Cloning and functional analysis
of the B-carotene hydroxylase of Arabidopsis thaliana. J. Biol. Chem. 271, 24349-24352.

Tardy, F., and Havaux, M. (1996). Photosynthesis, chlorophyll fluorescence, light-

harvesting system and photoinhibition resistance of a zeaxanthin-accumulating mutant of
Arabidopsis thaliana. J. Photochem. Photobiol. B: Biol. 34, 87-94.

184

Tian, L., and DellaPenna, D. (2001). Characterization of a second carotenoid B-
hydroxylase gene from Arabidopsis and its relationship to the LUT1 locus. Plant Mol.
Biol. 47, 379-388.

185

CHAPTER 5

Xanthophyll Biosynthetic Mutants of Arabidopsis thaliana: Altered
Nonphotochemical Quenching of Chlorophyll Fluorescence is due to Changes in

Photosystem II Antenna Size and Stability

The work presented in this chapter has been published:
Lokstein, H., Tian, L., Polle, J.E.W., and DellaPenna, D. (2002). Biochim. Biophys.

Acta 1553, 309-319.

186

Abstract

Xanthophylls (oxygen derivatives of carotenes) are essential components of the
plant photosynthetic apparatus. Lutein, the most abundant xanthophyll, is attached
primarily to the bulk antenna complex, LHC II. We have used mutations in Arabidopsis
thaliana that selectively eliminate (and substitute) specific xanthophylls in order to study
their function(s) in viva. These include two lutein-deficient mutants, 1m] and lut2, the
epoxy xanthophyll-deficient aba] mutant and the IuIZabaI double mutant. Photosystem
stoichiometry, antenna sizes and xanthophyll cycle activity have been related to
alterations in non-photochemical quenching of chlorophyll fluorescence (NPQ). Non-
denaturing PAGE indicates reduced stability of trimeric LHC II in the absence of lutein
(and/or epoxy xanthophylls). Photosystem (antenna) size and stoichiometry is altered in

all mutants relative to wild type. Maximal ApH-dependent NPQ (qE) is reduced in the

following order: WT > aba] > [w] z lut2 > Iut20ba1, paralleling reduction in PS 11
antenna size. Finally, light-activation of NPQ shows that zeaxanthin and antheraxanthin
present constitutively in lut mutants are not qE active, and hence, the same can be
inferred of the lutein they replace. Thus, a direct involvement of lutein in the mechanism
of qE is unlikely. Rather, altered NPQ in xanthophyll biosynthetic mutants is explained
by disturbed macro-organization of LHC II and reduced PS II-antenna size in the absence
of the optimal, wild-type xanthophyll composition. These data suggest the evolutionary
conservation of lutein content in plants was selected for due to its unique ability to
optimize antenna structure, stability and macro-organization for efficient regulation of

light-harvesting under natural environmental conditions.

187

Introduction

Carotenoids (carotenes and their oxygenated derivatives, the xanthophylls) are
essential components of the photosynthetic apparatus in higher plants and their
composition comprising lutein, B-carotene, violaxanthin (V) as well as its deepoxidation
products and neoxanthin, is highly conserved throughout the plant kingdom. Lutein is the
most abundant carotenoid and is bound primarily to LHC II, the major light-harvesting
chlorophyll (Chl) a/b-binding complex of photosystem 11 (PS II). LHC II is composed of
mixed trimers of the closely related Lhcb1-3 gene products (Jansson, 1994). A 3.4 A
resolution structural model for trimeric LHC II reveals that in addition to 12 Chls, two
xanthophylls (assumed to be luteins) are centrally located between two membrane-
spanning or-helices in each monomeric subunit (Kflhlbrandt et al., 1994). Biochemical
analyses further indicate the presence of one neoxanthin and up to one V molecule per
LHC II monomer (Croce et al., 1999; Ruban et al., 1999). The minor LHCs of PS 11
(comprising the Lhcb4-6 gene products, also termed CP29, CP26 and CP24, respectively)
as well as LHC I proteins also bind xanthophylls to a varying extent (Thayer et al., 1992;
Bassi et al., 1993; Jansson, 1994; Ruban et al., 1999). The various xanthophylls play
pivotal, synergistic roles in the photosynthetic apparatus (Siefermann-Harms, 1987;
Paulsen, 1995; Niyogi, 1999). Xanthophylls are required for stable assembly of pigment-
protein complexes in vitro and in viva (reviewed in Paulsen, 1995), act as accessory light-
harvesting pigments (Siefermann-Harms, 1987), and along with B-carotene are of

paramount importance for photoprotection (for a recent review see Niyogi, 1999).

188

The primary photoprotective process in plants is the nonradiative dissipation of
energy absorbed in excess of what can be utilized in photosynthesis (reviewed in Niyogi,
1999). The phenomenon can be readily visualized as nonphotochemical quenching of Chl
fluorescence (N PQ). NPQ is a cooperative phenomenon - which despite intensive study -
is not yet fully understood at the mechanistic level. Several processes contribute to NPQ.
The major component (often termed qE, for energization-dependent quenching) is rapidly
reversible and associated with acidification of the thylakoid lumen, or trans-thylakoidal
ApH (for a review see Horton et al., 1994). Studies of plants with altered antenna
composition and isolated LHCs in vitro indicate that qE arises in the photosystem II
antenna system (Rees et al., 1989; Lokstein et al., 1993; Briantais, 1994; Harte] et al.,
1996)

Transthylakoidal ApH induces a variety of alterations to LHC II pigments and
proteins that collectively contribute to qE. A major consequence is protonation of specific
LHC amino acid residues that is proposed to cause conformational changes in antenna
organization facilitating qE in viva. ApH-mediated quenching can be mimicked in vitro
with isolated LHCs and has been used to establish the link between pH-induced LHC II
aggregation and quenching (Horton et al., 1994). In the same system, added xanthophylls,
most notably zeaxanthin (Z), antheraxanthin (A) and V, have differing effects on pH-
induced aggregation/quenching that are consistent with models proposed for qE.
Zeaxanthin production by the xanthophyll cycle in response to high-light stress has been
correlated with NPQ (for reviews see Demmig-Adams, 1990; Pfiindel and Bilger, 1990).
Violaxanthin deepoxidase (VDE) is activated by low pH and converts V (di-epoxide), via

A (mono-epoxide) into Z (epoxide-free) (see Pfilndel and Bilger, 1990). Z epoxidase

189

catalyzes the reverse reactions (in the dark or under low light) and is inhibited by low pH
(Pfiindel and Bilger, 1990). Thus, a second consequence of the ApH generated in high
light is the net formation of Z (+A) by the xanthophyll cycle. The combined protonation
of LHCs and production/interaction of Z (+A) with LHCs are two major factors that
cooperatively facilitate qE in viva.

Genetic analyses have substantiated and firrther defined the in-viva-roles of
specific xanthophylls and other, novel components of NPQ in algae and plants (N iyogi,
1999). Mutants defective in VDE activity (e.g., the Arabidopsis thaliana npq1 mutant)
are unable to form Z in response to high light. The npq1 mutant retains a rapid, initial
pH-dependent component of NPQ but is defective in a slower phase and has reduced
maximal amplitude compared to the wild type (WT) (N iyogi et al., 1997; Niyogi et al.,
1998). Conversely, mutations disrupting Z epoxidase activity (e.g., the A. thaliana aba]
mutant) accumulate constitutively high levels of Z, display accelerated NPQ induction
and, surprisingly, have decreased NPQ amplitude. Together aba] and npq1 have
confirmed and delineated the role for Z (+A) and the xanthophyll cycle in NPQ in viva. A
third mutation in A. thaliana , npq4, eliminates the qE component of NPQ despite having
an antenna pigment and protein composition, xanthophyll cycle activity and
photosynthetic rate indistinguishable from WT. The NPQ4 locus encodes the psbS
(CP22) protein and defines this protein as essential for NPQ (Li et al., 2000).

Another class of mutations shown to affect NPQ, the lat mutants, are defective in
specific aspects of lutein synthesis (Pogson et al., 1996). Lutein levels are reduced by
approximately 80% in [w] due to a disruption in e-ring hydroxylation and as a result

zeinoxanthin, the immediate biosynthetic precursor of lutein, accumulates. The lut2

I90

mutant is defective in e-ring cyclization and devoid of lutein and all other or-carotene-
derived xanthophylls. In addition to their lutein deficiency, both [at mutants also
accumulate elevated levels of B-carotene-derived xanthophyll cycle pigments (V, A and
Z). Similar to npq1, NPQ induction is retarded in both lut mutants and maximal
amplitude is decreased, suggesting a previously unsuspected role for lutein in NPQ
(Pogson et al., 1998).

As a group, the A. thaliana xanthophyll biosynthetic mutants are particularly well
suited for probing the function(s) of individual xanthophylls in viva and for providing
insight into the mechanistic contribution of specific xanthophylls to NPQ. The current
study was performed to provide a more detailed understanding of the various functions of
xanthophylls within the light-harvesting system, the ability and limits of various
xanthophylls to functionally complement for each other and the mechanistic basis for
altered NPQ in the mutants. The common theme emerging from this work is that altered
NPQ in xanthophyll biosynthetic mutants is due primarily to significant changes in LHC

Il-stability, PS II antenna size and superstructure.

Materials and methods

Plant Materials and Pigment Analysis

Wild type A. thaliana thaliana and the [w] , lut2, aba] and luIZabaI mutants

(Pogson et al., 1998) were grown in soilzvermiculite (3:1) at 100 pmol photons m'2 s'l

under a 10:14-h Ii ght/dark cycle (21/19 °C) in growth chambers. Leaves from 6-8-week

I91

old plants (prior to bolting) were used for all experiments. Pigment isolation and analyses
were performed as described (Pogson et al., 1996) except that tissue was frozen in liquid

nitrogen immediately following treatments and stored at -80 °C until extraction.

Photosystem Stoichiometry and Antenna Size Measurements

Isolation of thylakoid membranes and non-denaturing PAGE ("green gels") was
performed as described (Dormann et al., 1995). PS 11 and PS I concentrations were
assessed from the amplitude of light minus dark absorption changes of freshly prepared
thylakoid membrane suspensions at 320 nm (QA) and 700 nm (P700) (Melis and Brown,
1980). Functional Chl a/b-antenna sizes of PS I and PS 11 were obtained from analyses of

the kinetics of P700 photooxidation and QA-photoreduction, respectively (Melis, 1989).

Chlorophyll Fluorescence Measurements

In viva Chl fluorescence was determined using a pulse amplitude modulation
fluorometer (FMS 2, PP Systems, Haverhill, MA, USA) from attached leaves as
previously described (Lokstein et al., 1993). Fluorescence parameters are according to
(van Kooten and Snel, 1990). F.,/Fm = (Fm-Fo)/Fm is the maximum photochemical
efficiency of PS 11, in the dark-adapted state. NPQ was quantified as (F m/F m')-l. Photon
flux densities (PF Ds) were measured with a quantum sensor (Li-Cor, LI-189A; Lincoln,

NE, USA).

192

Results

Photochemical Efficiency and Nonphotochemical Chlorophyll Fluorescence

Quenching

Maximum photochemical efficiency of PS 11 (measured as F v[F m) was similar in
WT and the lutl and lut2 mutants, slightly lower in aba] and significantly lowered in
Iut2aba1 (Table 5.1). The lower Fv/Fm in aba] and lutZabaI most likely arise from a
higher PS I contribution to F0 and/or the presence of Z as a constitutive quencher of PS 11
fluorescence (refer also to later sections). NPQ induction kinetics for all genotypes at 500

2 s'1 actinic light are shown in Figure 5.1A. As has been previously

pmol photons m'
observed (Pogson et al., 1998), when compared to WT (half-rise time, t1/2 = 73 s) NPQ
induction is retarded in [w] (t1/2 = 87 s) and lut2 (t1/2 = 105 s) and considerably
accelerated in aba] (t1/2 = 11 s) and lut2aba1 (t1/2 = 14 s). Maximal NPQ levels attained
at 500 pmol photons m'2 s'1 are reduced in the order: WT > aba] > [w] z lut2 > IutZabaI .
With the exception of lut2aba1, NPQ induction in all genotypes can be described by a
single exponential rise, indicating that only one process, qE, contributes to NPQ at this
actinic light level. However, at 500 pmol photons In2 5'1 the lutZabaI curve shows, in
addition to the fast initial rise, a second slower component. This second component is
associated with a significant increase in F0 (relative to the initial dark-adapted F0 value)

after relaxation of qE (data not shown) and hence is attributable to photoinhibition. This

photoinhibitory component is not observed in the other genotypes following illumination

with 500 pmol photons m'2 s".

193

Table 5.1. Chl a/b-ratios, photosystem stoichiometries, antenna sizes and maximum photosystem

II efficiencies (FV/FM) in A. thaliana WT and xanthophyll mutants.

 

 

WT Iut1 lut2 aba1 lutZaba1

cm a/b-ratio 2.72 2.888 2.663 3.04 3.23
Fur... 0.840 0.843 0.846 0.805 0.783
cm a+b/P700 774 616 548 663 627

cm a+b/oA 527 409 351 466 319

PS ll/l (GA/P700) 1.47 1.51 1.56 1.42 1.97
cm a+b/PS II 3776' 275 213 3738 203

Chl a+b/PS | 220" 202‘ 202‘ 289 229b

 

Unless otherwise indicated all values are significantly different between genotypes (t-test,

significance level = 0.05). N > 30 for Chl a/b-ratio and Fv/Fm. N = 3 - 9 for all other

measurements. a"’not significantly different.

I94

_, ’13
2.01 A D’D’D D

- V-v—W‘v—v—v—v—v;
J ,
1.5 9’6-

 

 

g ‘ ff A’é
21.0-
01 / —13—w1’
05- A/ -O—Iut1
0/ —A—Iut2
—V—aba1
0.01 -<>-lut23ba1
0 ' 100 ' 200 ' 300 f 400
Time [8]

/0/ -El- WT
0 5 . o —O— Iut1
' / —A— lut2

. —-V- aba1

Iut2aba1
0.0 1 -<>-

 

 

fi

0 200 400 600 800 1000
Actinic Light [pmol photons/mas]

Figure 5.1. (A) Nonphotochemical fluorescence quenching (NPQ) in leaves of A.

thaliana WT and xanthophyll biosynthesis mutants, as a function of time of exposure to

500 pmol photons m'2 s". (B) NPQ as a fimction of actinic light intensity. All plants were

dark-adapted for 30 min prior to light exposure. Each data point is the mean of at least

three separate experiments. S.D. are within the symbol size.

195

Figure 5. 1 B displays the maximal NPQ levels reached after 6 min of illumination
as a function of incident PFD for all genotypes. Wild type, 11411 and lut2 curves start to

2 5'1 while the aba] and IutZabaI curves continuously

saturate at ~500 pmol photons m'
increase and are not saturated even at 1000 pmol photons rn'2 s". The maximal NPQ
levels reached at 1000 umol photons m'2 5'1 decrease in the order: WT z aba] > lut2 >
[w] > lut2aba1. This altered order (relative to 500 pmol photons m"2 s") is due to a
genotype-dependent, differential increase in the photoinhibitory component of NPQ
(increase in F0) at higher light levels as the capacity of qE build-up is exhausted (Figure
5.1B and data not shown). At higher PFDs this additional slower inducible NPQ-

component, indicative of susceptibility to photoinhibition, increases in the order: WT <

[w] < lut2 < aba] < luIZabaI.

Kinetics of Violaxanthin Deepoxidation in WT, [at] and lut2

A pivotal role for Z (and A) in NPQ is well-established (Demmig-Adams, 1990;
Horton et al., 1994; Pfiinder and Bilger, 1994; Harte] et al., 1996; Niyogi, 1999).
Therefore, we first investigated whether the compromised NPQ development in Int
mutants might be due to altered xanthophyll cycle activity and/or V+A+Z pool sizes.
Time-dependent changes in the relative amounts of V, A and Z in leaves of plants dark-
adapted for 24 h and then exposed to high light (1000 umol photons rn’2 s") are shown in
Figure 5.2. All data are expressed as mmol of the respective xanthophyll mol'l Chl a+b
instead of the more conventional % V+A+Z pool, as the V+A+Z pools in lutI and IutZ

were 2.5 and 3.8 fold larger, respectively, than in WT (Table 5.2 and Figure 5.2).

196

Table 5.2. Carotenoid composition of A. thaliana WT and xanthophyll mutant leaves.

 

 

WT Iut1 lut2 aba1 lutZeba1
Luteina 139 26 - 117 -
Neoxanthin 38 35° 35° - -
B-Carotene 81 91 114° 98 112°
Zeinoxanthin - 50 - - -
V+A+Z° 33 81 124° 122° 191
z Carotenoids 291° 284° 273 337 303

 

Unless otherwise indicated all values are significantly different between genotypes (t-test,

significance level = 0.05). N > 30 for all measurements.

a All carotenoid contents are expressed as mmoles carotenoid/moi Chl a+b.

" Not significantly different.

° Pool size of the xanthophyll-cycle pigments V, A and Z.

197

Figure 5.2. Kinetics of V deepoxidation via A to Z in leaves of A. thaliana WT (A)

2 l

and the [at] (B) and lut2 (C) mutants exposed to a PFD of 1000 pmol photons m' s' .
All plants were dark-adapted for 24 h prior to light exposure. 1], violaxanthin; A,

zeaxanthin; O, antheraxanthin. Each data point represents the mean of at least five

separate experiments. S.D. are within the symbol size. Note the different abscissa scales.

I98

Figure 5.2

[Xanthophyll Cycle Pigements][mmollmol Chl a+b]

35.0 1

17.5 -

 

 

 

 

0.0 .

 

 

A
AA/A A
r . -
goo-co 0— 4°
'1' ' I ' I T r f I
B
' fifl
/_______——
AA/A

 

 

 

 

 

3 a
A/A/

 

 

. (56028480 0 o

A?

 

 

‘— ' 1

0102030405060

Time [min]

199

In dark-adapted WT plants (Figure 5.2A), Z was absent and only traces of A were
detectable. The 60 min light treatment resulted in deepoxidation of 21.5 mmol V to 18.2
and 3.3 mmol Z and A, respectively, with 11.6 mmol V (35%) being recalcitrant to
deepoxidation. Both lut mutants (Figure 5.2B,C) have larger V+A+Z pools than WT and
contained high levels of A and Z in the dark-adapted state (17.1 and 24.9 mmol A and 6.5
and 6.2 mmol Z for [at] and lut2, respectively). The 60 min light treatment had little
effect on A levels but caused deepoxidation of 26 and 44.8 mmol V (mainly to Z) with
36.6 and 40.7 mmol V being recalcitrant to deepoxidation in [at] and lut2, respectively.
While these deepoxidation levels as attained afier 60 min illumination are much higher
than in WT, when one considers the time dependent product formation (Z+A) and that
NPQ is near maximal after 6 minutes of light treatment in all genotypes (Figure 5.1), the
xanthophyll cycle activity that may be relevant to NPQ is remarkably similar in all
genotypes. During the 6 min treatment leading to maximal NPQ 14.4, 11.1 and 22.1
mmol Z+A are formed with 50% product formation times of 2.5, 3.1 and 2.7 min, for
WT, [w] and lut2, respectively. In the light of these minor differences in rates and
product formation, it seems unlikely that altered xanthophyll cycle activity is the primary
cause of compromised NPQ in [at mutants.

The data in Figure 5.2 suggest that the large differences in deepoxidized and
recalcitrant V between Iut mutants and WT originate from the existence of at least two
functionally distinct sub-pools of V (and A+Z for the mutants) with differential
accessibility by VDE and Z epoxidase. One sub-pool is bound to the more peripherally
located xanthophyll (V, A, Z) binding sites of LHC-proteins and hence, is accessible by

the xanthophyll cycle enzymes. A second "protected" sub-pool (11.6 mmol V in WT and

200

60.2 and 71.8 mmol V+A+Z in [ml and lut2, respectively) is bound to sites within LHC
II and inaccessible by the xanthophyll cycle enzymes. In the case of the [w mutants, a
significant proportion of this recalcitrant V+A+Z sub-pool is apparently substituting for
lutein in the internally located lutein-binding sites of LHC II (Kiihlbrandt et al., 1994)
and unavailable to the xanthophyll cycle during light stress. Pigment compositional
analysis of isolated LHCs from WT and the mutants support this conclusion (refer to later
sections and Table 5.3). Finally and most significantly, the large amount of protected,
internally located Z+A (23.6 and 31.1 mol [at] and lut2, respectively) appears to not be
NPQ-active. To further test this hypothesis we performed NPQ light-activation

experiments in WT and the lut mutants.

Light-Activation of Nonphotochemical Quenching in WT, [at] and lut2

Light-activation of NPQ and its correlation with xanthophyll cycle activity is a
well-established phenomenon (Rees et al., 1989). Leaves of dark-adapted WT and lut
mutants were subjected to 500 pmol photons rrr'2 s'l for 6 min (inducing V deepoxidation
and NPQ build-up), subsequently dark-adapted for 2 min (sufficient to relax the qE
component of NPQ but not to re-epoxidate the newly formed Z+A) and then subjected to
a second high-light treatment (Figure 5.3).

NPQ induction in [w] and lut2 during the initial illumination period is delayed
compared to WT, consistent with the data in Figure 5.1A. However, during the second
illumination period, NPQ induction in both Iut mutants is similar to WT and is complete

in all three genotypes in less than 30 sec. This observation reveals the fiinctional

201

Table 5.3. Pigment composition of A. thaliana WT and xanthophyll mutant light-harvesting

complexes as resolved by non-denaturing electrophoresis as displayed in Figure 5.4.

 

 

Chl ba lutein neoxanthin VAZ XCarotenoids
WT-LHC 113 5.4 1.8 1.2 0.2 3.2
WT-LHCmono 5.4 1.7 1.1 0.2 3.0
lut1-LHCmono 5.3 0.3 1.0 1.5 2.8
rutz-LHCmono 5.5 n.d.” 1.3 2.0 3.3
aba1-LHCmono 5.8 1.5 n.d. 15" 3.0
lutZaba1-LHCmono 5.9 n.d. n.d. 35° 3.5

 

'All pigments are expressed as per 7 molecules Chl a; n = 3. SD. are about 10 %.
" Not detectable.

° Only zeaxanthin.

202

 

 

 

 

 

 

1.0; Actinic Light on E
0.8- W
0.6-

4
0.4-
0.2% U—i-L
0.0 . r 4 . . . . . . .
1.0-

* M1

0.8-
0.6-

l
04-

0:” MIN] LL44

0.0i '1-

1.0 -
0.8- ""2

 

 

 

 

Fluorescence (norm)

0.6 ~
0.4 ~

~ ll|_l|l Lu;

0-2 1" M
0.0

 

 

 

 

 

 

0 '1o'o izdo'ado V400 'so'o'
Time[s]

Figure 5.3: Light-activation of NPQ in leaves of A. thaliana WT, the [at] and lut2
mutants. Dark-adapted leaves were illuminated with a PFD of 500 pmol photons m"2 8"

for 6 min followed by 2 min darkness and a second illumination.

203

significance of the different pools of xanthophyll cycle pigments (V+A+Z) in the lut
mutants. Only Z (and possibly A) formed by xanthophyll cycle activity during the first
illumination period is NPQ active. This xanthophyll cycle (and NPQ-) active V+A+Z is
bound most likely to peripheral pigment binding sites of the LHCs. The large amounts of
constitutively present Z (and A) in dark-adapted lut mutants (bound to the lutein-binding
sites of LHC II) do not contribute to fast NPQ induction. Moreover, the same can be
logically inferred for lutein normally bound to these sites in WT. Thus, the hypothesis
that lutein is directly involved in the mechanism of NPQ (N iyogi et al., 1997; Pogson et
al., 1998) is not supported. An alternative explanation for compromised NPQ in the lut
mutants (and in xanthophyll-deficient A. thaliana mutants in general) would be more

indirect and result from alterations in LHC assembly/stability.

Assembly, Composition and Stability of Light-Harvesting Complexes in

Xanthophyll Mutants

In vitro reconstitution studies have shown that xanthophylls are indispensable for
stable assembly of monomeric LHCs (Paulsen, 1995). In order to investigate
assembly/stability of pigment-protein complexes in the various xanthophyll mutants,
thylakoid membranes were solubilized in mild (non-ionic) detergents and subjected to
non-denaturing PAGE ("green gels", Figure 5.4). Most striking is the considerably
reduced stability of trimeric LHC II in all xanthophyll mutants. Whereas the lat] mutant
retains a minor fraction of LHC ll-trimers (10-20 % of the WT level, corresponding

approximately to its residual lutein content), only a very faint trimeric LHC II band is

204

WT Iut1 lut2 Iut2aba1 aba1

w w PSI+PSI|
(LHC II)3
CP43I47

LHCmono

   

 

FP

Figure 5.4. Non-denaturing ("green") gel electrophoretic separation of pigment-
protein complexes from A. thaliana thylakoid membranes of the indicated
genotypes. Note that the LHCmono band also contains LHC I and minor LHC II
complexes, FP designates free pigment. The pigment composition of the LHC bands is

given in Table 5.3.

205

observed for aba1. The LHC II trimer band is absent in lut2 and luIZabaI, both of which
are completely deficient in lutein. Disappearance of trimeric LHC II is accompanied by a
concomitant increase of a band attributable to LHC II monomers (including also
monomeric LHC I and minor LHCs II, i.e., CP24, CP26 and CP29) in all mutant
genotypes. The free pigment band (almost exclusively carotenoids, V+A+Z as well as
zeinoxanthin in lutl) is least pronounced in WT, markedly stronger in all xanthophyll
mutants and most prominent in the lut2aba1 double mutant. Attempts to improve the
yield of mutant trimeric complexes in green gels through the use of alternate gel systems
(Peter and Thomber, 1991) and/or by altering detergent concentrations and/or
compositions (e.g., using SDS, n-decyl-B-D-maltoside, n-undecyl-B-D-maltoside, n-
dodecyl-B-D-maltoside or n-octyl-B-D-glucopyranoside and various combinations
thereof) were unsuccessful (not shown). Moreover the isolation of pure (trimeric) LHC II
with standard procedures (Burke et al., 1978; Krupa et al., 1987) also varying
solubilization conditions (e.g., 0.2 to 0.7 % Triton X-100 or 0.05 to 1.0 % n-dodecyl-B-
D-maltoside) proved to be impossible, too.

The bands corresponding to trimeric LHC 11 (WT) and monomeric LHCs (all
genotypes, note that the latter contains also the minor LHCs II and LHC I) were excised
and their pigment content analyzed by HPLC. The pigment composition of the bands is
given in Table 5.3. WT trimeric LHC II, WT monomeric LHC and aba1 monomeric
LHC pigment compositions were similar to those previously reported (Connelly et al.,
1997). In the lutl and lut2 monomeric LHCs lutein is replaced on a nearly equirnolar
basis by xanthophyll cycle pigments (V +A+Z). In lut2aba1 lutein, V and neoxanthin are

replaced by Z.

206

Photosystem Stoichiometry and Antenna Size in Xanthophyll Mutants

It has been shown previously that the extent of NPQ can be correlated with PS II
antenna size (Hartel et al., 1996). The xanthophyll mutants showed an increasing Chl a/b
ratio in the order: WT <lut1 ~ lut2 < aba1 < lut2abal (Table 5.1). Since Chl b is
associated exclusively with LHCs (mainly LHC II), the Chl a/b ratios are consistent with
a reduction in LHC II levels. However, alterations in photosystem stoichiometry alone or
in combination with altered antenna sizes could bring about similar changes in Chl a/b
ratios. To further address these possibilities, photosystem stoichiometry and antenna size
were analyzed in all genotypes (Table 5.1).

Light-induced absorption difference spectra of thylakoid membrane preparations
were used to determine the concentrations of QA and P700 as a measure of PS 11 and PS I
reaction centers, respectively (Melis and Brown, 1980). The corresponding PS antenna
sizes were derived from the kinetics of QA-photoreduction and Pmo-photooxidation
(Melis, 1989).

Lack of lutein leads to an increase in the ratio of PS 11 to PS I in the order:
lut2aba1 > lut2 > lutl > WT. Conversely, the lack of epoxidized xanthophylls
(neoxanthin and V) in aba1 leads to a decrease in the ratio of PS II to PS I relative to the
WT. With regard to PS I antenna size (Chl a+b/PS I), no clear tendency is observed in
relation to a specific xanthophyll deficiency. However, the lack of lutein is strongly
correlated with a reduction in PS 11 antenna size (Chl a+b/PS H) in the following order:

WT ~ aba1 > [at] > lut2 > lut2aba1. Remarkably, this reduction in PS 11 antenna size

207

parallels the decrease in maximum attainable NPQ (as measured at 500 pmol photons m’2

s'1 actinic illumination (cp. Figure 5.1 A) to avoid interference by photoinhibition)

Discussion

Xanthophylls are indispensable, integral components of plant photosynthetic
pi gment-protein complexes. In addition to their accessory light-harvesting function, they
are required for stable assembly of both antenna and reaction center complexes as well as
for effective photoprotection (see Paulsen, 1995). The ability of plants to dissipate excess
excitation energy (measurable as NPQ) is of paramount importance for photoprotection
and has been correlated with light-induced trans-thylakoidal ApH and Z formation (Rees
et al, 1989; Demmig-Adams, 1990; Hartel et al., 1996; Niyogi, 1999). It has also been
proposed that the xanthophyll cycle intermediate, A, may be involved in NPQ (Gilmore
and Yamamoto, 1993).

Mutations in A. thaliana have been useful in advancing our understanding of NPQ
in plants and fall into two general classes: those that affect NPQ without altering the
pigment compositions of unstressed leaves (e. g. npq1 and npq4) and those that affect
NPQ but also alter pigment compositions in unstressed leaves (e.g. lut1,lut2, aba] and
their double mutants). In aba] , NPQ induction is more rapid but attains lower amplitude
than in WT, while in [at] and lut2 both the induction and amplitude of NPQ are
negatively impacted (Figure 5.1A and Pogson et al., 1998). These effects are additive in
lutl aba1 and luIZabaI double mutants indicating the aba1 and lut mutations affect

independent processes related to NPQ (Figure 5.1A and Pogson et al., 1998). These

208

general observations can be extended to green algae - analogous mutations in
C hlamydamanas reinhardtii show similarly compromised NPQ (N iyogi et al., 1997).
Such studies have led to the proposal that in addition to light-induced trans-thylakoidal
ApH and Z formation, lutein is also required for efficient NPQ, either directly or
indirectly, by an unknown mechanism (N iyogi et al., 1997; Pogson et al., 1998). In the
current study we have compared and contrasted aspects of photosystem structure and
function across four A. thaliana xanthophyll mutant lines (lutI, lut2, aba] and lut2abal)
and WT in an attempt to identify a common mechanism(s) whereby xanthophyll
modifications differing from WT - in particular lutein deficiency - influence NPQ.

Given the importance and effects of the xanthophyll cycle (in particular Z
formation) in NPQ build-up, it would seem logical that any alterations in xanthophyll
cycle pigment pool size (V+A+Z), especially the levels of Z and A prior to light stress,
would have an impact on NPQ induction. This is indeed the case; however, instead of
positively affecting NPQ as one might expect, the high constitutive levels of Z (and A)
and ~3-fold increase in V+A+Z pools in lut mutants coincide with compromised NPQ
induction kinetics and amplitudes. Perhaps the compromised NPQ in the mutants is due
to an alteration in xanthophyll cycle activity/kinetics, rather than absolute xanthophyll
cycle pigment levels. This appears not to be the case either, as xanthophyll cycle activity
(V to Z conversion rates and amounts) during the timeframe of NPQ induction was
remarkably similar in Iut mutants and WT.

The experiments in Figure 5.2 show that approximately 65 % of the enlarged
V+A+Z pool in lut mutants is "protected" from xanthophyll cycle activity. This is most

likely due to these pigments occupying binding sites in LHCs that would normally (in

209

WT) bind lutein. Given the role of Z (and possibly A) as NPQ-active xanthophylls, one
would again anticipate that any Z (or A) would increase NPQ or at least predispose the
system to NPQ. However, NPQ light activation experiments clearly show that the large
amounts of Z+A (occupying lutein sites) in the lut mutants do not contribute to NPQ
indicating that, as in WT, only Z (and A) produced by the xanthophyll cycle is active in
NPQ in lut mutants. The observations are significant for several reasons. First, they
provide direct evidence that the location of a particular xanthophyll (e.g. Z and A) within
the pigment-protein complexes is as important - if not more important - than the identity
of the pigment for determining its role in NPQ. Second, since Z (and A) incorporated into
lutein-binding sites do not directly participate in NPQ, it follows that the lutein that
normally occupies these sites in WT also does not directly participate in NPQ. Still, the
question remains: If lutein is not directly involved in NPQ and substituted Z (and A) is
not "NPQ-active", by what mechanism does the absence of lutein in lut mutants affect
NPQ?

Although xanthophyll biosynthetic mutations affect many aspects of the plant
photosynthetic apparatus, when taken together, our studies of LHC stability, antenna size
and photosystem stoichiometry indicate that the effect of xanthophyll compositional
changes on NPQ results primarily from alterations to the PS 11 antenna. Non-denaturing
PAGE analysis of pigment-protein complexes indicates LHC II trimer stability is
substantially reduced across the spectrum of A. thaliana xanthophyll mutants studied,
though most severely in those that lack lutein (Figure 5.4). This is consistent with
previous reports of LHC II instability in a or-carotenoid-free mutant of the green alga

Scenedesmus abliquus (Bishop, 1996; Heinze et al., 1997). It is most likely that the

210

destabilization of LHC II trimers in A. thaliana xanthophyll mutants is brought about by
subtle structural changes in monomeric LHC II units due to the binding of V+A+Z
instead of lutein.

A clear inverse correlation exists between the maximal attainable NPQ and the
reduction of functional PS II antenna size across the full spectrum of mutants studied
(Figure 5.1 and Table 5.1). We propose that this is the primary, underlying lesion that
affects NPQ in all xanthophyll biosynthetic mutants. The lack of a WT xanthophyll
composition leads to instability of LHC II trimers. This reduction of LHC II trimer
stability is likely translated into a disruption of optimal, higher-order, macro-organization
of the PS II-antenna system (and probably less connectivity between PS II centers) and
hence, is responsible for the compromised NPQ. This is consistent with previous findings
that NPQ decreased in parallel with the reduction of PS 11 antenna size in a Chl b-less
barley mutant (Lokstein et al., 1993; Briantais, I994; Harte] et al., 1996). Interestingly,
Chl b-less barley mutants grown under an intermittent light regime (Hartel et al., 1996)
show a NPQ phenotype that closely resembles the A. thaliana npq4 mutant (Li et al.,
2000). Intermittent light grown Chl b-less barley mutants are devoid of all major and
minor LHCs but retain WT levels of the psbS protein (Bossmann et al., 1997), while the
npq4 mutant is unaffected in PS II-antenna composition but completely lacks psbS (Li et
al., 2000). This indicates that both psbS and an optimally organized and functional PS 11
antenna system are required for maximum NPQ build-up.

The xanthophyll composition of plant photosynthetic pigment-protein complexes
is remarkably conserved across evolution. This is presumably due to strong selection

pressure for retention of these particular xanthophylls in order to enable specific

211

ftmction(s) and/or the resultant fitness of land-based photosynthetic organisms. In
particular, lutein is the most abundant thylakoid carotenoid and its presence is diagnostic
for land-based photosynthetic organisms. Moreover, lutein is present in the branch of
algae that presumably gave rise to land plants. Thus, the initial identification of lutein-
deficient A. thaliana mutants, which lacked an obvious phenotype, and the subsequent
generation of xanthophyll double mutants that lacked all WT xanthophylls with relatively
minor observable whole plant phenotypes were both surprising and puzzling (Pogson et
al., 1996). If, as the single and double A. thaliana xanthophyll mutations imply, the plant
photosynthetic apparatus is so flexible with regard to xanthophyll composition, why do
all plants produce lutein, have such a highly conserved xanthophyll composition and
finally, why doesn't one find plants in nature with xanthophyll profiles similar to the A.
thaliana mutants?

The observation that all single and double A. thaliana xanthophyll biosynthetic
mutants are defective in aspects of NPQ suggested that plants with "non-WT"
xanthophyll compositions are impaired in their ability to cope with light in excess of what
is needed for photosynthesis (Pogson et al., 1998). The current study clearly shows that
although other xanthophylls (in particular V, A and Z) are able, to a certain degree, to
assume the structural role of lutein in the mutants, the resulting "non-WT" xanthophyll
compositions severely compromise antenna size, stability and function. These
"replacement" xanthophylls do not provide the same unique combination of antenna
stability, flexibility and function that lutein does to fulfill the requirement of land plants
for fast and effective regulatory and protective responses under rapidly fluctuating

environmental conditions. In this light, the strong selection for maintenance of a specific

212

xanthophyll composition in plant thylakoids during evolution is easily understood. This
may also explain why only descendents of green algae with their seemingly redundant
manifold of Chl a/b-binding and xanthophyll containing antenna complexes that confer
unique adaptive flexibility (the xanthophyll cycle and NPQ) have ever successfirlly

conquered land.

Acknowledgements

We would like to thank Prof. Anastasios Melis (U C Berkeley) for providing

access to his photosystem stoichiometry/antenna size measuring facilities. The authors

also acknowledge Erica Rust and Aimee Snell for assistance with the HPLC analysis.

213

References

Bassi, R., Pineau, B., Dainese, P., and Marquardt, J. (1993). Carotenoid-binding
proteins of photosystem 11. Eur. J. Biochem. 212, 297-303.

Bishop, NJ. (1996). The [Le-carotenoid, lutein, is specifically required for the formation
of the oligomeric forms of the light harvesting complex in the green alga, Scenedesmus
abliquus. J. Photochem. Photobiol., B: Biol. 36, 279-283.

Bossmann, B., Knoetzel, J., and Jansson, S. (1997). Screening of chlorina mutants of
barley (Hardeum vulgare L.) with antibodies against light-harvesting proteins of PS I and
PS 11: Absence of specific antenna proteins. Photosynth. Res. 52, 127-136.

Briantais, J .M. (1994). Light harvesting chlorophyll a/b complex requirement for
regulation of photosystem II photochemistry by nonphotochemical quenching.
Photosynth. Res. 40, 287-294.

Burke, J.J., Ditto, CL, and Arntzen, C.J. (1978). Involvement of the light-harvesting
complex in cation regulation of excitation energy distribution in chloroplasts. Arch.
Biochem. Biophys. 187, 252-263.

Connelly, J.P., Miiller, M.G., Bassi, R., Croce, R., and Holzwarth, AR. (1997).
Femtosecond transient absorption study of carotenoid to chlorophyll energy transfer in
the light-harvesting complex 11 of photosystem II. Biochemistry 36, 281-287.

Croce, R., Remelli, R., Varotto, C., Breton, J., and Bassi, R. (1999). The neoxanthin
binding site of the major light harvesting complex (LHCII) from higher plants. FEBS
Lett. 456, 1-6.

Demmig-Adams, B. (1990). Carotenoids and photoprotection: a role for the xanthophyll
zeaxanthin. Biochim. Biophys. Acta 1020, 1-24.

Diirmann, P., Hoffmann-Benning, S., Balbo, I., and. Benning, C. (1995). Isolation and
characterization of an Arabidopsis mutant deficient in the thylakoid lipid digalactosyl
diacylglycerol. Plant Cell 7, 1801-1810.

Gilmore, A.M., and Yamamoto, H.Y. (1993). Linear models relating xanthophylls and
lumen acidity to non-photochemical fluorescence quenching. Evidence that
antheraxanthin explains zeaxanthin-independent quenching. Photosynth. Res. 35, 67-78.

Hfirtel, H., Lokstein, H., Grimm, B., and Rank, B. (1996). Kinetic studies on the

xanthophyll cycle in barley leaves - Influence of antenna size and relations to
nonphotochemical chlorophyll fluorescence quenching. Plant Physiol. 110, 471-482.

214

Heinze, 1., Pfiindel, E., Hiihn, M., and Dau, H. (1997). Assembly of light harvesting
complexes 11 (LHC II) in the absence of lutein: A study on the alpha carotenoid free
mutant C-2A'-34 of the green alga Scenedesmus abliquus. Biochim. Biophys. Acta 1320,
188-194.

Horton, P., Ruban, A.V., and Walters, RC. (1994). Regulation of light-harvesting in
higher plants - indication of nonphotochemical quenching of chlorophyll fluorescence.
Plant Physiol. 106, 415-420.

Jansson, S. (1994). The light-harvesting chlorophyll a/b-binding proteins. Biochim.
Biophys. Acta 1184, 1-19.

Krupa, Z., Huner, N.P.A., Williams, J.P., Maissan, E., and James, D.R. (1987).
Development at cold-hardening temperatures - the structure and composition of purified
rye light harvesting complex II. 84, 19-24.

Kiihlbrandt, W., Wang, D.N., and Fujiyoshi,Y. (1994). Atomic model of plant light-
harvesting complex by electron crystallography. Nature 367, 614-621.

Li, X.P., Bjr’irkman, 0., Shih, C., Grossman, A.R., Rosenquist, M., Jansson, S., and
Niyogi, K.K. (2000). A pigment-binding protein essential for regulation of
photosynthetic light harvesting. Nature 403, 391-395.

Lokstein, H., Hfirtel, H., Hoff'mann, P., and Renger, G. (1993). The role of light
harvesting complex 11 in excess excitation energy dissipation - an in viva fluorescence
study on the origin of high energy quenching. J. Photochem. Photobiol., B: Biol. 19, 217-
225.

Melis, A., and Brown, J.S. (1980). Stoichiometry of System I and System 11 reaction
centers and of plastoquinone in different photosynthetic membranes. Proc. Natl. Acad.
Sci. USA 77, 4712-4716.

Melis, A. (1989). Spectroscopic methods in photosynthesis - photosystem stoichiometry
and photosystem antenna size. Phil. Trans. R. Soc. Lond. B 323, 397-409.

Niyogi, K.K., Bjiirkman, 0., and Grossman, AR. (1997). The roles of specific
xanthophylls in photoprotection. Proc. Natl. Acad. Sci. USA 94, 14162-14167.

Niyogi, K.K., Grossman, A.R., and Bjr'irkman, O. (1998). Arabidopsis mutants define
a central role for the xanthophyll cycle in the regulation of photosynthetic energy
conversion. Plant Cell 10, 1121-1134.

Niyogi, K.K. (1999). Photoprotection revisited: genetic and molecular approaches. Annu.
Rev. Plant Physiol. Plant Mol. Biol. 50, 333-359.

215

Paulsen, H. (1995). Chlorophyll a/b-binding proteins. Photochem. Photobiol. 62, 367-
382.

Peter, G.F., and Thomber, J.P. (1991). Biochemical composition and organisation of
higher plant Photosystem 11 light harvesting proteins. J. Biol. Chem. 266, 16745-16754.

Pfiindel, E. and Bilger, W. (1994). Regulation and possible function of the violaxanthin
cycle. Photosynth. Res. 42, 89-109.

Pogson, B., McDonald, K.A., Truong, M., Britton, G., and DellaPenna, D. (1996).
Arabidopsis carotenoid mutants demonstrate that lutein is not essential for photosynthesis
in higher plants. Plant Cell 8, 1627-1639.

Pogson, B.J., Niyogi, K.K., leirkman, O., and DellaPenna, D. (1998). Altered
xanthophyll compositions adversely affect chlorophyll accumulation and
nonphotochemical quenching in Arabidopsis mutants. Proc. Natl. Acad. Sci. USA 95,
13324-13329.

Rees, D., Young, A., Noctor, G., Britton, G., and Horton, P. (1989). Enhancement of
the ApH-dependent dissipation of excitation energy in spinach chloroplasts by light
activation: Correlation with the synthesis of zeaxanthin. FEBS Lett. 265, 85-90.

Ruban, A.V., Lee, P.J., Wentworth, M., Young, A.J., and Horton, P. (1999).
Determination of the stoichiometry and strength of binding of xanthophylls to the
photosystem II light harvesting complexes. J. Biol. Chem. 274, 10458-10465.

Siefermann-Harms, D. (1987). The light-harvesting and protective functions of
carotenoids in photosynthetic membranes. Physiol. Plantarum 69, 561-568.

Thayer, S.S., and Bjr’irkman, O. (1992). Carotenoid distribution and deepoxidation in
thylakoid pigment-protein complexes from cotton leaves and bundle-sheath cells of
maize. Photosynth. Res. 33, 213-225.

van Kooten, O., and Snel, J.F.H. (1990). The use of chlorophyll fluorescence
nomenclature in plant stress physiology. Photosynth. Res. 25, 147-150.

216

CHAPTER 6

Future directions

217

Over the past decade, most of the carotenoid biosynthetic genes have been cloned
from various organisms and some of which were further characterized for enzyme
activity and function (reviewed in Cunningham and Gantt, 1998; Hirschberg, 2001).
However, much remains to be learned about the mechanisms of carotenoid biosynthetic
enzymes and reactions. This dissertation focused on cloning of the e-hydroxylase gene
(Chapter 2 and 3) and has addressed the overlapping functions of the carotenoid
hydroxylases in Arabidopsis (Chapter 4) as well as xanthophyll functions (especially
lutein) in photosystem structure and photoprotection (Chapter 5). During the course of
this dissertation research, many more questions and hypotheses were generated than
could be answered and tested. In order to understand the mechanism of carotenoid
hydroxylation reactions, the developmental responses of plants towards carotenoid
hydroxylase deficiency, and the fitness of hydroxylase mutants under various stress
conditions, the following future experiments are proposed to ftu'ther our knowledge in
these areas. Experiments proposed in section 6.1.1 and 6.1.2 have been performed and
preliminary results obtained during the course of this dissertation research. Future
research should be focusing on the mechanism of hydroxylation reactions (6.3.) and the

response of various hydroxylase mutants to the stress conditions (6.4.).

6.1. Developmental responses to carotenoid deficiency in Arabidopsis hydroxylase

mutants

6.1.1. ABA biosynthesis in the hydroxylase deficient mutants

218

ABA (an apocarotenoid) is a sesquiterpene derived from C40 carotenoids. ABA is
a plant growth regulator involved in induction of seed dormancy and adaptation to a
variety of stresses (reviewed in Giraudat et al., 1994). The regulation of the stress
adaptation process is partially due to de nova synthesis of ABA. Therefore, understanding
ABA synthesis is essential for understanding plant stress accommodation. A direct
correlation between ABA production and epoxy-carotenoids, violaxanthin and
neoxanthin, has been established by labeling experiment and characterization of the aba

mutant (Creelman et al., 1987; Rock and Zeevaart, 1991).

ABA deficient mutants have played critical roles in studying ABA biosynthesis
and function (Milborrow, 2001). Some of these mutants were isolated due to their wilting
or viviparous phenotypes. Previously identified mutants in the carotenoid biosynthetic
pathway have also shown viviparous phenotypes, e. g. VPS mutant in maize is ABA
deficient and is a lesion in the phytoene desaturase gene; VP3 and VP7 may encode
lycopene cyclases in maize, etc. In addition, the aba1 mutation in Arabidopsis is a
mutation of the zeaxanthin epoxidase gene.

The B-hydroxylase double mutant b1b2 and hydroxylase triple mutant lut1b1b2
isolated during this dissertation research produce very low amounts of the ABA
precursors, violaxanthin and neoxanthin (less than 5% and 2% of total carotenoids,
respectively; refer to Table 4.1). Unlike the ABA deficient mutants, the hydroxylase
mutants do not show wilting or germination defective phenotypes. This result suggests
that although produced in very limited amount, the violaxanthin and neoxanthin pool in
the hydroxylase mutants is able to produce sufficient ABA for seed germination and

other ABA-related physiological processes under normal growth conditions.

219

 

In order to test this hypothesis, the ABA level in wild type WS and Col, as well as
b1 b2 and lut1b1b2 mutants were measured in collaboration with Dr. Jan Zeevaart (Plant
Research Laboratory, Michigan State University). It has been previously observed that
the wilting of leaves induces a sudden and rapid increase in ABA content ("stress" ABA;
Milborrow, 1981). Therefore, ABA levels in both normal growth and water stressed
plants (for 5 hr) were measured. Our preliminary data showed that b1 b2 and lut1b1b2
mutants accumulate similar amount of ABA compared to the wild type (~0.2 ug/g tissue)
under normal growth condition. When these plants were subject to drought stress, the
ABA level in wild type plants were induced up to 10 fold, while only a very moderate
increase (~0.1 ug/g tissue) was observed in the hydroxylase mutants. These results
indicate that although the hydroxylase mutants can synthesize sufficient ABA for normal
growth, stressed induced ABA production is greatly reduced in these mutants. The
carotenoid compositions of drought-stressed plants (wild type, hydroxylase mutants, and

aba1-5 control plant) are currently being characterized.

6.1.2. Chloroplast ultrastructures in the hydroxylase mutants

The increased Chl a/b ratios in the hydroxylase double and triple mutants (Table
4.3, Chapter 4) suggested an alternation in photosystem stoichiometry due to the presence
of dramatically decreased amounts of xanthophylls. In order to determine whether the
xanthophyll deficiency affects the structure of the photosynthetic apparatus, chloroplasts
in the wild type and mutants mesophyll cells were visualized by transmission electron

microscopy (TEM).

220

TEM study revealed abnormal chloroplast ultrastructures in both types of
hydroxylase mutants. b1b2 mutant showed reduced thylakoid grana stacking, which
indicates that the structure of the photosynthetic apparatus has been adversely affected by
xanthophyll deficiency and xanthophylls are involved in grana stacking either directly or
indirectly. Xanthophylls mainly bind to the light harvesting complex proteins in the
thylakoid membrane. Chl b deficient mutants grown under intermittent light lack LHC
proteins and have thylakoid membranes organized into large parallel arrays, indicating
that LHC proteins are necessary for grana stacking (Krol et al., 1995). Therefore, the
relative levels of LHC proteins in the wild type and hydroxylase mutant plants should be
quantified with specific antibodies.

Unlike the b1 b2 double mutant, lut1b1b2 triple mutant has a normal grana
stacking pattern, however there is a large increase in starch accumulation (as indicated by
the size of the starch grains) compared to wild type plants. Starch accumulation is
frequently observed under stress conditions such as phosphate deficiency, magnesium
deficiency, high light etc (Fischer et al., 1998; Weston et al., 2000; Zakhleniuk et al.,
2001). The triple mutant was grown under moderate light with nutrients identical to wild
type control plants, it is possible that the triple mutant was more stressed than wild type
plants under such grth conditions. Starch accumulation could be the consequence of
decreased degradation or increased synthesis. Therefore, starch content during a
prolonged dark period will be measured in the triple mutant. To further analyze the cause
of starch accumulation in the hydroxylase triple mutant, the enzymes involved in starch
synthesis and degradation, soluble sugar levels, and anion contents (e.g. phosphate)

should also be tested.

22]

6.1.3. Lipid composition in the hydroxylase mutants

Membrane lipids form bilayers in the cell. Chloroplast membranes primarily
contain galactolipids. MGDG accounts for more than 50% of the lipids in the thylakoid
membrane and LHC II has a high affinity for MGDG. As mentioned in section 6.1.2, the
structure of the thylakoid membrane has been altered in the hydroxylase mutants, which
suggests that lipid levels, esp. MGDG, may be correspondingly altered in these mutants.
In addition, MGDG is also essential for binding of violaxanthin and zeaxanthin produced
by the xanthophyll cycle to the LHC proteins (Morosinotto et al., 2002). The carotenoid
composition change in the hydroxylase mutants, drastically decreased violaxanthin and
neoxanthin in particular, may affect the membrane lipid composition and membrane
fluidity; therefore, these parameters in the hydroxylase mutants should be further

analyzed.

6.2. Mutants screening

T-DNA knockout (null) mutants for B-hydroxylase l and 2 (b1 and b2,
respectively) were isolated and crossed to generate b1 b2 double mutant, in which both 8-
hydroxylase activity were eliminated (Chapter 4). The b1 b2 double mutant was further
introduced into the lutl (EMS mutation, now known to be leaky) background in order to

address the possible overlapping function of the hydroxylases.

222

The LUT1 gene was cloned during this dissertation research and a corresponding
knockout mutant was isolated. In the leaves of the lutI knockout mutant, there is no
lutein accumulation indicating it is a null allele of the s-hydroxylase gene. Although the
previously generated lut1b1b2 triple mutant was informative as to the overlapping
hydroxylase functions, we were unable to eliminate the possible residual function of the
s-hydroxylase due to the potential leakiness of the lutl EMS mutation. It is therefore
critical to cross the lutI knockout mutant to b1 b2 and screen for putative triple
hydroxylase knockout mutant. Both HPLC and PCR based screening are currently
employed to isolate the putative homozygous triple mutant.

In the lut1b1b2 triple mutant, the production of lutein and B-carotene derived
xanthophylls was greatly reduced. However, the amount of zeinoxanthin, a
monohydroxyl precursor of lutein, was more than doubled in the triple mutant compared
to wild type. Zeinoxanthin bears one hydroxylated B-ring and may partially compensate
for the lack of lutein and B-carotene derived xanthophylls. In order to address this
question, lut2 mutation was introduced to the b1 b2 background. lut2 is a mutation in the
e-cyclase gene, in which zeinoxanthin and lutein production is entirely abolished. [3-
hydroxylase 2 gene (At5g52570) and 8-cyclase gene (At5g57030) locate to the bottom
arm of chromosome 5 and are tightly linked (only 7-8 cM apart), therefore they are not
independently segregating. Out of 288 F2 plants from lut2 and b1 b2 cross, only
heterozygous mutants are isolated due to this tight linkage. The heterozygous mutants are
currently growing on soil and the F 3 generation will be screened for possible homozygous

lut2b1 b2 mutant.

223

6.3. Structural and mechanistic basis of the B- and e-hydroxylases

The substrates for B- and 8- ring hydroxylase are very similar except for the
double bond position on their ring structures. The previously cloned B-hydroxylases are
non-heme diiron type monooxygenase (Bouvier et al., 1998). Like the membrane fatty
acid desaturases, they require ferredoxin, NADPH, NADPH-ferredoxin oxido-reductase
for activity. e-hydroxylase gene was cloned during this dissertation research and shown
to be a cytochrome P450 type monooxygenase that requires NADPH, O2, and CPR for
activity. However, the underlying mechanisms for these different types of hydroxylation
reactions are still not known. In the future, biochemical analyses of substrate specificity
and enzyme kinetics should be carried out to further characterize [3- and 8- ring
hydroxylase activity. In addition, crystallization of both types of hydroxylases will
provide valuable information on substrate binding, enzyme conformation and will further
our knowledge on the reaction mechanisms. A better understanding of the hydroxylation
reactions will also add insights into possible carotenogenic complexes proposed by
Cunningham and Gantt (1998).

e-hydroxylase can be engineered into bacteria, such as cyanobacteria, to produce
lutein that is not normally present. Metabolic engineering in plants can be carried out by
either abolishing c-hydroxylase gene expression to divert carbon flow into B-carotene
production. Lutein is only synthesized in land plants and green algae. Putative homologs

of e-hydroxylases have been identified in monocots, dicots, and diatoms. An important

question remains to be addressed is how the e-hydroxylase evolved.

224

 

6.4. Hydroxylase mutants response to environmental stresses

In chapter 4 and 5 of this dissertation, the function of xanthophylls in NPQ was
addressed using mutants deficient in carotenoid biosynthetic genes. Carotenoids have
also established their roles as antioxidants to scavenge reactive oxygen species, especially
under stress conditions (Mittler, 2002). When plants are exposed to high light, free
radicals and oxidizing species are generated and released within the chloroplasts. Future
studies on the hydroxylase mutants should focus on characterizing the responses of these
plants to high light stress. Other abiotic stresses, such as salt, drought and desiccation,

temperature etc. should also be applied and mutant responses characterized.

225

References

Bouvier, F., Keller, Y., D'Harlingue, A., and Camara, B. (1998). Xanthophyll
biosynthesis: molecular and functional characterization of carotenoid hydroxylases from
pepper fruits (Capsicum annuum 1..) Biochim. Biophys. Acta. 1391, 320-328.

Creelman, R.A., Gage, D.A., Stuits, J.T., and Zeevaart, J.A.D. (1987). Abscisic acid
biosynthesis in leaves and roots of Xanthium strumarium. Plant Physiol. 85, 726-732.

Cunningham, F.X. Jr., and Gantt, E. (1998). Genes and enzymes of carotenoid
biosynthesis in plants. Ann. Rev. Plant Physiol. Plant Mol. Biol. 49, 557-583.

Fischer, E.S., Lohaus, G., Heineke, D., and Heldt, H.W. (1998). Magnesium
deficiency results in accumulation of carbohydrates and amino acids in source and sink
leaves of spinach. Physiol. Plant. 102, 16-20.

Giraudat, J., Parcy, F., Bertauche, N., Gosti, F., Leung, J., Morris, P.C., Bouvier-
Durand, M., and Vartanian, N. (1994). Current advances in abscisic acid action and
signalling. Plant Mol. Biol. 26, 1557-1577.

Hirschberg, J. (2001). Carotenoid biosynthesis in flowering plants. Curr. Opin. Plant
Biol. 4, 210-218.

Krol, M., Spangfort, M.D., Huner, N.P.A., Oquist, G., Gustafsson, P., and Jansson,
S. (1995). Chlorophyll a/b-binding proteins, pigment conversions, and early light-induced
proteins in a chlorophyll b-less barley mutant. Plant Physiol. 107, 873.

Milborrow, B.V. (1981). Abscisic acid and other hormones. In: Paleg, L.C., Aspinall, D.
eds. The physiology and biochemistry of drought resistance in plants. Sydney: Academic
Press, 348-388.

Milborrow, B.V. (2001). The pathway of biosynthesis of abscisic acid in vascular plants:
a review of the present state of knowledge of ABA biosynthesis. J. Exp. Bot. 52, 1145-
1 164.

Mittler, R. (2002). Oxidative stress, antioxidants and stress tolerance. Trends Plant Sci.
7, 405-410.

Morosinotto, T., Baronio, R., and Bassi, R. (2002). Dynamics of chromophore binding
to the Lhc proteins in viva and in vitro operation of the xanthophyll cycle. J. Biol. Chem.
277, 36913-36920.

Rock, C.D., and Zeevaart, J.A.D. (1991). The aba mutant of Arabidopsis thaliana is
impaired in epoxy-carotenoid biosynthesis. Proc. Natl. Acad. Sci. USA 88, 7496-7499.

226

 

Weston, E., Thorogood, K., Vinti, G., and Lopez-Juez, E. (2000). Light quantity
controls leaf-cell and chloroplast development in Arabidopsis thaliana wild type and
blue-light-perception mutants. Planta 211, 807-815.

Zakhleniuk, O.V., Raines, C.A., and Lloyd, J.C. (2001). pha3: a phosphorus-deficient
mutant of Arabidopsis thaliana (L.) Heynh. Planta 212, 529-534.

227

APPENDIX A

Analysis of carotenoid biosynthetic gene expression during marigold petal

development.

The work presented in this appendix has been published:
Moehs, C.P., Tian, L., Osteryoung, K.W., and DellaPenna, D. (2001). Plant Mol. Biol.

45, 281-293.

228

Abstract

Marigold (T agetes erecta L.) flower petals synthesize and accumulate carotenoids
at levels greater than 20 times that in leaves and provide an excellent model system to
investigate the molecular biology and biochemistry of carotenoid biosynthesis in plants.
In addition, marigold cultivars exist with flower colors ranging from white to dark orange
due to >100-fold differences in carotenoid levels, and presumably similar changes in
carbon flux through the pathway. To examine the expression of carotenoid genes in
marigold petals, we have cloned the majority of the genes in this pathway and used these
to assess their steady state mRN A levels in four marigold cultivars with extreme
differences in carotenoid content. We have also cloned genes encoding early steps in the
biosynthesis of isopentenyl pyrophosphate (IPP), the precursor of all isoprenoids,
including carotenoids, as well as two genes required for plastid division. Differences
among the marigold varieties in the expression of these genes suggest that differences in
mRNA transcription or stability underlie the vast differences in carotenoid synthesis and

accumulation in the different marigold varieties.

Introduction

Carotenoids are a large family of C40 isoprenoid pigments that are found in all
higher plants as well as many bacteria and fungi. Over 600 different carotenoid structures
have been identified (Starub, 1987). Carotenoids play essential roles in photosynthesis in
the green tissues of higher plants. They have structural and functional roles in the light
harvesting antennae and serve as photoprotective compounds by quenching triplet
chlorophyll and singlet oxygen derived from excess light energy (Demmig-Adams and
Adams, 1996). The role of the carotenoids that accumulate to high levels in many fruits
and flowers is thought to be to serve as visual attractants for insects and animals to aid in
pollination and seed dispersal.

Interest in carotenoid synthesis has also been stimulated by the various roles that
carotenoids play in the human diet (DellaPenna, 1999; Hirschberg, 1999). Some B-ring
containing carotenoids are precursors of vitamin A in the human diet. Vitamin A
deficiency is a significant cause of blindness and early mortality, particularly in countries
where rice, which is low in vitamin A, is the main staple in the diet. In addition, diets rich
in carotenoids, such as lycopene, that lack B rings and thus do not have provitamin A
activity have other health-promoting effects (Mayne, 1996). The nutritional benefits of
carotenoids have led to metabolic engineering of crops deficient in carotenoids for
increased production of these compounds (Burkhardt et al., 1997; Shewmaker et a1.,
1999; Ye et al., 2000).

The first committed step in the biosynthesis of carotenoids is the head to head

condensation of two molecules of geranylgeranyl pyrophosphate to form the colorless

230

intermediate phytoene. This reaction is catalyzed by the enzyme phytoene synthase
(Harker and Hirschberg, 1998). In plants, four subsequent desaturation reactions are
catalyzed by two enzymes, phytoene desaturase and Q-carotene desaturase, that each
introduces two symmetrical double bonds to yield lycopene. Subsequently, the ends of
the linear carotenoid lycopene can be cyclized and various oxygen functions introduced
to form the xanthophylls.

Although the biochemical reactions that lead to the synthesis of the carotenoids
have been known since the mid-19605, the majority of the genes encoding enzymes of the
pathway have only been identified in the past decade (reviewed by Cunningham and
Gantt, 1998). Carotenoid biosynthesis occurs in the plastids, yet all of the genes are
nuclear encoded. Biochemical characterization of the enzymes responsible for
carotenogenesis has been hampered by the fact that many of them are membrane
associated and difficult to purify in an active state. In addition, the steady state mRNA
levels of many of the enzymes are very low in photosynthetic tissues and can only be
detected by very sensitive techniques such as RT-PCR (Giuliano et al., 1993).

During the development of fruits and flowers, plastids differentiate into organelles
referred to as chromoplasts that are specialized for the sequestration of lipophilic
molecules such as carotenoids (Camara et a], 1995). The process of carotenoid
accumulation in chromoplasts has been studied in a number of non-photosynthetic plant
tissues such as tomato and pepper fruits, and daffodil flowers. During chromoplast
development in tomato fruit, it is known that differential transcriptional regulation of
carotenoid biosynthetic genes plays a critical role in determining the amounts and types

of carotenoids that accumulate. While the mRN As encoding the genes phytoene synthase

231

and phytoene desaturase are strongly induced during fruit development (Giuliano et al.,
1993), the mRNAs encoding cyclization enzymes are down-regulated (Pecker et al.,
1996), consistent with the accumulation of lycopene in tomato fruits. Similar evidence for
the importance of phytoene synthase and phytoene desaturase in the regulation of
carotenoid accumulation in pepper fruits has also been obtained (Bouviet et a1., 1994).

In the current study, we characterize the carotenoid biosynthetic pathway in
marigolds, T agetes erecta L. Marigolds are a self-fertile, diploid species (2n=24) best
known for their large showy flowers. Marigold varieties exist that range in petal color
from white to dark orange. The pigmentation of the flowers is due to the massive
synthesis and accumulation of carotenoids during the 7-10 days of petal development.
The dark orange colored varieties contain concentrations of carotenoids that are up to 20-
fold greater than in marigold leaves and 20 times the concentration of carotenoids found
in ripe tomato fruit. For this reason, marigolds are grown commercially as a source of
carotenoids and a meal made from the petals is used as an animal feed supplement
(Delgado-Vargas and Paredes-Lopez, 1997).

The massive increase in the concentration of carotenoids in marigold petals in
orange varieties and the existence of numerous, genetically distinct mutants that cloning
from marigold and expression analysis of many of the genes required for carotenoid and
IPP biosynthesis as well as two of the genes required for plastid division. Differences in
the expression of these genes during petal development among four varieties that
accumulate vastly different amounts of carotenoids yields insights into the role transcript
abundance accumulate lower amounts of petal carotenoids makes marigold an excellent

system to examine the regulation of flux through this pathway. In tomatoes and peppers,

232

other experimental systems used to study carotenoid biosynthesis; the available mutants
generally do not affect flux but rather are blocked at individual biosynthetic steps of the
pathway. This highlights the advantage of using marigold petals as an experimental
system to study flux through the carotenoid pathway. In this work, we describe the plays

in carotenoid biosynthesis and chromoplast development in this system.

Materials and methods

Plant materials

Marigold seeds were from the W. Atlee Burpee Company, Clinton, Iowa. They
include Dark Orange Lady (wild type) and three mutant cultivars, Golden Lady (yellow
flowers), Primrose Climax (pale yellow flowers), and French Vanilla (white flowers).
Petal development was identical in the various cultivars and arbitrarily divided into six
stages from which RNA and carotenoids were extracted. Stage 1 petals were
approximately 2-4 mm in length, stage 2 petals were approximately 5-7 mm in length,
stage 3 and stage 4 petals were approximately 8-10 and 13-15 mm in length, respectively.
Stage 5 petals were defined as partially opened flowers (> 2 cm), while stage 6 petals
represented the petals from fully opened flowers. Color development in wild type

marigold petals is first visible at stage 3.

Molecular techniques

233

Because carotenoid accumulation in marigold petals increases most rapidly in
stages 3 and 4, a petal cDNA library was made from poly A RNA isolated from stage 3
and 4 petals from the variety Dark Orange Lady (cDNA library was constructed as a
service by Stratagene, La Jolla, CA). Total RNA was isolated by LiCl precipitation
(Sambrook et al., 1989) and poly A RNA was obtained by two passes of the RNA over an
oligo-dT cellulose column.

Hybridization of the marigold lambda library with Arabidopsis genes as
heterologous probes was performed at 50° C using 6X SSPE; filters were washed in 2X
SSC, 0.1% SDS at 50° C. Plaque screening and isolation was done according to standard
techniques (Sambrook et al., 1989). Single clone excision and mass excision of the
lambda library to yield phagemids were done as recommended the manufacturer
(Stratagene, La J olla, CA). Isolated plasmids were sequenced by primer walking on both
strands using an ABI dRhodamine cycle sequencing kit. Marigold cDNAs isolated using
Arabidopsis heterologous probes were phytoene synthase, phytoene desaturase, €-
carotene desaturase, lycopene e-cyclase and B-carotene hydroxylase. A marigold or
tubulin cDNA was identified by random sequencing of marigold library clones and used
as a constitutive control.

To analyze steady state mRNA levels for the described marigold genes, poly A
RNA was isolated as described above from pools of stage 1 and 2 petals, stage 3 and 4
petals, and stage 5 and 6 petals of the indicated cultivars. Tissue was harvested in liquid
nitrogen and either used immediately for RNA isolation or stored at —80° C. Poly A RNA

prepared by two passes of the RNA over an oligo-dT cellulose column was separated by

electrophoresis on 1% MOPS-forrnaldehyde agarose gels (2 ug/lane) and blotted to nylon

234

 

membranes (Micron Separations, Westborough, MA). Probes were prepared from the
isolated marigold gene fragments by random primer labeling (Gibco BRL, Rockville,
MD). Equivalent numbers of counts were used for all hybridizations. Northern blots were
hybridized overnight at 42° C in 30% formamide and washed under low stringency
condition using 2X SSC at 50° C. An exposure was made using a phosphorimager
(Molecular Bioimager, BioRad, Hercules, CA); subsequently, the blots were washed
using stringent conditions of 0.1X SSC at 60 ° C and an additional exposure was made.
The northern blots using the Arabidopsis l-deoxy-D-xylulose reductoisomerase EST
(ABRC EST clone 120E8T7) and the marigold or-tubulin genes as probes were washed at

1X SSC at 55° C.

Color complementation

Marigold genes encoding lycopene [3 cyclase, l-deoxy-D-xylulose synthase,
IPP isomerase, GGDP synthase, and the plastid division genes MinD and FtsZ were
isolated by functional expression in E. coli that had been engineered to express various
carotenoids (Cunningham et al., 1994). Mass excision of the marigold lambda library and
color complementation screening were performed as described (Cunningham et al.,
1996). E. coli (strain DHSor) was grown in LB supplemented with chloramphenicol to
maintain the carotenoid-gene containing plasmid and with arnpicillin (100ug/ml) to
maintain the marigold library plasmids. Marigold lycopene B cyclase was identified as a

yellow, B-carotene containing colony in a background of pink, lycopene accumulating

colonies due to the presence of pAC-LYC (Cunningham et al., 1994). Marigold l-deoxy-

235

D-xylulose synthase was identified in the same screen as a much darker red colony in a
background of paler pink colonies (Moehs et al., 1998). Marigold IPP isomerase and the
plastid division gene MinD were isolated during a low temperature screen. Zeaxanthin
producing E. coli cells (harboring pAC-ZEAX, Sun et al., 1996) grew more slowly than
control E. coli when incubated at 18 ° C. Marigold IPP isomerase and marigold MinD
were selected based on the observation that the presence of these genes allowed
zeaxanthin containing E. coli cells to grow more rapidly and accumulate more pigment
than in their absence. Marigold GGDP synthase was isolated using a derivative of pAC-
BETA (Cunningham et al., 1996) whose E. herbicola GGDP synthase had been deleted.
E. coli cells harboring this plasmid are colorless. When the marigold library was
introduced into this background, rare yellow, B-carotene producing colonies were
expressing a complementing marigold GGDP synthase. The marigold plastid division

gene F tsZ was isolated fortuitously during a color complementation screen.

Carotenoid extraction, quantification, and HPLC analysis

Carotenoids were extracted from leaves and from different stages and different
varieties of marigold petals according to Pogson et al. (1996). Following saponification,
the extract was washed free of alkali and the organic phase was dried under vacuum.
Absorption determinations were made in ethanol at 444 nm using an extinction
coefficient of 2500. HPLC separation of carotenoids was performed as described by

Norris et al. (1995).

236

 

 

Results
Isolation of marigold genes

Figure A.1 depicts the pathway and enzymes for carotenoid biosynthesis in plants
leading from glyceraldehyde-3-phosphate and pyruvate to lutein and zeaxanthin. The
genes in the pathway that we have cloned from marigold are listed in Table A1
Phytoene synthase, phytoene desaturase, Q-carotene desaturase, lycopene B-cyclase,
lycopene e-cyclase, and B-carotene hydroxylase are genes specific to the carotenoid.
Additional cDNAs cloned include l-deoxy-D-xylulose synthase (DXP synthase), IPP
isomerase, and GGDP synthase that catalyze steps of the central isoprenoid pathway
common to the synthesis of carotenoids and other plastid derived terpenoids such as
tocopherols, plastoquinone, and chlorophylls. Cloning of these cDNAs was accomplished
either by using the corresponding Arabidopsis cDNAs as heterologous probes, or by
using a function-based E. coli cloning technique termed color complementation
(Cunningham and Gantt, 1998). -

Marigold cDNAs encoding phytoene synthase, phytoene desaturase, C-carotene
desaturase, lycopene s-cyclase, and B-carotene hydroxylase were isolated using
Arabidopsis cDNAs as heterologous probes. In general, the marigold genes were highly
homologous to the corresponding Arabidopsis genes at the protein level with the percent
identity ranging between 70 and 80%. The identity is even higher when the less
conserved N-terrninal putative plastid targeting signals are not considered. Identity at the

DNA level was lower, averaging about 60%. Table A.1 lists characteristics of the isolated

237

Figure A.l. Non-mevalonate isoprenoid and carotenoid biosynthetic pathway in
plants. Enzymes are abbreviated as follows: DXPS, l-deoxy-D-xylulose-S-phosphate
synthase; DXR, 1-deoxy-D-xylulose-5-phosphate reductoisomerase; ?, additional
unknown enzymes of non-mevalonate pathway; IPP, isopentenyl pyrophosphate
isomerase; GGDP, geranylgeranyl diphosphate synthase; PSY, phytoene synthase; PDS,

phytoene desaturase; ZDS, Q-carotene desaturase; Beta cyc, B-ring cyclase; Eps eye, 8-

ring cyclase; Beta hydrox, B-ring hydroxylase; lEps hydrox, e-ring hydroxylase.

238

Pyruvate + Glyceraldehyde-3-P

1 DXPS
1-Deoxy-D-xylulose-5-P
' DXR
2-C-methylerythritoI-4-P
/ \ 7
IPP 3': DMAPP "’1’
1 GGDP
GGDP (x2)
I PSY
phytoene
I; PDS
C-carotene
1' ZDS
Eps cyc ly,cope:e Beta cyc
6-carotene y-carotene
Beta eye 1 5 Beta cyc
or-carotene B-carotene
Beta Hydrox Beta Hydrox
l ' l
Eps Hydrox 1 Beta Hydrox
lutein zeaxanthin

Figure A.1

239

 

Table A. 1. Information about the genes discussed in this report including the number of
nucleotides of each clone, the number of amino acids encoded by each cDNA, and the
accession number is presented. Abbreviations: DXPS, l-deoxy-D-xylulose-5-phosphate
synthase; IPPI, isopentenyl pyrophosphate isomerase; GGDPS, geranylgeranyl
diphosphate synthase; PSY, phytoene synthase; PDS, phytoene desaturase; ZDS, C-
desaturase; Beta Cyc, B-ring cyclase; Eps Cyc, e-ring cyclase; Beta Hydrox, B-ring

hydroxylase; FtsZ and MinD, plastid division genes FtsZ and MinD, respectively.

 

 

% Protein identity with Accession

cDNA Nucleotides Amino acids

Arabidopsis homologue number

DXPS 2291 725 67 AF251020
IPPI 987 232 83 AF25101 1
GGDPS 1176 361 66 AF251012
PSY 1371 422 70 AF251015
PDS 1861 551 79 AF251014
ZDS 2061 526 77 AF251013
Beta Cyc 1906 512 74 AF251017
Eps Cyc 1830 517 69 AF251016
Beta Hydrox 996 308 65 AF251018
FtsZ 1357 410 76 AF251346
MinD 1164 296 75 AF251019

 

240

marigold cDNAs.

Color complementation was used to isolate marigold genes encoding l-deoxy-D-
xylulose synthase (DXP synthase), IPP isomerase, GGDP synthase, and lycopene B-
cyclase, as well as the plastid division genes MinD and F tsZ. This technique relies on the
ability of E. coli to accumulate various carotenoids when carotenoid biosynthetic genes
are expressed from a plasmid (Cunningham et al., 1996). A second plasmid from the
library of interest can be introduced into this background with a different selectable
marker and a compatible replicon to enable the selection of colonies with the desired
phenotype. This method has proven to be effective for isolating carotenoid biosynthetic
genes from a number of organisms (Cunningham and Gantt, 1998). The gene encoding
lycopene B-cyclase was isolated using E. coli containing the plasmid pAC-LYC, which
caused the bacteria to accumulate lycopene, resulting in a pink appearance of the
colonies. The mass-excised marigold cDNA library was introduced into this background
and four colonies of approximately 3.5 x 105 colonies screened were identified based on

their yellow color. Subsequent sequencing of the marigold library plasmids confirmed

these cDNAs encode marigold lycopene B-cyclase.

Interestingly, several dark red colonies were also identified in the course of this
screen. These colonies accumulated greater than two fold more lycopene (data not
shown) than the pink pAC-LYC containing E. coli. Sequencing of several of these
colonies identified the presence of a gene with strong homology to Arabidopsis l-deoxy-
D-xylulose synthase (DXP synthase), the first committed step in an alternative, non-
mevalonate route to isopentenyl pyrophosphate (Lichtenthaler, 1999). This enzyme

catalyzes a transketolase-like reaction in which pyruvate and glyceraldehyde-3-phosphate

241

are condensed to form l-deoxy-D-xylulose. The fact that the presence of the marigold
DXP synthase more than doubled the amount of lycopene that accumulated in E. coli
screen suggests that the activity of this enzyme is limiting for carotenoid synthesis in E.
coli. Similar results have been reported by Harker and Brarnley (1999) who found that
overexpression of DXP synthases from Bacillus subtilis and Synechocystis sp.6803 in
lycopene producing E. coli likewise led to an increase in lycopene accumulation. It is
interesting to speculate that the activity of this enzyme catalyzes a rate-limiting step in
plant plastid isoprenoid biosynthesis, although this has not been definitively shown.
Certainly, its activity is necessary as evinced by the albino phenotype of Arabidopsis
mutants lacking this enzyme (Mandel et al., 1996).

Additional functional expression screens were used to isolate the remaining genes
discussed in this report. Marigold GGDP synthase was isolated using a deletion
derivative of pAC-BETA (Cunningham et al., 1996). pAC-BETA contains the E.
herbicola genes encoding GGDP synthase, phytoene synthase, phytoene desaturase, and
lycopene B cyclase, and consequently E. coli containing this plasmid accumulates B
carotene and is yellow in color. Colorless E. coli colonies result when they are
transformed with a derivative of pAC-BETA from which the GGDP synthase has been
deleted. We introduced the marigold library into this GGDP synthase deficient
background and identified two colonies that were yellow in a background of
approximately 3 x 105 non-pigmented colonies. These colonies contained marigold
library plasmids expressing a complementing GGDP synthase gene.

Finally, a low-temperature screen led to the isolation of marigold genes encoding

the isoprenoid enzyme IPP isomerase and the plastid division protein MinD. This screen

242

was based on the observation that E. coli engineered to accumulate zeaxanthin (Sun et al.,
1996) grow more slowly at 18° C than E. coli harboring the vector plasmid alone and
accumulated significantly less zeaxanthin than when they were grown at 37° C. The basis
of this temperature dependent phenotype is unknown. When the marigold cDNA library
was transformed into zeaxanthin containing E. coli, we identified numerous rapidly
growing, highly pigmented colonies in a background of pale, slow growing colonies.
Plasmids isolated from several of these colonies were sequenced, and similarity searches
against the publicly available databases revealed IPP isomerase homologous sequences,
as well as a marigold gene with similarity to the E. coli MinD protein which is required
for proper cell division in E. coli (de Boer et al., 1989). A marigold homologue of the
Arabidopsis plastid division gene AtFtsZI-I was also identified fortuitously during an
unrelated expression screen. Antisense inhibition of AtF tsZI -1 expression in Arabidopsis
has been shown to greatly reduce plastid division, leading in some cases to mesophyll
cells containing a single enlarged plastid (Osteryoung et al., 1998). This and other
evidence indicates that bacterial and plastid division are evolutionarily related
(Osteryoung and Pyke, 1998). The basis of increased growth of E. coli harboring the
marigold plastid division gene MinD may lie in its complementation of a cold-sensitive
aspect of E. coli cell division.

The majority of the genes we isolated appear to represent full-length cDNAs with
the exception of phytoene synthase which lacks a small number of amino acids including
an initiating methionine at its N-terminus when aligned with the corresponding
Arabidopsis thaliana protein. Analysis of the conceptual translation of all of the reported

cDNAs with the chloroplast targeting signal recognition program ChloroP (Emanuelsson

243

et al.,1999) indicated that each contained a high confidence plastid targeting signal with
the exception of the marigold e-cyclase and the marigold IPP isomerase. When the
corresponding Arabidopsis e-cyclase and IPP isomerase were similarly analyzed with
ChloroP, the e-cyclase was found to contain a high confidence plastid targeting signal
while the IPP isomerase did not. Since the products of these genes are expected to be
chloroplast targeted, the plastid targeting of these and the other enzymes will have to be

experimentally confirmed.

Analysis of carotenoid levels in different varieties of marigolds

Total carotenoids were extracted from the leaves and petals of marigold varieties
whose petals ranged in pigmentation from white to dark orange. Flowers of the varieties
examined are depicted in Figure A.2B; the stages of petal development are shown in
Figure A.2A. While the four varieties contained vastly different amounts of carotenoids
in their petals, all of the varieties contained approximately equal quantities of carotenoids
in their leaves (Figure A.3). An HPLC profile of the carotenoids found in the leaves of
the wild-type dark orange cultivar is shown in Figure A.4A.This profile is similar to that
observed in most higher plant chloroplasts (Ryberg et al., 1993) where lutein is the most
abundant carotenoid, followed in abundance by B-carotene, violaxanthin, and neoxanthin.
An identical profile is also observed for leaves of the other three marigold cultivars (data
not shown).This indicates that the mutation(s) that alter flux through the carotenoid
Figure A.2.pathway in petals of the different varieties do not affect carotenoid synthesis

in leaves.

244

Figure A.2. A. Stages of marigold petal development are depicted. Petal development is
identical in the various cultivars and was arbitrarily divided into six stages from which
RNA and carotenoids were extracted. B. Fully expanded flowers from the four marigold
cultivars discussed in this report are shown.

Images in this dissertation are presented in color.

245

 

12345 6

Stages of Marigold Petal Development

 

French Vanilla Primrose Climax

Golden Lady Dark Orange Lady

Marigold Cultivars
Figure A.2

246

In contrast to leaves, the cultivars differ dramatically in the accumulation of
carotenoids in their petals. During wild type petal development, there is both a
quantitative and a qualitative change in the carotenoids that accumulate. As shown in
Figure A.3, there is a sharp induction of carotenoid biosynthesis in wild type petals that is
evident by stage 3 and continues until the petals are fully expanded. This induction leads
to the accumulation of greater than 20—fold more carotenoids in petals than in the leaves
and more than a 100-fold increase during petal development. In contrast, the other
varieties show a much-reduced induction of carotenoid biosynthesis, and carotenoid
accumulation does not appear to increase beyond stages 3 and 4. The ratios of
carotenoids present in marigold petals are also different from those found in leaves.
Figure A.4B shows the HPLC profile of an unsaponified extract of carotenoids from
petals of the wild type cultivar. The large peaks eluting between 30 and 35 minutes are
the carotenoid lutein esterified on its ring hydroxyl groups to various fatty acids.
Saponification of this extract yields lutein almost exclusively as shown in Figure A.4C.
Esterification of carotenoids is frequently seen in chromoplasts (Camara et al.,1995) and
the identification of esterified carotenoids in marigold petals has been previously reported
(Quackenbush and Miller, 1972; Gau et al. ,1983; Rivas, 1989). Thus, in wild-type petals
there is both a strong induction of overall synthesis and an alteration of the carotenoid
pathway in favor of the or-carotene-derived xanthophylls, which contain a B and an a
cyclic end group, as opposed to the B-carotene-derived xanthophylls, which contain two
B cyclic end groups.

All of the varieties in this study accumulate primarily esterified lutein in their

petals, albeit in vastly different amounts. This indicates the mutations resulting in the

247

 

 
   

in 1:1 FV
0 _
3 200° I PC
2 . G.
._ _
3, DOL
.5
o 1000‘
5
2
8 _
D
:1.

04

 

 

 

 

r

leaf stages 1+2 stages 3+4 stage 6
petal development stages

Figure A.3. Carotenoid levels in leaves and during various stages of petal
development in the different marigold cultivars. FV, French Vanilla; PC, Primrose

Climax; GL, Golden Lady; DOL, Dark Orange Lady.

248

 

Ca

 

Esterified L

 

Absorbance (440 nm)

 

 

 

 

 

L
‘ AL Ar..—
C. L
0 10 20 30
Time (min)

Figure A.4. HPLC analysis of pigments from marigold leaves and petals of the
cultivar Dark Orange Lady. A. Profile of leaf pigments. B. Profile of an unsaponified
extract of stage 6 petals. C. Profile of a saponified extract of stage 6 petals.

Abbreviations: N, neoxanthin; V, violaxanthin; L, lutein; Esterified L, esterified lutein;

Cb, Chlorophyll b; Ca, Chlorophyll a; b, B-carotene.

249

Primrose Climax, Golden Lady and French Vanilla varieties are not only petal specific
but also primarily quantitative in their effect, that is they do not block individual steps of
the carotenoid pathway or cause accumulation of biosynthetic intermediates. This
suggests that differences in the expression pattern of carotenoid biosynthetic pathway
genes among the different varieties would relate primarily to the regulation of flux

through the pathway and not to the specific type of carotenoid produced.

Expression analysis of isolated genes

To determine the steady state mRNA levels of the genes we have cloned, poly A
RNA was isolated by oligo-dT cellulose chromatography from each of the four marigold
varieties and from different petal developmental stages. Equal amounts of poly A RNA
(2 jig/lane) were probed with equivalent numbers of counts from labeled probes made
from each of the genes. In Figures A5 and A6, the ratio of expression of each of the
isoprenoid and carotenoid biosynthetic genes at petal developmental stage 3 and 4
relative to their expression in wild type leaves is shown. The Northern blots are shown as
insets to allow an assessment of the relative expression levels of all of the genes to be
made. Based on the hybridization with a marigold or-tubulin, the expression of the
characterized genes in the French Vanilla and the Primrose Climax varieties may be
somewhat over-estimated since or-tubulin mRN As in petals of these cultivars are slightly
higher than wild type.

It is apparent that in the dark orange variety the majority of the genes described,

including early steps in the isoprenoid pathway as well as the genes specific for all steps

250

of the carotenoid pathway, accumulate to considerably higher levels during petal
development compared to their accumulation in leaves. In addition, a comparison among
the different varieties allows one to draw several conclusions regarding the contribution
of transcript abundance of particular pathway steps to regulating flux through the
carotenoid pathway. As expected from its position as the first committed enzymatic step
in the carotenoid pathway, the expression of phytoene synthase appears to be most
closely correlated with the levels of carotenoids and hence flux into the pathway. Despite
the nearly ten-fold difference between the amount of carotenoids in fully expanded petals
of the yellow flowers of Golden Lady and the orange flowers of Dark Orange Lady
(Figure A.3), the abundance of transcripts of pathway genes beyond phytoene synthase is
nearly identical. However, phytoene synthase steady state mRNA levels were reduced at
least four-fold relative to the wild type dark orange variety (Figure A6). The conclusion
that the expression and/or stability of phytoene synthase transcripts is a key requirement
for the accumulation of carotenoids is also supported by its virtually undetectable
expression in the varieties with pale yellow or white flowers. These varieties have much
reduced expression of most of the characterized genes, but of the carotenoid pathway-
specific genes; phytoene synthase mRNA levels are reduced to the greatest extent.
Interestingly, the expression of the DXP synthase, the first step in the non-mevalonate
route to IPP is also greatly reduced in the mutant cultivars whereas the expression of later
steps (deoxy-xylulose reductoisomerase and IPP isomerase) are much less affected
(Figure A.5). This suggests that the general production of IPP precursors may also be
reduced in these cultivars.

Among the other carotenoid pathway genes analyzed, it is particularly noteworthy

251

Figure A.5. Northern blots of RNA from stage 3 and 4 petals of each of the indicated
cultivars and from leaves of the Dark Orange Lady cultivar. Approximately 2 ug of
poly A RNA per lane were loaded. Bar graphs represent expression levels of the
indicated genes relative to their expression in leaf tissue. Abbreviations: DXPS, l-deoxy-
D-xylulose-S-phosphate synthase; DXR, 1-deoxy-D-xylulose-S-phosphate
reductoisomerase; IPP, isopentenyl pyrophosphate isomerase; GGDP, geranylgeranyl
diphosphate synthase; FV, French Vanilla; PC, Primrose Climax; GL, Golden Lady;

DOL, Dark Orange Lady.

252

 

 

 

 

 

 

 

 

 

 

 

 

 

8 .
DXPS i
liutrl DXR
IPP
GGDP
Leaf FV PC GL DOL
Figure A.5

253

Figure A.6. Northern blots of RNA from stage 3 and 4 petals of each of the indicated
cultivars and from leaves of the Dark Orange Lady cultivar. Approximately 2 pg of
poly A RNA per lane were loaded. Bar graphs represent expression levels of the
indicated genes relative to expression in leaf tissue. Abbreviations: PSY, phytoene
synthase; PDS, phytoene desaturase; ZDS, Q-desaturase; Beta cyc, B-ring cyclase; Eps
cyc, s-ring cyclase; Hydrox, B-ring hydroxylase; TUB, (at-tubulin; F V, French Vanilla;

PC, Primrose Climax; GL, Golden Lady; DOL, Dark Orange Lady.

254

listed

11‘
H; I
‘-

4”
I).

Figure A.6

PSY

PDS

ZDS

Beta cyc

Eps cyc

Hyd rox

TUB

 

 

O

 

I A l A
v v v v

 

I J

 

 

 

 

 

l _L__l l I

chanchcochco

 

 

120«
801
40“

 

 

 

 

 

 

d '1“ 2'. I ' ‘ ' I
‘... ' ,‘ ’ l
g . ..4.
iv . ”I . A \ -~‘_ 1 J‘”.—‘. 4,
-

 

 

o]_-_-__l__-__-_

Leaf FV PC GL DOL

255

that while the expression of the B-cyclase is not induced during petal development, the
expression of the 8-cyclase is the most strongly induced of the examined genes. Esterified
lutein, with one B and one 8 ring, is the predominant carotenoid synthesized by petals,
while carotenoids having two [3 rings are much reduced compared to leaves (Figure A.4).
Thus the increase in the abundance of the mRN A for e-cyclase is consistent with a recent
model (Cunningham et al., 1996) proposing that the level of e-cyclase activity controls
the degree to which the intermediate lycopene is diverted into the branch of carotenoids
having one 8 and one 8 ring. The e-cyclase gene is up regulated as well in the pale
varieties, although not to the same extent as in the yellow and orange flowered varieties.
This reinforces the notion that the pale varieties are altered in their flux through the
carotenoid pathway but do not differ qualitatively in the type of carotenoid that
accumulates in the petals from the yellow and orange varieties. Somewhat surprising,
however, is the strong increase in abundance of the B-hydroxylase gene in the yellow and
orange varieties, particularly since the transcript for the B-cyclase gene is not similarly
increased. An Arabidopsis mutant lacking B-hydroxylase activity has been isolated
(Pogson et al., 1996) however the gene has not yet been cloned. The Arabidopsis B-
hydroxylase gene will hydroxylate 8 rings but only at a low level compared to its natural
0 ring substrate (Sun et a1., 1996). The marigold B-hydroxylase is 76 % identical to the
Arabidopsis enzyme at the amino acid level excluding the first 50 amino acids that are
predicted to be a plastid-targeting signal. It has a similarly low activity towards Q rings
(data not shown). Attempts to clone a marigold gene related to the marigold l3-

hydroxylase that would have a high level of e-hydroxylase activity have been

256

 

unsuccessful.

Interestingly, transcripts encoding two proteins involved in plastid division also
increased in abundance during petal development in the yellow and orange flower
varieties. Recent characterization of the role of the Arabidopsis F tsZ proteins suggests
that both a cytosol-localized and a plastid-targeted form of these proteins are required for
plastid division in plants (Osteryoung et al., 1998). We have cloned a marigold
homologue of the Arabidopsis plastid-targeted F tsZ gene and Figure A.7 shows its
expression in the various marigold varieties. In addition, we show the expression of a
second gene also involved in plastid division that represents an eukaryotic ortholog of the
E. coli gene minD. In E. coli, MinD is required for proper placement of the F tsZ division
ring (de Boer et a1., 1989). Recent research suggests that the eukaryotic MinD ortholog
performs an analogous function in plastid division (Colletti et al., 2000). While the
expression of both of these genes is virtually undetectable in leaves, FtsZ is increased in
abundance about ten fold in yellow and orange flowers, while MinD is increased about 2-
fold in these varieties. In the French Vanilla and Primrose Climax varieties, the
accumulation of these transcripts is not increased. These results suggest that petal
development involves alterations in plastid division in marigold and that defects in the
expression of plastid division genes are correlated with decreased carotenoid biosynthesis

in pale flower varieties.

Discussion

Marigold flowers range in color from white to dark orange, yet these differences

257

 

 

FtsZ

 

 

 

 

 

MIND 4 4

 

 

 

leaf FV PC GL DOL

Figure A.7. Northern blots of RNA from stage 3 and 4 petals of each of the indicated
cultivars and from leaves of the Dark Orange Lady cultivar. Approximately 2 pg of
poly A RNA per lane were loaded. Bar graphs represent expression levels of the
indicated plastid division genes relative to expression in leaf tissue. Abbreviations: FV,

French Vanilla; PC, Primrose Climax; GL, Golden Lady; DOL, Dark Orange Lady.

258

 

in color are due not to the accumulation of different carotenoids or pathway intermediates
but rather due to the accumulation of vastly different amounts of the same carotenoid,
lutein. During petal development lutein is esterified on each of its hydroxyl groups with
various fatty acids, principally myristate and palmitate (Gau et al., 1983; Rivas, 1989)
and stored in globules in the chromoplast. In order to identify molecular correlates of
carotenoid accumulation and chromoplast development, we have cloned genes from
marigold involved in isoprenoid and carotenoid biosynthesis as well as genes involved in
plastid division. It is apparent that in the varieties French Vanilla and Primrose Climax,
which accumulate very low amounts of carotenoids in their petals, the accumulation of
transcripts of all carotenoid pathway genes, isoprenoid biosynthetic genes and at least
two genes involved in plastid division, are considerably reduced compared to the wild
type orange flowers of the variety Dark Orange Lady. Although the nature of the
mutations in these varieties is unknown, this wholesale down-regulation of an entire
pathway suggests that they are regulatory in nature and affect the transcription or stability
of the messages of multiple genes during petal development. In contrast, the variety
Golden Lady is distinguished from the variety Dark Orange Lady primarily by a reduced
accumulation of the transcripts for phytoene synthase and DXP synthase. It should be
noted that low levels of all of the enzymes are likely to be present even in the pale
varieties since low levels of esterified lutein (1-2% of wild type) are still found in these
varieties.

In contrast to petals, the synthesis of carotenoids in leaves appears to be
unaffected in all varieties. This raises the question whether there exist petal-specific

isoforrns of the enzymes required for carotenoid biosynthesis and plastid division or

259

 

whether the down-regulation of the pathway in petals is due to the tissue-specific mis-
regulation of single genes. At present, we cannot say, although studies in other organisms
suggests multiple genes or small gene families are common for several of the steps
involved in isoprenoid and carotenoid biosynthesis. Indeed, in Arabidopsis six different
GGDPS genes have been identified several of which are localized to different subcellular
compartments (Okata et al., 1999) and two differentially regulated phytoene synthase
genes exist in tomato (Bartley and Scolnik, 1993). Tomato Psy-l appears to be
responsible for carotenoid biosynthesis in fruits, while Psy-2 synthesizes foliar

carotenoids (Fraser et al., 1999). These two enzymes are 95% identical in their amino

 

acid sequence, however antisense inhibition of Psy-l abolishes carotenoid accumulation

 

in fruits only (Brarnley et al., 1992). In our experiments, low stringency hybridization
with the marigold s-cyclase gene revealed a lower molecular weight transcript
specifically in leaves, suggesting the presence of at least a second e-cyclase gene (data
not shown). Data from genomic southern analysis for several pathway steps are
consistent with the presence of small gene families in marigold (data not shown). Thus, it
is likely that multiple genes exist in marigold for many of the steps in the pathways we
have examined.

Somewhat surprisingly, the DXP synthase mRNA showed the lowest absolute
accumulation of the genes we examined. This gene encodes the first step in the recently
identified non-mevalonate route to plastid isoprenoids and is necessary for the synthesis
of carotenoids and other plastid synthesized isoprenoids (Mandel et al., 1996; Bouvier et
al., 1998). It remains to be seen for this and the other genes we isolated how closely the

level of gene expression is correlated with protein abundance and enzyme activity.

260

 

Attempts are in progress to raise antibodies to several of the marigold enzymes to address

these questions.

Cunningham et al. (1996) determined that Arabidopsis thaliana e-cyclase
catalyzes the introduction of a single 8 ring onto the linear lycopene, while the B-cyclase
introduced two [3 rings onto lycopene to produce B-carotene. Carotenoids with two
8 rings are rare in nature. These results have led Cunningham and Gantt (1998) to
propose that there may exist two multienzyrne complexes in plastid membranes, one
containing two B-cyclase enzymes, while the other contains one B- and one e-cyclase.

Since the s-cyclase mRNA is the most strongly up regulated of the studied genes and
lutein is the predominant petal carotenoid, we expect that marigold chromoplast
membranes would be enriched in carotenogenic complexes containing e-cyclase. Other
studies have suggested that there is a channeling of intermediates through a multienzyme
complex during carotenoid biosynthesis (Candau et al., 1991; Camara, 1993) and the
interaction of multiple enzymes has also been found in the synthesis of another abundant
class of plant pigments, the flavonols (Burbulis and Winkel-Shirley, 1999).

The up regulation of two genes required for plastid division in two of the studied
marigold varieties and the lack of induction of these genes in the pale-colored marigold
varieties suggests a correlation between carotenoid accumulation and chromoplast
replication. Given that plastids are the site for both synthesis and storage of carotenoids,
it is likely that plastid numbers or size would increase to accommodate or allow high
level synthesis and accumulation of carotenoids. F tsZ (filamentous temperature-sensitive)
was first identified in E. coli; mutants in this gene cause long undivided filamentous cells

to form (Lutkenhaus and Addinall, 1997). In Arabidopsis, both plastid-targeted and

 

presumed cytoplasmic FtsZ homologues exist and both are required for plastid division
(Osteryoung et al., 1998). MinD, likewise, is a bacterial division gene first identified in E.
coli and plays a role in determining the site of cell division (de Boer et al., 1989).
Antisense repression of the Arabidopsis homologue leads to asymmetric plastid division
in leaves and petals, indicating a role for the gene in placement of the plastid division
apparatus (Colletti et al., 2000). The process of plastid differentiation and division during
petal development has not been extensively characterized although one published study
(Pyke and Page, 1998) suggests that there is little plastid division during petal
development in Arabidopsis. However, Pyke (1997) has presented evidence for a large
increase in plastid number and reduction in plastid size during the chloroplast to
chromoplast differentiation process in ripening tomato fruit. Further work will be needed
to examine plastid differentiation and division in the petals of the different varieties of

marigolds and its relationship to carotenoid accumulation.

Acknowledgements

The authors would like to thank Dr. F .X. Cunningham Jr. for gifts of plasmids
used in color complementation screening and for helpful discussions. The authors would
also like to thank Kelly S. Colletti for sequencing the marigold FtsZ gene and Michelle

Bogoger for expert technical assistance.

262

L

'

 

 

References

Bartley, G.E., and Scolnik, RA. (1993). cDNA cloning, expression during development
and genome mapping of Psy-2, a second tomato gene encoding phytoene synthase. J.
Biol. Chem. 267, 5036-5039.

Bouvier, F., d'Harlingue, A., Suire, C., Backhaus, R.A., and Camara, B. (1998).
Dedicated roles of plastid transketolases during the early onset of isoprenoid biogenesis
in pepper fi'uits. Plant Physiol. 117, 1423-1431.

Bouvier, F., Hugueney, P., d'Harlingue, A., Kuntz, M., and Camara, B. (1994).
Xanthophyll biosynthesis in chromoplasts: isolation and molecular cloning of an enzyme
catalyzing the conversion of 5,6-epoxycarotenoids into ketocarotenoid. Plant J. 6, 45-54.

Bramley, P.M., Teulieres, C., Blain, 1., Bird, C., and Schuch, W. (1992). Biochemical
characterization of transgenic tomato plants in which carotenoid synthesis has been
inhibited through expression of antisense RNA to pTOMS. Plant J. 2, 343-349.

Burbulis, LE, and Winkel-Shirley, B. (1999). Interactions among enzymes of the
Arabidopsis flavonoid biosynthetic pathway. Proc. Natl. Acad. Sci. USA 96, 12929-
12934.

Burkhardt, P.K., Beyer, P., Wunn, J., Kloti, A., and Armstrong, G.A. (1997).
Transgenic rice (Oryriza sativa) endosperm expressing daffodil (Narcissus
pseudonarcissus) phytoene synthase accumulates phytoene, a key intermediate of
provitamin A biosynthesis. Plant J. 11, 1071-1078.

Camara, B. (1993). Plant phytoene synthase complex: component enzymes,
immunology, and biogenesis. Meth. Enzymol. 214, 352-365.

Camara, B., Hugueney, P., Bouvier, F., Kuntz, M., and Moneger, R. (1995).
Biochemistry and molecular biology of chromoplast development. Int. Rev. Cytol. 163,
175-246.

Candau, R., Bejarano, E.R., and Cerda-Olmedo, E. (1991). In vivo channeling of
substrates in an enzyme aggregate for B-carotene biosynthesis. Proc. Natl. Acad. Sci.
USA 88, 493 6-4940.

Colletti, K.S., Tattersall, E.A., Pyke, K.A., Froelich, J.E., Stokes, K.D., and
Osteryoung, K.W. (2000). Placement of the chloroplast division apparatus is mediated
by a homologue of the bacterial cell division site-deterrnining factor MinD. Curr. Biol. in
press.

263

 

"rn—I—I—n. m

 

Cunningham, F.X. Jr., and Gantt, E. (1998). Genes and enzymes of carotenoid
biosynthesis in plants. Annu. Rev. Plant Physiol. 49, 557-583.

Cunningham, F.X. Jr., Pogson, B., Sun, Z., McDonald, K.A., DellaPenna, D., and
Gantt, E. (1996). Functional analysis of the [3 and e lycopene cyclase enzymes of

Arabidopsis reveals a mechanism for control of cyclic carotenoid formation. Plant Cell 8,
1613-1626.

Cunningham, F.X. Jr., Sun, Z.R., Chamovitz, D., Hirschberg, J., and Gantt, E.
(1994). Molecular structure and enzymatic function of lycopene cyclase from the
cyanobacterium Synechococcus sp. strain PCC7942. Plant Cell 6, 1107-1121.

de Boer, P., Crossley, RE, and Rothfield, L.I. (1989). A division inhibitor and a
topological specificity factor coded for by the minicell locus determine proper placement
of the division septum in E. coli. Cell 56, 641-649.

Delgado-Vargas, F., and Paredes-Lopez, O. (1997). Effects of enzymatic treatments of
marigold flowers on lutein isomeric profiles. J. Agric.Food Chem. 45, 1097-1102.

DellaPenna, D. (1999). Nutritional genomics: manipulating plant micronutrients to
improve human health. Science 285, 375-379.

Demmig—Adams, B., and Adams, W.W. (1996). The role of xanthophyll cycle
carotenoids in the protection of photosynthesis. Trends Plant Sci. 1, 21-26.

Emanuelsson, 0., Nielsen, H., and von Heijne, G. (1999). ChloroP, a neural network-
based method for predicting chloroplast transit peptides and their cleavage sites. Prot.
Sci. 8, 978-984.

Fraser, P.D., Kiano, J.W., Truesdale, M.R., Schuch, W., and Bramley, P.M. (1999).
Pnytoene synthase-2 enzyme activity in tomato does not contribute to carotenoid
synthesis in ripening fruit. Plant Mol. Biol. 40, 687-698.

Gau, W., Ploschke, H.J., and Wuensche, C. (1983). Mass spectrometric identification
of xanthophyll fatty acid esters from marigold flowers (Tagetes erecta) obtained by high-
performance liquid chromatography and craig counter-current distribution. J.
Chromatogr. 262, 277-284.

Giuliano, G., Bartley, G.E., and Scolnik, P.A. (1993). Regulation of carotenoid
biosynthesis during tomato development. Plant Cell 5, 379-3 87.

Harker, M., and Bramley, P.M. (1999). Expression of prokaryotic l-deoxy-D-xylulose-

5-phosphatases in Escherichia coli increases carotenoid and ubiquinone biosynthesis.
FEBS Lett. 448, 115-119.

264

 

 

 

Harker, M., and Hirschberg, J. (1998). Molecular biology of carotenoid biosynthesis in
photosynthetic organisms. Meth. Enzymol. 297, 244-263.

Hirschberg , J. (1999). Production of high-value compounds: carotenoids and vitamin E.
Curr .Opin. Biotech. 10, 186-192.

Lange, B.M., Wildung, M.R., McCaskill, D., and Croteau, R. (1998). A family of
transketolases that directs isoprenoid biosynthesis via a mevalonate-independent
pathway. Proc. Natl. Acad. Sci. USA 95, 2100-2104.

Lichtenthaler, H.K. (1999) The 1-deoxy-D-xylulose-5-phosphate pathway of isoprenoid
biosynthesis in plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 50, 47-65.

Lutkenhaus, J., and Addinall, SC. (1997). Bacterial cell division and the Z ring. Annu.
Rev. Biochem. 66, 93-116.

Mandel, M.A., Feldmann, K.A., Herrera-Estrella, L., Rocha-Sosa, M., and Leon, P.
(1996). CLAl , a novel gene required for chloroplast development, is highly conserved in
evolution. Plant J. 9, 649-658.

Mayne, S.T. (1996). Beta-carotene, carotenoids, and disease prevention in humans.
FASEB J. 10, 690-701.

Moehs, C.P., Tian, L., Cunningham, F.X. J r., and DellaPenna, D. (1998).
Identification and characterization of a marigold 1-deoxy-D-xylulose-S-phosphate
synthase, an enzyme of the non-mevalonate isoprenoid pathway. 9th International
Conference on Arabidopsis Research, University of Wisconsisn-Madison, WI.

Norris, S.R., Barrette, T.R., and DellaPenna. D. (1995). Genetic dissection of
carotenoid synthesis in Arabidopsis defines plastoquinone as an essential component of
phytoene desaturation. Plant Cell 7, 2139-2149.

Okada, K., Kamiya, Y., Saito, T., Nakagawa, T., and Kaawamukai, M. (1999).
Localization and expression of geranylgeranyl diphosphate synthases in Arabidopsis
thaliana. Annual Meeting of the American Society of Plant Physiologists, Baltimore,
MD.

Osteryoung, K.W., and Pyke, KA. (1998). Plastid division: evidence for a
prokaryotically derived mechanism. Curr. Opin. Plant Biol. 1, 475-479.

Osteryoung, K.W., Stokes, K.D., Rutherford, S.M., Percival, A.L., and Lee, W.Y.

(1998). Chloroplast division in higher plants requires members of two functionally
divergent gene families with homology to bacterial ftsZ. Plant Cell 10, 1991-2004.

265

Pecker, 1., Gabbay, R., Cunningham, F.X. Jr., and Hirschberg, J. (1996). Cloning

and characterization of the cDNA for lycopene B-cyclase from tomato reveals decrease in
its expression during fruit ripening. Plant Mol. Biol. 30, 807-819.

Pogson, B., McDonald, K.A., Truong, M., Britton, G., and DellaPenna, D. (1996).
Arabidopsis carotenoid mutants demonstrate that lutein is not essential for photosynthesis
in higher plants. Plant Cell 8, 1627-1639.

Pyke KA (1997). The genetic control of plastid division in higher plants. Am. J. Bot. 84,
1017-1027.

Pyke, K.A., and Page, A.M. (1998). Plastid ontogeny during petal development in
Arabidopsis. Plant Physiol. 116, 797-803.

Quackenbush, F.W., and Miller, S.L. (1972). Composition and analysis of the
carotenoids in marigold petals. J. Assoc. Off. Anal. Chem. 55, 617-621.

Rivas, J.D.L. (1989). Reversed-phase high-performance liquid chromatographic
separation of lutein and lutein fatty acid esters from marigold flower petal powder. J.
Chromatogr. 464, 442-447.

Rohmer, M., Knani, M., Simonin, P., Sutter, B., and Sahm, H. (1993). Isoprenoid
biosynthesis in bacteria: a novel pathway for early steps leading to isopentenyl
diphosphate. Biochem. J. 295, 517-524.

Ryberg, H., Ryberg, M., and Sundqvist, C. (1993). Pigment-Protein Complexes in
Plastids: Synthesis and Assembly. Academic Press, London.

Sambrook, J ., Fritsch, E.F., and Maniatis, T. (1989). Molecular cloning: a laboratory
manual. 2nd ed. Cold Spring Harbor Laboratory Press, Plainview, N. Y.

Shewmaker, C.K., Sheey, J.A., Daley, M., Colburn, S., and Ke, D.Y. (1999). Seed-
specific overexpression of phytoene synthase: increase in carotenoids and other metabolic
effects. Plant J. 20, 401-412. '

Straub, O. (1987). List of carotenoids. Birkhauser Verlag, Basel, Switzerland.

Sun, Z.R., Gantt, E., Cunningham, F.X. Jr. (1996). Cloning and functional analysis of
the B-carotene hydroxylase of Arabidopsis thaliana. J. Biol. Chem. 271, 24349-24352.

Ye, X., Al-Babili, S., Kloti, A., Zhang, J., Lucca, P., Beyer, P., and Potrykus, I.

(2000). Engineering the provitamin A (ii-carotene) biosynthetic pathway into
(carotenoid-free) rice endosperm. Science 287, 303-305.

266

 

M IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

lllllilllllllllllll'llllllllllllll

I 3 1293 024 03

   

  

      

no