17
The influence of ascorbate on anthocyanin accumulation during high light acclimation in Arabidopsis thaliana: further evidence for redox control of anthocyanin synthesisMIKE PAGE*, NIGHAT SULTANA*, KONRAD PASZKIEWICZ, HANNAH FLORANCE & NICHOLAS SMIRNOFF Biosciences, College of Life and Environmental Sciences, University of Exeter, Geoffrey Pope Building, Stocker Road, Exeter EX4 4QD, UK ABSTRACT Ascorbate and anthocyanins act as photoprotectants during exposure to high light (HL). They accumulate in Arabidop- sis leaves in response to HL on a similar timescale, suggest- ing a potential relationship between them. Flavonoids and related metabolites were identified and profiled by liquid chromatography–tandem mass spectrometry (LC–MS/ MS). The ascorbate-deficient mutants vtc1, vtc2 and vtc3 accumulated less anthocyanin than wild-type (WT) during HL acclimation. In contrast, kaempferol glycoside accumu- lation was less affected by light and not decreased by ascor- bate deficiency, while sinapoyl malate levels decreased during HL acclimation. Comparison of six Arabidopsis ecotypes showed a positive correlation between ascorbate and anthocyanin accumulation in HL. mRNA-Seq analysis showed that all flavonoid biosynthesis transcripts were increased by HL acclimation in WT. RT-PCR analysis showed that vtc1 and vtc2 were impaired in HL induction of transcripts of anthocyanin biosynthesis enzymes, and the transcription factors PAP1, GL3 and EGL3 that activate the pathway.Abscisic acid (ABA) and jasmonic acid (JA), hormones that could affect anthocyanin accumulation, were unaffected in vtc mutants. It is concluded that HL induction of anthocyanin synthesis involves a redox-sensitive process upstream of the known transcription factors. Because anthocyanins accumulate in preference to kaempferol gly- cosides and sinapoyl malate in HL, they might have specific properties that make them useful in HL acclimation. Key-words: abscisic acid; cyanidin; jasmonic acid; kaempferol glycosides; LC–MS/MS; light stress; metabolite profiling; mRNA-Seq; oxidative stress; vtc mutants. INTRODUCTION Accumulation of purple anthocyanin pigments is one of the most obvious biochemical changes in Arabidopsis thaliana leaves exposed to high light intensity (HL). Anthocyanins are flavonoid compounds synthesized, along with the fla- vonols (kaempferol and quercetin), from phenylalanine and malonyl-CoA (Winkel-Shirley 2002). As well as being light-responsive, anthocyanin accumulation is inducible by sucrose (Teng et al. 2005), hormones such as abscisic acid (ABA) and jasmonates (Loreti et al. 2008; Shan et al. 2009), low temperature (Leyva et al. 1995), nutrient deficiency (Rubin et al. 2009) and wounding (Chalker-Scott 1999). Anthocyanins have various roles, most clearly in attracting pollinators and seed dispersers. Another function in some species is anti-herbivore defence (Karageorgou & Manetas 2006). Abiotic stress-induced accumulation in leaves and stems implies that flavonoids have a protective function. Two roles are proposed: UV-B screening and a direct anti- oxidant effect. Flavonoids have an absorption maximum in the UV-B region of the spectrum, and Arabidopsis mutants impaired in flavonoid biosynthesis are UV-B sensitive (Li et al. 1993; Landry, Chapple & Last 1995; Ormrod, Landry & Conklin 1995). The other main group of UV-B-absorbing phenolic compounds in Arabidopsis are the phenylpro- panoids. A coniferyl aldehyde 5-hydroxylase mutant, fah1, that lacks sinapic acid and sinapoyl malate, the predomi- nant sinapoyl ester in Arabidopsis leaves (Chapple et al. 1992), is even more UV-B sensitive than flavonoid mutants (Landry et al. 1995). There is strong evidence that anthocya- nin accumulation in leaves protects against photoinhibitory damage caused by high irradiance (Havaux & Kloppstech 2001; Gould, Dudle & Neufeld 2010; Zeng et al. 2010; Zhang et al. 2010). This could be attributed to their visible light and UV-B screening effect or to an antioxidant effect. Excess excitation energy can result in increased production of reac- tive oxygen species (ROS) and increases in antioxidants (Galvez-Valdivieso & Mullineaux 2010; Foyer & Noctor 2011; Murchie & Niyogi 2011). It has been suggested that flavonoids could act as antioxidants (Grace & Logan 2000; Hernández et al. 2009). They can react with hydrogen peroxide in peroxidase-catalysed reactions (Yamasaki, Sakihama & Ikehara 1997), as well as superoxide and per- oxynitrite (Rahman et al. 2006). Some flavonoids chelate iron and could therefore act as antioxidants by preventing Correspondence: N. Smirnoff. E-mail: [email protected] *These authors contributed equally to the research. Plant, Cell and Environment (2012) 35, 388–404 doi: 10.1111/j.1365-3040.2011.02369.x © 2011 Blackwell Publishing Ltd 388

The influence of ascorbate on anthocyanin accumulation during high light acclimation in Arabidopsis thaliana: further evidence for redox control of anthocyanin synthesis

Embed Size (px)

Citation preview

The influence of ascorbate on anthocyanin accumulationduring high light acclimation in Arabidopsisthaliana: further evidence for redox control ofanthocyanin synthesispce_2369 388..404

MIKE PAGE*, NIGHAT SULTANA*, KONRAD PASZKIEWICZ, HANNAH FLORANCE & NICHOLAS SMIRNOFF

Biosciences, College of Life and Environmental Sciences, University of Exeter, Geoffrey Pope Building, Stocker Road, ExeterEX4 4QD, UK

ABSTRACT

Ascorbate and anthocyanins act as photoprotectants duringexposure to high light (HL). They accumulate in Arabidop-sis leaves in response to HL on a similar timescale, suggest-ing a potential relationship between them. Flavonoids andrelated metabolites were identified and profiled by liquidchromatography–tandem mass spectrometry (LC–MS/MS). The ascorbate-deficient mutants vtc1, vtc2 and vtc3accumulated less anthocyanin than wild-type (WT) duringHL acclimation. In contrast, kaempferol glycoside accumu-lation was less affected by light and not decreased by ascor-bate deficiency, while sinapoyl malate levels decreasedduring HL acclimation. Comparison of six Arabidopsisecotypes showed a positive correlation between ascorbateand anthocyanin accumulation in HL. mRNA-Seq analysisshowed that all flavonoid biosynthesis transcripts wereincreased by HL acclimation in WT. RT-PCR analysisshowed that vtc1 and vtc2 were impaired in HL induction oftranscripts of anthocyanin biosynthesis enzymes, and thetranscription factors PAP1, GL3 and EGL3 that activatethe pathway. Abscisic acid (ABA) and jasmonic acid (JA),hormones that could affect anthocyanin accumulation, wereunaffected in vtc mutants. It is concluded that HL inductionof anthocyanin synthesis involves a redox-sensitive processupstream of the known transcription factors. Becauseanthocyanins accumulate in preference to kaempferol gly-cosides and sinapoyl malate in HL, they might have specificproperties that make them useful in HL acclimation.

Key-words: abscisic acid; cyanidin; jasmonic acid;kaempferol glycosides; LC–MS/MS; light stress; metaboliteprofiling; mRNA-Seq; oxidative stress; vtc mutants.

INTRODUCTION

Accumulation of purple anthocyanin pigments is one of themost obvious biochemical changes in Arabidopsis thaliana

leaves exposed to high light intensity (HL). Anthocyaninsare flavonoid compounds synthesized, along with the fla-vonols (kaempferol and quercetin), from phenylalanineand malonyl-CoA (Winkel-Shirley 2002). As well as beinglight-responsive, anthocyanin accumulation is inducible bysucrose (Teng et al. 2005), hormones such as abscisic acid(ABA) and jasmonates (Loreti et al. 2008; Shan et al. 2009),low temperature (Leyva et al. 1995), nutrient deficiency(Rubin et al. 2009) and wounding (Chalker-Scott 1999).Anthocyanins have various roles, most clearly in attractingpollinators and seed dispersers. Another function in somespecies is anti-herbivore defence (Karageorgou & Manetas2006). Abiotic stress-induced accumulation in leaves andstems implies that flavonoids have a protective function.Two roles are proposed: UV-B screening and a direct anti-oxidant effect. Flavonoids have an absorption maximum inthe UV-B region of the spectrum, and Arabidopsis mutantsimpaired in flavonoid biosynthesis are UV-B sensitive (Liet al. 1993; Landry, Chapple & Last 1995; Ormrod, Landry &Conklin 1995). The other main group of UV-B-absorbingphenolic compounds in Arabidopsis are the phenylpro-panoids. A coniferyl aldehyde 5-hydroxylase mutant, fah1,that lacks sinapic acid and sinapoyl malate, the predomi-nant sinapoyl ester in Arabidopsis leaves (Chapple et al.1992), is even more UV-B sensitive than flavonoid mutants(Landry et al. 1995).There is strong evidence that anthocya-nin accumulation in leaves protects against photoinhibitorydamage caused by high irradiance (Havaux & Kloppstech2001; Gould, Dudle & Neufeld 2010; Zeng et al. 2010; Zhanget al. 2010).This could be attributed to their visible light andUV-B screening effect or to an antioxidant effect. Excessexcitation energy can result in increased production of reac-tive oxygen species (ROS) and increases in antioxidants(Galvez-Valdivieso & Mullineaux 2010; Foyer & Noctor2011; Murchie & Niyogi 2011). It has been suggested thatflavonoids could act as antioxidants (Grace & Logan 2000;Hernández et al. 2009). They can react with hydrogenperoxide in peroxidase-catalysed reactions (Yamasaki,Sakihama & Ikehara 1997), as well as superoxide and per-oxynitrite (Rahman et al. 2006). Some flavonoids chelateiron and could therefore act as antioxidants by preventing

Correspondence: N. Smirnoff. E-mail: [email protected]

*These authors contributed equally to the research.

Plant, Cell and Environment (2012) 35, 388–404 doi: 10.1111/j.1365-3040.2011.02369.x

© 2011 Blackwell Publishing Ltd388

hydroxyl radical formation through the Fenton reaction(Perron & Brumaghim 2009). It is difficult to distinguishbetween UV-B screening and more direct antioxidanteffects in vivo because UV-B itself causes oxidativestress (Landry et al. 1995), so this role remains to be fullyestablished.

Like anthocyanins, ascorbic acid (AsA) accumulateswhen Arabidopsis plants are transferred from low to HLconditions (Smirnoff 2000; Bartoli et al. 2006; Golan,Muller-Moulé & Niyogi 2006; Zechmann, Stumpe & Mauch2010). It is presumed that AsA accumulation is related to itsrole in photosynthesis and photoprotection.AsA and ascor-bate peroxidase (APX) are involved in removing hydrogenperoxide produced by oxygen photoreduction and photo-respiration (Foyer & Noctor 2011). Cytosolic APX (cAPX)is required for modulating the hydrogen peroxide that leaksfrom chloroplasts and peroxisomes, and is involved in sig-nalling processes (Davletova et al. 2005; Galvez-Valdiviesoet al. 2009). AsA may also reduce tocopheroxyl radicalsproduced as a result of singlet oxygen-induced lipid peroxi-dation in photosystem II (PSII) (Havaux 2003). Thylakoidlumen AsA is involved in the xanthophyll cycle as a cofac-tor for violaxanthin de-epoxidase. AsA-deficient mutantsare impaired in non-photochemical quenching (Smirnoff2000; Muller-Moulé, Havaux & Niyogi 2003). It is also pos-sible that lumenal AsA can protect PSII by acting as anemergency electron donor when the oxygen-evolvingcomplex is inactivated by stress, particularly UV-B and hightemperature (Toth et al. 2009). AsA-deficient vtc mutantsare more susceptible to transfer to HL, although they canacclimate to long-term exposure to some extent (Muller-Moulé, Golan & Niyogi 2004). This proposal takes on sig-nificance if inactivation of the oxygen-evolving complexturns out to be a primary cause of HL damage to PSII(Takahashi et al. 2010; Takahashi & Badger 2011). Previ-ously, it has been noted that AsA-deficient Arabidopsisvtc2-1 accumulates much less anthocyanin under high irra-diance (Giacomelli, Rudella & van Wijk 2006; Giacomelliet al. 2007). In contrast, a mutant in cAPX accumulatesmore anthocyanin in response to low temperature (Asaiet al. 2004) or HL (Miller et al. 2007). Manipulation ofhydrogen peroxide-scavenging capacity by reducing cata-lase (CAT2) expression in Arabidopsis decreases both theexpression of flavonoid biosynthesis genes and anthocyaninaccumulation (Vanderauwera et al. 2005). The results there-fore suggest a role for photosynthetically produced ROSand antioxidants in HL-induced anthocyanin accumulation.

The flavonoid biosynthesis pathway appears to beco-ordinately controlled at the transcriptional level.Under conditions that induce flavonoid synthesis, thegenes encoding most of the enzymes from phenylalanineammonia lyase (PAL), through to the transferases thatdecorate the flavonoid and anthocyanin rings, are induced(Winkel-Shirley 2002; Vanderauwera et al. 2005). Fla-vonoid biosynthesis enzymes may also form complexesto channel pathway intermediates (Burbulis & Winkel-Shirley 1999). Three of the enzymes involved in flavonoidbiosynthesis are 2-oxoglutarate-dependent dioxygenases

(2-ODDs): flavanone 3-hydroxylase (F3H), flavonolsynthase (FLS) and leucoanthocyanidin dioxygenase(LDOX). 2-ODDs generally require AsA for efficientactivity. AsA prevents enzyme inactivation by reducingover-oxidized FeIV to FeII in the active site (Prescott &John 1996; Lukacin & Britsch 1997; Martens, Preuss &Matern 2010). 2-ODD activity in vtc mutants could there-fore be limited by AsA supply.

Given the decreased anthocyanin accumulation in vtc2-1(Giacomelli et al. 2006, 2007) and the observation thatanthocyanin accumulation in leaves exposed to HL is on asimilar timescale (days) to that of AsA, the relationshipbetween these processes merits further investigation.Therefore, the aim of the experiments reported here was toassess whether a link exists between AsA and anthocyaninsduring acclimation to HL by making use of AsA-deficientmutants, as well as natural variation in AsA in variousA. thaliana accessions.

MATERIALS AND METHODS

Plant material and growth conditions

The vtc mutants were obtained from Patricia Conklin (StateUniversity of New York, Cortland) and were all in the Col-0background (Conklin et al. 2000). The Col-0, GOT1, HR-5,Is-0, NFE1 and Old-2 ecotypes were obtained from TheEuropean Arabidopsis Stock Centre (NASC, Nottingham,UK). Surface-sterilized seeds were cold treated for 3 d at4 °C, then sown onto Levington F2 compost (Scotts, Marys-ville, OH, USA) (4:1, compost : vermiculite). Plants weregrown in controlled environment growth rooms undershort-day conditions (8 h light, 16 h dark) at a photosyn-thetic photon flux density (PPFD) of 100 mmol m-2 s-1,23 °C and 65% relative humidity (RH) for 7 weeks, result-ing in large rosettes, but preventing bolting. After this time,plants were acclimated to long days (16 h light, 8 h dark)in a controlled environment growth cabinet (23 °C, 65%RH, PPFD 100 mmol m-2 s-1) for 4 d (Microclima 1000E;Snijders, Tilburg, the Netherlands). HL treatments wereperformed in the same growth cabinets by subjectingthese plants to a PPFD of 550–650 mmol m-2 s-1, 16 h lightperiod for 4 d. Control low light (LL) plants were grownconcurrently in the same cabinet at an irradiance of100 mmol m-2 s-1.

Ascorbate assay

Three fully expanded rosette leaves were excised, weighed,then flash frozen in liquid nitrogen. Leaf samples wereground in 1 mL 1% metaphosphoric acid (MPA). Thehomogenate was centrifuged for 5 min at 15 500 g, 4 °C.Thesupernatant was retained and 125 mL of each sample com-bined with an equal volume of either 1% MPA (to assayreduced AsA) or 20 mm tris(2-carboxyethyl)phosphinehydrochloride (TCEP) in 1% MPA (to assay AsA +dehydroascorbate (DHA) – referred to as ‘total AsA’).DHA was reduced to AsA within 30 min under the

Influence of ascorbate on anthocyanin accumulation 389

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

conditions used, and the AsA remained stable for at least24 h at 21 °C. Assay mixes were passed through 0.2 mmPVDF syringe filters (Chromacol, Welwyn Garden City,UK) and AsA was measured by high-performance liquidchromatography (HPLC). Mobile phases consisted of 95%water, 5% acetonitrile, 0.1% formic acid (A) or 95% aceto-nitrile, 5% water, 0.1% formic acid (B). Then, 20 mL ofsample was injected onto a Phenomenex (Torrance, CA,USA) Luna C18 column (10 mm particle size, 250 ¥ 4.6 mm)and subjected to the following gradient using a DionexDX500 HPLC system (Dionex, Sunnyvale, CA, USA):0 min – 0% B; 4 min – 40% B; 7 min – 100% B; 9 min –100% B; 10 min – 0% B; 2 min post-time. Flow rate wasmaintained at 1 mL min-1 and the assay was carried out at21 °C. AsA was detected using an SPD-10A dual wave-length detector (Shimadzu, Tokyo, Japan) at 265 and280 nm, and had a retention time of approximately 3.5 min.Data were analysed with Chromeleon software (Dionex).The peak purity of AsA was determined by monitoring the265/280 nm signal ratio, which is ~10 for pure AsA. Inter-fering compounds were not found under the chromato-graphic conditions used.AsA was quantified by comparisonwith external standards using the peak areas at 265 nm.

Flavonoid, phenylpropanoid andhormone measurements

Total anthocyanins were determined in the 1% MPAextracts prepared for AsA. Samples were diluted 1:1 with1% MPA, and absorbance was measured at 530 and 657 nmusing an Infinite M200 plate reader (Tecan, Männedorf,Switzerland). [A530 – A657] values were normalized fortissue weight.

For metabolite profiling, 10 mg freeze-dried leaf powderwas extracted in 0.8 mL 10% methanol containing 1% aceticacid. After centrifugation (10 min at 16 100 g, 4 °C), thesamples were filtered through a 0.2 mm (PVDF) syringe filter(Chromacol). For LC–MS/MS analysis of anthocyanins andhormones, 10 mg freeze-dried leaf powder was extracted in0.8 mL 10% methanol + 1% acetic acid containing deuter-ated standards (Forcat et al. 2008). Metabolite profilingwas performed using a QToF 6520 mass spectrometer(Agilent Technologies, Palo Alto, CA, USA) coupled to a1200 series Rapid Resolution HPLC system.Five microlitresof sample extract was loaded onto a Zorbax StableBondC18 1.8 mm, 2.1 ¥ 100 mm reverse-phase analytical column(Agilent Technologies). Mobile phase A comprised 5%acetonitrile with 0.1% formic acid in water and mobile phaseB was 95% acetonitrile with 0.1% formic acid in water. Thefollowing gradient was used: 0 min – 0% B; 1 min – 0% B;5 min – 20% B; 20 min – 100% B; 25 min – 100% B; 26 min –0% B; 9 min post-time.The flow rate was 0.25 mL min-1 andthe column temperature was held at 35 °C for the duration.The source conditions for electrospray ionization were asfollows: gas temperature was 350 °C with a drying gas flowrate of 11 L min-1 and a nebulizer pressure of 55 psig. Thecapillary voltage was 3.5 kV in both positive and negativeion mode. The fragmentor voltage was 115 V and skimmer

70 V. Scanning was performed using the autoMS/MS func-tion at 3 scans s-1 with a sloped collision energy of 3.5 V/100 Da with an offset of 5 V.

Hormone and anthocyanin quantitative analysis was per-formed using an Agilent 6420B triple quadrupole (QQQ)mass spectrometer (Agilent Technologies). The HPLCsystem was the same as that used for QToF analyses. Fortymicrolitres of sample extract was loaded onto a ZorbaxEclipse Plus C18 3.5 mm, 2.1 ¥ 150 mm reversed-phase ana-lytical column (Agilent Technologies). The following gradi-ent was used: 0 min – 0% B; 1 min – 0% B; 5 min – 20% B;20 min – 100% B; 25 min – 100% B; 27 min – 0% B; 7 minpost-time. QQQ source conditions were as follows: gas tem-perature 350 °C, drying gas flow rate 9 L min-1, nebulizerpressure 35 psig, capillary voltage � 4 kV. The fragmentorvoltage and collision energies were optimized for each com-pound. Flavonoids were detected in positive ion mode usingmultiple reaction monitoring (MRM). The MRMs used forflavonoid analyses are shown in Table 1. These were basedon previous MS/MS identification of flavonoids in Arabidop-sis (Tohge et al. 2005; Stobiecki et al. 2006) and confirmed byaccurate mass QToF analyses. In the absence of authenticstandards, the flavonoids were quantified by peak area.Accurate quantification of ABA and jasmonic acid (JA) innegative ion mode used previously determined MRMs anddeuterated internal standards added during sample extrac-tion (Forcat et al. 2008).

Identification of phenylpropanoids

Phenylalanine and the phenylpropanoids were identifiedfrom their accurate masses and MS/MS spectra measured innegative ion mode. The data were extracted usingMassHunter software (Agilent Technologies). Theoreticalm/z values, isotope abundances and product ion m/z valuesfor the identified compounds are in brackets. MS/MSspectra were compared with ESI–QToF–MS/MS spectra ofknown compounds from the MassBank database (Horaiet al. 2010) (http://www.massbank.jp/en/database.html).Phenylalanine: retention time (RT) = 2.77 min; m/z164.0716 (164.0717); isotope 1 abundance 7.83% (10.29%),isotope 2 abundance 1.41% (0.89%); product ions m/z147.0456 (147.0458). Sinapoyl malate: RT = 8.99 min; m/z339.0709 (339.0722); isotope 1 abundance 16.35%(16.74%), isotope 2 abundance 0.32% (0.37%); productions m/z 223.0598 (223.0596), 208.0366 (208.0370), 164.0468(164.0467), 149.0235 (149.0235), 115.0030 (115.0032),71.0468 (71.0153). 5-Hydroxyferulic acid: RT = 7.87 min;m/z 209.0441 (209.0455); isotope 1 abundance 12.17%(11.11%), isotope 2 abundance 2.64% (1.57%); product ionm/z 165.0581 (165.0557) calculated from ferulic acid + 1oxygen atom).

Determination of transcript levels bysemi-quantitative RT-PCR and mRNA-Seq

RNA was extracted from three biological replicates whichwere then pooled for PCR analysis. Frozen tissue was

390 M. Page et al.

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

Tab

le1.

Fla

vono

ids

iden

tifie

din

Ara

bido

psis

thal

iana

leaf

tiss

ueby

liqui

dch

rom

atog

raph

y–el

ectr

ospr

ayio

niza

tion

–tan

dem

mas

ssp

ectr

omet

ry(E

SI–L

C–M

S/M

S)an

alys

is

Pea

kR

t(m

in)

Pre

curs

orio

n(m

/z)

Pro

duct

ion

(m/z

)C

ompo

und

A1

11.1

988

9[M

]+28

7[C

y]+

449

[Cy

+G

lc]+

727

[Cy

+G

lc+

Xyl

+C

ou]+

Cya

nidi

n3-

O-[

2″-O

-(xy

losy

l)6″

-O-(

p-co

umar

oyl)

gluc

osid

e]5-

O-g

luco

side

A2

10.0

494

9[M

]+28

7[C

y]+

449

[Cy

+G

lc]+

Cya

nidi

n3-

O-[

2″-O

-(2�

-O-(

sina

poyl

)xy

losy

l)gl

ucos

ide]

5-O

-glu

cosi

de

A3

11.4

697

5[M

]+28

7[C

y]+

535

[Cy

+G

lc+

Mal

]+

727

[Cy

+G

lc+

Xyl

+C

ou]+

Cya

nidi

n3-

O-[

2″-O

-(xy

losy

l)-6

″-O

-(p-

coum

aroy

l)gl

ucos

ide]

5-O

-mal

onyl

gluc

osid

e

A4

10.2

310

51[M

]+28

7[C

y]+

449

[Cy

+G

lc]+

889

[Cy

+2G

lc+

Xyl

+C

ou]+

Cya

nidi

n3-

O-[

2″-O

-(xy

losy

l)-6

″-O

-(p-

O-(

gluc

osyl

)-p-

coum

aroy

l)gl

ucos

ide]

5-O

-glu

cosi

de

A5

11.4

510

95[M

]+28

7[C

y]+

535

[Cy

+G

lc+

Mal

]+

975

[Cy

+2G

lc+

Xyl

+C

ou+

Mal

]+

Cya

nidi

n3-

O-[

2″-O

-(2�

-O-(

sina

poyl

)xy

losy

l)6″

-O-(

p-co

umar

oyl)

gluc

osid

e]5-

O-g

luco

side

A6

10.4

611

37[M

]+28

7[C

y]+

535

[Cy

+G

lc+

Mal

]+

889

[Cy

+2G

lc+

Xyl

+C

ou]+

Cya

nidi

n3-

O-[

2″-O

-(xy

losy

l)6″

-O-(

p-O

-(gl

ucos

yl)

p-co

umar

oyl)

gluc

osid

e]5-

O-[

6�-O

-(m

alon

yl)

gluc

osid

e]

A7

11.7

211

81[M

]+28

7[C

y]+

535

[Cy

+G

lc+

Mal

]+

933

[Cy

+G

lc+

Xyl

+Si

n+

Cou

]+

Cya

nidi

n3-

O-[

2″-O

-(2�

-O-(

sina

poyl

)xy

losy

l)6″

-O-(

p-O

-cou

mar

oyl)

gluc

osid

e]5-

O-[

6″″-

O-(

mal

onyl

)gl

ucos

ide]

A8

10.6

812

57[M

]+28

7[C

y]+

449

[Cy

+G

lc]+

1095

[Cy

+2G

lc+

Xyl

+C

ou+

Sin]

+

Cya

nidi

n3-

O-[

2″-O

-(2�

-O-(

sina

poyl

)xy

losy

l)6″

-O-(

p-O

-(gl

ucos

yl)

p-co

umar

oyl)

gluc

osid

e]5-

O-g

luco

side

A9

10.9

013

43[M

]+28

7[C

y]+

535

[Cy

+G

lc+

Mal

]+

1095

[Cy

+2G

lc+

Xyl

+C

ou+

Sin]

+

Cya

nidi

n3-

O-[

2″-O

-(6�

-O-(

sina

poyl

)xy

losy

l)6″

-O-(

p-O

-(gl

ucos

yl)-

p-co

umar

oyl)

gluc

osid

e]5-

O-(

6″″-

O-m

alon

yl)

gluc

osid

e

F1

11.4

843

3[M

+H

]+28

7[K

m+

H]+

Kae

mpf

erol

3-O

-rha

mno

side

F2

11.2

856

5[M

+H

]+28

7[K

m+

H]+

433

[Km

+R

ha+

H]+

Kae

mpf

erol

[(pe

ntos

ide)

-rha

mno

side

]

F3

11.4

757

9[M

+H

]+28

7[K

m+

H]+

433

[Km

+R

ha+

H]+

Kae

mpf

erol

3-O

-rha

mno

side

7-O

-rha

mno

side

F4

11.1

259

5[M

+H

]+28

7[K

m+

H]+

433

[Km

+R

ha+

H]+

Kae

mpf

erol

3-O

-glu

cosi

de7-

O-r

ham

nosi

de

F5

10.4

874

1[M

+H

]+28

7[K

m+

H]+

433

[Km

+R

ha+

H]+

595

[Km

+R

ha+

Glc

+H

]+

Kae

mpf

erol

3-O

-[6″

-O-(

rham

nosy

l)gl

ucos

ide]

7-O

-rha

mno

side

Pro

duct

ions

inbo

ldfo

ntw

ere

used

for

quan

tific

atio

n.C

y,cy

anid

in;G

lc,g

luco

se;X

yl,x

ylos

e;C

ou,p

-cou

mar

oyl

moi

ety;

Mal

,mal

onyl

moi

ety;

Sin,

sina

poyl

moi

ety;

Km

,kae

mpf

erol

;Rha

,rh

amno

se.

Influence of ascorbate on anthocyanin accumulation 391

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

ground in Z6 buffer [8 m guanidine hydrochloride, 20 mmMES, 20 mm ethylenediaminetetraacetic acid (EDTA),pH 7.0], then mixed with phenol:chloroform to removeDNA and protein. RNA was precipitated, then resuspendedin nuclease-free water. cDNA synthesis was carried outusing the RevertAid cDNA synthesis system (Fermentas,Vilnius, Lithuania). Reactions contained 2 mg total RNA ina total volume of 25 mL, with conditions as described in themanufacturer’s protocol.

PCR reactions were carried out with 1 mL cDNA tem-plate in a total volume of 20 mL, using ACT2 (At3g18780)as a control – 22 cycles with forward primer 5′-GCAGGAGATGGAAACCTCAAAG-3′ and reverse primer5′-CTGCTGGAATGGAGATCCACAT-3′. Genes of inter-est were amplified as follows: CHS (At5g13930) – forward5′-GCTGGTGCTTCTTCTTTGGATG-3′ and reverse 5′-CTGACTTCCTCCTCATCTCGTCTAG-3′ for 29 cycles;EGL3 (At1g63650) – forward 5′-CCGACACCGAGTGGTACTACTTAG-3′ and reverse 5′-CACAGATGATGATCGCTTCCACC-3′ for 33 cycles; FLS (At5g08640) –forward 5′-CACAACATTCCGAGGTCCAAC-3′ and re-verse 5′-GCTTGCGGTAACTGTAATCCTTG-3′ for 29cycles; GL3 (At5g41315) – forward 5′-GCTGCGGTTAAGACAGTGGTTTGC-3′ and reverse 5′-CATCTCTGGCTTCTGGTGAGTCC-3′ for 33 cycles; LDOX(At4g22880) – forward 5′-GTCCTCAAGTTCCCACAATCG-3′ and reverse 5′-CACATTTTGCAGTGACCCATTTGC-3′ for 29 cycles; PAP1 (At1g56650) – forward5′-GGCACCAAGTTCCTGTAAGAGC-3′ and reverse5′-CCCTTTTCTGTTGTCGTCGC-3′ for 26 cycles; TTG1(At5g24520) – forward 5′-GTCACATACGACTCACCATATCCAC-3′ and reverse 5′-CAATCCAATCAGGCTGCGAAG-3′ for 26 cycles. Band intensities were quantifiedusing ImageJ software and normalized to ACT2 intensities.Inspection of mRNA-Seq data confirmed that this gene isnot light responsive. It should be noted that because semi-quantitative RT-PCR cycle number varied between ampli-fications, relative expression values between genes are notdirectly comparable.

For mRNA-Seq, leaf tissue was pooled from four bio-logical replicates and RNA extracted as above for semi-quantitative RT-PCR. RNA quality was assessed usingthe Agilent 2100 BioAnalyser (Agilent Technologies),with high-quality samples concentrated and re-quantifiedbefore library preparation. Library preparation wascarried out according to the Illumina mRNA-Seq manual,with the following changes included: centrifugation stepsinvolving spin columns were performed at a slower speedof 6000–10 000 g; during spin column purifications,columns were incubated with EB buffer for 15 min atroom temperature; agarose was replaced with polyacryla-mide for gel matrices; excised polyacrylamide gel sliceswere incubated in 0.3 m sodium acetate, 2 mm EDTApH 8.0 at 37 °C overnight, then subjected to an ethanolprecipitation to recover DNA. Samples were run on anIllumina GAIIx DNA sequencer (Illumina Inc., SanDiego, CA, USA) to generate 76 bp single-end reads.Raw reads were trimmed to remove the typical variable

sequence over the first 12 bases, and then filtered toremove reads containing the adaptor sequence. Alignmentof reads to the A. thaliana genome (TAIR9 release) wasperformed with Tophat 1.0.13 and Bowtie 0.12.5 – onlyreads with a mapping quality of Q > 30 were accepted.Resulting files were processed with HTSeq to extractgene level counts, which were subsequently analysedwith DESeq (Anders & Huber 2010) to obtain gene-level differential expression calls. Differential expressiondata were visualized using MapMan 3.5.1 (Thimm et al.2004).

Statistical analysis

All experiments (except mRNA-Seq) were repeated atleast twice, and data from a representative experiment areshown. Metabolite data were subjected to Student’s t-test oranalysis of variance (anova) using SPSS v. 16 (IBM,Chicago, IL, USA). Where significant effects (P < 0.05)were found by anova, the treatment means were comparedwith the Tukey HSD test.

RESULTS

AsA and anthocyanin accumulation duringacclimation to HL

It has previously been shown that foliar AsA pool sizeincreases in response to HL intensity in Arabidopsis(Smirnoff 2000; Bartoli et al. 2006; Dowdle et al. 2007). Weinvestigated how AsA and anthocyanin accumulationwere affected by smaller adjustments in light intensityover a 4 d period. At the beginning of the time course,there was no difference in the foliar AsA content ofany of the plants (Fig. 1a). On all subsequent days, thoseleaves that had received the highest light intensity(730 mmol m-2 s-1) had significantly more AsA than controlplants (70 mmol m-2 s-1). The AsA content of plants grownat intermediate intensities was not statistically differentfrom the control plants, although a general irradiance-dependent increase was observed, supporting the hypoth-esis that AsA pool size is tightly linked to light intensity.Anthocyanins started to accumulate after transferringArabidopsis plants from LL to HL conditions (Fig. 1b).Initially, they accumulated in the lower epidermis, fol-lowed by the epidermal cells above the vascular bundlesfrom where accumulation spread to the epidermal layersabove the mesophyll cells. Accumulation was more exten-sive in fully expanded leaves than in young leaves, theformer being sampled in these experiments. Plants grownat the three highest light intensities visibly accumulatedmore anthocyanins than those plants grown at the twolowest intensities. At 540 and 730 mmol m-2 s-1, leaves hada larger anthocyanin pool after 24 h of HL treatment com-pared to control plants. Similarly to AsA pool size, foliaranthocyanin content tended to increase in line with incre-mental changes in light intensity.

Transcript levels of genes related to flavonoid and phe-nylpropanoid metabolism were determined by mRNA-Seq

392 M. Page et al.

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

from leaves acclimated to LL and HL for 4 d (Fig. 2; Sup-porting Information Table S1). This showed that, with theexception of cinnamate 4-hydroxylase (C4H), the transcriptlevels of all the genes involved in flavonoid biosynthesiswere increased by HL acclimation. Phenylpropanoids aresynthesized from coumaroyl-CoA. Transcripts of genes inthe first three steps of this pathway were not light respon-sive, although some of the transcripts encoding enzymes ofthe later stages of the pathway showed modest increases(Fig. 2). In contrast to flavonoids, transcripts of enzymes inthe d-mannose/l-galactose AsA biosynthesis pathway, with

the exception of VTC2 (GDP-l-galactose phosphorylase)did not increase after HL acclimation (Supporting Informa-tion Table S1).

HL-induced anthocyanin accumulation isimpaired in vtc mutants

Nine anthocyanins based on cyanidin and five flavonol(kaempferol) glycosides were identified. Their identitieswere based on previous analysis of the major flavonoids inArabidopsis (Tohge et al. 2005), along with accurate mass

abab

ab

c

c

ab

0 1 2 3 4

0 1 2 3 4

abab ab

ab

ab

Days

Tot

al A

sA (mm

ol g

–1 fr

esh

wt)

4

6

8

10

12

aa a

a

aa a

ab

bab

a

b

a

b

a

ab

b

Days

Tot

al r

elat

ive

anth

ocya

nin

0

0.4

0.8

1.2

1.6

2

2.4

a a a a a a bc

ca

ab b

c

c

a a

ab

ab

b

b

b

c

(a)

(b)

70

190

310

540

730

70

190

310

540

730

Figure 1. The effect of light intensity on ascorbate and anthocyanin accumulation. (a) Total foliar ascorbate accumulation in wild-type(WT) Arabidopsis thaliana plants during a 4 d light treatment. Plants were subjected to five different light intensities (mmol m-2 s-1) andsampled every 24 h (at day 0, all plants had previously been grown at 70 mmol m-2 s-1 and were sampled for ascorbate before transferringto the stated light intensities). Error bars represent +1 SEM (n = 4), lowercase letters represent significant differences on each day(P < 0.05) calculated using Student’s t-test. (b) Foliar anthocyanin content in the same plants over the same 4 d light treatment.Absorbance values were normalized for tissue weight, then made relative to the day 4, 730 mmol m-2 s-1 value. Error bars represent +1SEM (n = 3), lowercase letters represent significant differences on each day (P < 0.05) calculated using Student’s t-test.

Influence of ascorbate on anthocyanin accumulation 393

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

Fig

ure

2.P

heny

lpro

pano

idan

dfla

vono

idbi

osyn

thes

ispa

thw

aysh

owin

gch

ange

sin

gene

expr

essi

onan

dph

enyl

prop

anoi

d/fla

vono

idac

cum

ulat

ion

inre

spon

seto

high

light

(HL

).A

rabi

dops

isth

alia

napl

ants

wer

esu

bjec

ted

to4

dof

low

light

(LL

)(1

00mm

olm

-2s-1

)or

HL

(600

mmol

m-2

s-1).

Eac

hsp

otre

pres

ents

mR

NA

-Seq

diff

eren

tial

expr

essi

onda

tafo

ron

ege

ne.

Gen

esfo

rw

hich

raw

coun

tsw

ere

too

low

tope

rmit

relia

ble

infe

renc

esar

ede

note

dby

grey

squa

res.

Pat

hway

prod

ucts

/inte

rmed

iate

sar

esh

own

inov

albo

xes.

Dif

fere

ntia

lex

pres

sion

/acc

umul

atio

nva

lues

are

ona

log 2

fold

chan

gesc

ale.

Enz

ymes

show

nin

red

font

are

2-ox

oglu

tara

te-d

epen

dent

diox

ygen

ases

(2-O

DD

s)w

hich

may

requ

ire

asco

rbic

acid

(AsA

)to

mai

ntai

nth

eir

acti

vity

.Abb

revi

atio

ns:P

AL

,phe

nyla

lani

neam

mon

ialy

ase;

4CL

,4-c

oum

arat

eC

oAlig

ase;

C4H

,cin

nam

ate-

4-hy

drox

ylas

e;C

HS,

chal

cone

synt

hase

;CH

I,ch

alco

nefla

vano

neis

omer

ase;

F3H

,flav

anon

e3-

hydr

oxyl

ase;

F3′

H,fl

avon

oid

3′hy

drox

ylas

e;F

LS,

flavo

nols

ynth

ase;

DF

R,d

ihyd

rofla

vono

l4-r

educ

tase

;LD

OX

,leu

coan

thoc

yani

din

diox

ygen

ase;

3GT,

anth

ocya

nidi

n3-

O-g

luco

sylt

rans

fera

se;5

GT,

anth

ocya

nidi

n5-

O-g

luco

sylt

rans

fera

se;G

L3,

glab

ra3;

EG

L3,

enha

ncer

ofgl

abra

3;PA

P,pr

oduc

tion

ofan

thoc

yani

npi

gmen

t;T

T,tr

ansp

aren

tte

sta;

TT

G1,

tran

spar

ent

test

agl

abra

1;5M

AT,

mal

onyl

coA

:ant

hocy

anid

in5-

O-g

luco

side

-6″-

O-m

alon

yltr

ansf

eras

e;1,

hydr

oxyc

inna

moy

l-C

oA:s

hiki

mat

e/qu

inat

ehy

drox

ycin

nam

oylt

rans

fera

se;2

,p-c

oum

aroy

lshi

kim

ate/

quin

ate

3′-h

ydro

xyla

se;3

,hyd

roxy

cinn

amoy

lCoA

:shi

kim

ate/

quin

ate

hydr

oxyc

inna

moy

ltra

nsfe

rase

;4,c

affe

oyl-

CoA

O-m

ethy

ltra

nsfe

rase

;5,c

inna

moy

l-C

oAre

duct

ase;

6,si

napy

l/con

ifer

yl/c

inna

myl

alco

hold

ehyd

roge

nase

;7,c

onif

eryl

alde

hyde

5-hy

drox

ylas

e;8,

sina

pald

ehyd

ede

hydr

ogen

ase;

9,si

napa

te1-

gluc

osyl

tran

sfer

ase;

10,s

inap

oylg

luco

se:m

alat

esi

napo

yltr

ansf

eras

e;11

,sin

apoy

lglu

cose

:cho

line-

O-s

inap

oylt

rans

fera

se.

394 M. Page et al.

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

and MS/MS spectra determined by LC–QToF–MS/MS(Table 1). The relative concentrations of these flavonoidswere then determined by LC–QQQ–MS/MS using MRM.All nine anthocyanins showed an identical pattern ofaccumulation across the vtc mutants after transfer to HL(Fig. 3). Each compound accumulated in wild-type (WT)and all were significantly lower in the vtc mutants(P < 0.001). Anthocyanin accumulation in vtc3-1 was inter-mediate between WT and the other mutants, being signifi-cantly different from both (P < 0.001). In strong contrast tothe anthocyanins, the five kaempferol glycosides showed nosignificant difference between HL and LL, or between WTand vtc mutants (Fig. 4). It is therefore clear that anthocya-nin accumulation is far more responsive to light than thekaempferol glycosides, and the vtc mutants are collectivelyimpaired in anthocyanin accumulation.The AsA concentra-tions in the vtc mutants were significantly lower than WTand, unlike WT, did not increase significantly in HL (Fig. 5).

The results show a close relationship between AsA andanthocyanin accumulation, both following a very similartime course in WT plants after transfer to HL (Fig. 1).

Both flavonoids and the other major group of phenoliccompounds in Arabidopsis, the phenylpropanoids, are syn-thesized from phenylalanine (Fig. 2). To determine if phe-nylpropanoid synthesis is also affected in the vtc mutants,several representatives of this class of compounds wereidentified from a LC–QToF–MS/MS metabolite profilingexperiment and their relative concentrations determined(Fig. 6). Phenylalanine concentration was the same in allstrains and was significantly lower in HL (P < 0.01) only inthe case of vtc2-1. In LL sinapoyl malate was significantlylower than WT in vtc1 and vtc2-1, but not vtc3-1. Sinapoylmalate was also significantly decreased in HL in all strainsexcept vtc2-1. 5-Hydroxyferulic acid was identified in WTand was significantly decreased in HL. It was below thelimit of detection in the vtc mutants. Sinapoyl malate is the

Figure 3. Foliar anthocyanin profiles of Arabidopsis thaliana wild-type (WT) and vtc mutants acclimated to low light (LL)(100 mmol m-2 s-1) or high light (HL) (550–650 mmol m-2 s-1) for 4 d. Anthocyanins A1–A9 (see Table 1) were quantified by LC–ESI–QQQMS/MS. Mean values +1 SEM (n = 4) are shown. Columns labelled ‘a’ are significantly different (P < 0.05) from the LL treatment for thesame strain. Columns labelled ‘b’ are significantly different (P < 0.05) from the corresponding WT treatment. Significant effects are onlyshown for anthocyanin 1, and all other anthocyanins followed the same pattern.

Influence of ascorbate on anthocyanin accumulation 395

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

major phenylpropanoid in Arabidopsis (Chapple et al. 1992;Lorenzen et al. 1996; Nair et al. 2004), and in contrast toanthocyanins, its accumulation was generally decreased byHL. The inability to detect 5-hydroxyferulic acid in the vtcmutants suggests that phenylpropanoid accumulation islimited by AsA deficiency to some extent.

AsA and anthocyanin accumulation correlateacross Arabidopsis ecotypes

Natural variation present in Arabidopsis was used to helpestablish whether a relationship existed between AsA andanthocyanin pool size. Ecotypes with an altered AsAcontent compared to Col-0 were identified (Fig. 7). Is-0 andOld-2 had significantly less AsA under both LL and HL,while HR5 showed a significant reduction under HL only.GOT1 and NFE1 had WT AsA levels under both lightconditions. No differences in the redox state of AsA wereobserved across ecotypes. The anthocyanin profiles of theseecotypes were subsequently analysed through targetedLC–MS. Under LL conditions, six anthocyanins weredetected in Col-0 (A1, A2, A3, A4, A6 and A9), albeit atvery low concentrations (Table 2). The GOT1, HR5, Is-0and Old-2 ecotypes accumulated relatively large amountsof A6, while Old-2 accumulated a relatively large amountof A9, compared to Col-0. A1, A2, A3 and A4 were less

Figure 4. Foliar flavonol glycosideprofiles of Arabidopsis thaliana wild-type(WT) and vtc mutants acclimated to lowlight (LL) (100 mmol m-2 s-1) or high light(HL) (550–650 mmol m-2 s-1) for 4 d.Flavonols F1–F5 (see Table 1) werequantified by LC–ESI–QQQ MS/MS.Mean values +1 SEM (n = 4) are shown.Light treatments had no significant effectwithin strains, and within light treatmentsthere were no significant differencesbetween strains.

Figure 5. Total foliar ascorbate concentration in Arabidopsisthaliana wild-type (WT) and vtc mutants acclimated to low light(LL) (100 mmol m-2 s-1) or high light (HL) (550–650 mmol m-2 s-1)for 4 d. Mean values +1 SEM (n = 4) are shown. Columnslabelled ‘a’ are significantly different (P < 0.05) from the LLtreatment for the same strain. Columns labelled ‘b’ aresignificantly different (P < 0.05) from the corresponding WTtreatment.

396 M. Page et al.

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

abundant in the test ecotypes compared to Col-0. In leavesexposed to HL, anthocyanins were substantially moreabundant, with nine detected in each of the ecotypes. Accu-mulation was highest in Col-0 for all anthocyanins, whilelevels in the test ecotypes were markedly reduced in com-parison. Old-2 had the least total AsA and the least totalanthocyanins after HL treatment, highlighting a strong cor-relation between the two (Fig. 8).

ABA and JA in vtc mutants

Both ABA and JA have been implicated in mediatingstress-induced anthocyanin accumulation. Therefore, theeffect of light intensity on the concentrations of these hor-mones was measured in WT and vtc mutants (Fig. 9). ABAconcentration was unaffected in the vtc mutants comparedto WT, and was generally not significantly increased by HL.Similarly, JA concentrations were not significantly differentbetween strains and light treatments.

HL induction of genes involved in anthocyaninbiosynthesis is impaired in vtc mutants

Semi-quantitative RT-PCR analysis was used to determinewhether AsA content affected the expression level of genesinvolved in anthocyanin biosynthesis. Prior to HL treat-ment when anthocyanin accumulation was not directlyobservable, the expression of biosynthetic genes (CHS, FLSand LDOX) and signalling genes (EGL3, GL3, PAP1 andTTG1) was low in WT and vtc mutant plants (Fig. 10). Withthe exception of TTG1, expression of these genes wasgreatly increased after HL treatment in WT leaves, consis-tent with a deep-purple leaf colour. Transcript levelsalso tended to increase in the vtc mutants, but remainedsubstantially lower than WT, corresponding to a loweranthocyanin pool (Fig. 3). The expression of TTG1 showed

Figure 6. Foliar phenylalanine, sinapoyl malate and5-hydroxyferulate profiles of Arabidopsis thaliana wild-type(WT) and vtc mutants acclimated to low light (LL)(100 mmol m-2 s-1) or high light (HL) (550–650 mmol m-2 s-1)for 4 d. The compounds were identified and quantified byLC–ESI–QToF MS/MS. Mean values +1 SEM (n = 3) are shown.Columns labelled ‘a’ are significantly different (P < 0.05) fromthe LL treatment for the same strain. Columns labelled ‘b’ aresignificantly different (P < 0.05) from the corresponding WTtreatment.

Figure 7. Foliar ascorbate and dehydroascorbate (DHA) inCol-0 and five other Arabidopsis thaliana ecotypes acclimated toeither low light (LL) (100 mmol m-2 s-1) or high light (HL)(600 mmol m-2 s-1) for 4 d. Error bars represent +1 SEM of thetotal ascorbic acid (AsA), n = 3, * = significant difference versusCol-0 (P < 0.05) using Student’s t-test (LL and HL data analysedseparately).

Influence of ascorbate on anthocyanin accumulation 397

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

little change in any of the lines after HL treatment.These results are consistent across vtc mutants and showthat AsA deficiency strongly decreases the expression ofanthocyanin-related genes.

DISCUSSION

HL-induced AsA and anthocyanin accumulation

AsA and anthocyanins accumulated progressively inmature Arabidopsis leaves in an irradiance-dependentmanner in the 4 d following the transfer from LL. A closecorrespondence between AsA and anthocyanin accumula-tion was found in three different vtc mutants. VTC1 andVTC2 are involved in AsA biosynthesis (Conklin et al.1999; Laing et al. 2007; Linster et al. 2007), while the func-tion of VTC3 is not established. The results show that AsAdeficiency, rather than any other effects the mutations mayhave, is the cause of impaired anthocyanin accumulation.The fine tuning of this relationship is illustrated bythe correlation between AsA and anthocyanin across sixArabidopsis ecotypes. Furthermore, it was apparent thatanthocyanin accumulation varies in different leaves andinspection of the raw data shows that within a genotype,samples with high anthocyanin tended to have high AsA(data not shown). The close correspondence between irra-diance and concentration suggests that there is a lightsensing and signal transduction mechanism that finelyadjusts the levels of these metabolites during acclimation tothe prevailing irradiance level and, importantly, that thisis influenced by AsA. UV-B, blue light (cryptochrome)and red light (phytochrome) photoreceptors have beenproposed to act as the sensors for light-induced anthocya-nin accumulation at low irradiances (Chatterjee, Sharma &Ta

ble

2.Fo

liar

anth

ocya

nin

profi

les

ofsi

xA

rabi

dops

isth

alia

naec

otyp

esac

clim

ated

tolo

wlig

ht(L

L)

(100

mmol

m-2

s-1)

orhi

ghlig

ht(H

L)

(550

–650

mmol

m-2

s-1)

for

4d

Pea

k

LL

(pea

kar

ea)

HL

(pea

kar

ea¥

103 )

Col

-0G

OT

1H

R5

Is-0

NF

E1

Old

-2C

ol-0

GO

T1

HR

5Is

-0N

FE

1O

ld-2

A1

91�

63nd

ndnd

nd40

�21

32.9

�3.

218

.4�

3.1

14.2

�1.

711

.3�

0.5

3.2

�0.

310

.2�

6.0

A2

18�

18nd

ndnd

ndnd

7.7

�0.

26.

0�

0.3

4.9

�0.

53.

5�

0.3

4.2

�0.

31.

6�

0.4

A3

291

�19

058

�29

nd25

�25

48�

4821

�21

362.

1�

14.7

177.

7�

30.3

197.

2�

17.3

78.3

�4.

228

.8�

3.3

56.2

�16

.4A

417

9�

9951

�30

ndnd

ndnd

58.3

�3.

229

.6�

2.0

16.5

�0.

915

.4�

1.0

8.7

�1.

55.

7�

2.7

A5

ndnd

ndnd

ndnd

58.0

�13

.134

.2�

4.6

29.3

�4.

917

.8�

2.1

34.5

�3.

910

.6�

5.9

A6

18�

1820

1�

201

106

�86

66�

66nd

449

�39

018

2.2

�7.

111

5.6

�7.

110

8.7

�3.

012

3.4

�7.

340

.2�

6.7

20.1

�6.

4A

7nd

ndnd

ndnd

nd62

.8�

3.4

43.1

�5.

242

.4�

1.2

16.5

�2.

319

.2�

3.6

17.0

�5.

9A

8nd

ndnd

ndnd

nd69

.2�

6.5

33.7

�2.

423

.9�

3.5

13.7

�1.

735

.7�

2.9

6.7

�1.

8A

927

�27

nd40

�40

54�

5420

�20

314

�22

919

0.5

�4.

510

6.9

�7.

611

8.0

�14

.778

.0�

8.8

148.

1�

7.9

25.2

�5.

4To

tal

1312

821

333

434

206

1343

1023

.756

5.2

555.

135

7.9

285.

615

3.3

Mea

ns�

1SE

M(n

=3)

.nd,

not

dete

cted

.Ant

hocy

anin

sA

1–A

9(s

eeTa

ble

1)w

ere

quan

tifie

dby

LC

–ESI

–QQ

QM

S/M

S.

Figure 8. The relationship between high light (HL)-inducedascorbic acid (AsA) and anthocyanin pools in leaves of sixArabidopsis thaliana ecotypes. Mean values for each strain aretaken from Fig. 7 and Table 2. r = 0.8892, P < 0.02.

398 M. Page et al.

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

Khurana 2006; Shin, Park & Choi 2007). However, it iscurrently not clear if these photoreceptors are involved inadjusting anthocyanin synthesis during HL acclimation. It ispossible that chloroplast-derived signals that result fromthe sensing of excess excitation energy, via changes in redoxstate of photosynthetic components or ROS production,could be involved, as for example occurs with HL inductionof APX2 (Galvez-Valdivieso et al. 2009).

The light responsiveness of AsA in Arabidopsis andother species is well known (Grace & Logan 1996; Smirnoff2000; Muller-Moulé et al. 2004; Bartoli et al. 2006).AsA con-centration is finely tuned to irradiance but, as in the case ofanthocyanins, the mechanism is not fully understood. Theredox state of the photosynthetic electron transport chainhas been proposed as a signal for both ascorbate andanthocyanin accumulation (Yabuta et al. 2007; Jeong et al.2010; Das et al. 2011). The transcript levels and enzymeactivity of VTC2 (GDP-l-galactose phosphorylase), the

first dedicated enzyme in the AsA biosynthesis pathway vial-galactose, are rapidly increased by transfer from low toHL, while the other pathway enzymes show little or nochange (Dowdle et al. 2007; Muller-Moulé 2008). Transcrip-tome analysis by mRNA-Seq showed that VTC2 is the onlygene in the mannose/l-galactose pathway with higher tran-script levels after HL acclimation (Supporting InformationTable S1). VTC2 could therefore be a target for theirradiance-dependent transcriptional regulation of AsAbiosynthesis. The pattern of accumulation supports the pro-posed roles of AsA and anthocyanins in photoprotection(see Introduction).

The effect of HL and AsA deficiency onphenylpropanoid and flavonoid metabolism

LC–MS/MS analysis enabled the identification and profil-ing of five flavonols (kaempferol glycosides) and ninecyanidin-based anthocyanins. The anthocyanins all accumu-lated in a similar manner during HL acclimation in WTplants. Accumulation was greatly decreased in the vtcmutants with no differences in the profiles across thestrains. Although the kaempferol glycosides tended toincrease in HL in all the strains, the response was not sig-nificant. In strong contrast to the anthocyanins, the levels ofkaempferol glycosides were identical in WT and mutants.HL acclimation therefore specifically increases anthocya-nins, even though there is a strong increase in transcriptlevels of genes encoding enzymes in the early part of theflavonoid biosynthesis pathway and in the flavonol andanthocyanin branches (Fig. 2). The pathway for phenylpro-panoid biosynthesis branches off from the flavonoidpathway at coumaroyl-CoA, leading to the synthesis ofsinapoyl malate, the major sinapate ester in Arabidopsis(Chapple et al. 1992). HL acclimation decreased sinapoylmalate and 5-hydroxyferulic acid, a related metabolite.Although transcripts of some of the genes encodingsinapoyl malate had modest increases in HL, the initialstages of the phenylpropanoid pathway were unaffected.The results are consistent with a diversion from sinapoylmalate synthesis to anthocyanins under HL.This could arisefrom direct competition between the pathways forcoumaroyl-CoA and/or incorporation of sinapic and cou-maric acids into the anthocyanins. All of the nine anthocya-nins analysed contain coumaroyl or sinapoyl residues.Although both sinapoyl malate and flavonoids act as UV-Bscreens (Li et al. 1993; Landry et al. 1995), the preferentialaccumulation of anthocyanins over flavonols and sinapoylmalate suggests that the former may have specific proper-ties that are advantageous for acclimation to HL across thebiologically active spectrum (Havaux & Kloppstech 2001;Gould, McKelvie & Markham 2002; Gould et al. 2010;Zhang et al. 2010).

AsA deficiency specifically affects anthocyanin, but notflavonol accumulation. There was, however, also a signifi-cant decrease in sinapoyl malate in vtc1 and vtc2-1, while5-hydroxyferulic acid, a low abundance metabolite in WT,was not detectable in the vtc mutants. Low AsA therefore

(a)

(b)

Figure 9. Foliar abscisic acid (ABA) (a) and jasmonic acid (JA)(b) concentration in Arabidopsis thaliana wild-type (WT) and vtcmutants acclimated to low light (LL) (100 mmol m-2 s-1) or highlight (HL) (550–650 mmol m-2 s-1) for 4 d. Mean values +1 SEM(n = 4) are shown. Light treatments had no significant effectwithin strains, and within light treatments there were nosignificant differences between strains.

Influence of ascorbate on anthocyanin accumulation 399

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

Figure 10. Relative transcript levels of anthocyanin-related genes in leaves of Arabidopsis thaliana wild-type (WT) and vtc mutantsacclimated to low light (LL) (100 mmol m-2 s-1) at day 0 and after 3 d acclimation to high light (HL) (550–650 mmol m-2 s-1). Transcriptlevels were determined by semi-quantitative RT-PCR and staining intensities of PCR products normalized against ACT2. Abbreviations:FLS, flavonol synthase; CHS, chalcone synthase; LDOX, leucoanthocyanidin dioxygenase; GL3, glabra 3; EGL3, enhancer of glabra 3;PAP1, production of anthocyanin pigment 1; TTG1, transparent testa glabra 1.

400 M. Page et al.

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

reduces phenylpropanoid accumulation, although thiseffect is much less than on anthocyanin accumulation. Phe-nylalanine, the precursor of all the phenolics, was not sig-nificantly lower in the vtc mutants, so the supply of thisamino acid is unlikely to be limiting for flavonoid and phe-nylpropanoid synthesis. A number of reasons for decreasedanthocyanin accumulation can be envisaged. Firstly, AsAcould stabilize anthocyanins against oxidative degradation;AsA occurs in vacuoles and could inhibit oxidative degra-dation by peroxidases. Secondly, anthocyanin biosynthesisemploys a number of 2-ODDs (Fig. 2), which as notedabove require AsA to prevent over-oxidation of the activecentre iron. However, this class of enzyme is involved inboth flavonol and anthocyanin synthesis so differential AsArequirement or availability would need to be proposed.Thefinal possibility, discussed below, is that AsA deficiency dis-rupts the signalling processes required for anthocyaninaccumulation.

AsA influences signalling processes requiredfor HL-induced anthocyanin accumulation

Acclimation to HL increased expression of almost all theknown enzymes and transcription factors involved inanthocyanin synthesis (Vanderauwera et al. 2005). In con-trast, it is clear that transfer to HL failed to increase thetranscripts of selected transcription factors and enzymesinvolved in flavonoid biosynthesis in vtc1, vtc2-1 andvtc2-2, while a large increase occurred in WT plants. Thedecreased expression of transcription factors and theirtarget enzymes strongly suggests that the decreased antho-cyanin production by the vtc mutants is primarily causedby a lesion in the processes involved in the sensing orsignalling of HL. In this respect, it is already known thatvtc mutants are more resistant to biotrophic pathogens,most likely through activation of salicylic acid-mediateddefence responses (Pastori et al. 2003; Barth et al. 2004;Pavet et al. 2005; Colville & Smirnoff 2008; Mukherjeeet al. 2010). Therefore, AsA deficiency affects redox statusso as to prime pathogen defences. Increased hydrogen per-oxide (Mukherjee et al. 2010) or changes in the amount orredox state of glutathione (Pavet et al. 2005) have beenproposed to mediate the effect of AsA status. Previously, ithas been demonstrated that Arabidopsis plants with 7%of normal catalase activity, as a result of RNAi suppres-sion, are strongly impaired in HL-induced anthocyaninaccumulation and also show decreased expression of mostof the genes encoding flavonoid biosynthesis enzymes andtranscription factors (Vanderauwera et al. 2005). It appearsthat the increased hydrogen peroxide in these plantssomehow suppresses the induction of anthocyanin syn-thesis. The reason why HL-induced accumulation ofkaempferol glycosides is not affected in vtc mutants, eventhough chalcone synthase (common to both pathways)and FLS (specific to kaempferol synthesis) expressionare decreased, requires further investigation. The smallHL-induced kaempferol glycoside accumulation mustrequire a slow rate of synthesis that is therefore less

limited by enzyme capacity. In addition, if the kaempferolglycosides are more stable than the anthocyanins, an evenlower synthesis rate would be required to maintain poolsize. Catalase deficiency also has major effects on glu-tathione synthesis and redox state (Smith et al. 1984;Queval et al. 2007). These results are consistent with AsAand catalase deficiency affecting hydrogen peroxide orglutathione-mediated redox signals that are upstream oftranscription factors such as PAP1. Glutathione-deficientArabidopsis, produced by antisense suppression of g-ECsynthetase, also has decreased HL-induced anthocyaninaccumulation (Xiang et al. 2001). This result was inter-preted as indicating a requirement for transport of antho-cyanin into vacuoles as GSH conjugates. However, it isbecoming clear that anthocyanin transport may notexclusively involve GSH conjugates (Poustka et al. 2007;Zhao & Dixon 2010), leaving open the possibility that thealtered GSH levels affect anthocyanin biosynthesis moredirectly. However, this simple interpretation is compli-cated by results from APX mutants. Increased anthocya-nin accumulation occurs in an apx1/tapx double mutant inresponse to HL (Miller et al. 2007) and in response to lowtemperature in an apx1 mutant (Asai et al. 2004). Becausethese mutants are expected to have increased hydrogenperoxide levels, the results are currently not easilyreconciled with the opposite effect of decreased catalaseactivity.

ABA and JA have both been implicated in mediatingsucrose-induced anthocyanin accumulation (Loreti et al.2008; Shan et al. 2009).ABA-hypersensitive ELONGATORmutants have increased expression of anthocyanin genesand accumulate more anthocyanin (Zhou et al. 2009). It istherefore possible that AsA deficiency affects hormoneproduction or action. Leaf ABA and JA levels were notsignificantly different between WT and vtc mutants, andABA was marginally higher in HL. ABA was previouslyreported to be higher in vtc1 (Pastori et al. 2003). However,despite extensive ABA measurements additional to thoseshown here, higher levels have never been found in vtcmutants (Sultana & Smirnoff, unpublished results). Theunchanged ABA and JA levels show that AsA status doesnot affect synthesis of these hormones, but does not rule outdecreased sensitivity.

CONCLUSIONS

The ability to accumulate anthocyanins in Arabidopsis isfinely tuned by AsA status. Comparison of this responsewith catalase- and glutathione-deficient mutants, in whichanthocyanin synthesis is also impaired in HL, supports thehypothesis that the early signalling events controlling theexpression of the transcription factors that co-ordinatelyregulate the whole flavonoid biosynthesis pathway areredox sensitive. Transcriptional regulation of the flavonoidpathway is complex, with signals from light and carbohy-drate status requiring integration with negative regulationvia nitrogen status (Rubin et al. 2009). The targets andmechanism of the proposed redox regulation remain to

Influence of ascorbate on anthocyanin accumulation 401

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

be discovered. It is seemingly paradoxical that oxidativeconditions should suppress accumulation of a class ofmolecules that are potential antioxidants or photopro-tectants. Over-oxidation of a thiol/disulphide-dependentsignalling component in the mutant plants could provide anexplanation.

ACKNOWLEDGMENTS

Funding from the Biotechnology and Biological SciencesResearch Council (BBSRC), the Exeter University ScienceStrategy Fund, Hazara University, Mansehra (NWFP) Paki-stan and the Higher Education Commission (Pakistan) isgratefully acknowledged.

REFERENCES

Anders S. & Huber W. (2010) Differential expression analysis forsequence count data. Genome Biology 11, R106.

Asai N., Matsuyama T., Tamaoki M., et al. (2004) Compensation forlack of a cytosolic ascorbate peroxidase in an Arabidopsismutant by activation of multiple antioxidative systems. PlantScience 166, 1547–1554.

Barth C., Moeder W., Klessig D.F. & Conklin P.L. (2004) The timingof senescence and response to pathogens is altered in theascorbate-deficient Arabidopsis mutant vitamin C-1. PlantPhysiology 134, 1784–1792.

Bartoli C.G., Yu J.P., Gomez F., Fernandez L., McIntosh L. & FoyerC.H. (2006) Inter-relationships between light and respiration inthe control of ascorbic acid synthesis and accumulation in Ara-bidopsis thaliana leaves. Journal of Experimental Botany 57,1621–1631.

Burbulis I.E. & Winkel-Shirley B. (1999) Interactions amongenzymes of the Arabidopsis flavonoid biosynthetic pathway.Proceedings of the National Academy of Sciences of the UnitedStates of America 96, 12929–12934.

Chalker-Scott L. (1999) Environmental significance of anthocya-nins in plant stress responses. Photochemistry and Photobiology70, 1–9.

Chapple C.C.S., Vogt T., Ellis B.E. & Somerville C.R. (1992) AnArabidopsis mutant defective in the general phenylpropanoidpathway. The Plant Cell 4, 1413–1424.

Chatterjee M., Sharma P. & Khurana J.P. (2006) Cryptochrome 1from Brassica napus is up-regulated by blue light and controlshypocotyl/stem growth and anthocyanin accumulation. PlantPhysiology 141, 61–74.

Colville L. & Smirnoff N. (2008) Antioxidant status, peroxidaseactivity, and PR protein transcript levels in ascorbate-deficientArabidopsis thaliana vtc mutants. Journal of ExperimentalBotany 59, 3857–3868.

Conklin P.L., Norris S.R., Wheeler G.L., Williams E.H., Smirnoff N.& Last R.L. (1999) Genetic evidence for the role of GDP–mannose in plant ascorbic acid (vitamin C) biosynthesis. Pro-ceedings of the National Academy of Sciences of the United Statesof America 96, 4198–4203.

Conklin P.L., Saracco S.A., Norris S.R. & Last R.L. (2000) Identi-fication of ascorbic acid-deficient Arabidopsis thaliana mutants.Genetics 154, 847–856.

Das P.K., Geul B., Choi S.-B. & Park Y.-I. (2011) Photosynthesis-dependent anthocyanin pigmentation in Arabidopsis. Plant Sig-nalling & Behaviour 26, 23–25.

Davletova S., Rizhsky L., Liang H.J., Zhong S.Q., Oliver D.J., CoutuJ., Shulaev V., Schlauch K. & Mittler R. (2005) Cytosolic

ASCORBATE PEROXIDASE 1 is a central component of thereactive oxygen gene network of Arabidopsis. The Plant Cell 17,268–281.

Dowdle J., Ishikawa T., Gatzek S., Rolinski S. & Smirnoff N. (2007)Two genes in Arabidopsis thaliana encoding GDP-l-galactosephosphorylase are required for ascorbate biosynthesis and seed-ling viability. The Plant Journal 52, 673–689.

Forcat S., Bennett M.H., Mansfield J.W. & Grant M.R. (2008) Arapid and robust method for simultaneously measuring changesin the phytohormones ABA, JA and SA in plants following bioticand abiotic stress. Plant Methods 4, 16.

Foyer C.H. & Noctor G. (2011) Ascorbate and glutathione: theheart of the redox hub. Plant Physiology 155, 2–18.

Galvez-Valdivieso G. & Mullineaux P.M. (2010) The role of reac-tive oxygen species in signalling from chloroplasts to the nucleus.Physiologia Plantarum 138, 430–439.

Galvez-Valdivieso G., Fryer M.J., Lawson T., et al. (2009) The highlight response in Arabidopsis involves ABA signaling betweenvascular and bundle sheath cells. The Plant Cell 21, 2143–2162.

Giacomelli L., Rudella A. & van Wijk K.J. (2006) High lightresponse of the thylakoid proteome in Arabidopsis wild type andthe ascorbate-deficient mutant vtc2-2.A comparative proteomicsstudy. Plant Physiology 141, 685–701.

Giacomelli L., Masi A., Ripoll D.R., Lee M.J. & van Wijk K.J.(2007) Arabidopsis thaliana deficient in two chloroplast ascor-bate peroxidases shows accelerated light-induced necrosis whenlevels of cellular ascorbate are low. Plant Molecular Biology 65,627–644.

Golan T., Muller-Moulé P. & Niyogi K.K. (2006) Photoprotectionmutants of Arabidopsis thaliana acclimate to high light byincreasing photosynthesis and specific antioxidants. Plant, Cell &Environment 29, 879–887.

Gould K.S., McKelvie J. & Markham K.R. (2002) Do anthocyaninsfunction as antioxidants in leaves? Imaging of H2O2 in red andgreen leaves after mechanical injury. Plant, Cell & Environment25, 1261–1269.

Gould K.S., Dudle D.A. & Neufeld H.S. (2010) Why some stemsare red: cauline anthocyanins shield photosystem II against highlight stress. Journal of Experimental Botany 61, 2707–2717.

Grace S.C. & Logan B.A. (1996) Acclimation of foliar antioxidantsystems to growth irradiance in three broad-leaved evergreenspecies. Plant Physiology 112, 1631–1640.

Grace S.C. & Logan B.A. (2000) Energy dissipation and radicalscavenging by the plant phenylpropanoid pathway. Philosophi-cal Transactions of the Royal Society of London Series B – Bio-logical Sciences 355, 1499–1510.

Havaux M. (2003) Spontaneous and thermoinduced photon emis-sion: new methods to detect and quantify oxidative stress inplants. Trends in Plant Science 8, 409–413.

Havaux M. & Kloppstech K. (2001) The protective functions ofcarotenoid and flavonoid pigments against excess visible radia-tion at chilling temperature investigated in Arabidopsis npq andtt mutants. Planta 213, 953–966.

Hernández I., Alegre L., Van Breusegem F. & Munné-Bosch S.(2009) How relevant are flavonoids as antioxidants in plants?Trends in Plant Science 14, 125–132.

Horai H., Arita M., Kanaya S., et al. (2010) MassBank: a publicrepository for sharing mass spectral data for life sciences. Journalof Mass Spectrometry 45, 703–714.

Jeong S.W., Das P.K., Jeoung S.C., et al. (2010) Ethylene suppres-sion of sugar-induced anthocyanin pigmentation in Arabidopsis.Plant Physiology 154, 1514–1531.

Karageorgou P. & Manetas Y. (2006) The importance of being redwhen young: anthocyanins and the protection of young leaves ofQuercus coccifera from insect herbivory and excess light. TreePhysiology 26, 613–621.

402 M. Page et al.

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

Laing W.A., Wright M.A., Cooney J. & Bulley S.M. (2007) Themissing step of the l-galactose pathway of ascorbate biosynthe-sis in plants, an l-galactose guanyltransferase, increases leafascorbate content. Proceedings of the National Academy ofSciences of the United States of America 104, 9534–9539.

Landry L.G., Chapple C.C.S. & Last R.L. (1995) Arabidopsismutants lacking phenolic sunscreens exhibit enhancedultraviolet-B injury and oxidative damage. Plant Physiology 109,1159–1166.

Leyva A., Jarillo J.A., Salinas J. & Martinezzapater J.M. (1995)Low-temperature induces the accumulation of phenylalanineammonia-lyase and chalcone synthase messenger RNAs of Ara-bidopsis thaliana in a light-dependent manner. Plant Physiology108, 39–46.

Li J.Y., Oulee T.M., Raba R., Amundson R.G. & Last R.L. (1993)Arabidopsis flavonoid mutants are hypersensitive to UV-B irra-diation. The Plant Cell 5, 171–179.

Linster C.L., Gomez T.A., Christensen K.C., Adler L.N., YoungB.D., Brenner C. & Clarke S.G. (2007) Arabidopsis VTC2encodes a GDP-l-galactose phosphorylase, the last unknownenzyme in the Smirnoff–Wheeler pathway to ascorbic acid inplants. Journal of Biological Chemistry 282, 18879–18885.

Lorenzen M., Racicot V., Strack D. & Chapple C. (1996) Sinapicacid ester metabolism in wild type and a sinapoylglucose-accumulating mutant of Arabidopsis. Plant Physiology 112,1625–1630.

Loreti E., Povero G., Novi G., Solfanelli C., Alpi A. & Perata P.(2008) Gibberellins, jasmonate and abscisic acid modulate thesucrose-induced expression of anthocyanin biosynthetic genes inArabidopsis. New Phytologist 179, 1004–1016.

Lukacin R. & Britsch L. (1997) Identification of strictly conservedhistidine and arginine residues as part of the active site in Petuniahybrida flavanone 3-b-hydroxylase. European Journal of Bio-chemistry 249, 748–757.

Martens S., Preuss A. & Matern U. (2010) Multifunctional fla-vonoid dioxygenases: flavonol and anthocyanin biosynthesis inArabidopsis thaliana L. Phytochemistry 71, 1040–1049.

Miller G., Suzuki N., Rizhsky L., Hegie A., Koussevitzky S. &Mittler R. (2007) Double mutants deficient in cytosolic and thy-lakoid ascorbate peroxidase reveal a complex mode of interac-tion between reactive oxygen species, plant development, andresponse to abiotic stresses. Plant Physiology 144, 1777–1785.

Mukherjee M., Larrimore K.E., Ahmed N.J., Bedick T.S., Barg-houthi N.T., Traw M.B. & Barth C. (2010) Ascorbic acid defi-ciency in Arabidopsis induces constitutive priming that isdependent on hydrogen peroxide, salicylic acid, and the NPR1gene. Molecular Plant–Microbe Interactions 23, 340–351.

Muller-Moulé P. (2008) An expression analysis of the ascorbatebiosynthesis enzyme VTC2. Plant Molecular Biology 68, 31–41.

Muller-Moulé P., Havaux M. & Niyogi K.K. (2003) Zeaxanthindeficiency enhances the high light sensitivity of an ascorbate-deficient mutant of Arabidopsis. Plant Physiology 133, 748–760.

Muller-Moulé P., Golan T. & Niyogi K.K. (2004) Ascorbate-deficient mutants of Arabidopsis grow in high light despitechronic photooxidative stress. Plant Physiology 134, 1163–1172.

Murchie E.H. & Niyogi K.K. (2011) Manipulation of photoprotec-tion to improve plant photosynthesis. Plant Physiology 155,86–92.

Nair R.B., Bastress K.L., Ruegger M.O., Denault J.W. & Chapple C.(2004) The Arabidopsis thaliana Reduced Epidermal Fluores-cence1 gene encodes an aldehyde dehydrogenase involved inferulic acid and sinapic acid biosynthesis. The Plant Cell 16,544–554.

Ormrod D.P., Landry L.G. & Conklin P.L. (1995) Short-term UV-Bradiation and ozone exposure effects on aromatic secondary

metabolite accumulation and shoot growth of flavonoid-deficient Arabidopsis mutants. Physiologia Plantarum 93, 602–610.

Pastori G.M., Kiddle G., Antoniw J., Bernard S., Veljovic-JovanovicS., Verrier P.J., Noctor G. & Foyer C.H. (2003) Leaf vitamin Ccontents modulate plant defense transcripts and regulate genesthat control development through hormone signaling. The PlantCell 15, 939–951.

Pavet V., Olmos E., Kiddle G., Mowla S., Kumar S., Antoniw J.,Alvarez M.E. & Foyer C.H. (2005) Ascorbic acid deficiencyactivates cell death and disease resistance responses in Arabi-dopsis. Plant Physiology 139, 1291–1303.

Perron N.R. & Brumaghim J.L. (2009) A review of the antioxidantmechanisms of polyphenol compounds related to iron binding.Cell Biochemistry and Biophysics 53, 75–100.

Poustka F., Irani N.G., Feller A., Lu Y., Pourcel L., Frame K. &Grotewold E. (2007) A trafficking pathway for anthocyaninsoverlaps with the endoplasmic reticulum-to-vacuole protein-sorting route in Arabidopsis and contributes to the formation ofvacuolar inclusions. Plant Physiology 145, 1323–1335.

Prescott A.G. & John P. (1996) Dioxygenases: molecular structureand role in plant metabolism. Annual Review of Plant Physiol-ogy and Plant Molecular Biology 47, 245–271.

Queval G., Issakidis-Bourguet E., Hoeberichts F.A., Vandorpe M.,Gakiére B., Vanacker H., Miginiac-Maslow M., Van BreusegemF. & Noctor G. (2007) Conditional oxidative stress responses inthe Arabidopsis photorespiratory mutant cat2 demonstrate thatredox state is a key modulator of daylength-dependent geneexpression, and define photoperiod as a crucial factor in theregulation of H2O2-induced cell death. The Plant Journal 52,640–657.

Rahman M.M., Ichiyanagi T., Komiyama T., Hatano Y. & Konishi T.(2006) Superoxide radical- and peroxynitrite-scavenging activityof anthocyanins; structure–activity relationship and their syner-gism. Free Radical Research 40, 993–1002.

Rubin G., Tohge T., Matsuda F., Saito K. & Scheible W.R. (2009)Members of the LBD family of transcription factors repressanthocyanin synthesis and affect additional nitrogen responsesin Arabidopsis. The Plant Cell 21, 3567–3584.

Shan X.Y., Zhang Y.S., Peng W., Wang Z.L. & Xie D.X. (2009)Molecular mechanism for jasmonate-induction of anthocyaninaccumulation in Arabidopsis. Journal of Experimental Botany60, 3849–3860.

Shin J., Park E. & Choi G. (2007) PIF3 regulates anthocyaninbiosynthesis in an HY5-dependent manner with both factorsdirectly binding anthocyanin biosynthetic gene promoters inArabidopsis. The Plant Journal 49, 981–994.

Smirnoff N. (2000) Ascorbate biosynthesis and function in photo-protection. Philosophical Transactions of the Royal Society ofLondon Series B – Biological Sciences 355, 1455–1464.

Smith I.K., Kendall A.C., Keys A.J., Turner J.C. & Lea P.J. (1984)Increased levels of glutathione in a catalase-deficient mutantof barley (Hordeum vulgare L.). Plant Science Letters 37, 29–33.

Stobiecki M., Skirycz A., Kerhoas L., Kachlicki P., Muth D.,Einhorn J. & Mueller-Roeber B. (2006) Profiling of phenolicglycosidic conjugates in leaves of Arabidopsis thaliana usingLC/MS. Metabolomics 2, 197–219.

Takahashi S. & Badger M. (2011) Photoprotection in plants: a newlight on photosystem II damage. Trends in Plant Science 16,53–60.

Takahashi S., Milward S.E., Yamori W., Evans J.R., Hillier W. &Badger M.R. (2010) The solar action spectrum of photosystem IIdamage. Plant Physiology 153, 988–993.

Teng S., Keurentjes J., Bentsink L., Koornneef M. & Smeekens S.(2005) Sucrose-specific induction of anthocyanin biosynthesis in

Influence of ascorbate on anthocyanin accumulation 403

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404

Arabidopsis requires the MYB75/PAP1 gene. Plant Physiology139, 1840–1852.

Thimm O., Bläsing O., Gibon Y., Nagel A., Meyer S., Krüger P.,Selbig J., Müller L.A., Rhee S.Y. & Stitt M. (2004) MAPMAN: auser-driven tool to display genomics data sets onto diagrams ofmetabolic pathways and other biological processes. The PlantJournal 37, 914–939.

Tohge T., Nishiyama Y., Hirai M.Y., et al. (2005) Functional genom-ics by integrated analysis of metabolome and transcriptome ofArabidopsis plants over-expressing an MYB transcription factor.The Plant Journal 42, 218–235.

Toth S.Z., Puthur J.T., Nagy V. & Garab G. (2009) Experimentalevidence for ascorbate-dependent electron transport in leaveswith inactive oxygen-evolving complexes. Plant Physiology 149,1568–1578.

Vanderauwera S., Zimmermann P., Rombauts S., Vandenabeele S.,Langebartels C., Gruissem W., Inzé D. & Van Breusegem F.(2005) Genome-wide analysis of hydrogen peroxide-regulatedgene expression in Arabidopsis reveals a high light-induced tran-scriptional cluster involved in anthocyanin biosynthesis. PlantPhysiology 139, 806–821.

Winkel-Shirley B. (2002) Biosynthesis of flavonoids and effects ofstress. Current Opinion in Plant Biology 5, 218–223.

Xiang C.B., Werner B.L., Christensen E.M. & Oliver D.J. (2001)The biological functions of glutathione revisited in Arabidopsistransgenic plants with altered glutathione levels. Plant Physiol-ogy 126, 564–574.

Yabuta Y., Mieda T., Rapolu M., Nakamura A., Motoki T., MarutaT., Yoshimura K., Ishikawa T. & Shigeoka S. (2007) Light regu-lation of ascorbate biosynthesis is dependent on the photosyn-thetic electron transport chain but independent of sugars inArabidopsis. Journal of Experimental Botany 58, 2661–2671.

Yamasaki H., Sakihama Y. & Ikehara N. (1997) Flavonoid-peroxidase reaction as a detoxification mechanism of plant cellsagainst H2O2. Plant Physiology 115, 1405–1412.

Zechmann B., Stumpe M. & Mauch F. (2010) Immunocytochemicaldetermination of the subcellular distribution of ascorbate inplants. Planta 233, 1–12.

Zeng X.Q., Chow W.S., Su L.J., Peng X.X. & Peng C.L. (2010)Protective effect of supplemental anthocyanins on Arabidopsisleaves under high light. Physiologia Plantarum 138, 215–225.

Zhang K.M., Yu H.J., Shi K., Zhou Y.H., Yu J.Q. & Xia X.J. (2010)Photoprotective roles of anthocyanins in Begonia semperflorens.Plant Science 179, 202–208.

Zhao J. & Dixon R.A. (2010) The ‘ins’ and ‘outs’ of flavonoidtransport. Trends in Plant Science 15, 72–80.

Zhou X.F., Hua D.P., Chen Z.Z., Zhou Z.J. & Gong Z.Z. (2009)Elongator mediates ABA responses, oxidative stress resistanceand anthocyanin biosynthesis in Arabidopsis. The Plant Journal60, 79–90.

Received 3 February 2011; accepted for publication 2 May 2011

SUPPORTING INFORMATION

Additional Supporting Information may be found in theonline version of this article:

Table S1. The effect of high light (HL) acclimation on thetranscript levels of flavonoid, phenylpropanoid and ascor-bate biosynthesis genes in wild-type (WT) Arabidopsisdetermined by mRNA-Seq.

Please note: Wiley-Blackwell are not responsible for thecontent or functionality of any supporting materials sup-plied by the authors. Any queries (other than missing mate-rial) should be directed to the corresponding author for thearticle.

404 M. Page et al.

© 2011 Blackwell Publishing Ltd, Plant, Cell and Environment, 35, 388–404