GLYCOALKALOID METABOLISM1 Is Required for Steroidal Alkaloid Glycosylation and Prevention ...

20 Pages·2012·2.19 MB·English

ThePlantCell,Vol.23:4507–4525,December2011,www.plantcell.orgã2011AmericanSocietyofPlantBiologists.Allrightsreserved.
GLYCOALKALOID METABOLISM1 Is Required for Steroidal
Alkaloid Glycosylation and Prevention of Phytotoxicity
in Tomato W
MaximItkin,a,1IlanaRogachev,aNoamAlkan,b,2TallyRosenberg,cSergeyMalitsky,aLauraMasini,dSagitMeir,a
YokoIijima,e,3KohAoki,eRicdeVos,dDovPrusky,bSaulBurdman,cJulesBeekwilder,dandAsaphAharonia,4
aDepartmentofPlantSciences,WeizmannInstituteofScience,Rehovot76100,Israel
bAgriculturalResearchOrganization,TheVolcaniCenter,BetDagan50250,Israel
cDepartment of Plant Pathology and Microbiology, The Robert H. Smith Faculty of Agriculture, Food, and Environment, The
HebrewUniversityofJerusalem,Rehovot76100,Israel
dPlantResearchInternational,Wageningen6700AA,TheNetherlands
eKazusaDNAResearchInstitute,Kisarazu292-0818,Japan
Steroidal alkaloids (SAs) are triterpene-derived specialized metabolites found in members of the Solanaceae family that
provide plants with a chemical barrier against a broad range of pathogens. Their biosynthesis involves the action of
glycosyltransferasestoformsteroidalglycoalkaloids(SGAs).ToelucidatethemetabolismofSGAsintheSolanaceaefamily,
weexaminedthetomato(Solanumlycopersicum)GLYCOALKALOIDMETABOLISM1(GAME1)gene.Ourfindingsimplythat
GAME1isagalactosyltransferase,largelyperformingglycosylationoftheaglyconetomatidine,resultinginSGAproduction
in green tissues. Downregulation of GAME1 resulted in an almost 50% reduction in a-tomatine levels (the major SGA in
tomato) and a large increase in its precursors (i.e., tomatidenol and tomatidine). Surprisingly, GAME1-silenced plants
displayedgrowthretardationandseveremorphologicalphenotypesthatwesuggestoccurasaresultofalteredmembrane
sterollevelscausedbytheaccumulationoftheaglyconetomatidine.Together,thesefindingshighlighttheroleofGAME1in
theglycosylationofSAsandinreducingthetoxicityofSAmetabolitestotheplantcell.
INTRODUCTION In plants, SAs serve as phytoanticipins, providing the plant
with a preexisting chemical barrier against a broad range of
Thesteroidalalkaloids(SAs),alsoknownassolanumalkaloids,
pathogens(ChanandTam,1985;Guntheretal.,1997;Sandrock
arecommonconstituentsofnumerousplantsbelongingtothe
and Vanetten, 1998; Hoagland, 2009). For example, tomato
Solanaceaefamily,inparticularmembersofthegenusSolanum a-tomatine acts via disruption of membranes, followed by the
(Rahmanetal.,1998),whichcomprises1350species.SAshave
leakageofelectrolytesanddepolarizationofthemembranepo-
beenextensivelyinvestigatedfortheirdiversebiologicalactivi-
tential(McKee, 1959; SteelandDrysdale, 1988;Keukensetal.,
ties and occurrence in important crop plants (e.g., tomato
1992,1995).However,itwassuggestedthattomatoplantsarenot
[Solanumlycopersicum],potato[Solanumtuberosum],andegg-
affected by its presence, possibly due the existence of sterol
plant[Solanummelongena])(Eich,2008).ThesynthesisofSAs,
glycosidesandacetylatedsterolglycosidesintomatocellmem-
whichispresumedtostartfromcholesterol,likelyoccursinthe
branes(Roddick,1976a;SteelandDrysdale,1988;Blankemeyer
cytosolandinmostcasesinvolvesfurtherglycosylationofthe
etal.,1997).GlycosylationofSAsisbelievedtoreducethetoxicity
alkaminesteroidalskeleton(aglycone)atC-3btoformsteroidal ofa-tomatinetotheplantcell,astreatmentofleafdisksfromfour
glycoalkaloids(SGAs)(Bowles,2002;Friedman,2002;Arnqvist plantspeciesdemonstratedthata-tomatinecausedmoresevere
etal.,2003;Kalinowskaetal.,2005;Bowlesetal.,2006).
electrolyte leakage than did its aglycone tomatidine (Hoagland,
2009). In addition, studies of triterpene saponins in oat (Avena
1Current address: Agricultural Research Organization, The Volcani sativa) and Medicago truncatula pointed to the same role of
Center,POBox6,BetDagan50250,Israel. glycosylation(Mylonaetal.,2008;Naoumkinaetal.,2010).
2Currentaddress:DepartmentofPlantSciences,WeizmannInstituteof Recently,>50differentSAswereputativelyidentifiedintuber
Science,POBox26,Rehovot76100,Israel.
extractsfromsevengenotypes(bothwildandcultivatedspecies)
3Currentaddress:FacultyofAppliedBioscience,KanagawaInstituteof
(Shakya and Navarre, 2008). In cultivated potato, a-chaconine
Technology,1030Shimo-ogino,Atsugi,Kanagawa243-0292,Japan.
[email protected]. and a-solanine comprise>90%ofthetotal SAs.Threegenes,
The author responsible for distribution of materials integral to the encoding putative glycosyltransferases (GTs) involved in the
findingspresentedinthisarticleinaccordancewiththepolicydescribed biosynthesisofa-solanineanda-chaconinefromtheaglycone
in the Instructions for Authors (www.plantcell.org) is: Asaph Aharoni
solanidine,havebeenidentifiedinpotato.TheSolanumtuber-
([email protected]).
WOnlineversioncontainsWeb-onlydata. osumsterolalkaloidglycosyltransferase1(SGT1)genebehaves
www.plantcell.org/cgi/doi/10.1105/tpc.111.088732 in vitro as a UDP-Gal:solanidine galactosyltransferase (Moehs
4508 ThePlantCell
etal.,1997).SilencingSGT1inpotatoresultedinredirectionof Hierarchical clustering of the profiling data revealed several
themetabolicflux,causingstrongreductioninaccumulationof clusters of tissue-/developmental stage–specific SAs. For ex-
a-solanineandsignificantaccumulationofa-chaconine,inwhich ample, 25 metabolites, including a-tomatine, were associated
the primary glycosyl unit is Gal and Glc, respectively (McCue withgreentissues,astheyweredetectedinunripefruit,leaves,
etal.,2005).SGT2wasshowntoencodeaUDP-Glc:solanidine buds,andflowers(bothcontainingsepals).Wealsodetected32
glucosyltransferase (McCue etal.,2006),whichcouldmediate SAsassociatedwithtissuesoftheRRfruitstage(inpeel,flesh,
a-chaconine biosynthesis. Finally, SGT3 encodes a UDP-Rha: andseeds).ThelevelsofSAswereverylowinpollen,consisting
b-solanine/b-chaconine rhamnosyltransferase (McCue et al., of only one SA (hydroxy-dehydrotomatidine trihexoside plus
2007). deoxyhexose),whichwasalsofoundinpollen-containingbuds
Approximately 100 different SAs have been described in andflowers.Apartfromthe32SAsuniquetotheRRfruitstage
various tomato tissues, particularly in fruit (Moco et al., 2006; tissues, seeds harvested at that stage contained nine unique
Iijimaetal.,2008;Kozukueetal.,2008;Mintz-Oronetal.,2008; SAs. In roots, we identified two unique isomers of dehydroto-
Yamanakaetal.,2008).ThemajorSAintomato,a-tomatine,was matine.Budsandflowersaccumulated14SAsthatwereunique
reportedtobepresentinthegreentissuesoftheplanttogether tothesetissues,includingfiveisomersofhydroxytomatine.The
withdehydrotomatine(Kozukueetal.,2004).Whereasa-tomatine alkaminetomatidine,theprecursorofa-tomatine,wasfoundat
levelsdecreaseasthefruitmaturesandripens,recentstudies verylowlevelsinmostgreentissuesandinroots.
suggest a ripening-dependent conversion of a-tomatine into
esculeosideA,themostabundantSAinthered-ripe(RR)tomato
GAME1IsPartofaCladeofSAandSteroidalSaponinGTs
fruit(Fujiwaraetal.,2004;Mocoetal.,2007;Iijimaetal.,2008;
Mintz-Oron et al., 2008; Yamanaka et al., 2009). The levels of GTs of the plant Group 1 multigene family (120 members in
esculeoside A appear to be ripening and ethylene dependent Arabidopsis thaliana; Paquette et al., 2003) transfer the sugar
(Iijima et al., 2009). In tomato, very little is known about the moiety fromUDP-sugar toavast arrayof lowmolecular mass
enzymesandgenescontributingtoSAbiosynthesis.Asoluble acceptors, including secondary metabolites and hormones
protein fraction from tomato leaves was found to exhibit ga- (Bowles,2002).AsglycosylationofSAsisthoughtbecrucialfor
lactosyltransferaseactivityandweakglucosyltransferaseactiv- theirbiologicalactivity(MorrisseyandOsbourn,1999),wesetout
ity (Zimowski, 1998). However, no experiments have been to discover tomato GTs that could act in the SA pathway. We
reportedaboutgenesthatarerelevanttoSAcontentandplay usedthenucleicacidsequenceoftheGTreportedtoglycosylate
aroleintomato. SAsinpotato(St-SGT1;McCueetal.,2005)andidentifiedthree
Here,identificationoftheGLYCOALKALOIDMETABOLISM1 similarGTsfromtomato(GAME1to3).Inaphylogeneticanalysis
(GAME1)geneprovidesinsightintotheSAbiosyntheticpathway based on the publicly available full-length GT amino acid se-
in tomato. Glycosylation by GAME1 appears to be crucial to quences(seeSupplementalFigure1andSupplementalDataSet
preventthetoxiceffectofSAstotheplantcell,andGAME1is 1online),thesethreetomatoproteinsformedaseparateclade
likelyinvolvedinattachingtheGalgrouptotheC-3bpositionof withGTsfrompotato(St-SGT1-3)thatuseSAsandtwoSolanum
tomatidine aspart of the lycotetraose moiety formation. When aculeatissimumproteins(Sa-GT4AandSa-GT4R)showntouse
GAME1 is silenced, the new composition of SAs results in both SAs and steroidal saponins as sugar acceptors (Moehs
toxicitytotheplantcell(mostlikelyduetotheincreasedlevels etal.,1997;Koharaetal.,2005;McCueetal.,2005,2006,2007).
of the aglycone tomatidine) and as a consequence causes In this clade, the three GAME proteins each clustered most
marked developmental defects, including growth retardation. closelywithoneofthepotatoSGTproteins.TomatoGAME1(Sl-
Ourresultssuggestthatthistoxicityisduetoalterationinsterol GT1; UGT73L5, according to the nomenclature guidelines;
metabolism. We envisage that this work will promote future Mackenzieetal.,1997)exhibits91%identityattheaminoacid
researchtounravelfurthertheSApathwayanditssignificanceto levelto St-SGT1 (aUDP-Gal solanidine galactosyltransferase).
plantfitness. Two additional tomato genes homologous to St-SGT2 (coding
for UDP-Glc:solanidine glucosyltransferase) and to St-SGT3
(coding for UDP-Rha:b-solanine/b-chaconine rhamnosyltrans-
RESULTS ferase)werenamedGAME3(UGT73L6)andGAME2(UGT73L4),
respectively(seeSupplementalFigure1andSupplementalTable
MetabolicProfilingRevealsUniqueClustersofSAsThatAre 2 online). Further searches in genomic DNA scaffolds and
DistributedacrossDiverseTomatoPlantTissues comparison of the results with available BAC sequences
(http://solgenomics.net) revealed that GAME1 is located on
To examine the occurrence of SAs in different tomato (cv chromosome7.
MicroTom)plantparts,weidentifiedandexaminedthedistribu-
tion of 85 putative SAs in 21 tissues and fruit developmental GAME1ExpressionIsNegativelyRegulatedbyEthylene
stages by ultraperformance liquid chromatography coupled to
duringFruitRipeningandIsPredominantinGreenTissues
quadrupoletime-of-flightmassspectrometry(UPLC-qTOF-MS)
analysis(Figure1;seeSupplementalTable1online).Amongthe We subsequently focused our interest on GAME1 since its
setofidentifiedSAs,47representedthosewithauniquechem- expressionpatternwassimilartotheprofileofSAsaccumulating
ical formula, while the remaining substances were putatively specificallyingreentissues,beingfoundprimarilyinyoungand
identified as their isomers (see Supplemental Table 1 online). matureleaves,peel,flesh,andseedsderivedfromimmatureand
TomatoGLYCOALKALOIDMETABOLISM1Activity 4509
Figure1. DiversityofSAsinTomatoandItsCorrespondencewithGAME1Expression.
(A)DiversityofSAsintomato.HierarchicalclusteringofSAsobtainedbyUPLC-qTOF-MSanalysis.Yellowframesencloseseveralmetaboliteclusters
discussed in the text, whereas white arrows indicate the possible conversion of green tissue–associated metabolites into RR tissue–associated
metabolitesduringfruitripening.NumberswithinparenthesescorrespondtoSAsinSupplementalTable1online.Relativelevelsoffiveadditional
putativeSAsthatwereidentifiedinthecourseofthisstudycouldnotbemeasured.ThesesubstancesarelistedinSupplementalTable1online.
(B)GAME1expressionintomatotissues,measuredbyquantitativereal-timePCR.Thestatisticalsignificanceofthegeneexpressiondataforeach
4510 ThePlantCell
maturegreen(MG)fruitandflowerbuds(Figures1and2A).The efficiency astomatidine (seeSupplementalTable 3 online).Ap-
tightassociationbetweenGAME1transcriptlevelandtheaccu- parently,thedoublebondbetweencarbon5andcarbon6inthe
mulationofparticularSAs,primarilya-tomatineanddehydroto- B-ringofthesolanidineanddemissidinemoleculesinterfereswith
matineisomersandderivativesinmostoftheexaminedtissues, efficientgalactosylation.Othersteroidsubstratestested(choles-
suggested that it could be involved in the metabolism of this terol,campesterol,b-sistostanol,b-sitosterol,24-epibrassinolide,
setofSAs.GAME1expressionappearedtobedownregulated and 24-epicastasterone) did not show any detectable product
duringfruitripening(Figure2A).WefurthermeasuredGAME1 formationinthepresenceofGAME1andUDP-Gal.Thus,GAME1
expressioninfruittreatedwith1-methylcyclopropene(1-MCP), likely acts as a UDP-Gal galactosyltransferase for SAs, with a
aninhibitorofethyleneperceptionthatnegativelyaffectsfruit preferencefortomatidineasasubstrate.
ripening(Yokotanietal.,2009),findingthatGAME1expression Wefurtherexaminedtomatidine-galactosyltransferaseactivity
isnegativelyregulatedbytheethylenesignalingcascadethat in five tissues and developmental stages of wild-type plants
typicallytriggersthefruitripeningprocess(Figure2B).GAME1 (Figure2D).TheactivitywasconsistentwithGAME1expression
expressionwasfurtherexaminedinfruitoftheripeninginhibitor levels as measured in these tissues, showing high levels in
(rin)andnon-ripening(nor)mutants,whicharealsoimpairedin immature green peel and leaves (0.10 and 0.36 nmol·mg dry
theethylenesignalingcascade(HernerandSink,1973;Thompson weight21·min21,respectively)andrelativelylowactivityintheRR
etal.,1999).GAME1transcriptlevelsweresignificantlyhigherin peelandinroots(0.007and0.031nmol·mgdryweight21·min21,
thenororange(Or)andRRfruitstagesandintherinRRstage respectively).
(albeit a trend of increase was observed also at the Or stage;
Figure2C).Furthermore,GAME1expressionseemedtobemore
GAME1-SilencedPlantsDisplaySevere
affectedinthenorbackgroundthanintherinmutant.
DevelopmentalDefects
ToclarifytheroleofGAME1inSAbiosynthesis,wegenerated18
FunctionalCharacterizationoftheGAME1Recombinant
EnzymeProducedinEscherichiacoliCellsandin independenttransgenictomatoplants(incvMicroTom)wherein
GAME1wassilencedviaRNAinterference(RNAi)usingthe35S
VivoActivity
cauliflower mosaic virus promoter (hereafter referred to as
ThecorrelationbetweenGAME1expressionpatternsandcon- GAME1i plants). Interestingly, GAME1 silencing had a strong
tents of a-tomatine and its derivatives across the 21 different effectontomatoplantgrowthanddevelopment,whichisunex-
tomatoplanttissuessuggestedthattheenzymeitencodesmight pectedwhensilencingageneassociatedwithspecializedme-
beactingintheformationofthea-tomatinelycotetraoseglycosyl tabolism. Seven GAME1i lines exhibited a severe phenotype
chain.Ina-tomatine,thismoietycontainstwomoleculesofD-Glc (SPh) with strongly retarded growth compared with wild-type
and one each of D-Xyl and D-Gal, the latter attached to the plants(Figure4).Theseplantsdevelopeddeformedleavesand
tomatidine aglycone (see Supplemental Figure 2 online). We produced numerous small flower buds that in most cases
analyzed the specificity of GAME1 by in vitro enzyme assays aborted before fertilization (Figures 4B to 4Dand 4F). The low
usingarecombinantenzymeproducedinE.coli.UDP-Glcand amountoffruitthatoccasionallydevelopedintheSPhGAME1i
UDP-Galwerecomparedassugardonorswithtomatidineasa plants was parthenocarpic and shriveled (Figure 4K). Four
substrate. UDP-Gal was readily used as a donor, producing a additional GAME1i plants exhibited a mild phenotype (MPh).
novel product withamass-to-charge ratioof 578m/z ([M+H]), These plants displayed normal plant architecture except for
correspondingtogalactosylatedtomatidine(Figure3).Dehydro- theirfruit,whichdevelopedwoody,possiblysuberizedregions,
tomatidine(414m/z;Figure3),identifiedinthetomatidinestan- mostly around the fruit apex (Figures 4I and 4J). We further
dard as a contaminant (Kozukue et al., 2004), was also investigated one SPh and two MPh GAME1 lines. GAME1
galactosylated by GAME1 (product 576 m/z). UDP-Glc was transcript levels in GAME1i leaves were significantly reduced
alsoincorporatedtosomeextent,however,notexceeding5% (seeSupplementalFigure3online).Attheenzymaticlevel,the
of the Gal incorporation under identical concentrations and tomatidine-galactosyltransferaseactivitywasdeterminedinma-
conditions.ThissuggeststhatGAME1isprimarilyagalactosyl- tureleafextractsderivedfromtheGAME1i(SPh)line,constitut-
transferase. The K of GAME1 fortomatidine (in thepresence ing only 11% of the activity in the wild type (0.002 and 0.018
m
of8mMUDP-Gal)wasdeterminedtobe38612mMandthe nmol·mgfreshweight21·min21,respectively)(Figure5A).Thus,
k 1.860.4min21.Fromthesevalues,thecatalyticefficiency silencing of GAME1 resulted in a severe reduction in GAME1
cat
(k /K ) was calculated as 783 M21·s21. All other substrates expressionandreducedtomatidine-galactosyltransferaseactiv-
cat m
testedshowedlowercatalyticefficiencyornoturnoveratall.In ityintheleaves.
particular, solanidine and demissidine showed lower turnover TheGAME1iconstructwasalsointroducedintotheAilsaCraig
rates,whereassolasoninewasstillinthesamerangeofcatalytic background,atypicalindeterminatecultivar(seeSupplemental
Figure1. (continued).
examinedtissueispresentedinFigure2A.
For(A)and(B),thecolorindexreferstotherelativelevelsofaparticularmetaboliteortheGAME1transcriptacrossthedifferenttissuesexamined;the
highestlevelisdefinedas100%(n=3).Br,breaker;IG,immaturegreen;M,mature(fullyexpanded);Y,young.
TomatoGLYCOALKALOIDMETABOLISM1Activity 4511
Figure2. ExpressionofGAME1intheWildTypeandFruitoftherinandnorMutantsandTomatidine-GalactosyltransferaseActivityinWild-Type
TomatoTissues.
(A)Quantitativereal-timePCRrelativeexpressionanalysesofGAME1transcriptsin21tissuesofwild-typetomato(cvMicroTom).Differentlettering
abovethebarsdenotessignificantdifferencesingeneexpressionascalculatedbyaStudent’sttest(P<0.05;n=3).
(B)GAME1expressioniselevatedin1-MCP–treatedfruit(cvAilsaCraig)intheMGandOrdevelopmentalstagescomparedwithuntreatedcontrolfruit.
Student’sttestresultsforsignificance(P<0.05;n=3)areindicatedbyanasterisk.
(C)GAME1expressionintomatofruitoftherinandnorripeningmutants(cvAilsaCraig)ishigherthaninfruitofthewild-type(WT)plantsattheOrand
RRfruitstages.DifferentletteringabovethebarsdenotessignificantdifferencesinGAME1relativeexpressionlevelsascalculatedbyaStudent’sttest
(P<0.05;n=3).
(D)Tomatidine-galactosyltransferaseactivityinfivetissuesoftomato(cvMicroTom).Activitywasmeasuredinfreeze-driedsamplesandisexpressed
onadryweight(DW)basis.DifferentletteringabovethebarsdenotessignificantdifferencesinactivityascalculatedbyaStudent’sttest(P<0.05;
n=3).
Inallpanels,thebarsrepresentSE.Br,breaker;IG,immaturegreen;ML,mature(fullyexpanded)leaves;YL,youngleaves.
Figure4online).Theseplantswerealsodeformedwithstrongly levelsofa-tomatineanditsisomer,threeisomersofdehydroto-
retarded growth (see Supplemental Figure 4B online) and dis- matine, hydroxy-dehydrotomatine, isomer #7 of hydroxytoma-
playedbrowningandsuberizedregionsalongthestem(Figure tine,acetoxy-dehydrotomatine,dehydrotomatinetetrahexoside,
4L;seeSupplementalFigure4Conline).AlthoughtheAilsaCraig andtheunidentifiedglycoalkaloid(UGA)4.Ontheotherhand,
GAME1iplantsproducedseveralflowers,theyfailedtobearfruit. hydroxytomatine isomer #4, acetoxy-hydroxytomatine isomer
#1, and UGA 5 levels were increased approximately twofold
in SPh GAME1i leaves relative to the wild type (Table 1; see
LeavesofGAME1-SilencedPlantsExhibitAlteredLevels
Supplemental Figure 6 and Supplemental Table 4 online). In
ofSAs
addition, we observed strong and significant accumulation of
MetabolicprofilingofSAsusingUPLC-qTOF-MSwasperformed the a-tomatine precursors: tomatidine and dehydrotomatidine
onextractsfromGAME1iandwild-typeleaves.Principalcom- (alsonamedtomatidenol).
ponent analysis (PCA) indicated clear differences between the a-Tomatineandtomatidineconcentrationsweremeasuredin
SA composition of extracts derived from SPh GAME1i leaves GAME1i and wild-type leaves. Interestingly, a-tomatine levels
andthoseofthewildtype(seeSupplementalFigure5Aonline). werereducedmorethan1.6-foldinSPhGAME1icomparedwith
The SPh GAME1ileaf extracts contained significantly reduced thewildtype(Figure5B).Moreover,thetomatidineaglyconewas
4512 ThePlantCell
immature green to the RR stage) was generated based on
previouslypublisheddata(Friedman,2002;Kozukueetal.,2004;
Moco et al., 2007; Iijima et al., 2008, 2009; Mintz-Oron et al.,
2008)andthedatafromthisstudy(Figure6).
GAME1-SilencedPlantsExhibitAlteredSterolComposition
As the severe developmental phenotype observed in the SPh
GAME1iplantswassimilartophenotypesofbrassinosteroidor
gibberellicacid(GA )biosynthesisandsignalingmutants(Bishop
3
etal.,1999;Ueguchi-Tanakaetal.,2005),weattemptedtorestore
theGAME1iphenotypebyapplicationof24-epibrassinolide,
24-epicastasterone,orGA .AlthoughapplicationofGA restored
3 3
theinternodeelongationofSPhGAME1iplantstosomeextent,
treatmentswithbothbrassinosteroidsubstanceswereineffective
(datanotshown).
Alteration of sterol metabolism was also considered as a
reason for the morphological changes, since SA and sterol
biosynthesisaretightlylinked(Arnqvistetal., 2003). Moreover,
tomatidine,whichaccumulatestohighlevelsinGAME1iplants,
wasreportedtomimictheeffectofsterolbiosynthesisinhibitors
Figure 3. The Recombinant GAME1 Catalyzes the Galactosylation of (Simonsetal.,2006).Therefore,theconcentrationsof10different
Tomatidine. sterolsandtriterpenoidsweremeasuredviagaschromatography
(GC)–MS analysis of leavesfromSPh GAME1i, a setof segre-
EnzymeactivityassaysoftherecombinantGAME1producedinE.coli
cells. Thechromatograms (total ion count inUPLC-qTOF-MS) display gating F1 plants that displayed a normal wild-type phenotype,
analysisoftheGAME1recombinantproteinincubatedfor1hwith0.25 andthewildtype.Majordifferenceswereobservedinthecom-
mMtomatidineinthepresence(top)orabsence(bottom)of8mMUDP- positionofsterolsandtriterpenoids,bothinpathwayintermedi-
Gal(UDP-gal).Them/zofelutingcompoundsisindicated.Arrowspoint atesaswellasinendproducts,particularlyinthetwobranches
totomatidenol(dehydrotomatidine)(right)anddehydrotomatineplusGal downstream of 24-methylenelophenol (Figure 7A; see Supple-
(left). mental Figure7 online). Whereas campesterol levels increased
1.8-fold,thelevelsofstigmasterolanditsprecursorb-sitosterol
dramatically increased, from near baseline levels in wild-type were reduced. Small but significant reductions in cycloartenol
leaves up to 290-fold in SPh GAME1i (Figure 5B). In the MPh and b-amyrin were measured in SPh GAME1i leaves. Interest-
GAME1i leaves, we observed an almost 1.6-fold reduction of ingly, a-amyrin, (S)-2,3-oxidosqualene, cholesterol, lanosterol,
a-tomatinerelativetothewildtype(seeSupplementalFigure5C andcholestanolweredetectedattheirnativelevels.Segregating
online). However, tomatidine and dehydrotomatidine levels in F1 plants bearing a wild-type phenotype had a sterol profile
MPh GAME1i leaves were not different from the basal levels comparabletothatofthewildtype(seeSupplementalFigure7
detectedinwild-typeleaves. online).TheGAME1iphenotypecouldnotberescuedbyappli-
cationofb-sitosterol,stigmasterol,orboth(datanotshown).
MisbalancedphytosterollevelsintheArabidopsiscycloartenol
MGFruitofGAME1iPlantsDisplayAlteredLevelsofSAs
synthasemutantresultinbleachingofinflorescenceshootsand
SAprofilesofMGfruitoftheMPhGAME1ilineswerecompared chloroplastslackingthethylakoidmembranesystemandstarch
with wild-type fruits using PCA (see Supplemental Figure 5B granules and accumulating large plastoglobuli-like vesicles
online).GAME1iplantswithaseverephenotypewereexcluded (Babiychuketal.,2008).Althoughnobleachingwasobservedin
fromthisanalysis,astheyproducedveryfewfruit.ProfilesofMG GAME1i plants, transmission electron microscopy (TEM) re-
fruitfromMPhGAME1iwereclearlydifferentfromthewildtype vealed major changes in plastid ultrastructure. Chloroplasts of
(seeSupplementalFigure5Bonline).InMGfruitoftheGAME1i SPh GAME1i leaves possessed large plastoglobuli (0.4 mm in
#1andGAME1i#2lines,wemeasuredareductioninnumerous diametercomparedwith0.2mminwild-typeplants)andaltered
green tissue–associated SAs, including a-tomatine, and in hy- starchgrainandgranastructures(Figures7Bto7E).
droxy-dehydrotomatidinetrihexosideplusdeoxyhexose,whereas
elevatedlevelswererecordedfortomatidine(5timesmorethanin
StressResponse–RelatedSymptomsandGeneExpression
wild-typeMGfruit)anditsisomer,isomer#4oflycoperosideG/F inGAME1-SilencedPlants
oresculeosideAplushexose,severalhydroxytomatineisomers,
a-tomatine plus C4H6O3, and UGA 9 (Figure 6). A detailed WeintroducedtheGAME1iconstructintotheindeterminateM82
descriptionofSAsthatwerealteredintheGAME1ifruittissues cultivarbycrossing,obtainingplantsthatdisplayedaphenotype
ispresentedinTable1,Figure6,andSupplementalTable4online. resemblingthatofthecvMicroTomandcvAilsaCraigGAME1i
A proposed pathway that represents the biosynthesis and lines,includingseveregrowthretardation,deformedleaves,and
metabolism of SAs during tomato fruit development (from the abortion of flower buds (see Supplemental Figures 8A and 8B
TomatoGLYCOALKALOIDMETABOLISM1Activity 4513
Figure4. GAME1-SilencedTomatoPlantsDisplaySevereMorphologicalPhenotypes.
(A)and(B)Three-week-old(3w)wild-type(WT)andGAME1i(SPh)tomatoplants(cvMicroTom),respectively.
(C)and(D)Six-week-old(6w)GAME1iSPhtomatoplants(cvMicroTom).
(E)and(F)Four-month-old(4m)wild-typeandGAME1i(SPh)tomatoplants(cvMicroTom),respectively.
(G)to(K)MGandRRwild-typetomatofruit(cvMicroTom;[G]and[H])comparedwithfruitinthecorrespondingstagesoftheMPhGAME1iplants[(I)
and(J)]andtotheSPhGAME1iRRstagefruit(K).
(L)SuberizedstemofaGAME1iplant(cvAilsaCraig).
online).Inaddition,leavesoftheseplantsexhibiteddarknecrotic PROTEINASE INHIBITOR1 [Pautot et al., 1991]) as well as to
spots resembling symptoms typically seen after infection of otherpathogens(CARBOXYESTERASE[Koetal.,2005],ISOCI-
tomatobypathogenicbacterialikeXanthamonascampestrispv TRATEDEHYDROGENASE[ICDH][Jangetal.,2003],PR5-LIKE
vesicatoria(Xcv)orPseudomonassyringaepvtomato(seeSup- PROTEIN/NP24 [Rodrigo etal., 1991],and PEROXIDASE [PER]
plementalFigures8Cand8Donline;Jonesetal.,1991).Asno [Veraetal.,1993])(seeSupplementalTable5online).Toverifythe
othertomatolinesgrowninthesamegreenhouseatthesame microarray results, we performed quantitative real-time PCR,
whichshowedthattheexpressionofsevenoutof10examined
time showed similar disease-like symptoms, we hypothesized
geneswasindeedsignificantlyaltered(seeSupplementalFigure
thatsilencingGAME1mightmimicinductionofdiseasesymp-
8E online). These included five genes that were upregulated in
tomsintheabsenceofapathogen.Indeed,wewerenotableto
GAME1-silencedplants(PER,ENOLASE,PR5-LIKE,DICYANIN,
isolateXcvorP.syringaepvtomato(orotherputativepathogens)
andGDSLLIPASE)andtwogenes(HAT22-LIKEandCLAVAMI-
fromGAME1ileavesthatdisplayednecroticspots.
NATESYNTHASE-LIKE)thatweredownregulated.Theseresults
Microarray analysis was performed to examine if the trans-
support a significant alteration in expression of multiple biotic
gene,andpossibly,thecorrespondingchangesinthemetabolic
stress–relatedgenesintheSPhGAME1iplants.
profile,inducedtheplantresponsesystematthetranscriptional
level.Elevenand24genesshowedsignificantdown-andupregu-
ExtractsofGAME1-SilencedPlantsandTomatidineItself
latedexpression,respectively,inGAME1ileaves(seeSupplemental
InhibitArabidopsisRootGrowth
Table5online).Thesetofupregulatedtranscriptswasenriched
;
(eightintotal, 33%)ingenesthatareupregulatedinresponseto As described above, the metabolic composition of GAME1-
bacterialattack(ENOLASE[Giblyetal.,2004],twoGDSLLIPASE silencedplantshadastrongeffectontomatoplantgrowthand
homologs [Hong et al., 2008], and ETHYLENE-RESPONSIVE development.Wetestedwhetherthistoxicityaffectsgrowthand
4514 ThePlantCell
50%shorterthanthoseonmediumwithwild-typeextracts(see
SupplementalFigures9A,9B,and9Eonline).
To examine the possible cytotoxic role of a-tomatine and
tomatidine, Arabidopsisseedsweregerminatedon media sup-
plementedwithrangeofa-tomatineortomatidineconcentrations.
Nodifferenceingerminationrateorfrequencywasobserved,but
there was significant root growth retardation that was posi-
tively correlatedwithtomatidinelevelsatconcentrations above
0.003mgpermLmedium(seeSupplementalFigures9Cand9D
online).Bycontrast,therewasnocorrelationbetweenrootlength
andtheamountofsupplementeda-tomatine(seeSupplemen-
talFigure9Donline).When1-week-oldtomatoseedlingswere
sprayedwithvariousconcentrationsoftomatidinefor3weeks,
theydidnotexhibitasimilarphenotype.
AnAlteredSAProfileImpactsPathogenicityand
GrowthofTomatoDiseaseAgents
Severalfungiandbacteriapossesstomatinasesthatcleavethe
lycotetraosechainatvariouspositionsofa-tomatine(Sandrock
andVanetten,1998;MorrisseyandOsbourn,1999;Kaupetal.,
2005).GivenourfindingthatGAME1silencingaffectsexpression
ofbioticstress–relatedgenes,weexaminedwhetheralteredSA
glycosylation levels impact pathogenicity or growth of several
tomato disease agents. SPh and MPh GAME1i leaf extracts
inhibited germination and growth of the pathogenic fungi C.
coccodestoalesserextentthandidwild-typeleafextracts(see
Supplemental Figures 10A to 10C online). In addition, SPh
GAME1i extracts gave rise to decreased halo inhibition com-
paredwithMPhGAME1iextracts(seeSupplementalFigures10A
and 10C online). The effect of GAME1 silencing on plant sus-
ceptibilitytothebacterialpathogenXcvwasevaluatedininoc-
ulation experiments, revealing altered bacterial growth in
GAME1ileavescomparedwiththewild-type(seeSupplemental
Figure11online).
Figure5. AlteringGAME1ExpressionHasaMajorEffectoninPlanta
Tomatidine-GalactosyltransferaseActivityandonLevelsofa-Tomatine
DISCUSSION
andTomatidine.
Inthisstudy,weidentifiedGAME1,aGT-typeenzymethatuses
(A)Tomatidine-galactosyltransferaseactivityinfullyexpandedleaves(cv
MicroTom,4-week-oldplants).Activitywasmeasuredinextractsderived UDP-Gal to attach the first sugar moiety to the alkamine
from fresh samples and is expressed on a fresh weight (FW) basis. tomatidine. This discovery not only provides important insight
Different lettering above the bars denotes significant differences in intotheSGAbiosyntheticpathwayintomatobutalsoshedslight
tomatidine-galactosyltransferaseactivitylevelsascalculatedbyaStu- on the significance of glycosylation in the prevention of self-
dent’sttest(P<0.05;n=3).WT,wildtype. toxicityinplants.
(B)GAME1iplantsexhibitreducedlevelsofa-tomatineandaccumula-
tion of its precursor tomatidine. Their absolute concentrations were
NegativeRegulationofGAME1ExpressionandGreen
measuredinfullyexpandedleavesofthewildtype(4-week-oldplants)
Tissue–RelatedSAsbyClimactericEthyleneSignaling
andcomparedwiththoseofSPhGAME1iplantsofsimilarage(using
UPLC-qTOF-MS).SignificantdifferencesaccordingtoaStudent’sttest duringTomatoFruitDevelopment
(P<0.05;n=3)areindicatedbyanasterisk.
Metabolic profiling assays performed in recent years have
In(A)and(B),thebarsrepresentSE.
revealedaplethoraofSAsintomatoandpotato,beyondthose
reported earlier in these two species (Moco et al., 2006; Iijima
development of other plant species as well. No differences in etal.,2008;Kozukueetal.,2008;Mintz-Oronetal.,2008;Shakya
germination rate or frequency were observed between Arabi- andNavarre,2008).Theexaminationofalmost90differentSAs
dopsisseedssownonplatescontainingleafextractsfromwild- inthisstudyhighlightsthisremarkablediversity,whiledescribing
typeorGAME1iplants.However,SPhGAME1iextractsshowed severalgroupsofrelatedchemicalsthatclustertogetheracross
amajoreffectonArabidopsisrootgrowth;after3weeks,rootsof the21differenttomatotissuesandorgansstudied.Onenotable
seedlings on medium containing SPh GAME1i extracts were cluster,largelyrepresentedbya-tomatineanddehydrotomatine
TomatoGLYCOALKALOIDMETABOLISM1Activity 4515
Table1.SAsAlteredintheGAME1RNAiTomatoLines
Downregulated Upregulated
No. PutativeMetabolite m/z(+) Leaf Fruit(MG) Leaf Fruit(MG)
1 a-Tomatine 1034.5536 + +
2 a-Tomatineisomer 1034.5564 + +
3 Acetoxy-dehydrotomatine 1090.5434 +
11 Acetoxy-hydroxytomatineisomer#1 1108.5540 +
18 Dehydrotomatine(tomatidenol) 414.3500 +
19 Dehydrotomatidinetetrahexoside 1062.5485 +
20 Dehydrotomatineisomer#1 1032.5379 + +
21 Dehydrotomatineisomer#2 1032.5379 + +
22 Dehydrotomatineisomer#3 1032.5379 + +
23 Dehydrotomatineisomer#4 1032.5379 +
24 Dehydrotomatineisomer#5 1032.5379 +
26 Didehydrotomatine 1030.5223 +
33 Hydroxy-dehydrotomatidinetrihexoside+deoxyhexose 1062.5485 +
37 Hydroxy-dehydrotomatine 1048.5327 +
48 Hydroxytomatineisomer#2 1050.5485 +
49 Hydroxytomatineisomer#3 1050.5485 +
50 Hydroxytomatineisomer#4 1050.5585 +
51 Hydroxytomatineisomer#5 1050.5485 +
53 Hydroxytomatineisomer#7 1050.5585 +
61 LycoperosideG/ForesculeosideA+hexoseisomer#4 1432.6596 + +
64 Tomatidine 416.3529 + +
65 Tomatidineisomer 416.3529 +
66 Tomatidinedihexosidedipentosideisomer#1 1004.5430 +
69 Tomatidinetetrahexoside 1064.5641 +
70 a-Tomatine+C4H6O3 1136.5861 +
75 UGA4 1044.5379 +
76 UGA5 1046.5536 +
77 UGA6 1052.5620 +
80 UGA9 1122.5696 +
A“+”indicatesasignificantchangeinatleastonetransgenicline;no“+”indicatesanonsignificantchange.
derivatives, is present in green tomato tissues, including buds signaling.Itisexpectedthatgenesencodingenzymescatalyzing
andflowers(containinggreensepals).Interestingly,thiscluster acetylationorglycosylationstepsdownstreaminthea-tomatine
showedasimilarpatterninfruitpeel,flesh,andseeds,suggest- catabolismpathwaywillbepositivelyregulatedbyripeningand
ingacommonregulatorynetworkforcoordinatedSAmetabo- theethylenecascade.
lism in the different fruit tissues. Such a mechanism could be A second cluster of SAs represents metabolites of tomatine
climacteric ripening, which typically brings about a myriad of thataccumulateatthelaterOrandRRstagesoffruitdevelop-
metabolicchangesinthefruitduringthetransitionfromimmature ment and are mainly represented by lycoperoside G/F or
green to RR fruit. However, while most studies on climacteric esculeosideAderivatives.Again,thesedisplayasimilarpattern
ripeningfocusonthepericarptissue,theseresultstietogether inflesh,peel,andseedtissues.Apartfromthepresenceofgreen
ripening and secondary metabolism in the peel and seeds as SAs that are produced in the sepals, a unique set of hydrox-
well. The link between the temporal and spatial production of ytomatineanditsderivativesispresentinbudsandflowers.The
SAsintomatoandtheripeningprocessisevenmoresignificant clusterofesculeosideAanditsderivativestypicallyproducedin
inlightoftheincreasedproductionofa-tomatineinrin,nor,and fruit during the ripening process (see above) could not be
Never-ripe(Nr)mutantsandthedecreaseina-tomatinecontent detectedinthesetissues.Thislatterobservationsuggeststhat
inethylene-treatedfruitreportedbyIijimaetal.(2009).Inagree- floralorganslackenzymeactivityforconversionofhydroxylated
ment with these findings, we also found increased GAME1 SAstoesculeosideAanditsisomersandderivatives.
expression in 1-MCP–treated fruit as well as in fruit of rin and
normutants.Transcriptionalregulationrelatedtofruitmaturation
GAME1ActsPredominantlyasaUDP-Gal:
has been investigated extensively; however, most of these
TomatidineGalactosyltransferase
studies focused on genes that are upregulated during fruit
maturation andripening (Rose etal.,1997; Bartley andIshida, The promiscuity of members of the GT superfamily to which
2003;Itkinetal.,2009;Karlovaetal.,2011).Bycontrast,GAME1 GAME1 belongs and the possibility that in vitro assays with a
exhibits downregulated expression during the course of fruit recombinant enzyme do not reflect native, in planta activities
maturation and is negatively regulated by climacteric ethylene complicates the functional characterization of such enzymes.
4516 ThePlantCell
Figure6. ChangestoSALevelsinMGFruitofGAME1-SilencedPlants.