Manipulating the accumulation of phenolics in maize cultured cells using transcription factors

Biochemical Engineering Journal14(2003)207–216Manipulating the accumulation of phenolics in maizecultured cells using transcription factorsAnusha P.Dias,Erich Grotewold∗Department of Plant Biology,Plant Biotechnology Center,206Rightmire Hall,1060Carmack Road,The Ohio State University,Columbus,OH43210,USAReceived20August2002;accepted after revision15October2002AbstractTranscription factors are emerging as powerful tools to manipulating plant metabolism.R2R3Myb genes have expanded dramatically in the plants,where they are involved in the regulation of plant form and metabolic diversity.However,the function of most plant R2R3 Myb genes remains to be determined.We have used a maize cell culture system to investigate the consequences on the accumulation of metabolites of expressing the novel R2R3Myb transcription factor ZmMyb-IF35.We show here that,despite the high identity in the Myb domain with the P regulator of3-deoxyflavonoid biosynthesis,ZmMyb-IF35does not induce the accumulation offlavonoids.However, similar to P,ZmMyb-IF35induces the accumulation of ferulic and chlorogenic acids as well as several other compounds not found in the control Black Mexican Sweet maize cell lines or in P-expressing lines.Together,our studies show that ZmMyb-IF35and P activate different biosynthetic pathways,and suggest a promising role of ZmMyb-IF35for engineering the accumulation of various phenolic compounds.©2003Elsevier Science B.V.All rights reserved.Keywords:Plant cell culture;Biosynthesis;Phenolic compounds;Transcription factors;Cereals;Metabolite overproduction;Maillard reaction1.IntroductionMetabolic engineering involves the deliberate modifi-cation of biochemical networks with the objective of al-tering specific metabolites.Over the past few years,plant metabolic engineering has been gaining momentum as a way to profit from the amazing diversity of compounds present in the plant kingdom[1,2].A favorite strategy to increase or alter product formation has been to over-express enzymes assumed to correspond to rate-limiting steps of particular pathways utilizing transgenic approaches.While this strategy has been successfully employed in a number of cases(e.g.[3]),a wealth of evidence suggests that the manipulation of multiple enzymes in a pathway is more likely to yield the desired alterations influx[4].Although the“stacking”of transgenes in plants presents some serious technical limitations[5],there have been some sensational examples of success in manipulating important nutritional compounds in crops by expressing multiple transgenes[6]. An exciting alternative for engineering metabolism is to express a single or a few transcription factors that si-multaneously activate many biosynthetic genes from one ∗Corresponding author.Tel.:+1-614-292-2483;fax:+1-614-292-5379. E-mail address:grotewold.1@(E.Grotewold).pathway.This approach,pioneered by the independent ma-nipulation of two branches offlavonoid biosynthesis by the expression of the corresponding regulators in maize cultured cells[7],has become more viable as additional regulators of plant metabolic pathways are identified[4,8]. Indeed,nature has already taken advantage of regulatory proteins in speciation,as evolution and selection often tin-ker with differences in the activity or expression levels of transcription factors,providing an important mechanism to explain variations in quantitative traits,particularly related to metabolic pathways[9]or plant form[10].R2R3Myb genes encode one of the largest families of regulatory proteins[11].The Arabidopsis genome contains about125R2R3Myb genes[12],and even more are present in maize[13].R2R3Myb transcription factors have been suggested to provide plasticity for plant metabolism and development[14].Consistent with this idea,many R2R3 Myb proteins have been implicated in the control of plant metabolic pathways.For example,maize C1[15]and re-lated proteins in other plants[16,17]control anthocyanin pigmentation,the maize P gene regulates the accumulation of3-deoxyflavonoids and phlobaphene pigments[18],TT2 controls the accumulation of tannins in Arabidopsis seed coats[19],Arabidopsi s ATR1controls tryptophan biosyn-thesis[20],and the snapdragon AmMyb305/AmMyb3401369-703X/03/$–see front matter©2003Elsevier Science B.V.All rights reserved. doi:10.1016/S1369-703X(02)00225-5208 A.P.Dias,E.Grotewold/Biochemical Engineering Journal14(2003)207–216[21]and AmMyb308/AmMyb330[22,23]control distinct aspects of phenylpropanoid biosynthesis.However,not all R2R3Myb genes are transcriptional activators.For exam-ple,the AtMYB4gene has been shown to transcriptionally repress phenolic compounds in Arabidopsis[24].As part of a large-scale analysis of R2R3Myb sequences in maize[13],we identified several groups of genes that appear to have expanded recently during the evolution of the grasses[25].One of these groups corresponds to the P-to-A clade,characterized by at least10genes in maize [13].The P-to-A clade contains the P1gene(termed P in this study),a regulator of3-deoxyflavonoid biosynthesis[26]. ZmMyb-IF35is another member of this group of recently duplicated genes.ZmMyb-IF35has a very high identity with P in the Myb domain(>90%),but less than50%identity in the non-Myb C-terminal region[27].ZmMyb-IF35maps to chromosome3,bin3.04,close to a minor QTL for the ac-cumulation of chlorogenic acid(CGA)in maize silks[27]. Interestingly,P also controls the accumulation of CGA[7] and is one of two major QTLs responsible for the accumula-tion of this important insecticidal compound in maize silks [28].However,P also controls the accumulation of3-deoxy flavonoids and the phlobaphene pigments,by activating a subset of maizeflavonoid biosynthetic genes[7,26].P is primarily expressed infloral tissues,including the pericarp, cob glumes,silks and husk tissues.ZmMyb-IF35expression, however,does not appear to be restricted tofloral parts and expression was detected in all organs investigated(unpub-lished observations).Here,we investigated the potential of ZmMyb-IF35as a tool to manipulate metabolism in Black Mexican Sweet (BMS)maize cultured cells.We describe the generation of BMS cultured cells expressing ZmMyb-IF35from the con-stitutive CaMV35S promoter.We show that ZmMyb-IF35 does not induce the accumulation offlavonoid compounds, despite the close evolutionary relationship of ZmMyb-IF35 to P.Similar to P,however,ZmMyb-IF35induces the ac-cumulation of CGA and other phenolic compounds in BMS cells.In addition,ZmMyb-IF35induces the accumulation of several compounds not present in cell lines expressing P, as evidenced by HPLC and GC–MS experiments.Taken to-gether,our results indicate a potential role of ZmMyb-IF35 for the metabolic engineering of phenylpropanoids and high-light the potential of metabolic profiling of plant transgenic cell lines as a powerful tool to establish the function of tran-scription factors involved in the control of natural products.2.Materials and methods2.1.Generation of maize BMS transgenic cell linesA cDNA containing the entire coding sequence for ZmMyb-IF35[27]was cloned into a previously described vector[7]containing the cauliflower mosaic virus(CaMV) 35S promoter,the tobacco mosaic virus(TMV) leader and the maizefirst Adh1-S intron in the5 untranslated(5 UTR),and the potato proteinase inhibitor(pinII)termina-tion signal,to yield the35S::IF35construct.The35S::BAR construct was previously described[7].The BMS cells used in this study were maintained in Murashige and Skoog medium containing1mg/l of2,4-dichloro-phenoxyacetic acid(2,4-D)as a suspension culture in liquid medium shaken at150rpm in the dark at27◦C.The stable trans-formation of BMS cells was described before[7].Briefly, BMS cells were pretreated for12–18h in liquid media con-taining3%polyethylene glycol(8000molecular weight) and transformed by microprojectile co-bombardment with (3␮g)35S::IF35and35S::BAR at a1:1mass ratio.After 48h cells were re-suspended in a small volume of liquid medium,and re-plated on solid media containing3mg/l BASTA.BASTA resistant calli were identified4–6weeks after transfer to selective media,and maintained on selec-tive media in the dark.BMS cell lines transformed with the ERE::P construct are described before[29].2.2.Methanol extraction of transgenic cell linesA sample of100–500mg of tissue(fresh weight)was ho-mogenized using10–15,2.3mm metal beads in HPLC grade methanol(25␮l/100mg)in a polypropylene microfuge tube using the Beadbeater TM.The extracts were centrifuged at 13,000×g for15min to pellet insoluble debris and the supernatant evaporated using the SpeedVac with tempera-tures maintained below37◦C.Acid hydrolysis was carried out by boiling samples in2M HCl for20min.Normaliza-tion of methanol extracts was performed by weighing dry methanol extracts.Pellets were completely re-dissolved in HPLC grade methanol at10␮l/mg.Methanol extracts were centrifuged at13,000×g for5min to remove any residual debris and the supernatant used for HPLC and GC–MS anal-yses.For spectral analysis,∼500mg of fresh or frozen cal-lus cells were ground in a tube with1ml of30%HCl/70% butanol(v/v)and incubated for60min at37◦.Samples were then spun for10min at14,000×g,and the absorption spec-tra of the supernatants were determined.2.3.Microprojectile bombardment and gene expression assayFor each microprojectile preparation,the mass of DNA was adjusted to10␮g with35S::BAR to equalize the amount of35S promoter in each bombardment and1␮g of 35S::IF35and3␮g of reporter plasmid(pA1Luc or pBz1Luc [26])were used.To normalize luciferase activity to GUS activity,3␮g of UBI::GUS[30]were included in each bom-bardment.35S::P was used as a positive control for pA1Luc and a negative control for pBz1Luc.Each bombardment was done in duplicate.The assays for luciferase and GUS and normalization of the data were performed as described [26]where data are expressed as the ratio of arbitrary light units(luciferase)to arbitrary units offluorescence(GUS).A.P .Dias,E.Grotewold /Biochemical Engineering Journal 14(2003)207–2162092.4.High performance liquid chromatography (HPLC)HPLC analysis were carried out using a Waters Alliance ®2690HPLC System (Waters Corporation,Milford,MA),in conditions similar as described [31],but with some modi-fications (J.T.Blodgett and P.Bonello,unpublished).Injec-tions of 5␮l were made on a 4.6mm ×150mm RP-18ODS 5␮m C18reverse phase packing column (Waters Corpora-tion,Milford,MA).A guard column of dimension3.9mm ×20mm containing the same packing was used to protect the analytical column.Two solvents were used:Solvent A was composed of 0.1%ammonium formate in formic acid,and Solvent B was 88%(v/v)methanol and 0.1%ammonium formate in formic acid.Elution conditions were as follows:100%A isocratically for 5min and a linear gradient from 0to 100%B for 40min followed by a gradient of 0–100%A for 5min,and finally 100%A isocratically for 10min.The flow rate was 0.5ml/min and the column eluate was moni-tored at 280and 325nm using a multi-channel detector (Wa-ters Corporation,Milford,MA).Identification of phenolic compounds was confirmed by co-chromatography on HPLC with authentic standards (Sigma-Aldrich,St.Louis,MO).2.5.Gas chromatography (GC)and mass spectroscopy (MS)GC–MS was performed using a Thermo-Finnigan Trace 2000system with an AS 2000auto injector (Thermo Elec-tron Corporation,Waltham,MA).The GC useselectronicFig.1.Generation and analysis of maize BMS cultured cells expressing ZmMyb -IF35.(A)Strategy for the generation of transgenic maize BMS cell lines using particle bombardment.(B)Verification of the presence of the 35S::IF35transgene in six stable transformed BMS lines.Primers used for PCR are indicated in ne 1;35S::BAR,lane 2;IF35-1,lane 3;IF35-2,lane 4;IF35-3,lane 5;IF35-4,lane 6;IF35-6,lane 7;IF35-7.pressure and flow control with a split/splitless injector.The mass spectrometer was used in electron impact mode with a dedicated EI ion source.Data was collected and analyzed using a Finnigan Xcalibur version 1.1GC–MS data sys-tem.A 30m ×0.32mm ID ×0.25␮m film Restek XTI-5GC column was used with a 10m deactivated pre-column and an 18cm ×0.10mm ID flow-restricting post-column (Restek Corporation,Bellefonte,PA).Helium was used as the carrier gas,and 1␮l of sample was injected in splitless mode using an injector temperature of 250◦C.The splitless time was held for 0.8min after which the split flow was set at 70ml/min using a continuous septum sweep of about 5ml/min.The initial column temperature was held at 40◦C for 1min,before being ramped at 11◦C/min up to 310◦C.Throughout this analytical portion of the run the carrier gas flow was set at 1.0ml/min.At the end of the analytical ramp,the flow was increased to 2.0ml/min and the temperature rapidly increased to 360◦C,for a few minutes to condition the column for the next run.3.Results and discussion3.1.Generation of 35S::IF35maize BMS cultured cellsA cDNA containing the coding sequence and 198bp of predicted 5 UTR of ZmMyb-IF35sequence was cloned in a plasmid containing the CaMV 35S promoter and the210 A.P .Dias,E.Grotewold /Biochemical Engineering Journal 14(2003)207–216maize Adh1-S intron as an enhancer of expression (see Section 2)to give the 35S::IF35construct (Fig.1A).The 35S::IF35construct was bombarded into BMS cells to-gether with a plasmid expressing the bar gene from the CaMV 35S promoter,conferring resistant to the BASTA herbicide.Six BASTA-resistant cell lines were selected for further analysis.Visually these lines look phenotypi-cally indistinguishable from the original BMS cells.The presence of the 35S::IF35transgene was verified by PCR using two sets of oligonucleotides as primers.In the first experiment,primers corresponding to the first and second exons of ZmMyb-IF35resulted in the formation of 276and 219bp in a BMS line transformed just with 35S::BAR (Fig.1B,lane 1).The 219bp band corresponds to the genomic copy of ZmMyb-IF35,which contains an 83bp intron.The 276bp band corresponds to ZmMyb -IF25,a duplicate of ZmMyb -IF35that arose during the segmen-tal allotetraploid origin of the maize genome [27].The six BAR-resistant lines contain an additional 136bplongFig.2.ZmMyb-IF35does not activate the P -regulated pathway.(A)Absorption spectra of acidic-butanol extracts from:(a)35S::IF35-2,(b)BMS,(c)ERE::P without estradiol,and (d)ERE::P 10days after estradiol addition.(B)Results of transient expression after co-bombardment of cultured maize cells with 35S::P and 35S::IF35together with pA1Luc (black)and pBz1Luc (gray)reporter constructs.A UBI::GUS construct was included in every bombardment as a normalization control.Each treatment was done in replicate and the data were normalized for GUS activity as described [30].The fold activation is calculated as the ratio between each treatment and the treatment with pA1Luc or pBz1Luc constructs without activator.The average values are shown and the error bars indicate the standard deviation of the sample.fragment,corresponding to the 35S::IF35transgene.When PCR is performed on genomic DNA with a primer in the Adh1-S intron (unique to the 35S::IF35transgene)and an-other in the second exon of ZmMyb -IF35,a 620bp band is detected in the 35S::IF35transgenic lines,but not in the untransformed lines (Fig.1B).RT-PCR and Northern exper-iments indicate that BMS cells may accumulate low levels of ZmMyb -IF35transcript ,and that the transgene increases the level significantly (data not shown).Together,these re-sults indicate that the six BAR-resistant lines contain and express the 35S::IF35transgene.3.2.ZmMyb-IF35has a function distinct from that of P 3.2.1.ZmMyb-IF35does not induce the accumulation of 3-deoxy flavonoidsBecause of the sequence identity between P and ZmMyb-IF35,we investigated whether maize BMS cells expressing 35S::IF35accumulate the 3-deoxy flavonoidsA.P .Dias,E.Grotewold /Biochemical Engineering Journal 14(2003)207–216211normally regulated by P [7].Absorption spectra of acidic-butanol extracts of 35S::IF35and control lines were compared with spectra from lines expressing P from the estradiol-inducible promoter (ERE::P)before (−)or 10days after (+)induction with estradiol (ES).The results shown in Fig.2A correspond to those obtained with line IF35-2,but identical results were obtained with the other five 35S::IF35lines (not shown).The P -regulated flavan 4-ols luteoforol or apiferol (or their polymeric products)are converted in acidic alcohol solutions to the corresponding red flavylium ions (Fig.2A)with absorption spectra in the 530–570nm range [7,32].These red compounds were only present in the ERE::P line after induction with ES,but not in the control or 35S::IF35lines (Fig.2A).These findings suggest that ZmMyb -IF35is not capable of inducing the accumulation of 3-deoxy flavonoids in BMS cells.3.2.2.ZmMyb-IF35does not activate the A1geneP controls transcription of the A1gene encoding a fla-vanone/dihydroflavonol NADPH-dependent reductase [26].Fig.3.Reverse-phase HPLC analysis of methanol extracts.(A)Profile of extracts from untransformed BMS cultured cells.(B)Profile of extracts from 35S::IF35-2cultured cells.Inset corresponds to the profile obtained from extracts from 35S::IF35-2after 2years of cultivation in solid media.Asterisk indicates compounds found only in the 35S::IF35cell lines.(C)Profile of extracts from ERE::P cultured cells prior to the addition of estradiol.(D)Profile of extracts from ERE::P cultured cells 10days after the addition of estradiol.Major peaks absorbing at 280nm are labeled as follows:(a)coumaric acid;(b)cinnamic acid;(c)vanillic acid;(d)ferulic acid;(e)CGA;(f)ferulic acid glucoside;(g)caffeic acid;(h)isoorientin;(i)isovitexin;(j)vanillin.AU/mg,absorption units/mg methanol extracted material.To determine whether ZmMyb -IF35was capable of acti-vating this gene required for the biosynthesis of 3-deoxy flavonoids,we conducted transient expression experiments on BMS cells using a reporter luciferase construct under the control of the A1gene promoter (A1::Luc).Fig.2B shows that 35S::P efficiently activates the A1gene pro-moter,but 35S::IF35does not.Neither P nor ZmMyb-IF35activates Bz1,which corresponds to a gene specific to the anthocyanin branch of the maize flavonoid pathway [33].Because of the very similar Myb domains of P and ZmMyb-IF35,we explored the possibility that ZmMyb-IF35would inhibit the activation of A1by P.Transient expression experiments carried out with constant amounts of 35S::P and A1::Luc,and increasing amounts of 35S::IF35showed no effect of 35S::IF35on the activation of A1::Luc by 35S::P (not shown).Together,these findings strongly suggest that P and ZmMyb -IF35have distinct regulatory functions,de-spite the high identity in their R2R3Myb DNA-binding domains.212 A.P.Dias,E.Grotewold/Biochemical Engineering Journal14(2003)207–216Table1List of phenolic compounds found in the methanol extracts of transgenic and control cell linesPeak number Standards RT BMS IF35-1IF35-2IF35-3IF35-4IF35-6IF35-7ERE::P−E ERE::P+E Vanillin glucoside8.89f Ferulic acid glucoside12.59++++Catechin12.58c Vanillic acid13.21+++j Vanillin13.27+e CGA14.08++++++Coniferyl alcohol14.34g Caffeic acid14.60+++Benzoic acid16.67a Trans-4-coumaric acid18.06+++++++++d Ferulic acid18.42+++++++Salicylic acid18.99h Isoorientin19.59+Taxifolin19.80Naringin21.07+i Isovitexin21.28+b Trans-4-cinnamic acid22.61+++Naringenin26.74Phloretin27.53+Quercetin dihydrate28.08Pinosylvin28.09Chalcone28.51Trihydroxy chalcone29.90Kaempferol30.25Pinosylvin monomethyl ester31.393.3.Metabolic profiling of35S::IF35transgenicmaize BMS cells3.3.1.Induction of phenolic compounds byZmMyb-IF35Because of the common recent evolutionary origin of ZmMyb-IF35and P,and encouraged by the genetic linkage of ZmMyb-IF35and a QTL for CGA[27],we examined whether ZmMyb-IF35regulates the accumulation of several known phenolic compounds in maize BMS cells.Methanol extracts of35S::IF35transgenic BMS cells were subjected to reverse-phase HPLC,and compared to extracts obtained from BMS lines expressing only the35S::BAR selectable marker(Fig.3A and B).Major peaks were compared against 25available standards corresponding to phenolic compounds often found in plant tissues(Table1).We also included in the comparison methanol extracts obtained from BMS cells expressing the P gene from an estradiol-inducible promoter, before(Fig.3C)and10days after induction with estradiol (Fig.3D).The presence of35S::IF35significantly alters sev-eral major peaks in the HPLC profiles,as shown for line IF35-2(Fig.3B).Peak e(Fig.3B),with a retention time and spectral characteristics identical to CGA,was found infive out of six35S::IF35lines investigated(Table1).A peak with similar properties was not detected in BMS cells although low levels of ZmMyb-IF35expression is detected in untrans-formed BMS cell lines.Maybe,a low level of ZmMyb-IF35 expression is not sufficient to induce CGA accumulation. CGA is also found at low levels in induced ERE::P extracts (Fig.3D).Interestingly,cinnamate and coumarate,two pre-cursors of CGA normally present in BMS cells(Fig.3A, peaks a and b),are significantly reduced in cells expressing ZmMyb-IF35and P.Thus,it is tempting to speculate that ZmMyb-IF35and P may induce the expression of any of the enzymes necessary to convert cinnamic or coumaric acids into CGA(Fig.4),some of which have only very recently been identified[35].Ferulic acid is also increased(2–3-fold)upon ZmMyb-IF35 expression(Fig.3B,peak d).In contrast to CGA,ferulic acid is already present at significant levels in untransformed BMS cells(Table1),and its accumulation is affected quite dramatically by the expression of P(Fig.3D,peak d)[7]. However,a ferulic acid glucoside(Fig.3B,peak f)was only found in the lines containing the35S::IF35transgene (Table1).ERE::P lines accumulate significant amounts of the C-glycosylflavones isoorientin(Fig.3D,peak h)and iso-vitexin(Fig.3D,peak i),absent in any of the35S::IF35 lines(Table1),further supporting the contention that ZmMyb-IF35and P have distinct regulatory functions.How-ever,a compound with retention time and spectra similar to naringin,aflavanone diglycoside,was identified as a minor component of one of the35S::IF35lines(Table1,IF35-6). While the possible accumulation of this compound in maize cells is of interest,its possible induction by ZmMyb-IF35 remains to be further investigated.ZmMyb-IF35induces also the accumulation of several compounds not found in BMS or ERE::P cells(Fig.3B,A.P.Dias,E.Grotewold/Biochemical Engineering Journal14(2003)207–216213Fig.4.Metabolic grid for the biosynthesis of the main phenolics observed in this pounds highlighted have been identified in35S::IF35 transgenic cell lines.214 A.P .Dias,E.Grotewold /Biochemical Engineering Journal 14(2003)207–216labeled with asterisk).A compound with a retention time of 2.2–2.3was detected in all six 35S::IF35lines investigated.Because this region of the chromatogram has several peaks (Fig.3),we compared the spectra of this peak with those of similar retention times in ERE::P and BMS and no matches were found.Similarly,peaks with retention times of13.3Fig.5.GC analysis of methanol extracts.(A)Profile of extracts from untransformed BMS cultured cells.(B)Profile of extracts from 35S::IF35-2cultured cells.Arrows indicate compounds identified only in 35S::IF35lines and not in BMS control cultured cells.(C)Extracted ion chromatograms (EICs)of peaks 1–3.Top panels show the EIC corresponding to RTs 7.05,8.23,and 9.51and the bottom panel showing the top matches in the NIST mass spectrometry database library search software.and 25.1were found in six and four 35S::IF35lines,re-spectively,but not in control or ERE::P lines.The identity of the compounds with these retention times remains to be established.Although all six cell lines were transformed with the same construct,not all the lines accumulate the same compounds (Table 1).This may reflect different levelsA.P.Dias,E.Grotewold/Biochemical Engineering Journal14(2003)207–216215of expression of the transgene in each line,or might be due to other phenomena that we do not yet understand.It was interesting to notice that after propagating and maintaining the35S::IF35BMS lines for up to2years,many of the major peaks that are induced by the expression of the trans-gene disappear.Indeed,the profile of methanol extracts of 35S::IF35-2after2years of cultivation in solid media(inset in Fig.3B)is almost identical to the profile of untrans-formed BMS(Fig.3A).Similar results were observed with the previously reported35S::P BMS cells[7],which after 2–3years of propagation in solid media lost their ability to accumulate3-deoxyflavonoids and C-glycosylflavones(not shown),hence why ERE::P cells are used this study.From a metabolic engineering perspective,the loss of the com-pounds controlled by35S::IF35after extended propagation of the transgenic cell lines,presumably a consequence of de-creased expression of the transgene,suggests the necessity to use inducible promoters to drive the transgenes.While the ERE::P cells also lose activity with time(not shown), the expression of the transgene can be extended for much longer times by propagating cells in the absence of estradiol. Unfortunately,very few inducible promoter systems have been developed for monocots,compared to dicot systems.3.3.2.GC–MS identifies additional compoundsinduced by ZmMyb-IF35To further determine whether the expression of ZmMyb-IF35in the induction of additional metabolites,we ana-lyzed methanol extracts of control and35S::IF35cultured cells by gas chromatography followed by mass spectrom-etry(GC–MS).More than40peaks were identified and examined in the six available35S::IF35transgenic lines and compared with profiles obtained with methanol extracts of BMS lines.Significant variations were observed in the GC profiles when different methanol extracts obtained from the same lines were analyzed.After careful analysis,three mi-nor peaks were uniquely associated with the presence of the 35S::IF35transgene and absent in BMS control lines(inset in Fig.5B).The mass spectra of peak1predicts a molecular weight of108,peak2of122and peak3of134,consistent with2,5dimethyl pyrazine,2-ethyl-3-methyl pyrazine and 2-methyl-5-(1-propenyl)-pyrazine,respectively(Fig.5C).A number of pyrazine derivatives have been identified in the Maillard reaction in the presence of either a free radical initiators or an antioxidant[36].Maillard reaction is a type of non-enzymetic browning which involves the reaction of simple sugars and amino acids.The initial step involves the reaction between the reducing sugar and the primary amino acid resulting in the formation of an N glycoside,which under high temperatures or pressures yields aminoketones, aldehydes and CO2.Condensation of two aminoketones yields pyrazine derivatives that are also powerful aroma compounds.Although a number of amino acids have been tested in the Maillard reaction,volatile compounds formed in the presence of glucose-lysine or xylose-lysine model systems have shown to accumulate a number of pyrazine compounds similar to what we have detected in our GC–MS analyses in addition to furanones,furancarboxaldehydes, and4H-pyran-4-ones which were often identified in cul-tured maize cells(data not shown)[37,38].Role of induc-tion of pyrazine derivatives in35S::IF35transgenic cell lines remains to be investigated.In summary,by combining the expression of transcrip-tion factors in maize cultured cells with metabolic profiling we have established an efficient system to examine the metabolic pathways that R2R3Myb transcription factors have the potential to activate.Ourfindings suggest that ZmMyb-IF35may provide a powerful tool to manipulate the accumulation of phenolic compounds in maize and other important crops.For further reading see[34].AcknowledgementsWe thank Dr.Enrico Bonello for access to the HPLC equipment and for sharing unpublished information with us,and Pioneer Hi-Bred Intl.and Dr.Wes Bruce for the cDNA for ZmMyb-IF35.We acknowledge the excellent tech-nical assistance of Anton Prajitna and J.Marcela Hernan-dez.We also appreciate the comments and suggestions of the three anonymous reviewers.This work was supported in part by grants from the National Science Foundation (MCB-0130062)and the United States Department of Agri-culture(NRICGP2002-01267).References[1]R.Verpoorte,R.van der Heijden,H.J.G.ten Hoopen,J.Memelink,Metabolic engineering of plant secondary metabolite pathways for the production offine chemicals,Biotechnol.Lett.21(1999)467–479.[2]R.Verpoorte,Pharmacognosy in the new millennium:leadfindingand biotechnology,J.Pharm.Pharmacol.52(2000)253–262. [3]D.Shintani,D.DellaPenna,Elevating the Vitamin E content of plantsthrough metabolic engineering,Science282(1998)2098–2100. [4]E.L.Braun,T.J.Matulnik,A.P.Dias,E.Grotewold,Transcriptionfactors and metabolic engineering:novel applications for ancient tools,Recent.Adv.Phytochem.35(2001)79–109.[5]M.J.Leech,K.May, D.Hallard,R.Verpoorte,V.De Luca,P.Christou,Expression of two consecutive genes of a secondary metabolic pathway in transgenic tobacco:molecular diversity influences levels of expression and product accumulation,Plant Mol.Biol.38(1998)765–774.[6]X.Ye,S.Al-Babili,A.Kloti,J.Zhang,P.Lucca,P.Beyer,I.Potrykus,Engineering the provitamin A(␤-carotene)biosynthetic pathway into (carotenoid-free)rice endosperm,Science287(2000)303–305. [7]E.Grotewold,M.Chamberlain,G.St.Claire,J.Swenson, B.A.Siame,L.G.Butler,M.Snook,B.Bowen,Engineering secondary metabolism in maize cells by ectopic expression of transcription factors,Plant Cell10(1998)721–740.[8]J.Memelink,J.W.Kijne,R.van der Heijden,R.Verpoorte,Genetic modification of plant secondary metabolite pathways using transcriptional regulators,Adv.Biochem.Eng.Biotechnol.72(2001) 103–125.。