a root-expressed l-phenylalanine:4 · pdf filea root-expressed...

A Root-Expressed L-Phenylalanine:4-HydroxyphenylpyruvateAminotransferase Is Required for Tropane AlkaloidBiosynthesis in Atropa belladonnaC W

Matthew A. Bedewitz,a,1 Elsa Góngora-Castillo,b,1 Joseph B. Uebler,a Eliana Gonzales-Vigil,a,2

Krystle E. Wiegert-Rininger,a Kevin L. Childs,b John P. Hamilton,b Brieanne Vaillancourt,b Yun-Soo Yeo,c

Joseph Chappell,c Dean DellaPenna,d A. Daniel Jones,d,e C. Robin Buell,b and Cornelius S. Barrya,3

a Department of Horticulture, Michigan State University, East Lansing, Michigan 48824bDepartment of Plant Biology, Michigan State University, East Lansing, Michigan 48824c Plant Biology Program and Department of Pharmaceutical Sciences, University of Kentucky, Lexington, Kentucky 40546dDepartment of Biochemistry and Molecular Biology, Michigan State University, East Lansing, Michigan 48824eDepartment of Chemistry, Michigan State University, East Lansing, Michigan 48824

The tropane alkaloids, hyoscyamine and scopolamine, are medicinal compounds that are the active components of severaltherapeutics. Hyoscyamine and scopolamine are synthesized in the roots of specific genera of the Solanaceae in a multisteppathway that is only partially elucidated. To facilitate greater understanding of tropane alkaloid biosynthesis, a de novo transcriptomeassembly was developed for Deadly Nightshade (Atropa belladonna). Littorine is a key intermediate in hyoscyamine andscopolamine biosynthesis that is produced by the condensation of tropine and phenyllactic acid. Phenyllactic acid is derivedfrom phenylalanine via its transamination to phenylpyruvate, and mining of the transcriptome identified a phylogeneticallydistinct aromatic amino acid aminotransferase (ArAT), designated Ab-ArAT4, that is coexpressed with known tropane alkaloidbiosynthesis genes in the roots of A. belladonna. Silencing of Ab-ArAT4 disrupted synthesis of hyoscyamine and scopolaminethrough reduction of phenyllactic acid levels. Recombinant Ab-ArAT4 preferentially catalyzes the first step in phenyllactic acidsynthesis, the transamination of phenylalanine to phenylpyruvate. However, rather than utilizing the typical keto-acid cosubstrates,2-oxoglutarate, pyruvate, and oxaloacetate, Ab-ArAT4 possesses strong substrate preference and highest activity with thearomatic keto-acid, 4-hydroxyphenylpyruvate. Thus, Ab-ArAT4 operates at the interface between primary and specializedmetabolism, contributing to both tropane alkaloid biosynthesis and the direct conversion of phenylalanine to tyrosine.

INTRODUCTION

Plants synthesize a variety of chemically diverse specializedmetabolites, many of which have defined roles in defense againstabiotic or biotic stresses and herbivory or are involved in chemicalattraction to facilitate predation, pollination, or seed dispersal(Tewksbury and Nabhan, 2001; Klee, 2010; Klahre et al., 2011;Weinhold and Baldwin, 2011; De Luca et al., 2012; Mithöfer andBoland, 2012; Dudareva et al., 2013). The bioactive properties ofplant specialized metabolites have led to their exploitation byhumans as flavors, fragrances, pigments, and medicines. Onegroup of compounds, plant alkaloids, are widely used as stimulants,narcotics, and therapeutic agents and includes caffeine, nico-tine, cocaine, morphine, scopolamine, vinblastine, and vincristine

(Aniszewski, 2007; Grynkiewicz and Gadzikowska, 2008; Kingston,2009). Alkaloids are a structurally heterogeneous family of;12,000compounds that are synthesized through diverse pathways anddefined by the presence of a heterocyclic nitrogen atom (Zieglerand Facchini, 2008; O’Connor, 2010). The broad utility of alkaloidsfor humankind has led to considerable interest in understandingthe pathways and regulatory mechanisms underlying their syn-thesis. The development of functional genomics technologies,particularly deep and quantitative transcriptome profiling, hasgreatly facilitated these efforts (Murata et al., 2008; Liscombe et al.,2009; Desgagné-Penix et al., 2010; Giddings et al., 2011; Farrowet al., 2012; Góngora-Castillo et al., 2012; Winzer et al., 2012).Tropane alkaloids are a class of alkaloids defined by the pres-

ence of a bicyclic nitrogen bridge across a seven-carbon ring.Tropane alkaloids are synthesized by several plant families, in-cluding the Erythroxylaceae, Proteaceae, and the Euphorbiaceae,but are particularly prevalent in the Atropa, Hyocyamus, Datura,Duboisia, and Brugmansia genera of the Solanaceae (Griffin andLin, 2000; Jirschitzka et al., 2013). Extracts of these Solanaceousspecies have been used for centuries as medicinals, and thebioactive component, atropine, which is a racemic mixture ofD- and L-hyoscyamine, together with scopolamine, are the activeingredients in modern therapeutics used to treat motion sick-ness, postoperative nausea, arrhythmia, and tremors associatedwith Parkinson’s disease (Grynkiewicz and Gadzikowska, 2008).

1 These authors contributed equally to this work.2 Current address: Department of Wood Science, The University ofBritish Columbia, Vancouver, British Columbia V6T 1Z4, Canada.3 Address correspondence to [email protected] author responsible for distribution of materials integral to the findingspresented in this article in accordance with the policy described in theInstructions for Authors (www.plantcell.org) is: Cornelius S. Barry([email protected]).C Some figures in this article are displayed in color online but in black andwhite in the print edition.W Online version contains Web-only data.www.plantcell.org/cgi/doi/10.1105/tpc.114.130534

The Plant Cell, Vol. 26: 3745–3762, September 2014, www.plantcell.org ã 2014 American Society of Plant Biologists. All rights reserved.

mailto:[email protected]

http://www.plantcell.org

mailto:[email protected]

http://www.plantcell.org/cgi/doi/10.1105/tpc.114.130534

http://www.plantcell.org

The biosynthesis of hyoscyamine and scopolamine (Figure 1)is not completely understood but is initiated by decarboxylationof the nonproteinogenic amino acid ornithine by ornithine de-carboxylase to yield the polyamine putrescine (Michael et al.,1996). Putrescine is methylated by putrescine methyltransferase(PMT) to yield N-methylputrescine, which is subsequently con-verted into 4-methylaminobutanal by N-methylputrescine oxi-dase (Heim et al., 2007; Katoh et al., 2007; Biastoff et al., 2009).4-Methylaminobutanal is postulated to undergo spontaneouscyclization to form an N-methyl-D1-pyrrolinium cation, which isultimately converted to tropinone through an unknown mecha-nism (O’Connor, 2010; Jirschitzka et al., 2013). Tropinone repre-sents a key branch point in the pathway (Hashimoto et al., 1992;Nakajima et al., 1993). Tropinone reductase II reduces tropinoneto pseudotropine, which directly leads to the synthesis of ca-lystegines, a class of polyhydroxylated alkaloids that resemblemonosaccharides in structure and act as glycosidase inhibitors(Dräger, 2004). In contrast, tropinone reductase I reduces tropi-none to tropine, which leads to the biosynthesis of hyoscyamineand scopolamine. Tropine is then condensed with a phenyllacticacid moiety to form littorine, which undergoes P450-mediatedrearrangement to hyoscyamine aldehyde (Robins et al., 1994; Liet al., 2006; O’Connor, 2010). An uncharacterized alcohol de-hydrogenase is proposed to convert hyoscyamine aldehyde tohyoscyamine, which is subsequently converted to scopolaminein a two-step epoxidation reaction catalyzed by hyoscyamine6-hydroxylase (H6H) (Matsuda et al., 1991; Hashimoto et al., 1993).

Several of the enzymes and genes involved in tropane alkaloidbiosynthesis in the Solanaceae have been identified and char-acterized, and in all cases, both transcripts and proteins accu-mulate preferentially or exclusively in roots, the primary site oftropane alkaloid synthesis (Hashimoto et al., 1991; Kanegae et al.,1994; Nakajima and Hashimoto, 1999; Suzuki et al., 1999a). Sev-eral steps in tropane alkaloid biosynthesis remain to be identified,including the conversion of 4-methylaminobutanal to tropinone,the formation of phenyllactic acid and its condensation with tropineto form littorine, and the identification of the putative alcohol de-hydrogenase that converts hyoscyamine aldehyde to hyoscyamine.In addition, the identity of transcriptional regulators and tropanealkaloid transporters remains unknown, although several suchproteins involved in the synthesis of the structurally related com-pound nicotine are now characterized (Shoji et al., 2009, 2010;Todd et al., 2010; Hildreth et al., 2011; Zhang et al., 2012).

As part of a project to develop genomics-based resources formedicinal plants (http://medicinalplantgenomics.msu.edu/), a denovo assembly of the Deadly Nightshade (Atropa belladonna)transcriptome representing 44,198 unigenes, covering a total33 Mb of the transcriptome, was generated from diverse tissuesby RNA sequencing (RNA-seq). In addition, a rapid, targeted liquidchromatography-tandem mass spectrometry (LC-MS/MS)-basedmethod for the simultaneous detection of hyoscyamine, scopol-amine, and selected precursors in a single extract was developed,together with a tobacco rattle virus (TRV)-based virus-inducedgene silencing (VIGS) system, to facilitate functional analysis ofcandidate genes (Ratcliff et al., 2001; Liu et al., 2002; Dong et al.,2007). These resources were used to identify a phylogeneticallydistinct, root-expressed aromatic amino acid aminotransferase(ArAT), designated Ab-ArAT4, which preferentially catalyzes the

transamination of phenylalanine (L-Phe) to phenylpyruvate, theinitial step leading to formation of littorine, a key intermediate inhyoscyamine and scopolamine biosynthesis.

RESULTS

De Novo Assembly of the A. belladonnaReference Transcriptome

Compared with the model crop species of the Solanaceae family,tomato (Solanum lycopersicum), potato (Solanum tuberosum),pepper (Capsicum annuum), and petunia (Petunia hybrida), thereis a paucity of genomics-based resources for species of theHyoscyameae and Datureae tribes that synthesize the tropanealkaloids hyoscyamine and scopolamine. To facilitate gene dis-covery efforts related to tropane alkaloid biosynthesis and gener-ate nucleotide resources for comparative genomic analysis acrossthe Solanaceae family, a de novo transcriptome assembly ofA. belladonna was generated using a combination of paired-end(PE) and single-end (SE) read RNA-seq generated on the Illuminaplatform (Table 1). RNA from leaves, stems, flowers, roots, andwhole fruits of mixed developmental stages from a single refer-ence plant were pooled and used to construct a normalized cDNAlibrary. To maximize recovery of the complete A. belladonnatranscriptome, 11 tissues were selected to augment the RNA-seqread pool with 36-nucleotide SE RNA-seq reads (Table 1). RNAsamples were derived from three pooled biological replicates.Notably, both the normalized cDNA pool for PE sequencing andthe individual tissue samples for SE sequencing contained samplesextracted from roots, the primary site of tropane alkaloid bio-synthesis (Hashimoto et al., 1991; Suzuki et al., 1999a, 1999b).Using the Oases transcriptome assembler (Schulz et al., 2012),

a de novo assembly of the A. belladonna transcriptome wasgenerated. The final assembly compiled from;110 million readscontained 81,183 assembled sequences with an N50 larger than1 kb and an average length of 900 bp (Table 2). The Oases al-gorithm generates isoforms of a locus, which represent true al-ternative splice forms, alleles, close paralogs, and close homologs.An analysis of the assembled transcriptome revealed 44,198unigenes (loci), of which, 31,258 loci have a single isoform and12,940 loci have more than one isoform (Table 2). Low-complexitysequences and possible contaminants were removed, resultingin a total of 80,624 high-quality transcripts, representing 43,951unigenes (loci). Out of 80,624 high-quality transcripts, a peptidecould be predicted from 67,848 using ESTscan (Iseli et al., 1999).Contaminants were defined as sequences derived from a king-dom other than Viridiplantae, such as bacteria, fungi, virus, orarthropods; only a small number of the sequences (539) wereremoved from the data set, suggesting the high quality of ourtissue samples and resulting assembly.

Functional Annotation and Quality Assessment ofthe Assembly

Gene Ontology Slim (GOSlim) terms (Ashburner et al., 2000)were assigned to 25,614 out of 67,848 predicted peptides.Of these, ;67% (17,336) were categorized into 32 BiologicalProcess categories, ;85% (21,944) into 25 Molecular Function

3746 The Plant Cell

http://medicinalplantgenomics.msu.edu/

categories, and ;31% (7918) into 23 Cellular Components cat-egories (Supplemental Figure 1). The most represented BiologicalProcesses were cellular process (17.5%), metabolic process(11.7%), and biosynthetic process (11.1%). Nucleotide binding(15.8%), hydrolase activity (14.4%), and catalytic activity (14.1%)were the most prevalent Molecular Functions. For the CellularComponent categories, we observed that over 40% of the se-quences were localized to membranes, 15% to the nucleus, andaround 12% to the intracellular compartment (SupplementalFigure 1). Overall, the relative composition of the A. belladonnatranscriptome over the various functional categories was similar totranscriptomes generated with other closely related Solanaceaespecies (Rensink et al., 2005).Functional annotation was assigned using a combination of

sequence similarity search results, including UniRef100 (Suzeket al., 2007), the Arabidopsis thaliana proteome (Lamesch et al.,2012), and Pfam domains (Punta et al., 2012). As shown in Figure 2,a total of 66,250 (82.2%) assembled transcripts had a significantmatch to UniRef100 and 63,522 (78.8%) aligned to the Arabidopsisproteome, whereas Pfam domains were assigned to 42,225 (52%)assembled transcripts. The assembly was further characterized bysearching sequence similarity against the 236 known A. belladonnapeptide sequences available in GenBank, of which 217 (91.9%)were detected at full length or near full length (>80% coverage)within the transcriptome assembly. Together, these data infer thehigh quality nature of the A. belladonna transcriptome assembly.

Orthologous and Paralogous Gene Families

Clustering of the predicted A. belladonna proteome with thepredicted proteomes of potato, tomato, and Arabidopsis (79,565sequences total) generated a total of 18,519 orthologous andclose paralogous clusters of two or more members (SupplementalData Set 1). Of these, 15,999 clusters contained a total of 19,036predicted A. belladonna peptides, while 14,271 transcripts werenot assigned to any cluster. The majority of the A. belladonnapredicted peptides (10,851) were identified in clusters includingthe four species (Supplemental Figure 2) and the gene families inthis group include four to 87 members (Supplemental Data Set 1).A total of 2800 clusters were restricted to Solanaceae species(A. belladonna, tomato, and potato) and are thus termed Solanaceaelineage-specific (Figure 3; Supplemental Data Set 1). TheSolanaceae lineage-specific families contained a total of 9702proteins in gene families of three to 53 members. A sizeableproportion of A. belladonna transcripts (3004) was grouped withinthe Solanaceae lineage-specific families, including those encod-ing the known tropane alkaloid biosynthesis enzymes putrescinemethyltransferase, ornithine decarboxylase, and tropinone reductaseI and II. This grouping is consistent with prior findings that report

Figure 1. The Proposed Tropane Alkaloid Biosynthetic Pathway inA. belladonna and Related Solanaceous Species.

Enzyme name abbreviations are shown. ODC, ornithine decarboxylase;PMT, putrescine methyl transferase; MPO, methyl putrescine oxidase;TR I, tropinone reductase I; TR II, tropinone reductase II; CYP80F1, littorinemutase; H6H, hyoscyamine-6-hydroxylase. Unknown steps between theN-methyl-D1-pyrrolinium cation and tropinone are shown as a dashedarrow. Ab-ArAT4, the subject of this study, is shown in bold.

Ab-ArAT4 and Tropane Alkaloid Biosynthesis 3747

http://www.plantcell.org/cgi/content/full/tpc.114.130534/DC1








a high degree of sequence conservation within the Solanaceaefamily (Rensink et al., 2005). The OrthoMCL algorithm identifies closeparalogs, and 1205 clusters were A. belladonna specific, containing2681 predicted proteins in gene families of two to 10 members. It isimportant to note that some of these clusters may represent alter-native splice forms from alleles and close paralogs as the Oasesassembler may not distinguish between these transcript variants.

Representation and in Silico Expression Analysis ofA. belladonna Transcripts Involved in TropaneAlkaloid Biosynthesis

Several genes encoding enzymes involved in the synthesis ofhyoscyamine and scopolamine have been isolated from Datura,Hyoscyamus, and Atropa spp (Figure 1, Table 3). tBLASTnsearches of the A. belladonna transcriptome assembly usingeach of the known tropane alkaloid biosynthetic enzymes as thequery sequence revealed the presence of highly homologousunigenes within the transcriptome (Table 3). For several of theseenzymes, predicted full-length transcripts were identified in theA. belladonna transcriptome, but genes whose predicted full-length sequence exceed 1.5 kb (e.g., N-methylputrescine oxidaseand littorine mutase) were represented by multiple transcripts,suggesting their incomplete assembly (Table 3).

Tropane alkaloid biosynthesis occurs in the roots of Solanaceaespecies with the transcripts of known biosynthetic genes prefer-entially or specifically accumulating in root tissues (Hashimoto et al.,1991; Suzuki et al., 1999a, 1999b). The A. belladonna transcriptomecontains three cDNA libraries derived from root tissues: secondaryroot, primary tap root, and sterile seedlings (Table 2). An analysis oftranscript abundance measured by fragments per kilobase of tran-script per million fragments mapped indicates that the known genesinvolved in tropane alkaloid biosynthesis are preferentially ex-pressed in root tissues, with the majority of transcripts displayinghighest abundance in secondary roots (Table 3). These data indicatethat the de novo assembled A. belladonna transcriptome containsthe known transcripts involved in tropane alkaloid biosynthesis andprovides an accurate quantitative view of their abundance, in-dicating that this resource should have utility in identifying novelgenes involved in tropane alkaloid biosynthesis and transport.

Identification of an Aromatic Amino Acid Aminotransferaseas a Candidate Tropane Alkaloid Biosynthetic Enzyme

The condensation of phenyllactic acid with tropine forms littorine,a precursor of hyoscyamine and scopolamine (Figure 1) (Humphreyand O’Hagan, 2001). The phenyllactic acid moiety is derived fromL-Phe in a two-step conversion in which L-Phe is transaminated to

Table 1. Description of A. belladonna Tissues and RNA-seq Data

Library Description Read Length (nt) No. Reads (M) No. Reads PF (M)

SE Callus 36 38.2 30.1SE Secondary root 36 37.3 29.1SE Stem 36 39.6 30.9SE Ripe fruit 36 37.5 30.2SE Green fruit 36 35.8 29.6SE Leaf 36 36.6 29.8SE Flower buds 36 34.1 28.3SE Flower 36 28.5 24.9SE Mature seed 36 29.2 24.6SE Primary tap root 36 32.7 27.8SE Sterile seedling 36 32.3 27.7PE Normalizeda 55 30.0 27.3PE Normalizeda 120 40.0 32.0

PF, purity filtered reads; M, million reads; nt, nucleotides.aNormalized libraries were made from a pool of tissues derived from mature fruit, leaf, secondary stem, entire root, and mature flower.

Table 2. Summary of de Novo Transcriptome Assembly Using the Oases Assembler

Assembly Metric Value

Total reads used for assembly 143,300,844Total reads in assembly 110,259,258 (76.9%)N50 1,186 bpAverage transcript size 900 bpMin. transcript size 251 bpMax. transcript size 5,264 bpGC content 40.53%Total transcript number 81,183Total unigene (loci) number 44,198Isoforms-loci with 1 isoform 31,258Isoforms-loci with >1 isoforms 12,940Median isoform no. for those with >1 isoform 3

3748 The Plant Cell

phenylpyruvate, which is then reduced to phenyllactic acid (Figure 1).The A. belladonna transcriptome was mined for the presence ofArATs with homology to enzymes known or predicted to utilize eitherL-Phe or tyrosine (L-Tyr) as substrates by a combination of directscreening of the transcriptome assembly annotations and sub-sequent BLAST searches using the recovered sequences (http://medicinalplantgenomics.msu.edu/). Eight unigenes were identified,and subsequent RT-PCR analysis indicated that these correspondto six genes designated Ab-ArAT1 to 6 (Table 4). Full-length tran-scripts were obtained for Ab-ArAT1 to 5, which encode proteins ofbetween 419 and 441 amino acids (46 to 48 kD) that share between50 and 72% amino acid identity.

The majority of putative plant ArATs remain uncharacterizedbut the closely related Arabidopsis proteins encoded by At5g36160and At5g53970, Ps-TyrAT from Papaver somniferum, Ph-PPY-ATfrom petunia, and Cm-ArAT1 from melon (Cucumis melo) all utilizeL-Phe and L-Tyr as amino donors (Prabhu and Hudson, 2010; Leeand Facchini, 2011; Riewe et al., 2012; Yoo et al., 2013). An aminoacid alignment of the predicted protein sequence of the Ab-ArATproteins and selected homologs indicates that each containsa conserved catalytic lysine that corresponds to residue 270 inAb-ArAT1, together with conserved residues required for thebinding of the pyridoxal-59-phosphate cofactor (SupplementalFigure 3). Ph-PPY-AT was previously shown to be localized inthe cytosol rather than the chloroplast (Yoo et al., 2013); similarly,the Ab-ArATs identified in this study are not predicted to be lo-calized to any organelles and lack the typical transit peptides thatare characteristic of chloroplast-localized proteins. Phylogeneticanalysis of ArATs from Arabidopsis, Atropa, and other Solanaceae,

including the fully sequenced tomato and potato genomes, in-dicated an expansion of the ArAT family in Solanaceae comparedwith Arabidopsis (Figure 4; Supplemental Table 1) and suggestedpotential orthologous relationships. For example, subcladesA, B, D, and E contain putative orthologs of tomato, potato, andA. belladonna ArATs that likely perform similar roles in thesethree species (Figure 4). Ab-ArAT4 is notably more divergent(Supplemental Table 2) and is the lone ArAT in A. belladonna thatlacks clear tomato or potato orthologs, as one might expect for anenzyme involved in the synthesis of a species-specific compoundclass like tropane alkaloids.The known genes involved in tropane alkaloid biosynthesis

possess distinct expression patterns across the A. belladonnatissue panel, with the highest transcript abundances observed inroot tissues and little to no expression in non-root tissues (Table 3).Thus, we hypothesized that other pathway genes would followa similar expression pattern. Examination of ArAT family memberexpression levels across eleven diverse A. belladonna tissues(Table 4) indicates that only ArAT4 closely matches the expressionpattern of known tropane alkaloid biosynthesis genes (Table 3),with moderately high expression in secondary root, primary root,and sterile seedlings and low expression in non-root tissue sam-ples. In contrast, the five other ArAT family members are eitherexpressed constitutively in all tissues (e.g., ArAT2 and 3) or pre-dominantly in above ground tissues (e.g., ArAT1, 5, and 6), and allare expressed at lower levels than ArAT4 in roots. The distinctphylogenetic relationship of Ab-ArAT4 and the similarity of its ex-pression pattern to other tropane alkaloid biosynthetic genes makeit a likely candidate for the transaminase leading to phenyllactate,a precursor of littorine (Figure 1).

Ab-ArAT4 Is Required for Tropane Alkaloid Biosynthesis

VIGS is a powerful approach for moderately high-throughputcharacterization of gene function in plants and is particularlyeffective in members of the Solanaceae family, including silencingof root expressed genes involved in alkaloid biosynthesis (Liuet al., 2002; Li et al., 2006; Todd et al., 2010). Here, the utility of

Figure 2. Functional Annotation of A. belladonna Transcripts.

A. belladonna unigenes were aligned with annotated A. belladonna se-quences in GenBank, UniRef, the predicted Arabidopsis proteome, and thePfam domain database. For the UniRef, Arabidopsis, and Pfam results, thepercentage of total A. belladonna unigenes with a match in these threedatabases is reported. For the A. belladonna GenBank results, the per-centage of GenBank entries with a match to the A. belladonna unigene setis reported.

Figure 3. Cluster of Orthologous and Paralogous Gene Families in FourPlant Species Identified by OrthoMCL.

Number of clusters (c) and genes (g) for each orthologous group arepresented. Original data are provided in Supplemental Data Set 1.









VIGS for functional genomics in A. belladonna was investigatedusing the TRV two-component system, with constructs assembledin the TRV2 vector by ligation-independent cloning (Dong et al.,2007). Initially, the effectiveness of VIGS in A. belladonna was as-sessed using constructs targeting PHYTOENE DESATURASE (PDS)and ACTIN. The characteristic photobleached phenotype of PDS-silenced plants (Liu et al., 2002) was consistently observed 10 dafter infiltration, with phenotypes increasing in severity in plants 3weeks postinfiltration (Supplemental Figure 4). Similarly, ACTIN-silenced plants displayed a stunted phenotype compared with theTRV2 empty vector controls similar to that previously described inNicotiana benthamiana (Ryu et al., 2004) (Supplemental Figure 4).In addition, VIGS was performed to silence PMT and H6H, whichare involved in catalyzing early and late steps in the biosynthesis ofscopolamine, respectively (Hashimoto et al., 1991; Matsuda et al.,1991; Biastoff et al., 2009; Figure 1). Silencing of PMT followed bya targeted LC-MS/MS analysis revealed a 58% reduction in sco-polamine levels in roots, together with reduced levels of key inter-mediates in the pathway that occur downstream of putrescine,including N-methylputrescine, tropinone, tropine, and hyoscya-mine (Supplemental Figure 4). Additionally, silencing of H6H uti-lizing two constructs targeting distinct genic regions led to either an81 or 95% reduction in root scopolamine levels compared withTRV2 empty vector controls, together with a 3-fold increase in thelevels of the H6H substrate, hyoscyamine (Supplemental Figure 4).

Together, these data illustrate the utility of VIGS as a tool forfunctional analysis of candidate genes in A. belladonna.VIGS of ArAT4 resulted in a 77 and 66% reduction in hyo-

scyamine and scopolamine levels, respectively (Figure 5A).Congruent with the reduction in hyoscyamine and scopolamine,phenyllactate levels were reduced by ;95% in the ArAT4-silencedplants, whereas tropine, which condenses with phenyllactate toform littorine (Figure 1), was increased;4-fold (Figure 5A). However,the levels of L-Phe and phenylpyruvate were not substantially al-tered by ArAT4 silencing. In addition, tropinone was reduced in theArAT4-silenced plants by ;50%. ArAT4 transcripts weremonitored in root tissues of the silenced plants by quantitativeRT-PCR analysis, revealing a 52% reduction in TRV2:AbArAT4plants compared with the TRV2 empty vector controls (Figure 5B).To assess the possibility that off-target cosilencing might affect

tropane alkaloid biosynthesis, expression of the other ArAT familymembers were assessed in the TRV2:ArAT4-silenced plantsrelative to TRV2 empty vector controls. Cross-silencing within theAb-ArAT gene family was found to be minimal to nonexistent in theTRV2:ArAT4-silenced plants. In particular, expression of ArAT2 andArAT3, which are the two most highly expressed family membersafter ArAT4 in the roots of A. belladonna (Table 4), was not altered byArAT4 silencing (Figure 5B). Of the remaining three weakly ex-pressed family members, ArAT1 expression was unaltered, whileArAT5 and ArAT6 expression was slightly but significantly reduced

Table 3. Identification and Expression Analysis of A. belladonna Unigenes Involved in Tropane Alkaloid Biosynthesis

Tissue Type

A. belladonna UnigeneIdentifier Annotation Callus Leaf Stem Seedling

PrimaryRoot

SecondaryRoot

FlowerBuds Flowers

GreenFruit

RipeFruit Seed

aba_locus_14626_iso_1_len_1408_ver_2

Orn decarboxylase 58.6 38.7 60.4 108.4 396.9 886.4 354.2 241.5 178.1 44.2 27.7


PutrescineN-methyltransferase

12.8 0.0 0.0 146.9 74.7 1362.7 0.0 0.0 0.0 0.0 0.0


N-methylputrescine oxidase 31.0 0.0 2.5 12.1 65.9 180.4 5.5 1.5 0.0 0.0 2.2






Tropinone reductase I 0.0 1.8 34.3 5.2 10.7 45.0 7.6 16.0 16.5 59.0 37.4


Tropinone reductase I 0.0 0.0 0.0 17.3 4.6 228.6 0.0 0.0 0.0 0.0 0.0


Tropinone reductase II 9.1 76.0 88.5 147.7 125.8 396.7 38.0 84.0 73.3 2.7 0.0


Littorine mutase 1.6 0.0 0.0 33.4 2.1 158.9 0.0 0.0 0.0 0.0 0.0






Hyoscyamine 6b-hydroxylase

5.3 0.0 0.0 540.5 0.0 216.0 0.0 0.0 0.0 0.0 0.0


Hyoscyamine 6b-hydroxylase

7.0 0.0 0.0 734.0 0.0 234.0 0.0 0.0 0.0 0.0 0.0

Transcript abundance was determined by RNA-seq analysis and data are presented as fragments per kilobase of transcript per million mapped reads.

3750 The Plant Cell





(Figure 5B; Table 4). To directly test whether these other ArAT familymembers are involved in tropane alkaloid biosynthesis, the impactof silencing each gene on scopolamine levels was investigated(Supplemental Figure 5). Whereas silencing of ArAT4 reducedscopolamine levels by 66% (Figure 5A), silencing of the other fiveArATs did not affect scopolamine levels in A. belladonna roots(Supplemental Figure 5). Together, these data allow us to con-clude that ArAT4 is the predominate ArAT involved in tropanealkaloid biosynthesis in A. belladonna.

Ab-ArAT4 Possesses Phenylalanine:4-Hydroxyphenylpyruvate Aminotransferase Activity

The reported activity of several ArATs indicates they can useL-Tyr, L-Phe, and L-Trp to varying degrees as amino donors, with2-oxoglutarate generally being the preferential amino acceptor(Prabhu and Hudson, 2010; Lee and Facchini, 2011; Riewe et al.,2012). More recently, Ph-PPY-AT and Cm-ArAT1 were shown topreferentially catalyze the formation of L-Phe using L-Tyr as theamino donor and phenylpyruvate as the amino acceptor (Yooet al., 2013). The kinetic parameters of purified recombinantAb-ArAT4 were assessed using the three aromatic amino acids asamino donors together with several potential oxoacid acceptorsas cosubstrates (Table 5). The potential substrates and productsof these reactions are presented in Supplemental Figure 6.

A. belladonna ArAT4 displayed typical Michaelis-Menten kineticswith Kms for the potential amino donors L-Phe, L-Tyr, and L-Trp of2.1, 4.1, and 19.5 mM, respectively (Table 5; Supplemental Figure 7).Of the oxoacid acceptors tested 4-hydroxyphenylpyruvate (4-HPP),oxaloacetate, and phenylpyruvate were similar at ;0.5 mM, while2-oxoglutarate and pyruvate were 5.2 and 18.7 mM, respectively. Itwas surprising that 2-oxoglutarate, which is typically the preferred

cosubstrate of many characterized ArATs, was such a poor co-substrate for ArAT4. When catalytic efficiency (kcat/Km) wasmeasured, that for L-Phe and 4-HPP was at least an order ofmagnitude higher than any other oxoacid and amino donor pair,indicating these are the preferred enzyme cosubstrates. Consis-tent with the primary role of ArAT4 being to channel L-Phe towardscopolamine biosynthesis, the relative maximum activity of theforward reaction using L-Phe as the amino donor and 4-HPP as theamino acceptor to generate phenylpyruvate and L-Tyr was ;250-fold higher than the reverse reaction. When utilizing L-Phe and4-HPP as substrates, the activity of ArAT4 was largely insensitiveto changes in pH, with a broad optimum between 6.0 and 9.0 anda t1/2 (temperature at which 50% of activity is lost) of ;59°C(Supplemental Figure 8). Overall, and congruent with its role intropane alkaloid biosynthesis, these data indicate that Ab-ArAT4preferentially catalyzes the transamination of L-Phe to form phe-nylpyruvate, which is diverted toward scopolamine biosynthesis.However, unexpectedly, ArAT4 has a strong preference for 4-HPPas the preferred amino acceptor rather than 2-oxoglutarate.

ArAT4 Silencing Disrupts in Vivo ArAT Activity but Does NotAlter Aromatic Amino Acid Levels in A. belladonna Roots

In light of the atypical substrate preference of ArAT4 (Figure 6A)and to further characterize the impact of silencing ArAT4, enzymeassays were performed using total protein extracts isolated fromroots of TRV2 and TRV2:ArAT4 lines. Consistent with the reducedalkaloid content in the roots of the ArAT4-silenced plants, totalL-Phe:4-HPP aminotransferase activity was reduced by ;20%compared with TRV2 controls (Figure 6B). Furthermore, as silencingof ArAT4 affects L-Phe-derived phenyllactic acid and ArAT activity,there is also the possibility that TRV2:ArAT4 might influence amino

Table 4. ArAT Transcript Abundance (Fragments per Kilobase of Transcript per Million Mapped Reads) in Select A. belladonna Tissues as Determinedby RNA-seq Analysis

A. belladonna Unigene IdentifierGeneName

Tissue Type

Callus Leaf Stem SeedlingPrimaryRoot

SecondaryRoot

FlowerBuds Flowers

GreenFruit

RipeFruit Seed


ArAT1 2.7 15.2 1.3 7.1 0.7 0.7 3.8 70.9 2.2 1.0 0.6


ArAT2 28.3 38.4 47.1 43.1 50.8 85.2 41.1 45.9 62.7 36.6 21.9


ArAT3 10.4 9.2 15.8 18.7 11.1 33.6 16.2 16.3 21.3 21.4 41.6


ArAT4 1.2 0.0 0.0 30.6 12.2 168.1 1.3 0.0 1.9 1.7 0.0


ArAT4 0.0 0.0 1.4 30.5 8.7 130.8 3.5 2.3 1.7 0.0 0.0


ArAT4 0.0 0.0 0.0 16.9 6.5 64.4 2.0 0.0 0.0 0.0 0.0


ArAT5 0.0 0.0 0.0 0.0 0.0 4.2 29.1 1.5 0.0 0.0 0.0


ArAT6 0.0 0.0 0.0 2.0 0.0 0.0 6.4 0.0 2.0 0.0 2.5


ArAT6 2.4 0.0 0.0 1.9 0.0 1.8 8.1 0.0 2.8 0.0 3.7







acid homeostasis, particularly the levels of the aromatic aminoacids. To test this hypothesis, the abundance of the proteinogenicamino acids was determined in TRV2 and the TRV2:ArAT4 VIGSlines (Supplemental Table 3). Although statistically significant in-creases in arginine and lysine levels were observed in the ArAT4-silenced lines compared with TRV2 empty vector controls, overallthere was little change in amino acid abundance and no statisticallysignificant change in the abundance of the aromatic amino acids.Together, these data indicate that although ArAT4 channels L-Phetoward tropane alkaloid biosynthesis and simultaneously synthe-sizes L-Tyr from 4-HPP, flux through this pathway is insufficient toperturb aromatic amino acid homeostasis.

DISCUSSION

The Solanaceae is a diverse plant family of ;90 genera and3000 species that occupy a broad geographic range and in-cludes plants of economic importance as food, ornamental, andmedicinal crops (Knapp et al., 2004). In addition, members of theSolanaceae serve as model systems for investigating fundamentalprocesses involved in plant growth and development, including theripening of fleshy fruit (tomato), tuber formation (potato), pollinationand petal senescence (petunia), plant pathogen and pest inter-actions (tomato, potato, and Nicotiana spp), and the biosynthesisof specialized metabolites (tomato, pepper, Nicotiana, Hyocyamus,Datura, and Atropa spp). The development of genomics tools, in-cluding genome sequences and transcriptome data sets for sev-eral of these model species, is facilitating characterization of thesebiological processes (Xu et al., 2011; Ashrafi et al., 2012; Bombarelyet al., 2012; Tomato Genome Consortium, 2012; Kim et al., 2014;Qin et al., 2014). In this study, a deep transcriptome of A. belladonna,a member of the Solanaceae family often used for investigatingtropane alkaloid biosynthesis (Suzuki et al., 1999a, 1999b; Rotheet al., 2003; Richter et al., 2005), was generated from diverse tis-sues, facilitating in silico quantitative analyses of gene expression(Tables 1 to 3). The A. belladonna transcriptome contains sequencesof all known genes involved in tropane alkaloid biosynthesis andconfirms that they are preferentially expressed in roots, with thehighest transcript abundance observed in secondary roots (Suzukiet al., 1999a, 1999b; Li et al., 2006). The utility of the A. belladonnatranscriptome for identifying enzymes involved in tropane alkaloidbiosynthesis was demonstrated through the identification and char-acterization of Ab-ArAT4, an ArAT that possesses a novel L-Phe:4-HPP transaminase activity and channels L-Phe toward the synthesisof scopolamine (Figure 7).In general, primary metabolism is largely conserved in plants,

whereas specialized metabolism, which evolved from primarymetabolism, is highly variable and gives rise to the myriad ofchemical compounds synthesized within the plant kingdom (Wenget al., 2012). In addition to their role in protein synthesis, in plants,the aromatic amino acids L-Phe and L-Tyr are incorporated into thestructural polymers lignin, cutin, and suberin and a wide range ofspecialized metabolites, including flavonoids, volatiles, alkaloids,and tocopherols (Maeda and Dudareva, 2012). In plants, L-Phe andL-Tyr are synthesized in plastids from prephenate, which isderived from chorismate formed through the shikimate pathway.Prephenate is aminated to form arogenate, which is subsequentlyconverted to L-Tyr through the action of arogenate dehydrogenase

Figure 4. Phylogenetic Relationship of A. belladonna ArATs and RelatedProteins.

A Maximum Likelihood phylogenetic tree was constructed as de-scribed in Methods. Shading corresponds to putative orthologousproteins of the Solanaceae family defined as clades A through E. Theidentities of the proteins are provided in Supplemental Table 1, andthe multiple sequence alignment used to construct the phylogeny isprovided as Supplemental Data Set 2. Bootstrap values derived from2000 replicates are shown on the nodes of the tree. Ab-ArAT4 ishighlighted in bold.

3752 The Plant Cell




and L-Phe by arogenate dehydratase (Dal Cin et al., 2011; Maedaet al., 2011; Maeda and Dudareva, 2012). However, plant genomeshave retained multiple ArATs that likely utilize L-Phe and L-Tyr assubstrates but, like Ab-ArAT4, are not predicted to be plastid lo-calized. Recently, a petunia ArAT, Ph-PPY-AT, was shown to belocalized to the cytoplasm and able to directly catalyze the for-mation of L-Phe from phenylpyruvate using L-Tyr as the aminodonor, yielding 4-HPP as a by-product (Yoo et al., 2013). Thisreaction is analogous to the formation of L-Phe in bacteria anddemonstrates the existence of a second, direct, extraplastidicroute to L-Phe biosynthesis in plants and led to a proposed role forPh-PPY-AT in modulating aromatic amino acid homeostasis(Whitaker et al., 1981; Fischer et al., 1993; Yoo et al., 2013). Incontrast, our data support a direct role for Ab-ArAT4 only in spe-cialized metabolism, as silencing of this gene reduces tropanealkaloid accumulation but does not affect the abundance of thearomatic amino acids (Figure 5; Supplemental Table 3).

The lack of impact on primary metabolism as a result of si-lencing ArAT4, and particularly the finding that L-Phe, L-Tyr, andphenylpyruvate levels are unaltered, despite a 95% reduction inphenyllactate abundance (Figure 5; Supplemental Table 3), maybe due to several factors. For example, at least 30% of photo-synthetically fixed carbon is devoted to the production of L-Phe,which is subsequently channeled toward the synthesis ofstructural polymers and specialized metabolites (Razal et al.,1996; Maeda and Dudareva, 2012). In contrast, tropane alkaloidstypically accumulate to <1% of dry weight in Solanaceous spe-cies (Griffin and Lin, 2000). Therefore, disruption of the relativelysmall amount of L-Phe flux that is typically channeled towardtropane alkaloid biosynthesis, by silencing of Ab-ArAT4, may beinsufficient to produce a difference in the steady state levels ofL-Phe or phenylpyruvate pools. Furthermore, aromatic amino acidbiosynthesis is highly regulated and compensatory mechanismsmay be enacted in the ArAT4 VIGS lines that subsequently limit

Figure 5. Silencing of ArAT4 Affects Tropane Alkaloid Biosynthesis.

(A) Abundance of the tropane alkaloids hyoscyamine and scopolamine, together with selected intermediates and precursors in TRV2 empty vectorcontrol lines and ArAT4 VIGS lines. Data are presented as the mean n = 17 or 21 6 SE. Asterisks denote significant differences (***P < 0.001) asdetermined by Student’s t test.(B) Relative expression level of ArATs in TRV2 empty vector control lines and ArAT4 VIGS lines. Data were obtained by quantitative RT-PCR andare presented as the mean of six biological and three technical replicates with the expression level of ArAT4 in TRV2 empty vector control lines set to1 6 SE. The six plants selected for gene expression analysis represent individuals with phenyllactic acid levels closest to the median values for thegiven genotypes presented in (A). Asterisks denote significant differences (*P < 0.05; **P < 0.01) as determined by Student’s t test.[See online article for color version of this figure.]




changes in the steady state pools of L-Phe and L-Tyr. Indeed, theproposed role of Ph-PPY-AT in aromatic amino acid homeostasiswas only uncovered in a genetic background in which the typicalchloroplast localized L-Phe biosynthesis pathway was silenced(Yoo et al., 2013). Feedback inhibition of L-Phe biosynthesis and/ordiversion of L-Phe to other metabolites also represent potentialmechanisms that could suppress free L-Phe accumulation inArAT4-silenced lines. For example, L-Phe biosynthesis is subjectto feedback inhibition of chorismate mutase and arogenate dehy-dratase (Jung et al., 1986; Eberhard et al., 1996). In addition, attemptsto overproduce L-Phe in transgenic lines and the characterization ofmutations in regulatory domains of L-Phe biosynthesis genes resultin diversion of L-Phe toward additional aromatic amino acid-derivedspecialized metabolites (Tzin et al., 2009; Huang et al., 2010). Thesedata are consistent with the high degree of regulation of aromaticamino acid synthesis in primary metabolism operating to minimizechanges to the free pools of L-Phe and L-Tyr in ArAT4-silenced lines.

Although silencing of ArAT4 did not affect the abundance ofthe aromatic amino acids in A. belladonna roots, a modest butsignificant increase in both arginine and lysine was observed(Supplemental Table 3). The biological significance of thesechanges remains to be elucidated, but elevated arginine levels mayreflect altered flux through to this metabolite pool due to the de-creased utilization of ornithine, which is produced from arginine, intropane alkaloid synthesis as a result of ArAT4 silencing (Slocum,2005). Similarly, while reduced levels of phenyllactate, hyoscya-mine, and scopolamine and increased tropine content wereexpected in ArAT4-silenced lines, an unexpected reduction intropinone levels was also observed (Figure 5A). This suggests thepresence of a previously unrecognized feedback loop in tropinonebiosynthesis. The mechanism through which this may occur isunknown but may involve tropine-mediated inhibition of enzymesinvolved in tropinone biosynthesis, the identity of which are stillunknown.

The mechanisms that underlie the evolution of specializedmetabolism are diverse but frequently involve gene duplicationor gene family expansion, coupled with subfunctionalization orneofunctionalization that may influence the expression patternof a given gene or the activity of a protein or enzyme (Picherskyand Lewinsohn, 2011; Shoji and Hashimoto, 2011; Niemülleret al., 2012; Weng et al., 2012; Kaltenegger et al., 2013; Matsubaet al., 2013; Kang et al., 2014). Relative to Arabidopsis, there is an

expansion of the ArAT gene family in A. belladonna, tomato, andpotato, with two or three ArAT genes present in these species foreach Arabidopsis gene (Figure 4). Phylogenetic analysis indicatesthat Ph-PPY-AT is likely orthologous to Ab-ArAT1. These twoproteins share ;90% amino acid identity and reside within cladeB, together with proteins from other members of the Solanaceae(Figure 4). In contrast, Ab-ArAT4 is phylogenetically distinct,lacks obvious orthologs in tomato and potato, and on averageshares ;67% identity to the proteins in clade B (Figure 4;Supplemental Table 2). Furthermore, Ab-ArAT4 possesses a dis-tinct root preferential expression pattern that closely matchesother tropane alkaloid biosynthesis genes (Tables 3 and 4). Theother five A. belladonna ArATs do not display the characteristicroot-preferential expression pattern of ArAT4 (Table 4) and eitherexhibit very low expression in roots (ArAT1, ArAT5, and ArAT6) orare constitutively expressed in all the tissues examined (ArAT2and ArAT3). Whereas ArAT4 is required for tropane alkaloid bio-synthesis (Figure 5), silencing of the five additional ArATs did notaffect scopolamine levels (Supplemental Figure 5).Despite the phylogenetic and sequence divergence of Ab-ArAT4

and Ph-PPY-AT, both enzymes utilize the same compounds, butwhereas Ph-PPY-AT preferentially utilizes L-Tyr and PPY as co-substrates to catalyze the formation of L-Phe and 4-HPP (Yooet al., 2013), Ab-ArAT4 preferentially catalyzes the reverse reaction,performing transamination of L-Phe to form phenylpyruvate with4-HPP as the cosubstrate to yield L-Tyr (Table 5). Diversity in thecatalytic properties of Ph-PPY-AT and Ab-ArAT4 also support theirproposed roles in primary and specialized metabolism, respec-tively. For example, in addition to their preferred substrates,Ph-PPY-AT and Ab-ArAT4 are also active with additional aromaticamino acids as amino donors and various keto-acids and aminoacceptors (Table 5; Yoo et al., 2013). However, while Ph-PPY-ATexhibits Vmax and kcat/Km values for different substrate pairingsthat are similar to one another (Yoo et al., 2013), Ab-ArAT4 hasa clear preference for L-Phe and 4-HPP and possesses Vmax andkcat/Km values with these substrates that are one to more than twoorders of magnitude greater than those observed with othersubstrate pairings (Table 5). In addition, the Vmax of the forwardand reverse reactions catalyzed by Ph-PPY-AT exhibit only a2-fold difference, whereas Ab-ArAT4 has ;250-fold greaterpreference for the forward over the reverse reaction (Yoo et al.,2013; Table 5). This increased potential for reversibility supports

Table 5. Kinetic Properties of Ab-ArAT4

Variable Substrate (Concentrationin Assay)

Cosubstrate (Concentrationin Assay) Km (mM) Vmax (nmol s21 mg21) kcat (s

21) kcat/Km (mM21 s21)

4-HPP (0.054 to 7 mM) L-Phe (10 mM) 0.5 6 0.01 868.6 6 3 44.4 6 0.2 87.2 6 2.12-Oxoglutarate (0.781 to 25 mM) L-Phe (10 mM) 5.2 6 0.2 26 6 0.3 1.33 6 0.02 0.26 6 0.01Pyruvate (1.95 to 250 mM) L-Phe (10 mM) 18.7 6 4.1 2.9 6 0.2 0.15 6 0.01 0.008 6 0.001Oxaloacetate (0.156 to 10 mM) L-Phe (10 mM) 0.5 6 0.3 1.4 6 0.3 0.07 6 0.02 0.21 6 0.16Phenylpyruvate (0.0078 to 1 mM) L-Tyr (4 mM) 0.5 6 0.1 3.5 6 0.3 0.18 6 0.02 0.35 6 0.04L-Phe (0.078 to 10 mM) 4-HPP (7 mM) 2.1 6 0.3 604.2 6 35.6 30.9 6 1.8 15.1 6 1.4L-Phe (0.078 to 10 mM) 2-Oxoglutarate (25 mM) 1.4 6 0.02 35.1 6 0.3 1.79 6 0.02 1.25 6 0.03L-Trp (0.093 to 12 mM) 4-HPP (7 mM) 19.5 6 2.3 81.1 6 9.9 4.14 6 0.51 0.21 6 0.007L-Tyr (0.031 to 4 mM) Phenylpyruvate (1 mM) 4.1 6 1.3 2.4 6 0.4 0.12 6 0.02 0.03 6 0.004

Data are presented as the mean 6 SD of three replicates.

3754 The Plant Cell




the proposed role of Ph-PPY-AT in maintaining aromatic aminoacid homeostasis, while the more specialized, predominantly uni-directional activity of Ab-ArAT4 channels L-Phe, and subsequentlyphenylpyruvate, toward tropane alkaloid biosynthesis (Figure 5).

Overall, the kinetic properties of Ab-ArAT4, together with itsdistinct transcript accumulation, divergent phylogenetic separation,and absence of a clear ortholog in members of the Solanaceae thatdo not produce scopolamine, suggest that this enzyme likelyevolved from an ancestral enzyme involved in primary metabolismto a highly specialized enzyme with a dedicated role in tropanealkaloid biosynthesis. The activity and in vivo role of the majority ofthe ArATs phylogenetically related to Ab-ArAT4 and Ph-PPY-AT(Figure 4) remain unknown, but in light of the characterization ofthese two enzymes, it is possible that their homologs also utilizearomatic keto-acids as substrates rather than traditional keto-acids,such as 2-oxoglutarate. Therefore, it may be prudent to reexamine

Figure 7. Ab-ArAT4 Acts at the Interface of Tropane Alkaloid Bio-synthesis and Aromatic Amino Acid Metabolism.

Ab-ArAT4 participates in the catabolism of L-Phe to direct phenyl-pyruvate (PPY) and phenyllactate toward tropane alkaloid biosynthesisand simultaneously catalyzes the anabolism of L-Tyr. The characterizedenzymes of aromatic amino acid metabolism are shown in bold. Ab-ArAT4, A. belladonna Aromatic Aminotransferase 4; CM, chorismatemutase, ADH; arogenate dehydrogenase; ADT, arogenate dehydratase;PPA-AT, prephenate aminotransferase.[See online article for color version of this figure.]

Figure 6. L-Phenylalanine:4-Hydroxyphenylpyruvate AminotransferaseActivity in ArAT4-Silenced Lines.

(A) Summary of the preferred reaction of ArAT4.(B) ArAT activity was determined through monitoring the production ofL-Tyr using L-Phe and 4-HPP as substrates in reaction mixes containingcrude protein extracted from TRV2 and TRV2:ArAT4 VIGS lines. Data arepresented as the mean of three technical replicates derived froma pooled set of three biological replicates for each genotype,6 SE. Asterisksdenote significant differences (**P < 0.01) as determined by Student’s t test.


the kinetic properties of ArATs previously determined to exhibitmaximal activity with 2-oxoglutarate as the cosubstrate to assesstheir activities with aromatic keto-acids (Prabhu and Hudson, 2010;Lee and Facchini, 2011; Riewe et al., 2012).

Tropane alkaloids are synthesized by several plant familiesand the structure of the tropane ring is conserved (Griffin andLin, 2000). However, recent evidence suggests independentevolution of pathways leading to the reduction of tropinone inthe Solanaceae and Erythroxylaceae (Jirschitzka et al., 2012).Decoration of the tropine skeleton by a phenyllactate moleculeto form littorine, and subsequently hyoscyamine and scopol-amine, is a reaction that predominantly occurs in specific generaof the Solanaceae, and the identity of the enzymes involved inthese conversions are largely unresolved (Griffin and Lin, 2000;Humphrey and O’Hagan, 2001). Transcriptomics, coupled withsilencing, metabolite, and biochemical analysis, identified aspecialized role for ArAT4 in diverting L-Phe, via phenyllactate,into hyoscyamine and scopolamine in A. belladonna, estab-lishing the identity of the first enzyme in this branch of thepathway. There have been many attempts to engineer increasedproduction of hyoscyamine and scopolamine, particularlythrough overexpression of PMT and H6H (Moyano et al., 2003;Rothe et al., 2003; Zhang et al., 2004), which participate in earlyand late steps of the pathway, respectively (Figure 1). Theidentification of Ab-ArAT4 as the first step in the pathway thatprovides the phenyllactate moiety for the formation of littorine(Figure 1) provides a tool to potentially manipulate tropane al-kaloid levels in the Solanaceae. For example, combining ArAT4overexpression with previous transgenic strategies may increasethe flux of metabolites through this branch of the pathway leadingto increased availability of phenyllactate for subsequent con-densation with tropine to form littorine (Figure 1). The tremendousflux that exists through L-Phe (Razal et al., 1996) suggests that itmay be possible to divert more of this primary metabolite towardthe production of phenyllactate to increase tropane alkaloid lev-els. In addition, the transcriptome resources generated throughthis research will facilitate additional characterization of tropanealkaloid biosynthesis in the Solanaceae.

METHODS

Plant Material and Growth Treatments

Seeds of Deadly Nightshade (Atropa belladonna) were purchased fromHorizon Herbs. Seeds were surface sterilized with 50%bleach solution for5 min followed by extensive washing with sterile water and were pre-treated with gibberellic acid ($90% gibberellin A3 at 0.1% w/v) for 24 hprior to germination on moistened filter paper. Following radicle emer-gence, seedlings were transplanted into Jiffy Peat Pellets and grown at22°C under 16-h-day/8-h-night cycle at 120 mmol until the emergence ofthe sixth true leaf, after which they were transplanted into C900 pots(Nursery Supplies) and grown in Baccto High Porosity ProfessionalPlantingMix (Michigan Peat Company) supplementedwith 125 ppm 14-3-14-7Ca-1Mg fertilizer (Greencare Fertilizer) at 20°C in greenhouses atMichigan State University, East Lansing, MI. Sterile seedlings were grownfor 14 d on half-strength Murashige and Skoog modified basal mediumwith Gamborg vitamins, pH5.7, containing 1.5%w/v sucrose and 0.6%w/vagar at 23°Cunder an 18-h-day/6-h-night cycle using a 50mmol fluorescentlight source. Callus was induced from hypocotyls of sterile seedlings on

TCIM2 media (Murashige and Skoog modified basal medium with Gamborgvitamins, pH 5.7, 1.5%w/v sucrose, 1.0mg/L 6-benzylaminopurine, 1.0mg/Lnaphthaleneacetic acid, and 0.6% w/v agar). Plates were incubated as de-scribed for growth of sterile seedlings and calluswas harvested after 4weeks.

RNA Isolation, cDNA Synthesis, and Normalization

Total RNAwas isolated using the RNeasy Plant Mini Kit (Qiagen), and RNAquality and quantity was determined using the 2100 Bioanalyzer (AgilentTechnologies); 1 mg of total RNA isolated from leaves, stems, roots,flowers of mixed developmental stages, and whole fruits of mixed de-velopmental stages containing seeds was pooled and used for cDNAsynthesis using the SMART cDNA Library Construction Kit (ClontechLaboratories). cDNA was normalized to reduce high abundance tran-scripts using the duplex-specific nuclease-based normalization technology(Evrogen) (Zhulidov et al., 2004).

Library Construction, Transcriptome Sequencing, andSequence Processing

Normalized cDNA was digested with SfiI to remove the adaptor se-quences used in the normalization process, sheared, and size selected(average ;408 bp); 2.5 mg was utilized for library construction using thePaired-End DNA Sample Prep Kit (Illumina). In order to capture quanti-tative information of the A. belladonna transcriptome, RNA samples (10 mg)pooled from three biological replicates extracted from 11 diverse tissuesampleswere converted to cDNAusing themRNA-Seq kit (Illumina). Clustergeneration was performed using the Illumina Cluster Station with sizeselected (range 235 to 410 bp) cDNA fragments and sequenced using theIllumina Genome Analyzer IIx platform at the Michigan State UniversityResearch Technology Support Facility Genomics Core. Sequence readswere cleaned and filtered using the FASTX toolkit (http://hannonlab.cshl.edu/fastx_toolkit/index.html); “fastx_trimmer,” “fastx_artifacts_filter,” and“fastq_quality_trimmer” programswere used to trim and remove low-qualitysequences using a minimum quality score of 20. After cleaning and filteringpaired-end sequences, custom Perl scripts were used to filter single readsfor use as single-end sequences.

De Novo Assembly Strategy

A de novo assembly using the quality filtered reads was conducted usingthe Oases transcriptome assembler v0.1.21 (Zerbino and Birney, 2008;Schulz et al., 2012). To optimize the assembly, Oases was run using k-merlengths of 27, 29, 31, and 33. Metrics such as number of transcripts,number of transcripts longer than 250 bp, N50 contig length, maximumtranscript length, and average transcript length were used to determinethe optimal assembly.

To avoid increasing the number of redundant sequences in the finalassembly, an iterative approach to de novo assembly was used. In theinitial de novo assembly, two lanes of 55 nucleotide PE reads derived fromthe normalized PE library were assembled (k-mer length = 31, coveragecutoff = 4, and minimum contig length = 250 bp). This assembly was thenused as a reference sequence for the A. belladonna core transcriptome byconstructing an artificial genome in which contigs were concatenated intoa single pseudomolecule. Redundant reads in the single tissue librarieswere identified by mapping the 36 nucleotide SE reads to the first as-sembly using TopHat (v 1.2) (Trapnell et al., 2009) with default parameters.Using the unmapped reads from the single tissue libraries, a second denovo assembly was then performed using the PE reads together with theunmapped SE reads (k-mer length = 27, coverage cutoff = 4, and mini-mum contig length = 250 bp).

To leverage the PE read information and join fragmented contigs,a final assembly was then made using the “-long” Velvet option (long,

3756 The Plant Cell

http://hannonlab.cshl.edu/fastx_toolkit/index.html

http://hannonlab.cshl.edu/fastx_toolkit/index.html

k-mer length = 31, coverage cutoff = 4, andminimum contig length = 250 bp)with the normalized PE library reads and the assembled contigs. Contigs inthe final assembly were filtered to remove low-complexity sequences. Pu-tative contaminants in the assembly were identified using BLAST searchesagainst UniRef100 (Suzek et al., 2007) and removing contigs that were similarto non-plant sequences.A. belladonnaprotein sequences (339peptides) fromGenBank were used to quantify the quality and coverage of the assembly.Transcriptswere concatenated into a single pseudomolecule, and expressionabundances (fragments per kilobase of transcript per million fragmentsmapped) were determined in the transcriptome assembly for the 11 singletissue libraries in Table 1 as described by Góngora-Castillo et al. (2012).

Functional Annotation

Functional annotation of assembled transcripts was assigned using WU-BLAST (Chao et al., 1992) by searching the Arabidopsis thaliana proteome(Lamesch et al., 2012), UniRef100 (Suzek et al., 2007), and peptidesdeposited in GenBank for A. belladonna using an E-value cutoff #1e-10.Pfam domains (Finn et al., 2010) were identified using HMMPfam from theHMMER package (v 3) (Krogh et al., 1994; Eddy, 1998, 2009) with anE-value cutoff of 1e-10. Gene Ontology (GO) terms (Harris et al., 2004)were assigned to the 80,636 high-quality assembled transcripts using theHmmPfam tool from the HMMER package (v 2) (Krogh et al., 1994; Eddy,1998; Quevillon et al., 2005) with an E-value # 1e-10. GO terms weremapped to GO slim terms using custom Perl scripts and the Map2Slimscript from the go-perl (v 0.13) package (http://www.geneontology.org/GO.slims.shtml).

Orthologous and Paralogous Gene Family Analysis

OrthoMCL was used to identify orthologous and close paralogous genesusing the default parameters (Li et al., 2003; Chen et al., 2007) and anE-value cutoff of 1e-10. To eliminate false grouping of paralogs due toalternative isoforms, a representative transcript defined as the model thatproduces the longest peptide sequence was used in the orthologousclustering. Transposable elements for each species were filtered out toavoid clusters composed solely of transposable elements.

Multiple Sequence Alignments and Phylogenetic Analyses

Sequence analysis was performed using MEGA version 5 (Tamura et al.,2011). Amino acid alignments were performed using MUSCLE (Edgar,2004) and non-rooted phylogenetic trees constructed using theMaximumLikelihood method and the Jones-Taylor-Thornton model using defaultparameters. A bootstrap test of 2000 replicates was used to assess thereliability of the phylogeny. Estimates of evolutionary distance, p-distance,were calculated froman alignment constructed inMUSCLE. The amino acidalignment used to construct the phylogenetic tree in Figure 4 is available asSupplemental Data Set 2.

VIGS

VIGS experiments were performed using the two-component TRV systemwith target gene constructs assembled by ligation independent cloning inthe TRV2-LIC vector (Ratcliff et al., 2001; Dong et al., 2007). PCR fragmentsof A. belladonna PDS, ACTIN, PMT, H6H, and ArATs were amplified fromcDNA synthesized from RNA extracted from either the secondary rootsor sterile seedlings of A. belladonna using primer pairs described inSupplemental Table 4. Recombinant clones were confirmed using a com-bination of PCR verification and DNA sequence analysis. Constructs weretransferred into Agrobacterium tumefaciens strain GV3101 and bacterialgrowth, processing, and infiltration conditions were identical to thosepreviously described except that silencing experiments were performedby infiltrating Agrobacterium cultures into the cotyledons of 3-week-old

A. belladonna seedlings (Velasquez et al., 2009). Following infiltration, plantswere grown in a growth room at 22°C with a 16-h photoperiod at a lightintensity of 120 mmol in Sunshine Redi-earth Plug and Seedling Mix (SunGro Horticulture). Tissue samples for alkaloid extractions were harvested4 weeks after infiltration, frozen in liquid N2, and stored at 280°C.

Quantitative RT-PCR

RNA was extracted from root tissues as described above with inclusion ofan on-column DNase treatment. Synthesis of cDNA was accomplishedusing the SuperScript III first-strand synthesis kit (Invitrogen) and 1 mg ofRNA template. Gene-specific primers for Ab-ArATs, and the endogenouscontrol ELONGATION FACTOR1 (EF1; Supplemental Table 4), weredesigned using Primer Express 3.0 software (Applied Biosystems). PCRefficiency was determined for each set of primers using standard curvesderived from serial dilutions of cDNA. Quantitative PCR was performed in10-mL reactions using FAST SYBR master mix (Applied Biosystems) with300 nM of each primer, 20 ng of cDNA template for ArAT4, and 80 ng ofcDNA template for all other ArATs tested. Reactions were assembledusing a Biomek 3000 liquid handler and DNAs amplified using an AppliedBiosystems ABI Prism 7900HT real-time PCR system using the followingprogram: 2 min at 50°C and 10 min at 95°C, followed by 40 cycles of 15 sat 95°C and 1 min at 60°C. Data were analyzed using StepOne software,calculating average threshold cycle (CT) values and DCT mean for allArATs-silenced and TRV2 controls using six biological and three technicalreplicates of each line. The EF1 gene was used as an internal standard fornormalization. Values were normalized and 2(2 DDCт) calculated to de-termine relative transcript levels of the genes.

Metabolite Extractions

Metabolites were extracted from frozen powdered root tissue using 0.1 gof tissue per 1 mL of 20% methanol, containing 0.1% formic acid and10 µM propyl 4-hydroxybenzoate as an internal standard. Samples werevortexed and placed horizontally on an orbital shaker for 3 h at roomtemperature. Extracts were centrifuged for 30 min at room temperatureand the supernatant transferred to vials for analysis by LC-MS/MS. Thetissue pellets were dried at 37°C for 24 h, and the dry weight of the plantmaterial calculated through retrieving the difference between the weightsof the extraction tubes before and after the extraction.

Identification and Quantification of Tropane Alkaloids andIntermediate Compounds

The identification and quantification of tropane alkaloids and selectprecursors was accomplished by LC-MS/MS using a modified version ofa previously described protocol for amino acid analysis (Gu et al., 2007). AWaters Quattro Micro API mass spectrometer coupled to a ShimadzuLC-20ADHPLCandSIL-5000 autosampler was utilized alongwith aWatersSymmetry C18 column (2.13 100mm, 3.5-mmparticle size) at a 50°C oventemperature. HPLC was performed using a 10-mL injection volume witha flow rate of 0.4 mL/min and a gradient ranging from 99:1 to 5:95 1 mMaqueous perfluoroheptanoic acid (solvent A) and acetonitrile (solvent B;Supplemental Table 5 e). Quantitative analyses were performed usingelectrospray ionization in positive ion mode with multiple reaction moni-toring (Supplemental Table 6). Capillary voltage, extractor voltage, and radiofrequency lens settings were 3.17 kV, 2 V, and 0.1, respectively. Flow ratesof cone gas and desolvation gas were 30 and 600 L/h, respectively with thesource temperature at 100°C and desolvation temperature at 350°C. Argonwas used as the collision gas for collision-induced dissociation at amanifoldpressure of 2 3 1023 mbar, with collision energies and source cone po-tentials optimized for each metabolite using Waters QuanOptimize software.A series of calibration standards and blanks were analyzed with eachsample set.


http://www.geneontology.org/GO.slims.shtml

http://www.geneontology.org/GO.slims.shtml






Phenylpyruvic acid and DL-3-phenyllactic acid were detected in ex-tracts of A. belladonna by LC-MS/MS in negative ionization mode usingan Applied Biosystems MDS SCIEX 3200 QTRAP mass spectrometerequipped with Shimadzu LC-20AD HPLC pumps. Compounds wereseparated by HPLC using an Ascentis Express C18 column (2.1 3 100mm, 2.7-mm particles) at a 40°C oven temperature with a 10-mL injectionvolume and a flow rate of 0.4 mL/min on a gradient ranging from 98:2 to0:100 of 0.15% aqueous formic acid (solvent A) to methanol (solvent B;Supplemental Table 7). Measurements were performed using electro-spray ionization in negative ion mode and optimized MRM transitions(Supplemental Table 8). Instrument parameters were as follows: ionspray voltage, 3750 V; source temperature, 550°C; ion source gas 1,13; and ion source gas 2, 10. Data were processed and quantified usingApplied Biosystems Analyst v. 1.4.2 software. The final concentrationof metabolites was adjusted for the internal standard and calculatedusing standard curves of authentic compounds and normalized to dryweight.

Phenylpyruvate produced in enzyme assays using purified recombi-nant ArAT4 was detected by LC-MS/MS in negative ionization modeusing a Waters Quattro Premier XE mass spectrometer equipped withAquity UPLC pump and autosampler. Phenylpyruvate was separated byHPLC using an Acquity BEH C18 column (2.1 3 100 mm, 1.7-mm par-ticles) at a 50°C oven temperature with a 10-mL injection volume anda flow rate of 0.4 mL/min on a gradient ranging from 98:2 to 0:100 0.1%aqueous formic acid (solvent A) to methanol (solvent B; SupplementalTable 9). Measurements were performed using electrospray ionization innegative ion mode using an MRM transition of precursor ion > product ionm/z of 163.05 > 91. Capillary, cone, and collision voltages were at23 kV,22 V, and 10 V, respectively. Data were acquired using Waters MassLynxsoftware and analytes processed for calibration and quantification usingQuanLynx software. The final concentration of phenylpyruvate was cal-culated using a standard curve and the data presented as nmoles perminute per mg recombinant enzyme.

Amino Acid Analysis

Amino acids were detected by LC-MS/MS as previously described (Guet al., 2007). Leucine and isoleucine were combined as a single analytedue to poor resolution of these isomeric compounds in plant extracts. Thefinal concentration of metabolites was calculated using standard curvesof authentic compounds and calculated as either mmoles per minute permg recombinant enzyme, nmoles formed per mg crude A. belladonna rootprotein, or as pmoles per mg dry weight.

Heterologous Protein Expression and Purification

The predicted full-length open reading frame of ArAT4was amplified fromroot cDNA with Kod DNA polymerase using the primers AbArAT4 ENT-Fand AbArAT4 ENT STOP-R (Supplemental Table 4) and cloned into thepET100/D-TOPO vector as an N-terminal His-TAG fusion. Followingsequence verification, the construct was transformed into Escherichia colistrain NiCo21 (DE3) (New England Biolabs). A fresh transformant wasselected and inoculated into selective Terrific Broth medium and grownovernight at 37°C. The overnight culture was diluted 1:100 into 8 liters offresh selective Terrific Broth and shaken to an OD600 of 0.6, prior to in-duction by the addition of isopropyl b-D-thiogalactopyranoside to a finalconcentration of 1 mM. The expression cultures were grown at 16°C for20 h, and cells were harvested by centrifugation for 15 min at 4750g at 4°C.Cells were resuspended in a volume equal to 20% of the original culturevolume of lysis buffer (50 mM potassium maleate, 50 mM sodiumphosphate, pH 6.5, 100 mM potassium chloride, and 10% glycerol, with0.1mMpyridoxal 5ʹ-phosphate (PLP), 1 mMDTT, and 10mg/mL lysozymeadded just prior to use) containing one tablet of Complete ProteaseInhibitor Cocktail (Roche) per 50 mL of buffer. Resuspended cells were

incubated at 4°C for 4 h and lysed by sonication. After adding 2-oxoglutarateto a final concentration of 5 mM, the crude lysate was incubated in a 70°Cwater bath until the lysate temperature reached 60°C. This temperature wasmaintained for 15min, followed by cooling in an icewater bath until the lysatetemperature was below 10°C. The lysate was centrifuged for 10 min at20,000g at 4°C to removed cell debris, insoluble protein, and denaturedprotein. The supernatant was decanted and chitin bead slurry (New EnglandBiolabs) was added at a rate of 0.5mL of slurry to each 50mL of supernatantand incubated at room temperature for 30 min. The chitin beads werepelleted by centrifugation at 4750g for 10 min, and the supernatant waspassed over a Poly-Prep ChromatographyColumn to remove any remainingchitin beads. Imidazole was added to this supernatant to a concentration of10 mM, MgCl2 was added to 5 mM to chelate the EDTA contained in theprotease inhibitor cocktail, and the pH was adjusted to 8.0 with potassiumhydroxide. The supernatant was broken into four 450-mL batches. His-tagged ArAT4was then purified using a 0.1mL bed volume of Ni-NTA beads(Qiagen) according to the manufacturer’s instructions. The tagged proteinwas washed with 13 PBS, pH 7.4, containing 20 mM imidazole and elutedwith 13 PBS, pH 7.4, containing 250 mM imidazole. The eluate was ex-changed for a solution of 13 PBS, pH 7.4, 13 Complete Protease In-hibitor Cocktail, 1 mM DTT, and 0.1 mM PLP using an Amicon Ultra-30module (EMDMillipore). Protein fractions were quantified by the Bradfordassay (Bio-Rad Laboratories), separated by SDS-PAGE using 10%polyacrylamide gels, and visualized with Coomassie Brilliant Blue R 250stain. The enzyme was diluted to 0.1 mg/mL in a final buffer of 40%glycerol, 0.53 PBS, pH 7.4, 0.53 Complete Protease Inhibitor Cocktail,0.5 mM DTT, and 0.05 mM PLP. Enzyme solution was stored at 220°Cfor up to 10 d. This enzyme stock solution was diluted as needed intocold 13 PBS immediately prior to use.

Enzyme Activity Assays and Kinetics

For determination of the optimum pH of ArAT4, enzyme assays wereperformed at 30°C for 15 min in 50-mL reaction mixes of 25 mM buffer atthe reported pH, containing 0.1 mM PLP, 0.1 mM EDTA, 10 mM L-Phe,7 mM4-HPP, and 10 ng of purified enzyme. Reactions were stopped by theaddition of 10 mL of 5% formic acid. For determination of the t1/2 of ArAT4,purified recombinant enzyme was diluted 1:100 into 13 PBS and in-cubated at the temperatures indicated for 15 min and returned to ice priorto utilizing in standard assays containing 10 ng of purified enzyme, 10mML-Phe, 7 mM 4-HPP, 25 mM HEPES, pH 7.5, and the reagents notedabove at 30°C. The kinetic parameters of ArAT4 were determined usinga standard assay containing 25 mMHEPES buffer, pH 7.5, 0.1 mM EDTA,0.1 mM PLP, and specified amounts of aromatic amino acid or 2-oxoacidin a total volume of 100 mL. Reactions were terminated by addition of20mL 5% formic acid. Amounts of recombinant ArAT4 used in kinetic assaysand the length of time the assays were incubated are indicated in pa-rentheses. Kinetic assays contained 0.78 to 100 mM 2-oxoglutarate(100 ng enzyme for 1 h), 0.156 to 20mMoxaloacetate (100 ng enzyme for1 h), 1.953 to 250 mM pyruvate (500 ng enzyme for 2 h), 0.09 to 1 mMphenylpyruvate (500 ng enzyme for 1 h), 0.055 to 7 mM 4-hydroxy-phenylpyruvate (10 ng enzyme for 15 min), 0.078 to 10 mM L-Phe (10 ngenzyme for 15 min with 4-HPP or 100 ng for 1 h with 2-oxoglutarate), 0.031to 4 mM L-Tyr (500 ng for 1 h with phenylpyruvate), or 0.094 to 12 mM L-Trp(100 ng for 15 min). For all assays, reaction products were monitored andquantified by LC-MS/MS by methods previously indicated. Apparent Vmax

and Km values were determined using nonlinear regression of the Michaelis-Menten equation using the Solver add-on of Microsoft Excel 2010. A cal-culated molecular mass of 51.09 kD was used for determination of kcat.

Crude Protein Extraction and Analysis of ArAT Activity

Extraction of crude A. belladonna root proteins was performed usingpreviously described methods (Maeda et al., 2011), with the exception

3758 The Plant Cell






that filtration was performed using an Amicon Ultra-30 module (EMDMillipore). Powdered root tissue from three biological replicates waspooled for the extraction. In vivo assays of ArAT activity were performedusing 50 mg of protein at 30°C for 1 h in 100-mL reaction mixes containing25mMHEPES, pH 7.5, 0.1mMPLP, and 0.1mMEDTA together with 10mML-Phe and 3.5 mM 4-HPP. Reactions were terminated by the addition of20mL 5% formic acid and production of L-Tyrwas determined by LC-MS/MSas described above.

Accession Numbers

The Illumina sequencing reads are available in the National Center forBiotechnology Information Sequence Read Archive under accessionnumbers SRX046282 to SRX046294, SRX060267, and SRX060269.Transcriptome assemblies and quantitative analysis of transcript abun-dance for the A. belladonna transcriptome are available at http://medicinalplantgenomics.msu.edu/. Full-length sequences correspondingto Ab-ArAT1 to 5 are deposited in GenBank under accession numbersKC954703 to KC954707.

Supplemental Data

The following materials are available in the online version of this article.

Supplemental Figure 1. Assignment of Gene Ontology Terms toA. belladonna Transcripts.

Supplemental Figure 2. Number of A. belladonna Unigenes in theDifferent Clusters as Identified by OrthoMCL.

Supplemental Figure 3. Amino Acid Alignment of Ab-ArATs andSelected Homologs.

Supplemental Figure 4. The Utility of Virus-Induced Gene Silencingfor Functional Analysis of Alkaloid Biosynthesis in A. belladonna.

Supplemental Figure 5. The Impact of Silencing Additional ArATs onTropane Alkaloid Biosynthesis in A. belladonna.

Supplemental Figure 6. Potential Forward and Reverse ReactionsCatalyzed by ArATs.

Supplemental Figure 7. Kinetic Analysis of ArAT4.

Supplemental Figure 8. Physical Properties of the ArAT4 Enzyme.

Supplemental Table 1. Proteins Utilized for Phylogenetic TreeConstruction.

Supplemental Table 2. Estimates of the Average Evolutionary Di-vergence of Solanaceae ArATs.

Supplemental Table 3. Amino Acid Levels in ArAT4-Silenced Lines.

Supplemental Table 4. Oligonucleotide Primers Used in This Study.

Supplemental Table 5. HPLC Mobile Phase Gradients Utilized forLC-MS/MS Analyses of Alkaloids and Select Precursors in PositiveMode.

Supplemental Table 6. Multiple Reaction Monitoring ParametersUtilized for LC-MS/MS Analyses of Tropane Alkaloids and SelectPrecursors in Positive Mode.

Supplemental Table 7. HPLC Mobile Phase Gradients Utilized for LC/MS/MS Analyses of Phenylpyruvic Acid and DL-3-Phenyllactic AcidUsing an Applied Biosystems MDS SCIEX 3200 QTRAP MassSpectrometer.

Supplemental Table 8. Multiple Reaction Monitoring ParametersUtilized for LC/MS/MS Measurements of Phenylpyruvic Acid andDL-3-Phenyllactic Acid Using an Applied Biosystems MDS SCIEX3200 QTRAP Mass Spectrometer.

Supplemental Table 9. HPLC Mobile Phase Gradients Utilized for LC-MS/MS Analyses of Phenylpyruvic Acid Using a Waters Quattro PremierXE Mass Spectrometer.

Supplemental Data Set 1. OrthoMCL Clusters within the A. belladonnaTranscriptome.

Supplemental Data Set 2. Multiple Sequence Alignment Used toConstruct the Phylogeny Presented in Figure 4.

ACKNOWLEDGMENTS

Funding was provided by National Institutes of General Medical SciencesAward 1RC2GM092521 to J.C., D.D.-P., C.R.B., A.D.J., and C.S.B.,a Michigan State University (MSU) Foundation Strategic PartnershipGrant to C.S.B. and C.R.B., and by National Science Foundation AwardIOS-1025636 to C.S.B. and A.D.J. D.D.-P., C.R.B., A.D.J., and C.S.B. aresupported in part by Michigan AgBioResearch and through USDANational Institute of Food and Agriculture, Hatch project numbersMICL02265 and MICL02143. M.A.B. was supported by a SummerUndergraduate Research Fellowship from the American Society of PlantBiologists, the MSU College of Agriculture and Natural Resources Un-dergraduate Research Program, and a fellowship from the MSU PlantBreeding, Genetics, and Biotechnology Graduate Program. The QuattroPremier mass spectrometer was purchased with funds from NationalScience Foundation Major Research Instrumentation Grant DBI-0619489.We thank Shari Tjugum-Holland (Research Technology Support Facility,MSU) for assistance with cDNA library preparation and operation of theIllumina Genome Analyzer IIx instrument and Scott E. Kinison (University ofKentucky) for technical assistance.

AUTHOR CONTRIBUTIONS

C.S.B., C.R.B., A.D.J., D.D.-P., and J.C. conceived and designed theresearch. M.A.B., E.G.-C., K.E.W.-R., E.G.-V., J.B.U., B.V., Y.-S.Y., andC.S.B. designed and performed research. M.A.B., E.G.-C., and C.S.B.analyzed the data. E.G.-C., K.L.C., and J.P.H. developed the bioinformaticspipeline. A.D.J. established mass spectrometry methods. M.A.B., E.G.-C.,C.R.B., and C.S.B. wrote the article. All authors read and approved thearticle.

Received July 29, 2014; revised July 29, 2014; accepted August 21,2014; published September 16, 2014.

REFERENCES

Aniszewski, T. (2007). Alkaloids Secrets of Life: Alkaloid Chemistry,Biological Significance, Applications and Ecological Role. (Amsterdam:Elsevier).

Ashburner, M., et al., The Gene Ontology Consortium (2000). Geneontology: tool for the unification of biology. Nat. Genet. 25: 25–29.

Ashrafi, H., Hill, T., Stoffel, K., Kozik, A., Yao, J., Chin-Wo, S.R.,and Van Deynze, A. (2012). De novo assembly of the pepper tran-scriptome (Capsicum annuum): a benchmark for in silico discovery ofSNPs, SSRs and candidate genes. BMC Genomics 13: 571.

Biastoff, S., Brandt, W., and Dräger, B. (2009). Putrescine N-methyltransferase—the start for alkaloids. Phytochemistry 70: 1708–1718.

Bombarely, A., Rosli, H.G., Vrebalov, J., Moffett, P., Mueller, L.A.,and Martin, G.B. (2012). A draft genome sequence of Nicotiana























benthamiana to enhance molecular plant-microbe biology research.Mol. Plant Microbe Interact. 25: 1523–1530.

Chao, K.M., Pearson, W.R., and Miller, W. (1992). Aligning two sequenceswithin a specified diagonal band. Comput. Appl. Biosci. 8: 481–487.

Chen, F., Mackey, A.J., Vermunt, J.K., and Roos, D.S. (2007). As-sessing performance of orthology detection strategies applied toeukaryotic genomes. PLoS ONE 2: e383.

Dal Cin, V., et al. (2011). Identification of genes in the phenylalaninemetabolic pathway by ectopic expression of a MYB transcriptionfactor in tomato fruit. Plant Cell 23: 2738–2753.

De Luca, V., Salim, V., Atsumi, S.M., and Yu, F. (2012). Mining thebiodiversity of plants: a revolution in the making. Science 336:1658–1661.

Desgagné-Penix, I., Khan, M.F., Schriemer, D.C., Cram, D., Nowak, J.,and Facchini, P.J. (2010). Integration of deep transcriptome andproteome analyses reveals the components of alkaloid metabolism inopium poppy cell cultures. BMC Plant Biol. 10: 252.

Dong, Y., Burch-Smith, T.M., Liu, Y., Mamillapalli, P., and Dinesh-Kumar,S.P. (2007). A ligation-independent cloning tobacco rattle virus vector forhigh-throughput virus-induced gene silencing identifies roles forNbMADS4-1 and -2 in floral development. Plant Physiol. 145: 1161–1170.

Dräger, B. (2004). Chemistry and biology of calystegines. Nat. Prod.Rep. 21: 211–223.

Dudareva, N., Klempien, A., Muhlemann, J.K., and Kaplan, I.(2013). Biosynthesis, function and metabolic engineering of plantvolatile organic compounds. New Phytol. 198: 16–32.

Eberhard, J., Ehrler, T.T., Epple, P., Felix, G., Raesecke, H.R.,Amrhein, N., and Schmid, J. (1996). Cytosolic and plastidic chorismatemutase isozymes from Arabidopsis thaliana: molecular characterizationand enzymatic properties. Plant J. 10: 815–821.

Eddy, S.R. (1998). Profile hidden Markov models. Bioinformatics 14:755–763.

Eddy, S.R. (2009). A new generation of homology search tools basedon probabilistic inference. Genome Inform. 23: 205–211.

Edgar, R.C. (2004). MUSCLE: multiple sequence alignment with highaccuracy and high throughput. Nucleic Acids Res. 32: 1792–1797.

Farrow, S.C., Hagel, J.M., and Facchini, P.J. (2012). Transcript andmetabolite profiling in cell cultures of 18 plant species that producebenzylisoquinoline alkaloids. Phytochemistry 77: 79–88.

Finn, R.D., et al. (2010). The Pfam protein families database. NucleicAcids Res. 38: D211–D222.

Fischer, R.S., Bonner, C.A., Boone, D.R., and Jensen, R.A. (1993).Clues from a halophilic methanogen about aromatic amino acidbiosynthesis in archaebacteria. Arch. Microbiol. 160: 440–446.

Giddings, L.-A., Liscombe, D.K., Hamilton, J.P., Childs, K.L.,DellaPenna, D., Buell, C.R., and O’Connor, S.E. (2011). A stereo-selective hydroxylation step of alkaloid biosynthesis by a uniquecytochrome P450 in Catharanthus roseus. J. Biol. Chem. 286: 16751–16757.

Góngora-Castillo, E., et al. (2012). Development of transcriptomicresources for interrogating the biosynthesis of monoterpene indolealkaloids in medicinal plant species. PLoS ONE 7: e52506.

Griffin, W.J., and Lin, G.D. (2000). Chemotaxonomy and geographicaldistribution of tropane alkaloids. Phytochemistry 53: 623–637.

Grynkiewicz, G., and Gadzikowska, M. (2008). Tropane alkaloids asmedicinally useful natural products and their synthetic derivativesas new drugs. Pharmacol. Rep. 60: 439–463.

Gu, L., Jones, A.D., and Last, R.L. (2007). LC-MS/MS assay forprotein amino acids and metabolically related compounds for large-scale screening of metabolic phenotypes. Anal. Chem. 79: 8067–8075.

Harris, M.A., et al., Gene Ontology Consortium (2004). The GeneOntology (GO) database and informatics resource. Nucleic AcidsRes. 32: D258–D261.

Hashimoto, T., Hayashi, A., Amano, Y., Kohno, J., Iwanari, H.,Usuda, S., and Yamada, Y. (1991). Hyoscyamine 6 beta-hydroxylase,an enzyme involved in tropane alkaloid biosynthesis, is localized at thepericycle of the root. J. Biol. Chem. 266: 4648–4653.

Hashimoto, T., Matsuda, J., and Yamada, Y. (1993). Two-step epoxi-dation of hyoscyamine to scopolamine is catalyzed by bifunctional hy-oscyamine 6 beta-hydroxylase. FEBS Lett. 329: 35–39.

Hashimoto, T., Nakajima, K., Ongena, G., and Yamada, Y. (1992).Two tropinone reductases with distinct stereospecificities from culturedroots of Hyocyamus niger. Plant Physiol. 100: 836–845.

Heim, W.G., Sykes, K.A., Hildreth, S.B., Sun, J., Lu, R.H., andJelesko, J.G. (2007). Cloning and characterization of a Nicotianatabacum methylputrescine oxidase transcript. Phytochemistry 68:454–463.

Hildreth, S.B., Gehman, E.A., Yang, H., Lu, R.H., Ritesh, K.C.,Harich, K.C., Yu, S., Lin, J., Sandoe, J.L., Okumoto, S., Murphy,A.S., and Jelesko, J.G. (2011). Tobacco nicotine uptake permease(NUP1) affects alkaloid metabolism. Proc. Natl. Acad. Sci. USA 108:18179–18184.

Huang, T., Tohge, T., Lytovchenko, A., Fernie, A.R., and Jander, G.(2010). Pleiotropic physiological consequences of feedback-insensitivephenylalanine biosynthesis in Arabidopsis thaliana. Plant J. 63: 823–835.

Humphrey, A.J., and O’Hagan, D. (2001). Tropane alkaloid biosynthesis.A century old problem unresolved. Nat. Prod. Rep. 18: 494–502.

Iseli, C., Jongeneel, C.V., and Bucher, P. (1999). ESTScan: a pro-gram for detecting, evaluating, and reconstructing potential codingregions in EST sequences. Proc. Int. Conf. Intell. Syst. Mol. Biol.1999: 138–148.

Jirschitzka, J., Dolke, F., and D’Auria, J.C. (2013). Increasing thepace of new discoveries in tropane alkaloid biosynthesis. Adv. Bot.Res. 68: 39–72.

Jirschitzka, J., Schmidt, G.W., Reichelt, M., Schneider, B.,Gershenzon, J., and D’Auria, J.C. (2012). Plant tropane alkaloidbiosynthesis evolved independently in the Solanaceae and Eryth-roxylaceae. Proc. Natl. Acad. Sci. USA 109: 10304–10309.

Jung, E., Zamir, L.O., and Jensen, R.A. (1986). Chloroplasts ofhigher plants synthesize L-phenylalanine via L-arogenate. Proc.Natl. Acad. Sci. USA 83: 7231–7235.

Kaltenegger, E., Eich, E., and Ober, D. (2013). Evolution of homo-spermidine synthase in the convolvulaceae: a story of gene dupli-cation, gene loss, and periods of various selection pressures. PlantCell 25: 1213–1227.

Kanegae, T., Kajiya, H., Amano, Y., Hashimoto, T., and Yamada, Y.(1994). Species-dependent expression of the hyoscyamine 6 beta-hydroxylase gene in the pericycle. Plant Physiol. 105: 483–490.

Kang, J.H., Gonzales-Vigil, E., Matsuba, Y., Pichersky, E., andBarry, C.S. (2014). Determination of residues responsible for sub-strate and product specificity of Solanum habrochaites short-chaincis-prenyltransferases. Plant Physiol. 164: 80–91.

Katoh, A., Shoji, T., and Hashimoto, T. (2007). Molecular cloning ofN-methylputrescine oxidase from tobacco. Plant Cell Physiol. 48:550–554.

Kim, S., et al. (2014). Genome sequence of the hot pepper providesinsights into the evolution of pungency in Capsicum species. Nat.Genet. 46: 270–278.

Kingston, D.G.I. (2009). Tubulin-interactive natural products as anti-cancer agents. J. Nat. Prod. 72: 507–515.

Klahre, U., Gurba, A., Hermann, K., Saxenhofer, M., Bossolini, E.,Guerin, P.M., and Kuhlemeier, C. (2011). Pollinator choice in Petuniadepends on two major genetic Loci for floral scent production. Curr.Biol. 21: 730–739.

3760 The Plant Cell

Klee, H.J. (2010). Improving the flavor of fresh fruits: genomics, bio-chemistry, and biotechnology. New Phytol. 187: 44–56.

Knapp, S., Bohs, L., Nee, M., and Spooner, D.M. (2004). Solanaceae—a model for linking genomics with biodiversity. Comp. Funct.Genomics 5: 285–291.

Krogh, A., Brown, M., Mian, I.S., Sjölander, K., and Haussler, D.(1994). Hidden Markov models in computational biology. Applica-tions to protein modeling. J. Mol. Biol. 235: 1501–1531.

Lamesch, P., et al. (2012). The Arabidopsis Information Resource(TAIR): improved gene annotation and new tools. Nucleic Acids Res.40: D1202–D1210.

Lee, E.J., and Facchini, P.J. (2011). Tyrosine aminotransferasecontributes to benzylisoquinoline alkaloid biosynthesis in opiumpoppy. Plant Physiol. 157: 1067–1078.

Li, L., Stoeckert, C.J., Jr., and Roos, D.S. (2003). OrthoMCL: iden-tification of ortholog groups for eukaryotic genomes. Genome Res.13: 2178–2189.

Li, R., Reed, D.W., Liu, E., Nowak, J., Pelcher, L.E., Page, J.E., andCovello, P.S. (2006). Functional genomic analysis of alkaloid bio-synthesis in Hyoscyamus niger reveals a cytochrome P450 involvedin littorine rearrangement. Chem. Biol. 13: 513–520.

Liscombe, D.K., Ziegler, J., Schmidt, J., Ammer, C., and Facchini, P.J.(2009). Targeted metabolite and transcript profiling for elucidating en-zyme function: isolation of novel N-methyltransferases from threebenzylisoquinoline alkaloid-producing species. Plant J. 60: 729–743.

Liu, Y., Schiff, M., and Dinesh-Kumar, S.P. (2002). Virus-inducedgene silencing in tomato. Plant J. 31: 777–786.

Maeda, H., and Dudareva, N. (2012). The shikimate pathway and aromaticamino acid biosynthesis in plants. Annu. Rev. Plant Biol. 63: 73–105.

Maeda, H., Yoo, H., and Dudareva, N. (2011). Prephenate amino-transferase directs plant phenylalanine biosynthesis via arogenate.Nat. Chem. Biol. 7: 19–21.

Matsuba, Y., et al. (2013). Evolution of a complex locus for terpenebiosynthesis in Solanum. Plant Cell 25: 2022–2036.

Matsuda, J., Okabe, S., Hashimoto, T., and Yamada, Y. (1991).Molecular cloning of hyoscyamine 6 beta-hydroxylase, a 2-oxoglutarate-dependent dioxygenase, from cultured roots of Hyoscyamus niger.J. Biol. Chem. 266: 9460–9464.

Michael, A.J., Furze, J.M., Rhodes, M.J.C., and Burtin, D. (1996).Molecular cloning and functional identification of a plant ornithinedecarboxylase cDNA. Biochem. J. 314: 241–248.

Mithöfer, A., and Boland, W. (2012). Plant defense against herbi-vores: chemical aspects. Annu. Rev. Plant Biol. 63: 431–450.

Moyano, E., Jouhikainen, K., Tammela, P., Palazón, J., Cusidó,R.M., Piñol, M.T., Teeri, T.H., and Oksman-Caldentey, K.M. (2003).Effect of pmt gene overexpression on tropane alkaloid production intransformed root cultures of Datura metel and Hyoscyamus muticus.J. Exp. Bot. 54: 203–211.

Murata, J., Roepke, J., Gordon, H., and De Luca, V. (2008). The leafepidermome of Catharanthus roseus reveals its biochemical spe-cialization. Plant Cell 20: 524–542.

Nakajima, K., and Hashimoto, T. (1999). Two tropinone reductases,that catalyze opposite stereospecific reductions in tropane alkaloidbiosynthesis, are localized in plant root with different cell-specificpatterns. Plant Cell Physiol. 40: 1099–1107.

Nakajima, K., Hashimoto, T., and Yamada, Y. (1993). Two tropinonereductases with different stereospecificities are short-chain de-hydrogenases evolved from a common ancestor. Proc. Natl. Acad.Sci. USA 90: 9591–9595.

Niemüller, D., Reimann, A., and Ober, D. (2012). Distinct cell-specificexpression of homospermidine synthase involved in pyrrolizidine al-kaloid biosynthesis in three species of the boraginales. Plant Physiol.159: 920–929.

O’Connor, S.E. (2010). Alkaloids. In Comprehensive Natural ProductsII: Chemistry and Biology, L. Mander and H.-W. Liu, eds (Oxford,UK: Elsevier), pp. 977–1007.

Pichersky, E., and Lewinsohn, E. (2011). Convergent evolution inplant specialized metabolism. Annu. Rev. Plant Biol. 62: 549–566.

Prabhu, P.R., and Hudson, A.O. (2010). Identification and partialcharacterization of an L-tyrosine aminotransferase (TAT) from Arab-idopsis thaliana. Biochem. Res. Int. 2010: 549572.

Punta, M., et al. (2012). The Pfam protein families database. NucleicAcids Res. 40: D290–D301.

Qin, C., et al. (2014). Whole-genome sequencing of cultivated andwild peppers provides insights into Capsicum domestication andspecialization. Proc. Natl. Acad. Sci. USA 111: 5135–5140.

Quevillon, E., Silventoinen, V., Pillai, S., Harte, N., Mulder, N.,Apweiler, R., and Lopez, R. (2005). InterProScan: protein domainsidentifier. Nucleic Acids Res. 33: W116–W120.

Ratcliff, F., Martin-Hernandez, A.M., and Baulcombe, D.C. (2001).Technical Advance. Tobacco rattle virus as a vector for analysis ofgene function by silencing. Plant J. 25: 237–245.

Razal, R.A., Ellis, S., Singh, S., Lewis, N.G., and Towers, G.H.N.(1996). Nitrogen recycling in phenylpropanoid metabolism. Phyto-chemistry 41: 31–35.

Rensink, W.A., Lee, Y., Liu, J., Iobst, S., Ouyang, S., and Buell, C.R.(2005). Comparative analyses of six solanaceous transcriptomesreveal a high degree of sequence conservation and species-specifictranscripts. BMC Genomics 6: 124.

Richter, U., Rothe, G., Fabian, A.K., Rahfeld, B., and Dräger, B. (2005).Overexpression of tropinone reductases alters alkaloid composition inAtropa belladonna root cultures. J. Exp. Bot. 56: 645–652.

Riewe, D., Koohi, M., Lisec, J., Pfeiffer, M., Lippmann, R., Schmeichel,J., Willmitzer, L., and Altmann, T. (2012). A tyrosine aminotransferaseinvolved in tocopherol synthesis in Arabidopsis. Plant J. 71: 850–859.

Robins, R.J., Woolley, J.G., Ansarin, M., Eagles, J., and Goodfellow, B.J.(1994). Phenyllactic acid but not tropic acid is an intermediate in thebiosynthesis of tropane alkaloids in Datura and Brugansia transformedroot cultures. Planta 194: 86–94.

Rothe, G., Hachiya, A., Yamada, Y., Hashimoto, T., and Dräger, B.(2003). Alkaloids in plants and root cultures of Atropa belladonnaoverexpressing putrescine N-methyltransferase. J. Exp. Bot. 54: 2065–2070.

Ryu, C.M., Anand, A., Kang, L., and Mysore, K.S. (2004). Agrodrench:a novel and effective agroinoculation method for virus-induced genesilencing in roots and diverse Solanaceous species. Plant J. 40: 322–331.

Schulz, M.H., Zerbino, D.R., Vingron, M., and Birney, E. (2012).Oases: robust de novo RNA-seq assembly across the dynamicrange of expression levels. Bioinformatics 28: 1086–1092.

Shoji, T., and Hashimoto, T. (2011). Recruitment of a duplicatedprimary metabolism gene into the nicotine biosynthesis regulon intobacco. Plant J. 67: 949–959.

Shoji, T., Kajikawa, M., and Hashimoto, T. (2010). Clustered tran-scription factor genes regulate nicotine biosynthesis in tobacco.Plant Cell 22: 3390–3409.

Shoji, T., Inai, K., Yazaki, Y., Sato, Y., Takase, H., Shitan, N.,Yazaki, K., Goto, Y., Toyooka, K., Matsuoka, K., and Hashimoto,T. (2009). Multidrug and toxic compound extrusion-type trans-porters implicated in vacuolar sequestration of nicotine in tobaccoroots. Plant Physiol. 149: 708–718.

Slocum, R.D. (2005). Genes, enzymes and regulation of argininebiosynthesis in plants. Plant Physiol. Biochem. 43: 729–745.

Suzek, B.E., Huang, H., McGarvey, P., Mazumder, R., and Wu, C.H.(2007). UniRef: comprehensive and non-redundant UniProt refer-ence clusters. Bioinformatics 23: 1282–1288.


Suzuki, K., Yamada, Y., and Hashimoto, T. (1999a). Expression ofAtropa belladonna putrescine N-methyltransferase gene in rootpericycle. Plant Cell Physiol. 40: 289–297.

Suzuki, K., Yun, D.J., Chen, X.Y., Yamada, Y., and Hashimoto, T.(1999b). An Atropa belladonna hyoscyamine 6beta-hydroxylasegene is differentially expressed in the root pericycle and anthers.Plant Mol. Biol. 40: 141–152.

Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., andKumar, S. (2011). MEGA5: molecular evolutionary genetics analysisusing maximum likelihood, evolutionary distance, and maximumparsimony methods. Mol. Biol. Evol. 28: 2731–2739.

Tewksbury, J.J., and Nabhan, G.P. (2001). Seed dispersal. Directeddeterrence by capsaicin in chilies. Nature 412: 403–404.

Todd, A.T., Liu, E., Polvi, S.L., Pammett, R.T., and Page, J.E. (2010).A functional genomics screen identifies diverse transcription factorsthat regulate alkaloid biosynthesis in Nicotiana benthamiana. PlantJ. 62: 589–600.

Tomato Genome Consortium (2012). The tomato genome sequenceprovides insights into fleshy fruit evolution. Nature 485: 635–641.

Trapnell, C., Pachter, L., and Salzberg, S.L. (2009). TopHat: discoveringsplice junctions with RNA-Seq. Bioinformatics 25: 1105–1111.

Tzin, V., Malitsky, S., Aharoni, A., and Galili, G. (2009). Expression ofa bacterial bi-functional chorismate mutase/prephenate dehydratasemodulates primary and secondary metabolism associated with aromaticamino acids in Arabidopsis. Plant J. 60: 156–167.

Velasquez, A.C., Chakravarthy, S., and Martin, G.B. (2009). Virus-induced gene silencing (VIGS) in Nicotiana benthamiana and tomato.J. Vis. Exp. 28: e1292.

Weinhold, A., and Baldwin, I.T. (2011). Trichome-derived O-acylsugars are a first meal for caterpillars that tags them for predation.Proc. Natl. Acad. Sci. USA 108: 7855–7859.

Weng, J.K., Philippe, R.N., and Noel, J.P. (2012). The rise ofchemodiversity in plants. Science 336: 1667–1670.

Whitaker, R.J., Byng, G.S., Gherna, R.L., and Jensen, R.A. (1981).Diverse enzymological patterns of phenylalanine biosynthesis inpseudomonads are conserved in parallel with deoxyribonucleic acidhomology groupings. J. Bacteriol. 147: 526–534.

Winzer, T., et al. (2012). A Papaver somniferum 10-gene cluster forsynthesis of the anticancer alkaloid noscapine. Science 336: 1704–1708.

Xu, X., et al., Potato Genome Sequencing Consortium (2011). Genomesequence and analysis of the tuber crop potato. Nature 475: 189–195.

Yoo, H., Widhalm, J.R., Qian, Y., Maeda, H., Cooper, B.R., Jannasch, A.S.,Gonda, I., Lewinsohn, E., Rhodes, D., and Dudareva, N. (2013). Analternative pathway contributes to phenylalanine biosynthesis in plantsvia a cytosolic tyrosine:phenylpyruvate aminotransferase. Nat. Com-mun. 4: 2833.

Zerbino, D.R., and Birney, E. (2008). Velvet: algorithms for de novoshort read assembly using de Bruijn graphs. Genome Res. 18: 821–829.

Zhang, H.-B., Bokowiec, M.T., Rushton, P.J., Han, S.-C., andTimko, M.P. (2012). Tobacco transcription factors NtMYC2a andNtMYC2b form nuclear complexes with the NtJAZ1 repressor and reg-ulate multiple jasmonate-inducible steps in nicotine biosynthesis. Mol.Plant 5: 73–84.

Zhang, L., et al. (2004). Engineering tropane biosynthetic pathway inHyoscyamus niger hairy root cultures. Proc. Natl. Acad. Sci. USA101: 6786–6791.

Zhulidov, P.A., Bogdanova, E.A., Shcheglov, A.S., Vagner, L.L.,Khaspekov, G.L., Kozhemyako, V.B., Matz, M.V., Meleshkevitch,E., Moroz, L.L., Lukyanov, S.A., and Shagin, D.A. (2004). SimplecDNA normalization using kamchatka crab duplex-specific nucle-ase. Nucleic Acids Res. 32: e37.

Ziegler, J., and Facchini, P.J. (2008). Alkaloid biosynthesis: metab-olism and trafficking. Annu. Rev. Plant Biol. 59: 735–769.

3762 The Plant Cell

DOI 10.1105/tpc.114.130534; originally published online September 16, 2014; 2014;26;3745-3762Plant Cell

Chappell, Dean DellaPenna, A. Daniel Jones, C. Robin Buell and Cornelius S. BarryWiegert-Rininger, Kevin L. Childs, John P. Hamilton, Brieanne Vaillancourt, Yun-Soo Yeo, Joseph Matthew A. Bedewitz, Elsa Góngora-Castillo, Joseph B. Uebler, Eliana Gonzales-Vigil, Krystle E.

Atropa belladonnaTropane Alkaloid Biosynthesis in A Root-Expressed l-Phenylalanine:4-Hydroxyphenylpyruvate Aminotransferase Is Required for

This information is current as of May 21, 2018

Supplemental Data /content/suppl/2014/09/03/tpc.114.130534.DC1.html

References /content/26/9/3745.full.html#ref-list-1

This article cites 100 articles, 29 of which can be accessed free at:

Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists

https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

http://www.plantcell.org/cgi/alerts/ctmain

http://www.plantcell.org/cgi/alerts/ctmain

http://www.aspb.org/publications/subscriptions.cfm

a root-expressed l-phenylalanine:4 · pdf filea root-expressed...

Documents