lable at ScienceDirect

Phytochemistry 131 (2016) 9e16

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Contents lists avai

Phytochemistry

journal homepage: www.elsevier .com/locate/phytochem

Identification of CYP82E21 as a functional nicotine N-demethylasein tobacco flowers

Verena Liedschulte*, Joanne Deborah Schwaar, H�el�ene Laparra, Aline Vuarnoz,B�erang�ere Philippon, Nicolas Bakaher, Nicolas Sierro, Lucien Bovet, Gerhard Lang,Simon GoepfertPMI R&D, Philip Morris Products SA, (part of Philip Morris International Group of Companies), Neuchatel, Switzerland

a r t i c l e i n f o

Article history:Received 13 May 2016Received in revised form8 July 2016Accepted 14 August 2016Available online 3 September 2016

Keywords:Nicotiana tabacumSolanaceaeTobaccoGene functionExpression studyNicotine N-demethylaseNicotine conversionNornicotine biosynthesis

* Corresponding author.E-mail addresses: Verena.Liedschulte@pmi.com

Schwaar@pmi.com (J.D. Schwaar), Helene.Laparra@pVuarnoz@pmi.com (A. Vuarnoz), Berangere.PhilippNicolas.Bakaher@pmi.com (N. Bakaher), Nicolas.SLucien.Bovet@pmi.com (L. Bovet), Gerhard.Lang@pGoepfert@pmi.com (S. Goepfert).

http://dx.doi.org/10.1016/j.phytochem.2016.08.0040031-9422/© 2016 Philip Morris Products S.A. Publishby-nc-nd/4.0/).

a b s t r a c t

In the tobacco plant, nicotine N-demethylase enzymes (NND) belonging to the cytochrome P450 familycatalyse the conversion of nicotine to nornicotine, the precursor of the carcinogenic tobacco-specificN-nitrosamine, N-nitrosonornicotine. To date three demethylase genes, namely CYP82E4, CYP82E5 andCYP82E10, have been shown to be involved in this process, while the related CYP82E2 and CYP82E3 genesare not functional. We have identified a further gene named CYP82E21 encoding a putative nicotineN-demethylase closely related to the CYP82E genes. The CYP82E21 gene was found in all Nicotianatabacum cultivars analysed and originates from the tobacco ancestor Nicotiana tomentosiformis. We showthat, in contrast to all other previously characterized NND genes, CYP82E21 is not expressed in green orsenescent leaves, but in flowers, more specifically in ovaries. The nicotine N-demethylase activity ofCYP82E21 was confirmed by ectopic expression of the coding sequence in a tobacco line lacking func-tional CYP82E4, CYP82E5 and CYP82E10 genes, resulting in an eightfold increase of nicotine demethyla-tion compared to the control plants. Furthermore, nornicotine formation can be reduced in ovaries byintroducing a CYP82E21-specific RNAi construct. Together, our results demonstrate that the CYP82E21gene encodes a functional ovary-specific nicotine N-demethylase.© 2016 Philip Morris Products S.A. Published by Elsevier Ltd. This is an open access article under the CC

BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

The pyridine alkaloids nicotine, nornicotine, anatabine andanabasine are found in many species of the genus Nicotiana atvarying relative concentrations. The alkaloid fraction of thecommercially most relevant plant, tobacco (Nicotiana tabacum),consists predominantly (>90%) of nicotine (Eich, 2008).

Due to their pronounced neurotoxic activity, these alkaloidsserve as an efficient protection of the plants against herbivores. Inaddition, their presence in floral rewards has been shown tomodulate pollinator behaviour. In line with their ecological role as

(V. Liedschulte), Joanne.mi.com (H. Laparra), Aline.on@pmi.com (B. Philippon),ierro@pmi.com (N. Sierro),mi.com (G. Lang), Simon.

ed by Elsevier Ltd. This is an open

antifeedants, the production of alkaloids in the roots and theiraccumulation in the shoot increase upon leaf damage (Baldwin,1988; Baldwin and Ohnmeiss, 1993; Halpern et al., 2010). In to-bacco production, the biosynthesis of nicotine is also induced bythe common agricultural practice of “topping” (i.e., the removal ofthe flowering head) (Qi et al., 2012).

In cultivated tobacco, alkaloids with a secondary amine moietysuch as nornicotine and anabasine are undesired as they are directprecursors to carcinogenic tobacco-specific N-nitrosamines(TSNAs) (reviewed by Hoffmann et al., 1994). TSNA formationmostly takes place during the curing of the tobacco leaf (Andersenet al., 1987; Burton et al., 1994), a drying process in which the to-bacco leaf is changed both physically and chemically, and isessential for both tobacco aroma and flavour. During this curingprocess nitrate is microbially reduced to nitrite which may reactthen with alkaloids to yield nitrosamines. Further quantities ofTSNAs may be formed during tobacco pyrolysis (Moldoveanu andBorgerding, 2008).

Besides its role as a precursor to N-nitrosonornicotine (NNN),

access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/

Fig. 1. Unrooted tree of CYP82E21 and related CYP82E coding sequences, generated bythe neighbour-joining method (Ntab: N. tabacum; Ntom: N. tomentosiformis; Nsyl:N. sylvestris).

V. Liedschulte et al. / Phytochemistry 131 (2016) 9e1610

nornicotine itself may cause adverse physiological effects inhumans by inducing aberrant glycation of proteins and covalentbinding to commonly used steroid drugs (Dickerson and Janda,2002).

In tobacco, nornicotine is formed by oxidative demethylation ofnicotine. This conversion of nicotine to nornicotine is catalysed byN-demethylase enzymes (NND) belonging to the family of cyto-chrome P450monooxygenases (CYP). Several genes coding for suchNNDs have been identified in tobacco, of which three NNDs havebeen shown to be involved in nornicotine formation in tobacco,namely CYP82E4, CYP82E5 and CYP82E10 (Gavilano and Siminszky,2007; Lewis et al., 2010; Siminszky et al., 2005). Of these NNDs,CYP82E4 contributes most to nornicotine formation and differsfrom CYP82E5 and CYP82E10. Indeed, CYP82E4 is specificallyexpressed in senescent leaves and plays a major role in nicotine N-demethylation during curing (Chakrabarti et al., 2008). Interest-ingly, the fraction of the total alkaloids in tobacco leaves that isnornicotine can rise from normally <5% to as much as 98% due to anunstable “converter locus” that stimulates transcript accumulationof CYP82E4 (Gavilano et al., 2006; Mann et al., 1964). Two furtherrelated genes identified in tobacco, CYP82E2 and CYP82E3, do notcode for functional NNDs due to point mutations (Chakrabarti et al.,2007; Gavilano et al., 2007; Siminszky et al., 2005).

In order to decrease the NNN content in tobacco, it is possible toreduce the plant nitrate content or its metabolisation to nitrite bymodifying the curing process or by microbial treatment (Lu et al.,2016; Wei et al., 2014). However, the knowledge of the nornico-tine biosynthesis pathway in tobacco provides an alternativestrategy the targeting of the three identified NND genes CYP82E4,CYP82E5 and CYP82E10 by RNA interference or by introgression ofnon-functional variant forms. RNAi-induced silencing of CYP82E4and its close homologues efficiently reduced nicotine to nornico-tine conversion to about 0.8% (Gavilano et al., 2006) and conse-quently leads to cured tobacco with reduced NNN levels (Lewiset al., 2008). Knockout mutations in all three genes likewisereduced conversion of nicotine to nornicotine to about 0.5% (Lewiset al., 2010). The remaining nornicotine level suggests that eitherthe presence of further functional NND genes or a contribution ofother, hitherto unidentified, processes to nornicotine formationexist.

In this study, we report the identification of a novel member ofthe nicotine N-demethylase family, its temporal and organ-specificexpression pattern, and its contribution to nicotine demethylationin different plant organs.

2. Results

2.1. The nicotine N-demethylase CYP82E21 is most closely related toCYP82E4

During a genomic DNA sequencing effort aimed at building areference tobacco genome, we identified a novel member of nico-tine N-demethylases that is very closely related to the describedCYP82E genes and that was termed CYP82E21 according to DavidNelson's classification (Nelson, 2006). The CYP82E21 gene could beidentified in all N. tabacum varieties analysed. Its sequence wasconfirmed by PCR amplification and Sanger sequencing.

The N. tabacum CYP82E21 gene is of Nicotiana tomentosiformisorigin since related sequence and transcripts have been observed inN. tomentosiformis, but not in Nicotiana sylvestris (Sierro et al.,2013). At the protein sequence level, CYP82E21 is most closelyrelated to CYP82E4 (94% identity) and to CYP82E3 and CYP82E2 (94and 93% identity, respectively) and it shares 91% identity withCYP82E5 and CYP82E10 (Fig. 1 and Fig. S1).

2.2. CYP82E21 is not expressed in leaves but in flowers

As previously reported, CYP82E4 and CYP82E2 are stronglyinduced upon air-curing, whereas CYP82E5, CYP82E10 and CYP82E3are already expressed in green tobacco leaves (Chakrabarti et al.,2007).

The CYP82E2 gene was inherited from N. sylvestris where it isstill functional and leads to almost complete metabolism ofnicotine to nornicotine during curing, whereas the CYP82E3 geneis derived from N. tomentosiformis where it is likewise still func-tional and leads to high conversion of nicotine to nornicotine ingreen leaves (Chakrabarti et al., 2007). The CYP82E2 and CYP82E3genes in N. tabacum are not functional nicotine N-demethylasesdue to point mutations leading to amino acid exchange in crucialregions of the enzyme (Chakrabarti et al., 2007; Siminszky et al.,2005; Fig. S1).

To determine the function of CYP82E21 and its role in nicotineconversion, the expression of CYP82E21 and of the other functionalCYP82E genes was analysed. For this purpose, different organs ofgreenhouse grown N. tabacum var. TN90 were harvested, RNA wasextracted and analysed via quantitative PCR. The PCR productswere sequenced to confirm the specificity of the reaction.

The results are shown in Fig. 2 and confirm the expression ofCYP82E5 and CYP82E10 in green leaves (Gavilano et al., 2007).They were also expressed in all other organs that were tested.CYP82E5 and CYP82E10, inherited from N. tomentosiformis andfrom N. sylvestris, respectively, are most likely homeologs as theyshow a very high sequence identity and a similar expressionpattern. Due to the sequence identity of CYP82E21 with CYP82E4,which is like CYP82E21 of N. tomentosiformis origin, we expected asimilar expression pattern for CYP82E4 and CYP82E21. CYP82E4 isreported to be expressed during senescence and also to showhigher expression in mature, compared to immature flowers(Chakrabarti et al., 2008), an observation which we also made.While CYP82E4 and CYP82E21 were both expressed in flowers,CYP82E21was not expressed at detectable levels in leaves whereasCYP82E4 showed expression exclusively in senescent leaves.CYP82E4 and CYP82E21 were both expressed in roots, albeit to anextremely low extent.

Fig. 2. Expression pattern of CYP82E21 and related functional CYP82E genes indifferent organs of N. tabacum cv. TN90. For each gene, relative expression levelscompared to the housekeeping gene NtACT9 were determined and then calculated inrelation to each other with all organ expression values per gene summing up to 1. Barsindicate mean ± SD of three biological replicates taken from three greenhouse grownplants.

V. Liedschulte et al. / Phytochemistry 131 (2016) 9e16 11

2.3. CYP82E21 is not expressed during leaf curing

As CYP82E4 expression is known to be induced during leafsenescence (Chakrabarti et al., 2008), we investigated whetherCYP82E21 is expressed at any stage during tobacco curing. Relativeexpression levels of CYP82E21 and related functional CYP82E geneswere analysed at different curing time points in N. tabacum cv.Stella leaves (Fig. 3). Samples were taken in mature leaves (upperstalk position) and in senescent leaves (lower stalk position) atharvest time. The leaves were transported to an air curing barn anda sample was taken when curing started and after 12, 24 and 48 h.CYP82E5 and CYP82E10 were expressed to a similar extent in all

Fig. 3. Expression pattern of CYP82E21 and related functional CYP82E genes inN. tabacum cv. Stella leaves at different curing time points. Samples were taken inmature leaves (upper stalk position) and in senescent leaves (lower stalk position) atharvest time. During air-curing, the first sample was taken when curing starts (“0 h”)and then again after 12 h (“12 h”), 24 h (“24 h”) and 48 h (“48 h”). Samples are takenfrom pools of several leaves. Two pools were analysed as biological replicates, thesecond replicate gave similar results (data not shown). For each gene, relativeexpression levels compared to the housekeeping gene NtACT9 were determined andthen calculated in relation to each other with all organ expression values per genesumming up to 1. Bars indicate mean ± SD of three technical replicates.

samples, including green leaves which corresponds to literaturedata (Gavilano and Siminszky, 2007). On the contrary, expression ofCYP82E4 was activated only in senescent leaves and the level ofexpression increased during the first 12 h of air-curing prior todeclining at later time points. During this curing time-courseexperiment, CYP82E21 was not expressed in leaves. Thus, weconcluded that CYP82E21 does not contribute to nicotine conver-sion in leaves, but may play a more specific role in flowers.

2.4. CYP82E21 is specifically expressed in ovaries

In order to analyse the organ-specific expression of CYP82E21 inflowers, mature flowers of N. tabacum cv. TN90 were dissected fromthree plants (biological replicates). Six organs were isolated: sepals,petals, ovaries, stigma including style, filaments and anthers. Theorgans were extracted and CYP82E21 expression analysed (Fig. 4).CYP82E21 expression was found almost exclusively in ovaries, withvery low expression levels observed in other flower organs.

2.5. CYP82E21 encodes a functional nicotine demethylase

Although the predicted CYP82E21 protein sequence presents ahigh level of sequence similarity with other demethylases such asCYP82E4, CYP82E5 and CYP82E10, its precise function cannot beinferred by homology alone, since it has been shown that a fewamino acids difference is sufficient to alter substrate specificity inthe cytochrome P450 family (Gavilano et al., 2007; Fig. S1).

The function of CYP82E21 was confirmed by ectopic expressionin tobacco. For this purpose, an expression construct was designedwith the CYP82E21 coding sequence under the control of a consti-tutive promoter (mirabilis mosaic virus promoter, MMV) in anultra-low nicotine to nornicotine converter Burley plant back-ground (TN90 tobacco cultivar) carrying mutations in CYP82E4(W329Stop), CYP82E5 (W422Stop) and CYP82E10 (P381S) genes(Lewis et al., 2010). From twenty independent primary trans-formants (T0) as well as from five control plants, leaves of matureplants were analysed for CYP82E21 expression as well as for

Fig. 4. CYP82E21 expression in flower organs of N. tabacum cv. TN90. The NtACT9 genewas used as normalizer, the immature flower sample as calibrator. Bars indicatemean ± SD of three biological replicates taken from three greenhouse grown plants.The different flower organs of each plant were pooled from several flowers in order toobtain enough material for extraction.

V. Liedschulte et al. / Phytochemistry 131 (2016) 9e1612

nicotine and nornicotine levels to calculate nicotine conversion(Fig. 5). Mutations in CYP82E4, CYP82E5 and CYP82E10 wereconfirmed in all transgenic and control lines using the CAPS/dCAPS

Fig. 5. (a) Relative CYP82E21 expression and (b, c) nicotine to nornicotine conversionlevels in (b) young and (c) mature plants, measured in leaves of the nine highestCYP82E21 overexpressing T0 lines in an ultra-low converter N. tabacum cv. TN90cyp82e4/cyp82e5/cyp82e10 background and in untransformed N. tabacum cv. TN90cyp82e4/cyp82e5/cyp82e10 plants (Ctrl). CYP82E21 expression levels measured in thedifferent T0 plants were expressed relative to plant #3 which was randomly chosen.Bars indicate mean of three technical replicates for CYP82E21 expressing T0 plants andmean ± SD of 5 biological replicates for the controls.

assay (Li et al., 2012), which was modified for CYP82E5.Expression data confirmed that CYP82E21 was not expressed in

control TN90 cyp82e4/cyp82e5/cyp82e10 green leaves (Fig. 5a). Sixof the transgenic lines that showed the highest levels of CYP82E21expression (#10, #3, #16, #15, #17, and #20) exhibited elevatednicotine to nornicotine conversion, indicating a correlation be-tween CYP82E21 expression and conversion (R2 ¼ 0.79; Fig. S2a).Nicotine to nornicotine conversion was 0.5% in control plants, andwas increased up to fourfold in young and up to eightfold (~4%) inmatureMMV:CYP82E21 high-expressing lines (Fig. 5b, c). Low levelsof nicotine to nornicotine conversion indistinguishable from thecontrol plant level were observed in plants showing the lowestexpression of the transgene. These data demonstrate that the newlyidentified CYP82E21 gene encodes a functional nicotine N-deme-thylase. Interestingly, in the best expressing lines we did not ach-ieve any conversion higher than 4% which is in contrast tooverexpression studies with CYP82E4 in which 90% conversionwere observed (Siminszky et al., 2005).

2.6. Silencing of CYP82E21 leads to up to 90% reduced conversion inflower ovaries

A silencing construct was designed to specifically targetCYP82E21, using a relatively short (100 bp) specific sequence ofCYP82E21 in forward and reverse direction, separated by an intronsequence. The construct was expressed constitutively under thecontrol of the MMV promoter in an ultra-low converter Burley to-bacco background (cv. TN90) lacking functional CYP82E4, CYP82E5and CYP82E10 genes (as described in 2.5). From nine independentprimary transformants (T0) as well as from five control plants,ovaries of mature flowers were collected from several flowers andanalysed for CYP82E21 expression as well as for nicotine and nor-nicotine levels in order to calculate conversion (Fig. 6). Mutations inCYP82E4, CYP82E5 and CYP82E10 were again confirmed in alltransgenic and control lines using the modified CAPS/dCAPS assay(Li et al., 2012).

CYP82E21 expressionwas reduced in all plants carrying the RNAiconstruct (Fig. 6a). In low CYP82E21 expressing lines, conversion ofnicotine to nornicotine was reduced by up to ten-fold to 0.5%(Fig. 6b). Expression levels and nornicotine conversion rates arecorrelated (R2 ¼ 0.85; Fig. S2b). These results suggest thatCYP82E21 is functional and accounts for about 90% of nicotineconversion in ovaries of tobacco TN90 cyp82e4/cyp82e5/cyp82e10plants.

3. Discussion

Based on the high sequence similarity between CYP82E21 andCYP82E4, both derived from the N. tomentosiformis ancestor, it islikely that CYP82E21 and CYP82E4 were formed by a duplicationevent. Although both genes are expressed in flowers, theirexpression pattern at different flower stages is not identical.Overexpression and silencing experiments of CYP82E21 showingrespectively increased and reduced conversion compared to controlplants demonstrate that this gene encodes a functional nicotine N-demethylase.

In the TN90 cyp82e4/cyp82e5/cyp82e10mutant background, thebest CYP82E21 overexpressing lines showed a maximum conver-sion of around 4% in leaves. This contrasts with the results ofoverexpression of CYP82E4 which had previously led to 90% con-version (Siminszky et al., 2005). One explanation could be thatCYP82E21 carries out an additional catalytic role and that deme-thylation is only a low efficiency or secondary activity. It is knownthat various levels of specificity can be observed throughout theP450 class of enzymes (Cochrane and Vederas, 2014; Diaz-Chavez

Fig. 6. (a) Relative CYP82E21 expression and (b) nicotine to nornicotine conversionlevels in ovaries of mature flowers of nine CYP82E21 RNAi T0 plants in an ultra-lowconverter N. tabacum cv. TN90 cyp82e4/cyp82e5/cyp82e10 background and in un-transformed N. tabacum cv. TN90 cyp82e4/cyp82e5/cyp82e10 plants (Ctrl). CYP82E21expression levels of Ctrl plants were set to 1. Bars indicate mean of three technicalreplicates for CYP82E21 expressing T0 plants and mean ± SD of 5 biological replicatesfor the controls.

V. Liedschulte et al. / Phytochemistry 131 (2016) 9e16 13

et al., 2013). It has been shown, for example, that methylanabasinecan be a substrate for the nicotine demethylases (Cai et al., 2012).

Alternatively, the comparatively overall weak conversion couldbe explained by a stereospecific conversion of (R)-nicotine. Such astereospecific activity was demonstrated for CYP82E5 andCYP82E10 (Cai et al., 2012, 2013). Within these demonstrations itwas shown that the leaves of the triple mutant exhibited conver-sion level of 0.5% and that the (R)-nicotine fraction representedaround 4% of the total nicotine content. Concomitant decrease ofthe (R)-nicotine fraction and increase of the (R)-nornicotine in theplants containing WT alleles of either CYP82E5 or CYP82E10 or bothsupported this conclusion: CYP82E5 and CYP82E10 perform thesame function and show no additive effect on nicotine demethy-lation (Cai et al., 2013). Conversely, CYP82E4 has been distinctivelyshown to catalyse the demethylation of both enantiomers ofnicotine and therefore it increases the overall nornicotine pool insenescent leaves (Cai et al., 2012, 2013).

In an additional experiment where CYP82E21 was overex-pressed in TN90 WT background, we could not identify any trans-genic lines with increased conversion rate (data not shown). Thiscould suggest non additivity of CYP82E5, CYP82E10 and CYP82E21.

While demethylation being only an accessory activity catalysed byCYP82E21 cannot be ruled out, we suspect that the activity ofCYP82E21 could be stereospecific and limited by the 4e5% (R)-nicotine present in the TN90 cyp82e4/cyp82e5/cyp82e10.

The observation of 5% conversion in the ovaries of the triplemutant while only 0.5% was observed in the leaves of the sameplant is equally surprising. Special nornicotine storage/transport orde novo synthesis could be reasonable hypotheses. As silencing ofCYP82E21 reduced conversion from 5% to 0.5% in ovaries of thetriple mutant plants we assume its implication in the observedconversion of 5% in ovaries. The finding that level of conversion inovaries with functional CYP82E21 is similar to the one observed inthe leaves of the CYP82E21 overexpression lines and in linewith thereported fraction of (R)-nicotine is a further indication for the hy-pothesis that CYP82E21 might act as an (R)-specific demethylase.

In vitro assays for substrate specificity and analysis of enantio-meric fractions of nicotine and nornicotine in plants whereexpression of CYP82E21 is down- or up-regulated will be requiredto test this hypothesis. Additionally, grafting experiments as per-formed by Cai et al. (2013) using null alleles of CYP82E21might alsobe required to pinpoint the precise contribution of CYP82E21 inoverall nornicotine synthesis.

The ecological role of nicotine demethylation in flowers remainsto be elucidated. Increased alkaloid levels in flowers can lead toreduced pest damage and also play a role in the relationship withpollinators (Adler et al., 2001; McCall and Karban, 2006). Alkaloidsare also found in nectar and pollen, where they can favour out-crossing and attract pollinators (Adler et al., 2012; Irwin et al., 2014;Kessler and Baldwin, 2007; Kessler et al., 2012; Marlin et al., 2014;Singaravelan et al., 2005). Specific expression of nicotine N-deme-thylases in ovaries could contribute to modulation of flower ornectar composition with respect to nicotine and nornicotine con-tent and their enantiomer fraction. All these factors, including thetoxicity or attraction of (S)- and (R)- nicotine and (S)- and (R)-nornicotine towards certain pests or pollinators or also alternativesubstrate specificities might have played a role in the evolution ofthe different CYP82E members.

4. Conclusion

As a product of enzymatic nicotine N-demethylation, nornico-tine is widespread in the genus Nicotiana. Since nornicotine is aprecursor to the carcinogenic tobacco-specific N-nitrosamine NNN,efforts have been undertaken to understand nornicotine formationand to apply this knowledge to develop tobacco cultivars withreduced nornicotine content. This has resulted in the discovery ofthe genes CYP82E4, CYP82E5 and CYP82E10 which are responsiblefor most of the nornicotine formation in tobacco leaves (Gavilanoand Siminszky, 2007; Lewis et al., 2010; Siminszky et al., 2005)and that these genes can be useful targets for establishing low-nornicotine tobacco lines (Gavilano et al., 2006; Lewis et al., 2010).

In the present study, we describe a new nicotine N-demethylasegene, CYP82E21, which is closely related to other CYP82E groupgenes and encodes a functional nicotine N-demethylase. However,in contrast to the previously identified CYP82E enzymes, CYP82E21does not contribute to nornicotine formation in leaves, but mainlyin ovaries. The specific expression of CYP82E21 in flowers is likelythe reason that this gene had not been discovered previously.

We assume that CYP82E21 might have similar stereo specificityas CYP82E5 and CYP82E10 and is involved mainly in (R)-nicotinedemethylation. Upon confirmation of this hypothesis, CYP82E21would class with CYP82E5 and CYP82E10 as specific (R)-nicotineconverting N-demethylases apart from CYP82E4. The result wouldallow further analyses on the structural basis for this stereospecificconversion.

V. Liedschulte et al. / Phytochemistry 131 (2016) 9e1614

It might be speculated that the possibility for differentialexpression of the various nicotine N-demethylase genes in differentplant organs is beneficial for the plant. This would enable an organspecific regulation of alkaloid composition, thus addressing the twoknown chemo-ecological roles of nicotinic alkaloids in Nicotianasp., i.e., feeding deterrence and the modulation of pollinatorbehaviour. Further experiments are needed, notably, the determi-nation of the tissue specific localisation of CYP82E21 and the othernicotine N-demethylases and their respective effect on flower andnectar alkaloid composition and investigations on the influence onpollinator behaviour.

5. Experimental

5.1. General experimental procedures

Sequence alignment and phylogenetic tree. The phylogenetic treewas constructed using Bioedit v 7.2.5 software (Hall, 1999). Thesequences were aligned with the automated ClustalW module(Thompson et al., 1994). The distance between the sequences wascalculated using the DNAdist module using the default parametersand the tree was generated using the Neighbor Joining algorithm(Saitou and Nei, 1987). The tree was drawn with FigTree (Rambaut,2014).

Sequences. The following coding sequences (Genbank identi-fiers) were used for the phylogenetic tree as well as for primerdesign for expression analysis. N. tabacum: CYP82E2: DQ131887.2;CYP82E3: DQ131888.1; CYP82E4: DQ131885.1; CYP82E5:EU182719.1; CYP82E10: HM802352.1; CYP82E21: XM_016644847.N. tomentosiformis: CYP82E21: XM_009629684; CYP82E3:EF042306.1; CYP82E4: EF042307.1; CYP82E5: XM_009620923.N. sylvestris: CYP82E2: EF472002.1; CYP82E10: HQ259993.1. Thegenomic sequence of N. tabacum CYP82E21 was confirmed bySanger sequencing.

Generation of CYP82E21 expression and RNAi plants. For con-struction of a CYP82E21 RNAi construct, a 100 bp sequence ofCYP82E21 (50-GCGATCCTAAACTCTGGCCAAATCCTGA-TAATTTCGATCCAGAGA-GATTTGTCGCTGCAGGTATTGACTTTCGTGGTCAGCATTATGAGTA-TATCCCGTT-30) was used in sense and antisense direction,connected by a hairpin loop. The construct was designed withGateway sites and synthesized by Geneart™ (Thermo Fisher Sci-entific, Regensburg, Germany). It was transferred into an expres-sion vector using the Gateway technology and correspondingenzymes (Invitrogen™, Thermo Fisher Scientific, Carlsbad, CA,USA). Agrobacterium tumefaciens was transformed with theexpression vector and used for tobacco leaf disc transformationusing previously described methods (Horsch et al., 1985).

Sterile culture of tobacco plants. Plants were grown under axenicconditions in Murashige and Skoog medium containing vitaminsand MES-buffer, which was supplemented with 2% sucrose and0.8% Phyto Agar (Duchefa, Haarlem, The Netherlands). Plants weregrown in 16 h of light (24 �C) and 8 h of darkness (20 �C).

Expression analysis using qPCR. Total RNA was extracted fromtobacco using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany).RNAwas digested using RQ1 RNase-free DNase (Promega, Madison,WI, USA) and the DNase reaction was stopped using RQ1 DNasestop solution (Promega, Madison, WI, USA). RNA was reversetranscribed using an oligo dT15 primer, dTNPs, RNasin Plus RNaseInhibitor and M-MLV Reverse Transcriptase, RNase (H-), PointMutant (all from Promega, Madison, WI, USA). qRT-PCR was per-formed on the Mx3005P system (Stratagene, Agilent, Waldbronn,Germany). Amplification reactions were carried out using ABsoluteBlue SYBR Low Rox Mix (Thermo Fisher Scientific, Surrey, UK). Foreach target, different primer pairs were designed and tested for

primer dimer formation and their performance in a qPCR run. Theirefficiency was tested using a standard curve with a five-fold dilu-tion of cDNA. The PCR products were sequenced in order to verifythat the primers specifically amplify their target sequence. As thedifferent CYP82E genes are close in sequence, primer design wascomplex and the chosen primers do not always show optimal ef-ficiency. The discrepancies in efficiency mean that the qPCRexperiment values are gene specific and do not represent an ab-solute expression value. Therefore, the comparison in expressionvalues is valid for each gene between the various organs but notbetween genes. The following primer pairs were used: CYP82E21was amplified with forward primer (50-TTGATCCAGGGTTTCAAT-TACAGC -30) and reverse primer (50-AACGTACCAAATTA-GAAAAACGTGTACC-30), CYP82E4 with forward primer (50-TTTTCAGAATTGGTTAGAGGAACATATTAAT-30) and reverse primer(50- TGTGTCTATCTCTTCTTGTGCTTTCG -30), CYP82E5 with forwardprimer (50- AGAGATTCTTCGCTGATGATATTGACTAC -30) and reverseprimer (50- CCGTAATTGTCACTTCTACAGGATTTACT -30), CYP82E10with forward primer (50- GCTGATATTGACTTTCGTGGTCAA -30) andreverse primer (50- GCGAGGCGTAATTACCACTTCTAT -30). As an in-ternal standard, the NtACT9 gene encoding actin 9 was used withforward primer (50- CTATTCTCCGCTTTGGACTTGGCA -30) andreverse primer (50- AGGACCTCAGGACAACGGAAACG -30). ForCYP82E21 expression analysis of the CYP82E21 RNAi plants, forwardprimer (50- AATTTTGGTCTCATCGTGAAGATGATA-30) and reverseprimer (50- TCACTCTCTTCTACCCATCTATCCTTG -30) were employed.

LC-UV/MS analysis of ovaria. Weighed aliquots (ca. 50 mg) offrozen and pulverized ovaria were extracted with aq. NH4OAc(100mM; sample/solvent ratio 1:3 w/v) by shaking (1 h; 70 �C). Theextraction mixtures were centrifuged and the supernatants ana-lysed by LC-UV/MS on an UHPLC system (Ultimate 3000; ThermoFisher Scientific, Bremen, Germany) equipped with a photodiodearray detector and connected to a mass spectrometer (Q Exactive;Thermo Fisher Scientific, Bremen, Germany). Chromatographicseparation was performed on an Acquity UPLC BEH C18 column(1.7 mm, 50 � 2.1 mm; Waters, Milford, MA, USA); the columntemperature was set to 55 �C. Eluents were water with 0.1% NH3(eluent A) and MeOH (eluent B) applied as a gradient (0 mine25%B; 1.9 mine60% B; 2.0 mine100% B; 2.4 mine100% B; flow: 0.6 mL/min). The autosampler was programmed to inject 5 mL of eachsample together with the same volume of an internal standardsolution (nornicotine-d4, 2 mg/mL in MeOH). Nornicotine andnicotine were eluted after 1.00 and 1.58 min, respectively. For MSdetection, positive mode electrospray ionizationwas applied with acapillary voltage of 3.7 kV, full scan datawere acquired over a rangem/z of 145e200. Nornicotine was quantified using its [MþH]þ

signal at m/z 149.10732 trace and that of the internal standard (m/z153.13243). Nicotine concentrations were calculated from peakarea measurements (in the 260 nm trace) using external calibra-tion. Nicotine conversion was calculated as [nornicotine]/([nornicotine]þ[nicotine]) � 100.

LC-UV/MS analysis of freeze-dried plant material. Weighed ali-quots (ca. 25 mg) of lyophilized and pulverized plant material wereextracted with water (1.5 mL; 90 �C; 1 h). After centrifugation, al-iquots (66.7 mL) of the supernatants were mixed with MeCN(883 mL) and nornicotine-d4 solution (50 mL; 2 mg/mL acetonitrile),centrifuged again and the resulting supernatants analysed by LC-UV/MS on an Acquity UPLC system (Waters, Milford, MA, USA)coupled to a tandem quadrupole MS detector (Acquity TQD; Wa-ters, Milford, MA, USA). Chromatographic separation was per-formed on an Acquity UPLC BEH HILIC column (1.7 mm,150 � 2.1 mm; Waters); the column temperature was set to 50 �C.Eluents were aqueous HCOONH4 (33 mM) adjusted to pH 4.5 withHCOOH (eluent A) and MeCN (eluent B) applied as a gradient(0 mine17% A; 2.0 mine30% A; 2.1 mine50% A; 2.6 mine50% A;

V. Liedschulte et al. / Phytochemistry 131 (2016) 9e16 15

flow: 0.5 mL/min). The injection volume was 5 mL. Nicotine andnornicotine were eluted after 1.82 and 2.22 min, respectively. ForMS detection, positive mode electrospray ionization was appliedwith a capillary voltage of 1.3 kV. Nornicotine was quantified usingthe MRM transition 149.0/132.0 (153.0/136.0 for nornicotine-d4)with cone voltage and collision energy set to 30 V and 14 V,respectively. Nicotine concentrations were calculated from its peakarea (in the 250 nm trace) using external calibration. Nicotineconversion was calculated as [nornicotine]/([nornicotine]þ[nicotine]) � 100.

DNA extraction and CAPS/dCAPS assay. Leaf samples wereextracted using the BioSprint 96 (Qiagen, Hilden, Germany)together with the BioSprint 96 DNA plant kit (Qiagen, Hilden,Germany). DNA samples were analysed in a CAPS/dCAPS assay (Liet al., 2012) using the MnlI assay for CYP82E4, HaeIII for CYP82E5and HphI for CYP82E10. One modification was included in theCYP82E5 assay in order to improve the specificity of the reaction.The E5H1 primer pair can bind the CYP82E10 gene and produce aband comparable to the CYP82E5 wild type allele in the gel.Consequently, a first PCR reaction was performed using the E5S1sequencing primer pair and then the product of this PCR was usedin a second nested PCR using the described E5H1 primers. Thenested PCR product was then digested by HaeIII.

Plant material

TN90 (PI 543792, TC 586, USDA -GRIN database), Stella (SOTA,CH).

Funding

The study was funded by Philip Morris Products SA.

Declaration of interests

The authors are all employees of Philip Morris Products SA.

Acknowledgements

V.L. and S.G. developed the research plan, V.L. designed andconducted experiments and evaluated the results. J.D.S. conductedtobacco transformation and CYP82E21 expression analysis intransgenic plants. H.L. supervised and followed laboratory (tobaccotransformation) and greenhouse activities. A.V. conducted nicotineand nornicotine analysis, B.P. conducted the CAPS/dCAPS assay, N.B.provided the phylogenetic analysis, N.S. provided the CYP82E21sequence information, and G.L. set up the method for nicotine/nornicotine analysis. S.G. initiated the project; S.G., L.B., G.L. andH.L. advised on the project and contributed in discussion of results.V.L., S.G. and G.L. wrote the manuscript. We thank Anthony Trickerand James Battey for review of the manuscript.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.phytochem.2016.08.004.

References

Adler, L.S., Karban, R., Strauss, S.Y., 2001. Direct and indirect effects of alkaloids onplant fitness via herbivory and pollination. Ecology 82, 2032e2044.

Adler, L.S., Seifert, M.G., Wink, M., Morse, G.E., 2012. Reliance on pollinators predictsdefensive chemistry across tobacco species. Ecol. Lett. 15, 1140e1148.

Andersen, R.A., Burton, H.R., Fleming, P.D., Hamilton-Kemp, T.R., Gay, S.L., 1987.Effects of air-curing environment on alkaloid-derived nitrosamines in burleytobacco. IARC Sci. Publ. 84, 451e455.

Baldwin, I.T., 1988. Short-term damage-induced increases in tobacco alkaloidsprotect plants. Oecologia 75, 367e370.

Baldwin, I.T., Ohnmeiss, T.E., 1993. Alkaloidal responses to damage in Nicotiananative to North America. J. Chem. Ecol. 19, 1143e1153.

Burton, H.R., Dye, N.K., Bush, L.P., 1994. Relationship between tobacco-specific ni-trosamines and nitrite from different air-cured tobacco varieties. J. Agric. FoodChem. 42, 2007e2011.

Cai, B., Jack, A.M., Lewis, R.S., Dewey, R.E., Bush, L.P., 2013. (R)-nicotine biosynthesis,metabolism and translocation in tobacco as determined by nicotine demethy-lase mutants. Phytochemistry 95, 188e196.

Cai, B., Siminszky, B., Chappell, J., Dewey, R.E., Bush, L.P., 2012. Enantioselectivedemethylation of nicotine as a mechanism for variable nornicotine compositionin tobacco leaf. J. Biol. Chem. 287, 42804e42811.

Chakrabarti, M., Bowen, S.W., Coleman, N.P., Meekins, K.M., Dewey, R.E.,Siminszky, B., 2008. CYP82E4-mediated nicotine to nornicotine conversion intobacco is regulated by a senescence-specific signaling pathway. Plant Mol. Biol.66, 415e427.

Chakrabarti, M., Meekins, K.M., Gavilano, L.B., Siminszky, B., 2007. Inactivation ofthe cytochrome P450 gene CYP82E2 by degenerative mutations was a key eventin the evolution of the alkaloid profile of modern tobacco. New Phytol. 175,565e574.

Cochrane, R.V., Vederas, J.C., 2014. Highly selective but multifunctional oxygenasesin secondary metabolism. Acc. Chem. Res. 47, 3148e3161.

Diaz-Chavez, M.L., Moniodis, J., Madilao, L.L., Jancsik, S., Keeling, C.I., Barbour, E.L.,Ghisalberti, E.L., Plummer, J.A., Jones, C.G., Bohlmann, J., 2013. Biosynthesis ofsandalwood oil: Santalum album CYP76F cytochromes P450 produce santalolsand bergamotol. PLoS One 8, e75053.

Dickerson, T.J., Janda, K.D., 2002. A previously undescribed chemical link betweensmoking and metabolic disease. Proc. Natl. Acad. Sci. U. S. A. 99, 15084e15088.

Eich, E., 2008. Solanaceae and Convolvulaceae: Secondary Metabolites e Biosyn-thesis, Chemotaxonomy and Economic Significance. Springer Verlag, Berlin,Heidelberg.

Gavilano, L.B., Coleman, N.P., Bowen, S.W., Siminszky, B., 2007. Functional analysis ofnicotine demethylase genes reveals insights into the evolution of modern to-bacco. J. Biol. Chem. 282, 249e256.

Gavilano, L.B., Coleman, N.P., Burnley, L.E., Bowman, M.L., Kalengamaliro, N.E.,Hayes, A., Bush, L., Siminszky, B., 2006. Genetic engineering of Nicotiana taba-cum for reduced nornicotine content. J. Agric. Food Chem. 54, 9071e9078.

Gavilano, L.B., Siminszky, B., 2007. Isolation and characterization of the cytochromeP450 gene CYP82E5v2 that mediates nicotine to nornicotine conversion in thegreen leaves of tobacco. Plant Cell Physiol. 48, 1567e1574.

Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor andanalysis program for Windows 95/98/NT. Nucl. Acids. Symp. Ser. 41, 95e98.

Halpern, S.L., Adler, L.S., Wink, M., 2010. Leaf herbivory and drought stress affectfloral attractive and defensive traits in Nicotiana quadrivalvis. Oecologia 163,961e971.

Hoffmann, D., Brunnemann, K.D., Prokopczyk, B., Djordjevic, M.V., 1994. Tobacco-specific N-nitrosamines and Areca-derived N-nitrosamines: chemistry,biochemistry, carcinogenicity, and relevance to humans. J. Toxicol. Environ.Health 41, 1e52.

Horsch, R.B., Fry, J.E., Hoffmann, N.L., Wallroth, M., Eichholtz, D., Rogers, S.G.,Fraley, R.T., 1985. A simple and general method for transferring genes intoplants. Science 227, 1229e1231.

Irwin, R.E., Cook, D., Richardson, L.L., Manson, J.S., Gardner, D.R., 2014. Secondarycompounds in floral rewards of toxic rangeland plants: impacts on pollinators.J. Agric. Food Chem. 62, 7335e7344.

Kessler, D., Baldwin, I.T., 2007. Making sense of nectar scents: the effects of nectarsecondary metabolites on floral visitors of Nicotiana attenuata. Plant J. 49,840e854.

Kessler, D., Bhattacharya, S., Diezel, C., Rothe, E., Gase, K., Sch€ottner, M., Baldwin, I.T.,2012. Unpredictability of nectar nicotine promotes outcrossing by humming-birds in Nicotiana attenuata. Plant J. 71, 529e538.

Lewis, R.S., Bowen, S.W., Keogh, M.R., Dewey, R.E., 2010. Three nicotine demethylasegenes mediate nornicotine biosynthesis in Nicotiana tabacum L.: functionalcharacterization of the CYP82E10 gene. Phytochemistry 71, 1988e1998.

Lewis, R.S., Jack, A.M., Morris, J.W., Robert, V.J., Gavilano, L.B., Siminszky, B.,Bush, L.P., Hayes, A.J., Dewey, R.E., 2008. RNA interference (RNAi)-inducedsuppression of nicotine demethylase activity reduces levels of a key carcinogenin cured tobacco leaves. Plant Biotechnol. J. 6, 346e354.

Li, D., Lewis, R.S., Jack, A.M., Dewey, R.E., Bowen, S.W., Miller, R.D., 2012. Develop-ment of CAPS and dCAPS markers for CYP82E4, CYP82E5v2 and CYP82E10 genemutants reducing nicotine to nornicotine conversion in tobacco. Mol. Breed. 29,589e599.

Lu, J., Zhang, L., Lewis, R.S., Bovet, L., Goepfert, S., Jack, A.M., Crutchfield, J.D., Ji, H.,Dewey, R.E., 2016. Expression of a constitutively active nitrate reductase variantin tobacco reduces tobacco-specific nitrosamine accumulation in cured leavesand cigarette smoke. Plant Biotechnol. J. 14, 1500e1510.

Mann, T.J., Weybrew, J.A., Matzinger, D.F., Hall, J.F., 1964. Inheritance of the con-version of nicotine to nornicotine in varieties of Nicotiana tabacum L. andrelated amphidiploids. Crop Sci. 4, 349e353.

Marlin, D., Nicolson, S.W., Yusuf, A.A., Stevenson, P.C., Heyman, H.M., Krüger, K.,2014. The only Africanwild tobacco, Nicotiana africana: alkaloid content and theeffect of herbivory. PLoS One 9, e102661.

McCall, A.C., Karban, R., 2006. Induced defense in Nicotiana attenuata (Solanaceae)fruit and flowers. Oecologia 146, 566e571.

V. Liedschulte et al. / Phytochemistry 131 (2016) 9e1616

Moldoveanu, S.C., Borgerding, M., 2008. Formation of tobacco specific nitrosaminesin mainstream cigarette smoke; Part 1, FTC smoking. Beitr. Tabakforsch. Int.2008 (23), 19e31.

Nelson, D.R., 2006. Cytochrome P450 nomenclature, 2004. Methods Mol. Biol. 320,1e10.

Qi, Y., Guo, H., Li, K., Liu, W., 2012. Comprehensive analysis of differential genes andmiRNA profiles for discovery of topping-responsive genes in flue-cured tobaccoroots. FEBS J. 279, 1054e1070.

Rambaut, A., 2014. Figtree, a Graphical Viewer of Phylogenetic Trees. Internet.Available from. http://tree.bio.ed.ac.uk/software/figtree.

Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method for recon-structing phylogenetic trees. Mol. Biol. Evol. 4, 406e425.

Sierro, N., Battey, J.N., Ouadi, S., Bovet, L., Goepfert, S., Bakaher, N., Peitsch, M.C.,Ivanov, N.V., 2013. Reference genomes and transcriptomes of Nicotiana sylvestrisand Nicotiana tomentosiformis. Genome Biol. 14, R60.

Siminszky, B., Gavilano, L., Bowen, S.W., Dewey, R.E., 2005. Conversion of nicotine tonornicotine in Nicotiana tabacum is mediated by CYP82E4, a cytochrome P450monooxygenase. Proc. Natl. Acad. Sci. U. S. A. 102, 14919e14924.

Singaravelan, N., Nee'man, G., Inbar, M., Izhaki, I., 2005. Feeding responses of free-flying honeybees to secondary compounds mimicking floral nectars. J. Chem.Ecol. 31, 2791e2804.

Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensi-tivity of progressive multiple sequence alignment through sequence weighting,position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22,4673e4680.

Wei, X., Deng, X., Cai, D., Ji, Z., Wang, C., Yu, J., Li, J., Chen, S., 2014. Decreasedtobacco-specific nitrosamines by microbial treatment with Bacillus amyloli-quefaciens DA9 during the air-curing process of burley tobacco. J. Agric. FoodChem. 62, 12701e12706.