Overexpression of the PAP1 Transcription Factor Reveals a Complex Regulation of Flavonoid and Phenylpropanoid Metabolism in Nicotiana tabacum Plants Attacked by Spodoptera litura

Anthocyanin pigments and associated flavonoids have demonstrated antioxidant properties and benefits for human health. Consequently, current plant bioengineers have focused on how to modify flavonoid metabolism in plants. Most of that research, however, does not consider the role of natural biotic stresses (e.g., herbivore attack). To understand the influence of herbivore attack on the metabolic engineering of flavonoids, we examined tobacco plants overexpressing the Arabidopsis PAP1 gene (encoding an MYB transcription factor), which accumulated anthocyanin pigments and other flavonoids/phenylpropanoids. In comparison to wild-type and control plants, transgenic plants exhibited greater resistance to Spodoptera litura. Moreover, herbivory suppressed the PAP1-induced increase of transcripts of flavonoid/phenylpropanoid biosynthetic genes (e.g., F3H) and the subsequent accumulation of these genes' metabolites, despite the unaltered PAP1 mRNA levels after herbivory. The instances of down-regulation were independent of the signaling pathways mediated by defense-related jasmonates but were relevant to the levels of PAP1-induced and herbivory-suppressed transcription factors, An1a and An1b. Although initially F3H transcripts were suppressed by herbivory, after the S. litura feeding was interrupted, F3H transcripts increased. We hypothesize that in transgenic plants responding to herbivory, there is a complex mechanism regulating enriched flavonoid/phenylpropanoid compounds, via biotic stress signals.


Introduction
The production of specialized metabolites in plants has attracted great interest as a result of the successful development of mechanisms that enable metabolic enzymes to be produced [1][2][3]. To synchronously activate metabolic genes, plants use a large variety of transcription factors. These factors may control multiple genes or orchestrate entire metabolic pathways or their specific branches. PRODUCTION OF ANTHOCYANIN PIGMENT1 (PAP1) is a typical R2R3 MYB-type transcription factor in Arabidopsis thaliana that is able to ectopically activate the entire array of genes involved in the biosynthesis of phenolic acids and flavonoid compounds (including anthocyanins) in the flavonoid biosynthetic pathway in several plant species, including tobacco, salvia, petunia and rose [4][5][6][7][8]. The overexpression of Arabidopsis PAP1 in transgenic tobacco plants under the control of a constitutive promoter led to the induction of anthocyanins in the leaves, flowers, stems and roots; the deep red or purple appearance of the pants parts was evidence [5]. Moreover, PAP1-overexpressing petunia flowers have been shown to increase not only levels of anthocyanins but also emissions of floral volatiles (benzenoids) [7]. Gene expression profiling and measurements of the levels of pathway intermediates suggest that both increased metabolic flux and the transcriptional activation of scent and color genes underlie the enhancement of petunia flower color and scent production [7]. The coordinated regulation of metabolic steps within or between pathways involved in vital plant functions -for instance, pathways involved in floral display determining plantpollinator interactions -provides plants with a clear advantage. Indeed, bees and humans can discriminate between the floral scents of PAP1-overexpressing and control rose flowers [6]. Therefore, PAP1 overexpression may find applications in horticulture and human health.
Manipulating flavonoid metabolism, especially in response to biotic or abiotic stresses, may have unpredicted side effects, as these compounds belong to the largest group of specialized metabolites produced by plants [9]. Flavonoids serve as essential components of a number of structural polymers that provide protection; lignin, for instance, due to its antioxidant and free radical scavenging properties, protects plants from ultraviolet light and defends plants against herbivores and pathogens [10]. Anthocyanin accumulation in Arabidopsis is enhanced in response to a number of environmental stress factors, such as cold, drought, pathogen attack and nutrient depletion [11]. Moreover, it appears that transgenic tobacco expresses structural genes and regulatory genes involved in flavonoid biosynthesis, and these genes defend the plant against herbivores [12,13]. Therefore, PAP1-overexpressing plants that typically accumulate large amounts of flavonoids might possess increased plant resistance to biotic stresses, including herbivores, but to date no practical evidence of this prediction has been provided. Whether the PAP1 transgene interferes with endogenous TFs and processes involved in plants' defense and development has not been extensively investigated.
Because PAP1-overexpressing plants constitutively produce large amount of anthocyanins, we asked if the accumulation of other inducible flavonoids and phenylpropanoids was affected. PAP1 and MYBJS1 seem to share common gene targets in the pathways they regulate, and so these two transcription factors are likely to engage in complex interactions when PAP1 is introduced into plants under a constitutive promoter and also, in response to herbivory, when other MYBs are additionally activated upon. It is likely that when PAP1-overexpressing plants are damaged by herbivores, the dynamic activation of the metabolic flow via other MYB members competes with activated coumaric acid units used in the constitutive synthesis of anthocyanins by PAP1-promoted enzymes. Both transcriptional and post-transcriptional regulations of PAP1 have been shown to control anthocyanin levels in ubiquitous light/dark rhythms [22]. Moreover, when anthocyanin production is regulated by independent PAP-transcription and environmental stress, anthocyanin levels are reduced [23]. Given that the stress of herbivory could affect PAP1 function in PAP1overexpressing plants, we asked how efficiently transgenes such as PAP1 could be expected to perform under natural, uncontrolled environmental conditions.
In the current study, we found that the defensive properties of PAP1-overexpressing plants increased upon herbivore attack, and addressed the issue of metabolic fluxes and transcriptional regulation in the flavonoid/phenylpropanoid pathways that are co-regulated by powerful PAP1-activators and herbivory. Since jasmonic acid (JA) and its derivatives (referred as jasmonate [JAs]) coordinate the expression of regulatory and structural genes in the flavonoid/phenylpropanoid pathways under stress, we also assessed the role of JAs (and other phytohormones) in the herbivory-and PAP1-mediated regulation of phenylpropanoid metabolism. Finally, we discuss the implications of enhancing the production of specialized metabolites in plants under natural environmental conditions.

Production of transgenic tobacco plants
The full-length coding region of Arabidopsis PAP1 (At1g56650) cDNA was inserted into the binary vector pSMABR35SsGFP by replacing the GFP reporter gene of the vector [24]. The resulting plasmid, pSMABR-35SPAP1, was transformed into Agrobacterium tumefaciens strain EHA101 by electroporation. PAP1overexpressing tobacco (N. tabacum cv. SR1) was produced by Agrobacterium-mediated transformation as described previously [24]. After rooting and acclimatization, the regenerated plants were grown in a controlled greenhouse to set seeds. Nine lines of transgenic T 1 seeds were tested for germination on Murashige and Skoog medium supplemented with 5 mg l 21 bialaphos at 25uC under a 16-h photoperiod at a light intensity of ca. 30 mE m 22 s 21 . T 2 seeds harvested from each T 1 individual plant that showed ca. 3:1 segregation ratio were tested for bialaphosresistance again. Finally, T 2 and T 3 homozygous plant lines were used for further analyses. A homozygous tobacco T 2 line transformed with the binary plasmid pSMABR-35SpGUS (in which the reporter GFP gene of pSMABR35SsGFP was replaced by b-glucuronidase [GUS] gene), expressing bialaphos resistance [bar] and GUS genes, was used as a control.
Three representative PAP1-overexpressing plants (PAP1 lines), constitutively expressing Arabidopsis PAP1 under the control of the cauliflower mosaic virus (CaMV) 35S promoter, were used. The PAP1 lines exhibited substantial expression of PAP1 based on their increased pigmentation throughout their development, in comparison to wild-type (WT) and transgenic plants expressing uidA (coding for the GUS reporter gene and serving as a control) (Figures S1 and S2). Except for having increased pigmentation, PAP1 lines developed and matured in a similar manner to their control plants, as shown in previous studies [5,6] (Figure S2).

Plants and caterpillars
Tobacco plants were grown in plastic pots in a growth chamber at 25uC (16 h photoperiod at a light intensity of 80 mE m 22 s 21 ) for about 5 weeks. Eggs of Spodoptera litura (Fabricius) (Lepidoptera: Noctuidae) were obtained from Sumika Technoservice Co. Ltd. (Takarazuka, Japan), and the hatched larvae were reared on artificial diet (Insecta LF, Nihon Nosan Kogyo Ltd., Tokyo, Japan) in the laboratory at 25uC.

Chemical and herbivore treatments and leaf sample preparation
For chemical treatment, methyl jasmonate (MeJA, 0.1 mM, Wako Pure Chemical Industrials, Ltd., Osaka, Japan) in 2 mL of 0.1% ethanol solution was sprayed onto intact plants in plastic pots. Ethanol solution (0.1%) was used as the control. For herbivore treatment, four third-instar larvae were incubated with a whole tobacco plant in a pot ( Figure S3). All herbivory bioassays were carried out in a climate-controlled chamber at 25uC (16 h photoperiod). For the determination of phytochemical concentrations and RNA extractions, leaf tissue samples were harvested into the liquid nitrogen, flash-frozen, and kept at 280uC prior to extraction.

Performance of S. litura larvae
Larvae were reared on artificial diet until the assay started, as described above. Second-instar S. litura larvae were weighed prior to release onto a potted plant (ranging from 0.7 to 1.2 mg) and released one at a time. Each larva was allowed to move on the whole plant in a climate-controlled room at 25uC under a 16-h photoperiod. Twenty-six to thirty-two larvae were analyzed for each line.

Analysis of phenolic compounds
Phenylpropanoids and flavonoids in tobacco leaves were determined using high-performance liquid chromatography (HPLC) with a slight modification of the method described by Galis et al. (2013) and Matsuba et al. (2010) [25,26]. Replicated analyses were conducted with five independent leaf samples.
To determine phenylpropanoids (chlorogenic acid) and flavonoids (kaempferol 3-O-rutinoside and rutin), the ground leaf samples (100 mg each) were briefly homogenized in the liquid nitrogen at room temperature in 0.75 ml of 84 mM acetate buffer (adjusted to pH 4.8 with ammonia) in 40% methanol (v/v) using a FastPrep FP120 instrument (Thermo Savant, Thermo Fisher Scientific, Pittsburgh, PA, USA) in the presence of 2.3-mm diameter zirconia beads. Samples were centrifuged for 20 min at 16,100 g, 4uC, and supernatants were collected. Pellets were reextracted in 0.75 ml buffer containing 80% (v/v) methanol by shaking at RT for 15 min and centrifuged as before; supernatants were pooled with the first extracts. Aliquots (20 ml) of each sample were analyzed on a Shimadzu Prominence HPLC system equipped with a COSMOSIL 5C 18 -MS-II packed column (4.6 mm650 mm; Nacalai Tesque, Kyoto, Japan) in a gradient setup: solvent A (0.1% formic acid, 0.1% ammonia in water, pH 3.6) and solvent B (methanol); time (min)/B (%) gradient conditions: 0/2, 10/80, 15/80, 16/95, 21/95, 24/2, followed by a 6-min equilibration step. Flow rate was set to 1 ml/min and the column was maintained at 35uC. Chromatographic peaks of chlorogenic acid isomers and rutin were detected by an SPD-M20A diode array detector and compared with external calibration curves of authentic chlorogenic acid (Sigma-Aldrich) and rutin (Ishizu Pharmaceuticals Co. Ltd., Osaka, Japan) standards, respectively. Kaempferol 3-O-rutinoside content was estimated based on rutin external calibration.

Identification of phenolic acid and flavonoid compounds
Leaf extracts of PAP1 lines originally prepared for HPLC analyses were used to identify the major UV-absorbing peaks in chromatograms. A 20-ml aliquot of leaf sample was first separated on HPLC as above, and peak fractions of major UV-absorbing compounds were collected using an automated fraction collector (FRC-10A, Shimadzu, Kyoto, Japan). Fractions from 10 runs were pooled, evaporated, re-dissolved in 70% methanol in water (v/v) and spun shortly before analysis. A 10-ml aliquot of each fraction was subjected to measurement on a triple quadrupole liquid chromatography-tandem mass spectrometry (MS) system (LC-MS/MS 6410, Agilent Technologies, Santa Clara, CA, USA) equipped with a Zorbax SB-C18 column (2.1 mm id650 mm, (1.8 mm), Agilent Technologies), maintained at 40uC and coupled to an MS detector operating in positive scan mode (m/z 100-1000; fragmentor at 135 V). Samples were passed through a UV-PDA detector before entering the MS electrospray ion source for reference and later identification of the peaks. Solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile) were used in time (min)/B (%) gradient mode: 0/5, 0.5/5, 2/40, 6/40, 10/95, 15/95, 16/5, 20/5 at a constant flow rate of 0.4 ml/min. The mass spectra corresponding to UVabsorbing peaks in the chromatograms were analyzed and used to identify each compound. Only the rutin peak could be assigned against an authentic rutin standard (Ishizu Pharmaceutical Co., Ltd., Osaka, Japan), while the remaining peaks were putatively assigned based on their m/z, fragmentation properties and characteristic UV absorbance spectra.

Phytohormone analysis
Leaf hormone concentrations were determined according to the method described by

cDNA synthesis and quantitative PCR
Approximately 100 mg of the ground leaf samples was homogenized in liquid nitrogen, and total RNA was isolated from leaf tissues using NucleoSpin RNA Plant (Machery-Nargel, Düren, Germany) following the manufacturer's protocol. First-strand cDNA was synthesized using a ReverTra Ace qPCR RT Master mix (Toyobo, Osaka, Japan) and 0.5 mg of total RNA at 37uC for 15 min. Real-time PCR was performed on a Light Cycler Nano system (Roche Applied Science, Indianapolis, IN, USA) using FastStart Essential DNA Green Master (Roche Applied Science) and gene-specific primers (Table S1). The following protocol was used: initial polymerase activation: 60 s at 95uC; 40 cycles of 15 s at 95uC and 45 s at 60uC; and then melting curve analysis preset by the instrument. Relative transcript abundances were determined after normalization of raw signals with housekeeping transcript abundances of N. tabacum ELONGATION FACTOR 1a (NtEF1a; GenBank D63396). Replicated analyses were conducted with four to five independent leaf samples.

Statistical analysis
All experiments were repeated at least four times, and statistical analyses were performed using a Tukey's HSD test after a one-way ANOVA. A P value,0.05 was considered statistically significant.

Results
Transgenic tobacco plants constitutively expressing PAP1 genes are better able to defend themselves against herbivores than are non-transgenic tobacco plants We evaluated the resistance of PAP1 lines, constitutively expressing Arabidopsis PAP1, to the arthropod herbivore S. litura. Larvae were applied onto the leaves of potted WT, GUS and PAP1 lines for 2 days (Figure 1). Larvae on the PAP-5 and PAP-8 lines had markedly lower weight gain compared to those on WT or GUS lines (P,0.05). Larvae on another PAP1 line (PAP-2) exhibited slightly, but not significantly, less weight gain compared to those on WT or GUS lines (P.0.05).

Biosynthesis of flavonoids/phenylpropanoids in PAP1 lines
Since PAP1 has been shown to activate the phenylpropanoid/ flavonoid pathway [28], we measured endogenous levels of their metabolites in PAP1 lines. According to HPLC analysis supported by identification of the most abundant compounds by LC-MS/ MS, there was increased accumulation of an anthocyanin compound (cyanidin-3-O-rutinoside) as well as of two other flavonoid compounds (kaempferol 3-O-rutinoside and rutin) and of phenylpropanoid compounds (chlorogenic acid isomers 1 and 2) in leaves of PAP1 lines, relative to WT and GUS lines ( Figure 2). PAP-8 leaves accumulated all of the above five compounds at the highest levels, unsurprisingly as this line was best able to defend itself against S. litura larvae (see Figure 1). Interestingly, S. litura infestation substantially diminished the increase of the concentration of these compounds in PAP1 lines, whereas no significant change (either increase or decreased) of those compounds was observed in both WT and GUS lines (Figure 2).
We therefore assessed the transcriptional pattern of flavonoid/ phenylpropanoid pathway genes in leaves of PAP1 lines not infested or infested with S. litura for 2 days (Figure 3). In comparison to GUS lines, PAP-8 showed significantly increased transcript levels of eight genes (PALs, CHS, CHI, F3H, F39H, ANS and DFR) involved in phenylpropanoid and flavonoid biosynthesis. The increase of the transcript level was, however, diminished by 34, 60, 29, 39 and 29% of the initial increase in five cases, CHS, CHI, F3H, ANS and DFR, respectively, in PAP1 lines in response to herbivory. PAP1 was expressed similarly between uninfested and infested PAP-8 leaves. Moreover, we also determined the transcript levels of four genes responsible for producing endogenous transcription factors: MYBJS1 (another MYB transcription factor sharing transcriptional targets with PAP1 in the phenylpropanoid pathway [18]), An2 (PAP1 homologue MYB transcription factor [29]), and An1a and An1b (NtAn2-interacting, basic helix-loop-helix (bHLH) transcription factors [30]). An1a, An1b and An2 are known to be responsible for anthocyanin pigmentation in tobacco flowers [29,30]. The transcript of MYBJS1 was induced similarly between GUS and PAP-8 leaves indicates that the MYB gene was not under the direct control of PAP1. An2 was not expressed in either GUS or PAP-8 leaves, confirming what was previously known, namely, that this transcription factor is abundantly expressed in flowers but not in leaves [29]. On the other hand, although PAP-8 increased transcript levels of An1a and An1b transcript levels of GUS lines, the increase of the transcript levels was diminished by herbivory; this result was also observed in the structural genes mentioned above (see Figures 3 and S4 for 2 days and 4 days of exposure, respectively).

Upstream and downstream JA signaling for PAP1-activated genes
Since JA signaling along with other phytohormone signaling plays a central role in defense responses against chewing herbivores [31], it might be expected that induced JA production and signaling would be relevant to the low concentration of flavonoid/phenylpropanoid pathway products in PAP1 lines attacked by S. litura. To assess whether PAP1 overexpression affects JA biosynthesis and signaling, we analyzed endogenous levels of jasmonoyl-L-isoleucine (JA-Ile, an active form of JAs [32]) and of two other phytohormones: abscisic acid (ABA, involved in protective wound-healing processes [33]) and salicylic acid (SA, involved in pathogenesis and herbivore responses [34][35][36]). As shown in Figure 4A, JA and JA-Ile increased and accumulated at similar rates in the infested PAP-8 leaves compared to in the leaves of infested GUS lines. Those data indicate that PAP1 is not an upstream transcription regulator for the biosynthesis of JA. Moreover, neither ABA nor SA was increased in leaves of GUS or PAP-8 lines upon infestation, indicating that neither of those two hormones is directly linked by PAP1 actions.
Next, we investigated the gene expression levels in leaves of GUS and PAP-8 lines by applying an exogenous solution of MeJA (the commonly used active form of JAs [37]: 0.1 mM). Even though the PAP1-regulated transcript levels of An1a, An1b and F3H are negatively affected by herbivory (as shown in Figure 3), MeJA did not reduce their transcript levels to the same extent that it reduced transcript levels of PAP1 ( Figure 4B). The MeJA concentration used for our treatment was undoubtedly sufficient to induce defense responses in leaf tissues, as the transcript levels of a typical MeJA-inducible protease inhibitor (PIs [38]) and MYBJS1 genes were substantially up-regulated in the same treated samples. We therefore conclude that the down-regulation of PAP1activated genes after herbivory is closely linked not to the JA signaling and biosynthesis cascades but, rather, to other herbivoryassociated signals discussed below.

Reversible and transient regulatory patterns of F3H and PI2 during herbivory and post-herbivory periods
To clarify the temporal regulation of PAP1-regulated genes, we analyzed the temporal patterns of the expression levels of F3H (PAP1-regulated) and PI2 (JA-inducible) not only during 48 h after the first exposure to S. litura but also for up to 48 h after the larvae were removed (post-herbivory recovery; Figure 5). Consistent with the data shown in Figure 3, the regulation of the expression levels of An1a, An1b and F3H by PAP1 was suppressed 48 h after the first exposure to S. litura. After the larvae were removed, the expression levels remained low during the first 8 h but then increased 24 h and 48 h after the damage had ended. In contrast, PI2 and PAP1 displayed different expression profiles: the expression level of PI2 reached a maximum at 48 h after first exposure to S. litura but dropped to the basal level 8 h, 24 h and 48 h after the damage had ended, consistent with the defense function of the PI2 protein; the expression level of PAP1 was unchanged through the experimental period. Collectively, these results suggest that F3H might be regulated positively by an additional, JA-independent regulator(s) during the post-herbivory period.

Discussion
To date, light stimuli and abiotic stress have been reported to modulate anthocyanin concentrations in several plant taxa. For instance, in grape berries, and flowers of rose and Brunsfelsia calycina, anthocyanin accumulation appears to be suppressed in response to high temperature [39][40][41]. Most impressively, Rowan et al. (2009) reported that abiotic stress treatments of high temperature and low light cause reductions in the concentrations of anthocyanins and down-regulation of the genes involved in anthocyanin biosynthesis in leaves of PAP1-overexpressing Arabidopsis plants [23]. The inhibition of transcription occurs independently of PAP1, and the loss of anthocyanins is not the result of an obvious catabolic enzyme regulated at the transcription level. The authors addressed whether the decreased flux of anthocyanins into the vacuole occurred as a consequence of a decline in the expression of biosynthetic genes by reducing the expression of three genes (TT8, TTG1 and EGL3) of the PAP1 transcriptional complex and enhancing the expression of the potential transcriptional repressors AtMYB3, AtMYB6 and AtMYBL2. Those results especially reflect to our finding that transcript levels of both biosynthetic genes and endogenous transcription factor (positive regulator) genes were decreased in infested PAP1 plants compared to those in the uninfested PAP1 plants (Figure 3). An1a and An1b function by interacting with An2 (PAP1 homologue) to regulate anthocyanin biosynthesis [30], thus implying that those two factors can also interact with Arabidopsis PAP1. In fact, An1a and An1b transcripts were enhanced in the overexpression of PAP1, suggesting that those two bHLH proteins are able to interact with foreign PAP1 genes and are involved in the feedback regulation by the MYB/bHLH complexes as reported in Arabidopsis [42]. The diminished expression of those two genes suppress the expression of structural genes -genes that are also suppressed by herbivory -and regulates the flux of flavonoid/phenylpropanoid (Figures 2 and 3). Presumably, in return, plants gain energy and the resources required for induced defense responses (so-called homeostasis and feedback regulation), which might eventually make plants better able to withstand to generalist herbivores.
We present here two important findings, namely, that PAP1 lines that accumulated great amounts of flavonoid/phenylpropanoid pathway products defend against the generalist herbivore S. litura, and that the PAP1-stimulated accumulations of those compounds are attenuated by herbivory ( Figure 6). The first finding is in line with a previous report that transgenic tobacco plants expressing Arabidopsis transcription factor AtMYB12 exhibit more resistance to S. litura and Helicoverpa armigera than do non-transgenic plants, most probably due to these plants' enhanced accumulation of rutin [13], a chemical that is toxic for those herbivores [43]. As rutin is one of predominant products in PAP1 lines (Figure 2), this flavonoid compound is predicted to be a major anti-herbivore agent in our case as well.
The suppression of flavonoid/phenylpropanoid pathway products by herbivory is likely to reflect the phenomena described above, such as the response of Arabidopsis plants to abiotic stress, namely, the overexpression of PAP1 [23]. In addition to the proposed reason for suppression (see above), we should consider an alternative (or additional) explanation, such as that PAP1 protein could be post-transcriptionally degraded under herbivory even though normal transcript levels of PAP1 in PAP-8 leaves are maintained during herbivory (Figure 3). A similar mechanism is consistent with the repression of anthocyanin in the dark; in this case, the COP1/SPA ubiquitin ligase plays a critical role in the degradation of PAP1 proteins [22]. Although we conducted   immunoblot analysis using polyclonal antibodies against partial PAP1 peptide residues ([H]CKIKMKKRDITP[OH]) (produced by Sigma-Aldrich, Ishikari, Japan), our efforts unfortunately failed to detect sufficient PAP1 protein signals extracted from uninfested and infested PAP-8 leaves. Therefore, it remains to be clarified whether PAP1 degradation in response to herbivory is subject to similar mechanisms for light/dark responses observed in Arabidopsis.
Alternatively, the possibility that the general decrease observed in the concentration of phenolic compounds after two days of herbivory could be a result of an altered metabolic precursor allocation (competition) cannot be excluded. For example, in transgenic N. attenuata plants, the transient silencing of the expression of a hydroxycinnamoyl-CoA:shikimate/quinate hydroxycinnamoyl transferase (HCT) gene involved in lignin biosynthesis dramatically increased herbivory-induced phenolamides produced in the parallel metabolic branch [21]. It was proposed that a strong metabolic tension in the plant, exacerbated during herbivory, exists over the allocation of coumaroyl-CoA units among lignin and other phenolic compounds. This tension could be extended to anthocyanins and flavonoids, as shown in our experiments (see Figure 2), although the branches that are simultaneously enhanced have not yet been identified. In addition, the enhanced catabolism of some phenolics during herbivory stress could be responsible. Unfortunately, very little is known about the anthocyanins, kaempferol-3-O-rutinoside and rutin catabolism in plants, and the hypothesis that there is competition among metabolic precursors remains to be tested.
Generally, since JA-mediated defense signaling is predominantly activated when plants are attacked by chewing herbivores [31], it might be expected that induced JA production and signaling would be relevant to the defense responses to S. litura attack. Even though PI2 expression was highly induced during herbivory and exogenous MeJA application (see Figures 3 and 4), it was likely that the PAP1-activated genes were suppressed independently of the induction of JA and JA-Ile (an active form of JA) ( Figure 4A). Moreover, since MeJA application also failed to suppress PAP1activated genes in PAP-8 plants ( Figure 4B), JA signaling is not likely a part of the down-regulation pathway. Also, in our case SA and ABA signaling pathways seem to be irrelevant in our case (see Figure 4A). Those facts may reflect a recent report that a mitogenactivated protein kinase (MPK4) suppresses a JA-independent defense pathway in N. attenuata in response to Manduca sexta feeding [44].
Moreover, the reverse mechanism was also likely used, given that PAP1-regulated F3H expression was dramatically increased in PAP-8 leaves during the post-herbivory period, when JA signaling was already switched off ( Figure 5). This up-regulation might be caused by the fine-tuning of the PAP1 protein level after the cessation of herbivory, resulting in increased PAP1 levels beyond the basal level in the undamaged condition. Alternatively, additional signal(s) might be facilitated only when JA signaling is turned off, although these speculations require further investigation.
Horticultural ecosystems are very complicated and flexible. Hence, prior to setting genetically modified plants in the field, a wide range of systematic studies should be conducted to understand realistic horticultural ecosystems, where a myriad of plants, animals, and microorganisms interact and coevolve. However, it still remains to be elucidated how other multiple and simultaneous stresses can potentially influence flavonoid/ phenylpropanoid catabolism in real agricultural settings.  Table S1 RT-PCR primers used for this study. (XLSX) Figure 6. Schematic presentation of effect of Spodoptera litura feeding on the PAP1-regulated phenylpropanoid biosynthetic genes in a manner independent of jasmonate signaling. Arrows and bars indicate positive and negative interactions, respectively. TFs, transcription factors. doi:10.1371/journal.pone.0108849.g006